Deep Flexor Tendon Repairs Analysis of Current Concepts Introduction of Improvements Establishment of New Techniques

Deep Flexor Tendon Repairs Analysis of current concepts Introduction of improvements Establishment of new techniques

A Dissertation in Medicine by

Tim Sebastian Peltz

Submitted in Fulfilment of the Requirements for the Degree of Doctor of Philosophy March 2014

Prince of Wales Hospital University of New South Wales Copyright Statement

"I hereby grant to the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or patt in the University libraries in all f01ms of media, now or hereafter known, subject to the provisions of the Copytight Act 1968. I retain all proptietary tights, such as patent tights. I also retain the right to use in future works (such as atticles and books) all or part of this thesis or dissertation. I also authotise University Microfilms to use the abstract of my thesis in Dissertations Abstract Intemational. I have either used no substantial p01tions of copytight material in my thesis or I have obtained pennission to use copytight matetial; where petmission has not been granted I have applied I will apply for a pattial resttiction of the digital copy of my thesis or dissertation."

Date: 18.08.2014

Authenticity Statement

"I cettify that the Library deposit digital copy is a direct equivalent of the final officially approved version of my thesis. No emendation of content has occmTed and if there are any minor variations in formatting, they are the result of the conversion to the digital f01mat."

Date: 18.08.2014 II Originality Statement

"I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, or substantial propm1ions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgment is made in the thesis. Any cont1ibution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project's design and conception or in style, presentation and linguistic expression is acknowledged"

Date: 18.08.2014

III THE UNIVERSITY OF NEW SOUTH WALES

Thesis/Dissertation Sheet

Surname or Familv name: Peltz

First name: Tim Other name/s: Sebastian

Abbreviation for de~ree as given in the University calendar: PhD

School: Prince of Wales Clinical School Faculty: Medicine

Title: Deep Flexor Tendon Repairs: Analysis of current concepts, Introduction of improvements, Establishment of new techniques.

Abstract

The patienfs ability to master everyday living tasks is strongly dependent on the utility of healthy functional hands. Therefore the restoration of hand function after tendon injuries is of utmost importance. Tendon injuties are common, especially lacerations in zone II. In this area flexor tendons follow a complicated anatomy and are difficult to repair.

Consequently this thesis focuses on the research of tendon repairs in zone II and aims to improve repair quality and ultimately final functional outcome for the patient.

To approach this topic the author of this thesis firstly analysis current concepts of tendon repair research models, secondly investigates common tendon repair techniques and improvements of these techniques and thirdly introduces new techniques for repair of deep flexor tendons in zone II.

This thesis consists of ex vivo and in vivo experiments, which all build on another. Main results of these expetiments are as follows:

• In regards to comparability to human tendons, sheep tendons are better tendon surrogates as pig tendons if used in ex vivo laboratory experiments.

IV • When focusing on gapping resistance, "locking loop" repair configurations for tendon repairs are not substantially different to "grasping loop" configurations, and only "cross-locks", as used in the Adelaide repair technique, deserve the adjective description "locking".

• The cun·ent gold standard of tendon repairs, the Adelaide repair, produces better repair stability if performed with larger cross locks.

• The author's interlocking modification of the Adelaide repair can further improve the Adelaide repair's stability.

• In an ex vivo setting, the author's new tendon repair concept, the knotless 3D barbed suture tendon repair, produces superior repair stability than the Adelaide repair.

• The turkey tendon model is the first tendon model that replicates human anatomy and tendon sizes and can be used in ex vivo as well as in vivo tendon repair experiments.

• In an in vivo tendon repair scenario, the use of the knotless 3D barbed suture tendon repair with resorbable barbed sutures produces inferior repair stability compared to the Adelaide repair, but improves functional outcomes.

This thesis presents new insights into tendon repair research from a surgical and biomechanical point of view.

The use of the novel unknotted barbed suture repair method did show superior results in ex vivo experiments but barb resorbtion in the in vivo experiments caused high failure rates. Nevertheless, there is a probability that with the development of more stable small barbed suturing materials in the near future it will be possible to fUJther improve deep flexor tendon repairs using this novel repair technique.

Declaration relating to disposition of project thesis/dissertation I hereby grant to the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or in part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all property rights, such as patent rights. I also retain the right to use in future works (such as articles fr ~ ks all or part of this thesis or dissertation. I also authorise Unive i i r se the 350 word abstract of my thesis in Dissertation Abstracts International (this is applicable( fioc< S S/01 )/ "7 0 ~ / ,'- ...... ~ .. /.~ ...... ~:.0.!:.19 ·············~ . l;~;~ If··· W1tness Date The University rec~gnis~Ht'iere may be exceptional circumstances requiring restrictions on copying or conditions on use. Requests for restriction for a period of up to 2 years must be made in writing. Requests for a longer period of restriction may be considered in exceptional circumstances and require the approval of the Dean of Graduate Research.

FOR OFFICE USE ONLY

ACKNOWLEDGEMENTS

Acknowledgements and my honest thanks go out to all staff of the Surgical and

Orthopaedic Research Laboratories (SORL) at the University of New South Wales

(UNSW), staff of Prince of Wales Hospital (POW), St. Lukes Hospital, Sydney Hospital and staff of St. George Hospital Biological Research Centre (BRC). Primarily to my bosses, Prof. William R Walsh (or Bill), A/Prof Mark P Gianoutsos (or Mark) and my supervisor Dr Rema Oliver (or Rema). But also everybody else whom I met, annoyed, learned of and laughed with at SORL, BRC, POW, St Lukes and Sydney Hospital. Of course all would not have been possible without my amazing family and my dear friends over here and back home in Munich, who always gave me the biggest support no matter how impossible the task seemed. And I want to thank the ocean, for providing these beautiful waves for surfing, no matter what times, always energizing me with the power of nature and guiding me through life.

Dedicated to the coolest family on earth, the Peltz Family!

LIST OF PUBLICATIONS

Journal Articles

1. In reply to "Letter regarding performance of a knotless four-strand flexor tendon

repair with a unidirectional barbed suture device: a dynamic ex vivo

comparison. Peltz TS, Walsh WR. J Hand Surg Eur Vol. 2014 Jan;39(1):116-

2. Performance of a knotless four-strand flexor tendon repair with a unidirectional

barbed suture device: a dynamic ex vivo comparison. Peltz TS, Haddad R,

Scougall PJ, Gianoutsos MP, Bertollo N, Walsh WR. J Hand Surg Eur Vol.

2014 Jan;39(1):30-9.

3. In reply to “Letter regarding influence of locking stitch size in a four-strand

cross- locked cruciate flexor tendon repair”. Peltz TS, Walsh WR. J Hand Surg

Am. 2012 Jan;37(1):188-9.

4. Biomechanical evaluation of flexor tendon repair using barbed suture material: a

comparative ex vivo study. Haddad R, Peltz TS, Walsh WR. J Hand Surg Am.

2011 Sep;36(9):1565-6.

VII

5. Influence of locking stitch size in a four-strand cross-locked cruciate flexor

tendon repair. Peltz TS, Haddad R, Scougall PJ, Nicklin S, Gianoutsos MP,

Walsh WR. J Hand Surg Am. 2011 Mar;36(3):450-5. Epub 2011 Feb 17.

6. The change in three-dimensional geometry of the Kessler flexor tendon repair

under tension: a qualitative assessment using radiographs. Peltz TS, Haddad R,

Walsh WR. J Hand Surg Eur Vol. 2010 Oct;35(8):676-7.

7. The relationship between gap formation and grip-to-grip displacement during

cyclic testing of repaired flexor tendons. Haddad R, Peltz TS, Lau A, Bertollo

N, Nicklin S, Walsh WR. J Biomech. 2010 Oct 19;43(14):2835-8.

8. A biomechanical assessment of repair versus nonrepair of sheep flexor

tendons lacerated to 75 percent. Haddad R, Scherman P, Peltz TS,

Nicklin S, Walsh WR. J Hand Surg Am. 2010 Apr;35(4):546-51.

VIII

Conference Presentations

1. Mar 2010 Australian Hand Surgery Society Annual Conference

(Canberra/ACT). Podium presentation: A biomechanical comparison of

human, porcine and ovine deep flexor tendons. What is the ideal Model

for in vitro studies?

2. Mar 2010 Orthopaedic Research Society ORS Annual Meeting (New

Orleans/USA). Poster Presentation: Changes in Core Suture Geometry

within repaired Flexor Tendons: an X-ray Evaluation.

3. May 2010 Royal Australasian College of Surgeons Annual Scientific Congress

ASC (Perth/WA) Podium presentation: Changes in Core Suture

Geometry within repaired Flexor Tendons: an X-ray Evaluation. Poster

Presentation: A biomechanical comparison of human, porcine and ovine

deep flexor tendons. What is the ideal Model for in vitro studies?

4. Jun 2010 Federation of European Societies for Surgery of the Hand FESSH

Congress (Bucharest/Romania). Podium Presentation: 1. A

biomechanical comparison of human, porcine and ovine deep flexor

tendons. What is the ideal Model for in vitro studies? 2. Changes in

Core Suture Geometry within repaired Flexor Tendons: an X-ray

Evaluation.

5. Nov 2010 Combined NSW & QLD Hand Surgery Society Conference (Hamilton

Island/QLD). Podium presentation: New developments in tendon

repairs: The modified Adelaide repair and the knotless barbed suture

tenorrhaphy. A biomechanical comparison.

6. May 2011 International Confederation for Plastic Reconstructive & Aesthetic

Surgery (IPRAS) Conference in Vancouver, Canada: Podium

Presentation: New developments in tendon repairs: The modified

Adelaide repair and the knotless barbed suture tenorrhaphy. A

biomechanical comparison.

7. Jul 2011 2011 Plastic Surgery Congress (PSC) of the Australian Society of

Plastic Surgeons (ASPS) and the New Zealand Association of Plastic

Surgeons (NZAPS), Gold Coast, Australia. Podium Presentation: New

developments in tendon repairs: The modified Adelaide Repair and the

knot-less barbed suture tenorrhaphy.

8. Sep 2011 Combined Austrian and German Plastic Surgery Conference

(Innsbruck/Austria): Podium Presentation: Barbed sutures in tendon

repairs: A dynamic ex vivo comparison.

9. Feb 2012 Orthopaedic Research Society Annual Meeting (ORS) 2012

(San Francisco/USA): Poster Presentation: Changes in Core Suture

Geometry within Repaired Flexor Tendons: a X-ray Evaluation.

10. Mar 2012 Combined American and Australian Hand Surgery Societies

Conference (ASSH/AHSS) 2012 (Kauai, USA): Podium Presentations:

Changes in Core Suture Geometry within Repaired Flexor Tendons: A

X-ray Evaluation. And: New Developments in Tendon Repairs: The

Modified Adelaide Repair and the three-dimensional Knotless Barbed

Suture Tenorrhaphy.

11. Aug 2012 2012 New Zealand Hand Surgery Society Conference (Queenstown,

New Zealand): Podium Presentation: Changes in Core Suture

Geometry within Repaired Flexor Tendons: A X-ray Evaluation.

12. Oct 2012 9th Congress of Asian Pacific Federation of Societies for Surgery

of The Hand (APFSSH) (Nusa Dua, Indonesia): Podium Presentations:

“Changes in Core Geomentry within Repaired Flexor Tendons:

Investigation with A Novel Radiographic Method”. And: “ The Turkey

Deep Flexor Tendon: A New Animal Model For Multi-Stand Ex-Vivo and

In-Vivo Tendon Repair Experiments”.

13. Apr 2013 130th German Surgery Society conference (Munich, Germany):

Podium Presentation: "Changes in core suture geometry within repaired

flexor tendons: Investigations with a novel radiographic method".

14. May 2013 58th Plastic Surgery Research Conference (PSRC) (Los Angeles

USA): Podium Presentation: “New developments in tendon repairs:

Thecmodified Adelaide repair and the knotless barbed suture repair”.

And: Poster Presentation: "Geometrical and biomechanical stability of

currentctendon repair configurations: An investigation with a new

radiographic method."

15. Oct 2013 68th Annual meeting of the American Society for Surgery of the

Hand (ASSH) (San Francisco, USA)

- Moderator of the 2013 ASSH Tendon Repair Research Symposium.

- Poster Presentation: The turkey deep flexor tendon: A new animal

model for multi strand ex-vivo and in-vivo tendon repair experiments.

- Scientific paper podium presentation: The knotless tendon repair with

a resorbable unidirectional barbed suture device - An in vivo

comparison in the turkey foot.

XII

ABSTRACT

The patient's ability to master everyday living tasks is strongly dependent on the utility of healthy functional hands. Therefore the restoration of hand function after tendon injuries is of utmost importance. Tendon injuries are common, especially lacerations in zone II. In this area flexor tendons follow a complicated anatomy and are difficult to repair.

Consequently this thesis focuses on the research of tendon repairs in zone II and aims to improve repair quality and ultimately final functional outcome for the patient.

This thesis consists of ex vivo and in vivo experiments, which all build on another. Main results of these experiments are as follows: XIII

 In regards to comparability to human tendons, sheep tendons are better tendon

surrogates as pig tendons if used in ex vivo laboratory experiments.

 When focusing on gapping resistance, “locking loop” repair configurations for

tendon repairs are not substantially different to “grasping loop” configurations,

and only “cross-locks”, as used in the Adelaide repair technique, deserve the

adjective description “locking”.

 The current gold standard of tendon repairs, the Adelaide repair, produces better

repair stability if performed with larger cross locks.

 The author's interlocking modification of the Adelaide repair can further

improve the Adelaide repair's stability.

 In an ex vivo setting, the author's new tendon repair concept, the knotless 3D

barbed suture tendon repair, produces superior repair stability than the Adelaide

repair.

 The turkey tendon model is the first tendon model that replicates human anatomy

and tendon sizes and can be used in ex vivo as well as in vivo tendon repair

experiments.

 In an in vivo tendon repair scenario, the use of the knotless 3D barbed suture

tendon repair with resorbable barbed sutures produces inferior repair stability

compared to the Adelaide repair, but improves functional outcomes.

XIV

This thesis presents new insights into tendon repair research from a surgical and biomechanical point of view. The use of the novel unknotted barbed suture repair method did show superior results in ex vivo experiments but barb resorbtion in the in vivo experiments caused high failure rates. Nevertheless, there is a high probability that with the development of more stable barbed suturing materials in the near future it will be possible to further improve deep flexor tendon repairs using this novel repair technique.

TABLE OF CONTENTS

ACKNOWLEDGEMENTS ...... VI

LIST OF PUBLICATIONS ...... VII

ABSTRACT...... XIII

TABLE OF CONTENTS ...... XVI

LIST OF FIGURES ...... XXIII

LIST OF TABLES ...... XXXVII

NOMENCLATURE ...... XXXVIII

XVI

CHAPTER 1: Introduction ...... 1

1.1 Thesis Outline ...... 1

1.2 Background and Literature Review ...... 5

1.2.1 Tendon Anatomy and Physiology ...... 5 1.2.1.1 The role of the tendon in the human body ...... 5 1.2.1.2 Hand Anatomy ...... 7 1.2.1.3 Pulley System ...... 12 1.2.1.4 Flexor Tendon Blood Supply ...... 14 1.2.1.5 Zones ...... 17 1.2.1.6 Tendon biology and composition ...... 19 1.2.2 Flexor tendon repairs...... 24 1.2.2.1 History of tendon repairs ...... 24 1.2.2.2 Mechanism of injuries ...... 26 1.2.2.3 Repair techniques ...... 32 1.2.3 Biomechanical testing ...... 42 1.2.4 Flexor tendon healing and rehabilitation...... 43 1.2.4.1 Healing ...... 43 1.2.4.2 Rehabilitation ...... 45 1.2.5 Complications ...... 49 1.2.5.1 Gapping of repair constructs ...... 49 1.2.5.2 Increased gliding resistance ...... 50 1.2.5.3 Adhesion formation ...... 51 1.2.5.4 Repair failure ...... 53

XVII

1.3 Thesis Rationale and Aims ...... 54

1.3.1 Thesis Rationale ...... 54 1.3.2 Thesis Aims and Hypothesis ...... 55 Tendon research models ...... 56 Investigations of current tendon repair techniques ...... 56 Improvements to the Adelaide repair ...... 57 A new tendon repair method: The knotless barbed suture repair...... 57 A new tendon repair animal model ...... 58 The knotless barbed suture repair tested in an in vivo experiment ...... 59

CHAPTER 2: Ex vivo studies ...... 60

2.1 What is the ideal animal model for ex vivo tendon repair experiments? ..... 60

2.1.1 Introduction ...... 60 2.1.2 Methods ...... 62 2.1.2.1 Standardising site in zone II...... 67 2.1.2.2 Biomechanical testing ...... 69 2.1.2.3 Histology ...... 72 2.1.3 Results ...... 73 2.1.4 Discussion ...... 80

XVIII

2.2 Biomechanical and geometrical investigations of current flexor tendon techniques: Changes in core suture geometry within repaired flexor tendons under tension ...... 89

2.2.1 Two strand repairs or the Kessler Dilemma ...... 89 2.2.1.1 Introduction ...... 89 2.2.1.2 Pilot experiments ...... 93 2.2.1.3 Methods ...... 97 2.2.1.4 Results ...... 99 2.2.1.5 Discussion ...... 103 2.2.2 Multi strand repairs ...... 110 2.2.2.1 Introduction ...... 110 2.2.2.2 Methods ...... 111 2.2.2.3 Results ...... 116 2.2.2.4 Discussion ...... 118 2.3 The current Gold Standard in Flexor Tendon repairs, the “Adelaide

Repair” - Improvements and Modifications ...... 123

2.3.1 The influence of locking stitch size in an “Adelaide Repair” ...... 123 2.3.1.1 Introduction ...... 123 2.3.1.2 Materials and methods ...... 126 2.3.1.3 Results ...... 130 2.3.1.4 Discussion ...... 132 2.3.2 The interlocking Adelaide Repair: A stronger modification of the Adelaide Repair ...... 138 2.3.2.1 Introduction ...... 138 2.3.2.2 Materials and methods ...... 140 2.3.2.3 Results ...... 147 2.3.2.4 Discussion ...... 149

XIX

2.4 A new concept for flexor tendon repairs: The knotless 3D barbed suture

flexor tendon repair ...... 155

2.4.1 Introduction ...... 155 2.4.2 Materials and methods ...... 159 2.4.3 Results ...... 166 2.4.3.1 Gap formation ...... 166 2.4.3.2 Ultimate load...... 168 2.4.3.3 Cycles to 2 mm gap ...... 169 2.4.3.4 Mode of final failure ...... 171 2.4.4 Discussion ...... 171

CHAPTER 3: In vivo studies ...... 180

3.1 A new tendon model for ex vivo and in vivo experiments: The turkey foot 180

3.1.1 Introduction ...... 180 3.1.2 Methods ...... 198 3.1.2.1 Mechanical Testing ...... 200 3.1.2.2 Statistical Analyses ...... 201 3.1.3 Results ...... 202 3.1.4 Discussion ...... 203

3.2 Surgery on the turkey foot ...... 212

3.2.1 Introduction ...... 212 3.2.2 Methods ...... 214 3.2.2.1 Animal housing and welfare ...... 214 3.2.2.2 General Anaesthesia ...... 217 3.2.2.3 Surgery ...... 220 3.2.2.4 Postoperative Care ...... 231 3.2.2.5 Euthanization ...... 232 3.2.3 Results and Discussion ...... 233

3.3 The knotless barbed suture flexor tendon repair: An in vivo comparison in the turkey foot ...... 235

3.3.1 Introduction ...... 235 3.3.2 Methods ...... 237 3.3.2.1 Surgery ...... 238 3.3.2.2 Grouping of specimen ...... 246 3.3.2.3 Endpoints ...... 247 3.3.3 Results ...... 261 3.3.3.1 Operating time for core repair ...... 263 3.3.3.2 Repair failures per group after six weeks and failure mechanism ... 264 3.3.3.3 Range of motion per group after six weeks ...... 265 3.3.3.4 Classification of adhesion formation ...... 266 3.3.3.5 Repair bulk in mm ...... 267 3.3.3.6 Average gap in repair ...... 268 3.3.3.7 Failure force during biomechanical testing ...... 269 3.3.3.8 Mode of failure during testing ...... 270 3.3.3.9 Histology ...... 270 3.3.4 Discussion ...... 282

XXI

CHAPTER 4: Conclusions, limitations and future directions ...... 293

CHAPTER 5: References ...... 302

XXII

LIST OF FIGURES

Figure 1-1 Thesis structure ...... 4

Figure 1-2 The motor (left) and sensory (right) homunculus [8]...... 7

Figure 1-3 Homunculus [9] ...... 8

Figure 1-4 Radiographic anatomy of adult hand [11]...... 9

Figure 1-5 The muscles of the left hand [10]...... 11

Figure 1-6 Flexor tendon synovial sheath and pulleys including the palmar aponeurosis pulley, five annular pulleys and three cruciform pulleys [16]...... 13

Figure 1-7 Mechanics of the pulley system [17]...... 14

Figure 1-8 Vascular blood supply of FDS and FDP tendons [23] ...... 16

Figure 1-9 Flexor tendon zones [29]...... 17

Figure 1-10 Representative schematic of a multi-unit hierarchical structure of the tendon [36]...... 20

Figure 1-11 Fibroblast distribution within a collagen bundle. Collagen fibrils tightly surround the cells [42]...... 22

XXIII

Figure 1-12 Jersey finger. Typical closed tendon injury in sports: Finger gets caught in opponent’s jersey and deep flexor tendon tears of distal phalanx insertion while flexing the caught finger tip [68]...... 27

Figure 1-13 Typical multi level hand injury with tendon lacerations ...... 28

Figure 1-14 Typical single cut flexor tendon injury after work trauma with knife ...... 29

Figure 1-15 Simulation of an open tendon laceration in the act of grabbing a sharp object (left picture). Note the location of laceration of the deep flexor tendon changes after opening the grabbing hand or extending the finger (right picture) ...... 30

Figure 1-16 Zone II or no man's land with it’s complicated ...... 31

Figure 1-17 Example of a two-strand technique...... 34

Figure 1-18 Example of a true four-strand technique...... 34

Figure 1-19 Two-strand Kessler type repair performed with a single stranded suture

(above) and with a double stranded suture (below) to create a four-strand repair [85].

...... 35

Figure 1-20 : Double and Triple Kessler repair configuration [64]...... 35

Figure 1-21 Common peripheral suture techniques: (A) simple running; (B) cross-stitch;

Figure 1-22 Illustration of biological and biomechanical healing of repaired tendon. ... 44

Figure 1-23 Relative tolerance of repairs to early mobilisation. Active rehabilitation should not be used for a tendon repaired with a two-strand technique [79]...... 47 XXIV

Figure 1-24 Synergistic wrist and finger motion. (A) Wrist extension with finger flexion alternating with (B) wrist flexion and finger extension increases flexor tendon excursion without increasing force [111]...... 48

Figure 1-25: Gapped tendon catching beneath pulley [111]...... 50

Figure 2-1 Harvesting of pig deep flexor tendons...... 63

Figure 2-2 Harvesting of sheep deep flexor tendons...... 64

Figure 2-3 Harvesting of human deep flexor tendons...... 65

Figure 2-4 Harvested human deep flexor tendons...... 66

Figure 2-5 Harvested pig and sheep tendons...... 67

Figure 2-6 Dissection of zone II in the pig (A) and sheep (B), with identification of the A1 to A4 pulleys and the long vinculum (LV). Radiographs of the pig (C) and sheep (D) cadaver dissections with clamps marking long vinculum (LV) and A1 pulley...... 68

Figure 2-7: The tendon-loop construct fixed in the testing machine...... 70

Figure 2-8 Example graph of force displacement curve generated during mechanical testing...... 71

Figure 2-9: The loads generated with pulling the suture loop through the deep flexor tendon (splitting) of human, sheep, and pig tendon at start of splitting, 2mm splitting,

5mm splitting and 10mm splitting.* indicate significant difference (p<0.05)...... 74

Figure 2-10: The loads generated with pulling the suture loop through the deep flexor tendon (splitting) of human, sheep and pig illustrated graphically...... 74 XXV

Figure 2-11: Haematoxylin and Eosin light microscopy (x4 magnification) of human, sheep and pig flexor tendons at Zone II. In the cross sections, note the fascicle geometry, being most dense in the pig...... 76

Figure 2-12: Greater crimping of pig fibrils in the longitudinal projections compared to human and sheep (x4 magnification)...... 77

Figure 2-13 Greater crimping of pig fibrils in the longitudinal projections compared to human and sheep (x20 magnification)...... 78

Figure 2-14: Polarised microscopy x10 magnified ...... 79

Figure 2-15: Mean tendon widths of pig, sheep and human tendons 5mm distal to A1 pulley. Differences in width were significant (p<0.05)...... 80

Figure 2-16: Grasping Kessler Repair. Note, the transverse strand passes deep to the longitudinal strands...... 90

Figure 2-17: Locking Kessler repair. Note, the transverse strand passes superficial to the longitudinal strands...... 91

Figure 2-18: Modified Locking Kessler Repair. Note, same as Locking Kessler Repair, but longitudinal strands exit and re-enter the dorsal surface of the tendon to reliably produce a locking Kessler configuration...... 92

Figure 2-19 Anterior posterior (AP) X-rays of Kessler repair configuration performed with iodine contrasted braided polyester suture. Left pre tensioning, right post tensioning...... 94 XXVI

Figure 2-20 Lateral X-rays of Kessler repair configuration performed with iodine contrasted braided polyester suture. Left pre tensioning, right post tensioning...... 95

Figure 2-21 AP X-rays of Kessler repair configuration performed ...... 96

Figure 2-22 Lateral X-rays of Kessler repair configuration performed ...... 96

Figure 2-23 Illustration of measurements of "gapping" with Ligaclip markers...... 98

Figure 2-24: Pre and Post tension anterior-posterior radiographs of Grasping Kessler

(top), Locking Kessler (middle) and Modified Locking Kessler (bottom). Note the loss of the Kessler Configuration in the Grasping Kessler Repair post tension...... 100

Figure 2-25: Pre and Post tension lateral radiographs of Grasping Kessler (top), Locking

Kessler (middle) and Modified Locking Kessler (bottom). Note the different directions of unfolding rotation of Grasping Kessler compared to Locking and Modified Locking

Kessler post tension indicated with arrows...... 101

Figure 2-26 Measured gaps in mm post tensioning...... 102

Figure 2-27: Pre and post tension radiograph of faulty repair. Note, the transverse strand passes superficial to one of the longitudinal strands but deep to the other longitudinal strand. Therefore the unfolding rotation in opposite directions (see arrows)...... 104

Figure 2-28 Grasping (above) and locking (below) double Kessler repair techniques .. 112

Figure 2-29 The four-strand cruciate repair technique [89]...... 113

Figure 2-30 The cross locked cruciate or Adelaide technique [76]...... 114 XXVII

Figure 2-31 Example pictures of repairs from each tested group before and after tensioning...... 116

Figure 2-32 illustrates the average measured distances of the marker clips from the tendons end for each repair group. Gap in mm after tensioning (1 = Grasping Double

Kessler Repair; 2 = Locking Double Kessler Repair; 3 = Cruciate Repair; 4 = Adelaide

Repair)...... 117

Figure 2-33 Radiographs of Cruciate repair, pre and post tensioning, left AP and right lateral projections...... 120

Figure 2-34 Radiographs of Adelaide repair, pre and post tensioning, left AP and right lateral projections...... 121

Figure 2-35: Schematic drawing of the cross-locked cruciate repair technique...... 124

Figure 2-36: (I): Dissection of pig trotter: The deep flexor tendon is seen passing through the chiasma of the superficial tendon. The annular pulleys (A1-A4) and the long vinculum (LV) are labelled. (II): X-ray of porcine forelimb trotter: Zone II is adjacent to the proximal phalanx, as indicated between the left clamp clipped to the long vinculum

(LV) and the right clamp clipped to the distal margin of the A1 pulley (A1)...... 126

Figure 2-37: Schematic drawing of the repairs in group 1 (2mm cross-locks) and group 2

(4mm cross-locks)...... 128

Figure 2-38: Mean load to 2mm gap formation and mean load to failure by repair group...... 131 XXVIII

Figure 2-39 Pictures of Adelaide repair pre and post tensioning ...... 138

Figure 2-40 First drawings of the modification for the Adelaide repair ...... 139

Figure 2-41 Adelaide repair technique ...... 141

Figure 2-42 Interlocking or modified Adelaide repair technique ...... 142

Figure 2-43: Dynamic and static testing sequence: Testing was paused at 3N trough load and standardised pictures were taken at 10, 20, 30, 40, 50, 100, 150, 200, and 250 load cycles. Then the repair construct was pulled to failure at 20mm/min distraction rate...... 144

Figure 2-44: Testing setup with repaired tendon in hydraulic clamps held at 3N trough load. Caliper fixed at 10mm and mounted in same plane. Eight measurements were made using digital measurement software and averaged to determine average gap in each picture...... 145

Figure 2-45: Illustration of gapping propensities of repair techniques...... 147

Figure 2-46: Final load at failure of repair techniques...... 148

Figure 2-47: Illustration of the four-strand cross locked cruciate repair technique

(Adelaide Repair). Tendon purchase and cross-lock width indicated in mm...... 160

into vertical plane (three dimensional repair). Third strand (see 7) runs in the vertical plane. c) Fourth strand stays in the vertical plane and finishes with a triple zigzag. d)

Suture is cut flush with the tendon...... 161

Figure 2-49: Scanning Electron Microscopy (SEM) picture of the used 3/0 unidirectional barbed glycolic-carbonate suture (V-Loc 180TM; Covidien. Mansfield, MA)...... 162

Figure 2-50: Dynamic and static testing sequence: Testing was paused at 3N trough load and standardised pictures were taken at 10, 20, 30, 40, 50, 100, 150, 200, and 250 load cycles. Then the repair construct was pulled to failure at 20mm/min distraction rate...... 164

Figure 2-51: Testing setup with repaired tendon in hydraulic clamps held at 3N trough load. Caliper fixed at 10mm and mounted in same plane. Eight measurements were made using digital measurement software and averaged to determine average gap in each picture...... 165

Figure 2-52: Gapping in mm per load cycle...... 167

Figure 2-53: Ultimate load in Newton. Note the significant (p<0.05) higher load to failure in the barbed suture group.The conventinal knotted 3D repair could not improve final failure loads compared to the Adelaide repair...... 168

Figure 2-54 Example logarithmic regression calculation (black line). From the logarithmic regression for each group, the number of cycles to 2mm gap was interpolated...... 169 XXX

Figure 2-55: Average amount of loading cycles (3N to 30N) to cause a 2mm gap...... 170

Figure 3-1 Chicken tendon model ...... 181

Figure 3-2 Dissected rabbit rear paw...... 182

Figure 3-3 Dissected rabbit leg with flexor tendon and Achilles tendon marked...... 183

Figure 3-4 Achilles tendon with gastrocnemic and soleus tendon...... 184

Figure 3-5 Dissection of sheep trotter (left) and pig trotter (right)...... 185

Figure 3-6 Male Turkey (Meleagris Gallopavo) ...... 187

Figure 3-7 Male adult turkey leg ...... 188

Figure 3-8 Middle toe of an adult male turkey. Note similar size to human finger...... 188

Figure 3-9 Dissected adult turkey middle toe with paired neurovascular ...... 189

Figure 3-10 Dissected male turkey middle toe with FDS tendon, ...... 190

Figure 3-11 Dissected male turkey middle toe, short vinculum (SV), long ...... 191

Figure 3-12 AP and Lateral radiographs of male turkey middle toe, clamp indicates .. 192

Figure 3-13 Dissected PIP joint with collateral ligament ...... 193

Figure 3-14 Turkey PIP joint AP radiograph. Note lower bone density than in human phalanges...... 194

Figure 3-15 Turkey PIP joint in lateral projection. Again, note lower bone density than in human phalanges...... 194

Figure 3-16 Volar (plantar) access to tendon through Brunner incisions ...... 195

Figure 3-17 Transected deep flexor tendon in zone II ...... 196 XXXI

Figure 3-18 Unknotted 3D four-strand core repair with 4/0 ...... 196

Figure 3-19 Finished with a 6/0 Prolene simple running suture. Stable repair of deep flexor tendon , "no" bulk...... 197

Figure 3-20 Cross Locked Cruciate (Adelaide) tendon repair technique...... 199

Figure 3-21: Left: Harvested Turkey tendon. Right: Repaired Turkey tendon

...... 200

Figure 3-22 Dissection of a turkey gastrocnemius tendon with calcified area marked.208

Figure 3-23 H&E cross sections of sheep deep flexor tendon (left) and turkey deep flexor tendon (right) (2X magnified)...... 209

Figure 3-24 Comparison of human, sheep, pig and turkey tendon histology (H&E cross sections 4X magnified) ...... 210

Figure 3-25 Inhalation mask with oxygen and 2% Isoflurane for general anesthesia .. 218

Figure 3-26 Bird in prone position with scrubbed extremity ...... 220

Figure 3-27 Sterile operating field ...... 221

Figure 3-28 Sterile operating set up with bird in prone position...... 222

Figure 3-29 Application of tourniquet for middle digit...... 223

Figure 3-30 Marking of 2X1cm flap (arrow) over PIP joint...... 224

Figure 3-31 Dissection, fixation and transection of deep flexor tendon ...... 225

Figure 3-32 Cross locked cruciate tendon repair technique (Adelaide technique) ...... 226

Figure 3-33 Insertion of the four strand Adelaide core repair ...... 226 XXXII

Figure 3-34 Finished core repair (Adelaide technique) ...... 227

Figure 3-35 Finished composite repair: Adelaide core repair + ...... 228

Figure 3-36 Skin closure with 3/0 Vicryl sutures ...... 229

Figure 3-37 Dressed digit after operation ...... 230

Figure 3-38 Padding of extremity and application of dorsal plaster splint ...... 230

Figure 3-39 Operated turkey in recovery after operation...... 231

Figure 3-40 Illustration of the four-strand cross locked cruciate repair technique ...... 239

Figure 3-41: Illustration of our four-strand 3D knotless barbed suture repair technique.

Tendon purchase indicated in mm. a) First strand runs in the horizontal plane. Note, it passes through the suture loop at the end of the suture thread (see 3). b) Second strand

(see 5) runs from horizontal into vertical plane (three dimensional repair). Third strand

(see 7) runs in the vertical plane. c) Fourth strand stays in the vertical plane and finishes with a triple zigzag. d) Suture is cut flush with the tendon...... 241

Figure 3-42 Example image of completed repair in group 1 using a conventional knotted

Adelaide repair technique tendon reconstruction. Core suture knot and peripheral suture knot buried between tendon ends...... 243

Figure 3-43 Example image of completed repair in group 2 using an unknotted 3D barbed suture repair technique tendon reconstruction. No core suture knot needed, barbed suture end cut flush with tendon. Peripheral suture knot buried between tendon ends...... 244 XXXIII

Figure 3-44 Application of dorsal splint with padding...... 245

Figure 3-45 Grouping of animals ...... 246

Figure 3-46 Clamp clamped on middle toe deep flexor tendon...... 249

Figure 3-47 Measurement of total range of motion...... 250

Figure 3-48 Adhesion classification criteria [144] ...... 251

Figure 3-49 Adhesion classification: 0 (no adhesion) ...... 252

Figure 3-50 Adhesion classification: 2 (mild adhesion) ...... 253

Figure 3-51 Adhesion classification: 4 (moderate adhesion) ...... 254

Figure 3-52 Adhesion classification: 6 (advanced stage adhesion) ...... 255

Figure 3-53 Dissected tendon mounted between hydraulic clamps and ...... 258

Figure 3-54 Tendon specimen prepared for histology ...... 260

Figure 3-55 Broken splint after removal...... 262

Figure 3-56 Operating time of core repair per group ...... 263

Figure 3-57 Number of non intact tendon repairs after 6 weeks ...... 264

Figure 3-58 Toe range of motion in % of contra lateral toe range of motion ...... 265

Figure 3-59 Adhesion formation classification for groups: ...... 266

Figure 3-60 Average bulk in mm for groups ...... 267

Figure 3-61 Average remaining gap in mm after 6 weeks ...... 268

Figure 3-62 Average final failure forces in Newton for each group ...... 269

XXXIV

Figure 3-63 Longitudinal cut through middle toe of turkey tendon at level of PIP joint

...... 270

Figure 3-64 Failed repair in the Adelaide group. Note, suture is intact, but anchoring cross locks got pulled out of tendon substance...... 272

Figure 3-65 Histology cut from the same specimen. Clearly visible the pull out path. . 273

Figure 3-66 Disrupted barbed suture repair. Note suture is intact, but pulled out of the tendon...... 274

Figure 3-67 Pull out path of barbed suture. Note disruption of collagen tissue due to barb interaction...... 275

Figure 3-68 H&E slide of healed tendon from Adelaide repair group. See bridging of gap with fibrous tissue (*), demonstrating healing of tendon. Also note epitenon in continuity and not separated from tendon surface (arrow)...... 277

Figure 3-69 H&E slide of same specimen of healed Adelaide repair in higher magnification...... 278

Figure 3-70 H&E slide of healed tendon in barbed suture repair group. See fibrous scar tissue (left) and genuine tendon tissue (right). Again see intact epitenon (top and bottom) in continuity on tendon surface...... 279

Figure 3-71 Stump of failed repair with cellular fibrous tissue encapsulating the collagen tendon tissue. Left native H&E, right polarised H&E...... 280

XXXV

Figure 3-72 Moderate tissue reaction around resorbable (glycolic-carbonate) barbed suture. Left, native H&E, right polarised H&E...... 280

Figure 3-73 Minimal tissue reaction around non resorbable (polyester) suture in

Adelaide group...... 281

Figure 3-74 Quill barbed suture device. Note the single angle cut of barbs...... 283

Figure 3-75 Quill suture device. Note, double armed and barb orientation changes ... 283

Figure 3-76 V-Loc barbed suture device. Note the double angle cut of barbs...... 284

Figure 3-77 V-Loc barbed suture device. Note, the loop at the end of the suture...... 284

Figure 3-78 SEM pictures of V-Loc 180 suture after incubation in saline for 3 weeks. 287

Figure 3-79 Adelaide tendon repair (left) and barbed suture repair (right)...... 291

XXXVI

LIST OF TABLES

Table 1 Classification for macroscopic adhesion formation by Tang et al. [144] ...... 52

Table 2 Significances (p values) for forces needed to split tendons...... 75

Table 3 Overview of published reports on barbed suture tendon repair studies...... 173

Table 4 Overview of published tendon dimensions...... 205

Table 5 Overview of amount of failed repairs in histology group after six weeks...... 271

Table 6 Histological features of repairs after six weeks...... 276

XXXVII

NOMENCLATURE

ACEC Animal Care and Ethics Committee

ADL Adelaide Repair

ANOVA Analysis of Variances

AP Anterior Posterior

BRB Barbed Suture Repair

BRC Biological Research Centre

DIP Distal Inter Phalangeal

ECM Extra Cellular Matrix

FDP Flexor Digitorum Profundus

FDS Flexor Digitorum Superficialis

H&E Haematoxylin and Eosin

Hz Hertz

MCP Meta Carpo Phalangeal

N Newton

PIP Proximal Inter Phalangeal

POW Prince Of Wales

Psi Pound per square inch XXXVIII

p-value probability value

ROM Range Of Motion

SEM Scanning Electron Microscopy

SORL Surgical and Orthopeadic Research Laboratories

SPSS Statistical Program for Social Sciences

T-test statistical hypothesis test

UNSW University of New South Wales

UTS Ultimate Tensile Strength

XXXIX

CHAPTER 1: Introduction

1.1 Thesis Outline

This thesis is structured into four main chapters.

The first chapter provides an introduction into the science of tendons in the human body.

Based on a literature review, the function and anatomy of tendons in the hand are explained (1.2.1), followed by a discussion of past and current tendon repair techniques

(1.2.2) and biomechanical testing methods (1.2.3) including healing and rehabilitation issues (1.2.4), as well as an overview of clinical complications (1.2.5).

Following the introduction, the thesis overall rationale (1.3.1) and the aims for each study (1.3.2) are presented.

Thereafter the two main experimental chapters, "Ex Vivo Studies" (chapter 2) and "In

Vivo Studies" (chapter 3) are presented.

Chapter 2, (Ex vivo studies) starts with an analysis of animal tendon models for use in laboratory experiments investigating the question whether sheep or pig tendons are better reflecting human tendon biomechanics (2.1). Thereafter an experiment investigating current tendon repair techniques in regards to their geometrical stability is presented (2.2). Then two experiments investigating the authors recommendations for improvements of the current gold standard in flexor tendon repairs, the Adelaide repair, are depicted (2.3). The ex vivo chapter is finished with the introduction and biomechanical examination of the author's own new repair method, the unknotted 3D barbed suture tendon repair, distributing loads along the whole repair instead of relying on knotted concepts and therefore improving repair stability (2.4).

Chapter 3: Since no other available in vivo tendon model provides big enough tendons to test this new repair method, chapter three starts with the introduction of a completely new animal model for in vivo tendon experiments, the turkey tendon model. This new model is presented and validated by investigating the turkey's deep flexor tendon anatomy and biomechanics (3.1). Thereafter the turkeys husbandry and methods for in vivo surgery on the turkey foot are presented (3.2). This is a novelty because no other group used this model for surgical research previously. The chapter is finished with the report of the investigation of the author’s unknotted 3D tendon repair technique from chapter two, but this time tested in the vivo turkey model (3.3).

Introductions, methods, results and discussions including limitations for each study are presented within the individual chapters.

The final fourth chapter presents major conclusions drawn from the experimental studies, discusses overall limitations and future directions.

1. Introduction and thesis rationale

2. Ex Vivo Studies

2.1 What is the ideal animal model for ex vivo tendon repair experiments.

2.2 Biomechanical and geometrical investigations of current flexor tendon techniques: Changes in Core Suture Geometry within Repaired Flexor Tendons under tension.

2.3 The current Gold Standard in Flexor Tendon repairs, the “Adelaide Repair” - Author's recommendations for improvements and modifications.

2.4 A new concept for of flexor tendon repairs: The Knotless 3D Barbed Suture Flexor Tendon Repair.

3. In Vivo Studies

3.1 A new tendon model for ex vivo and in vivo experiments: The turkey foot.

3.2 Surgery on the turkey foot.

3.3 The knotless 3D barbed suture tendon repair in the turkey: An in vivo comparison.

4. Conclusions, limitations and future directions

4 Figure 1-1 Thesis structure

1.2 Background and Literature Review

This chapter provides the general background to the object of investigation, deep flexor tendons and deep flexor tendon repairs in the human hand. It comprises of five sections.

First, an explanation of the functions of tendons in the human body, including a description of the tendon anatomy and physiology in the hand; second, an overview of current methods of repairing a tendon in the hand, including a history of tendon repair practices and common mechanisms of tendon lacerations; third a short summary of biomechanical testing procedures, fourth a report on flexor tendon healing and rehabilitation; and fifth, a discussion of complications of tendon repair techniques.

1.2.1 Tendon Anatomy and Physiology

1.2.1.1 The role of the tendon in the human body

A tendon is a strong band of fibrous connective tissue connecting muscle to bone [2].

The primary function of a tendon is to transfer the force produced by muscle contraction to bones in order to stabilize and move body parts. Tendons exhibit viscous and elastic behaviour to varying degree, depending on their functional requirements. This leads to the differentiation between positional and energy-storing tendons [3]. More elastic energy-storing tendons such as the Achilles tendon in the human leg store and return kinetic energy. During the human stride, the Achilles tendon stores energy while stretching and releases the stored energy when the foot turns downwards, resembling a 5

spring. This makes the stride more efficient as the tendon reduces the muscles' work load. Not all tendons require this spring-like function. Compared to the Achilles tendon, positional tendons such as the deep flexor tendons in the hand do not need to store and release energy but mainly assist in precise limb placement [4]. In order to provide the fine control of joint movements, positional tendons are less elastic but stiffer. In this context, studies showed that the Achilles tendon proportional stretches up to 10.3%, while recorded maximum proportional stretch of positional tendons in the hand

(depending on the anatomical location) only 3.1% [5, 6].

1.2.1.2 Hand Anatomy

The hand is the most complex extremity of the human body [7]. Humans depend largely on the sophisticated functionality of the hand in almost every act of daily living. The cortical homunculus - a visual representation of “the body within the brain” (Figure 1-2)

- indicates the importance of the hand for the human.

Figure 1-2 The motor (left) and sensory (right) homunculus [8].

As Figure 1-2 illustrates, the relative amount of cerebral cortex surface area dedicated to the sensory inputs (left figure) and motor outputs (right figure) of the hand is disproportionately large. The area takes roughly one quarter of the motor cortex.

Figure 1-3 depicts another often found illustration of the homunculus that shows the hypothetical shape of the human body according to the amount of space the brains cortex assigns to each body part.

Figure 1-3 Homunculus [9]

The hand can execute movements with high precision but at the same time great strength. For example, grabbing a rope or the handle of a hammer requires great power.

On the other side, grabbing and moving a pencil to write or paint demands great precision. To achieve this combination of power and precision, the hand consists of a series of complex, delicately balanced bones, joints, muscles and tendons.

In total, the human hand contains 27 bones, not including ulna and radius (also not included the sesamoid bones) (Figure 1-4) [10].

A. Thumb B. Index C. Middle finger D. Ring finger E. Little finger I-V. Metacarpal bones 1,4. Distal phalanx 2. Middle phalanx 3,5. Proximal phalanx 6. Sesamoid bones 7. Distal interphalangeal joint (DIP) 8. Proximal interphalangeal joint (PIP) 9. Metacarpophalangeal joint (V.) 10. Carpometacarpal joints 11. Trapezium 12. Trapezoid 13. Capitate 14. Hamate 15. Scaphoid 16. Lunate 17. Triquetrum 18. Pisiform 19. Radius 20. Ulna

Figure 1-4 Radiographic anatomy of adult hand [11].

As can be seen in Figure 1-4, each finger consists of three phalanges, the proximal (3), middle (2) and distal phalanx (1), except the thumb which only has two, the proximal and distal phalanx. The body of the hand contains five metacarpal bones (I-V) that run from the carpal bones of the wrist to the base of each digit of the hand. Each of the phalange has a base, a body and a head. The joints between the phalanges are called

Proximal Inter Phalangeal joints (PIP joints) and Distal Inter Phalangeal joints (DIP joints). The thumb has only one joint, the Inter Phalangeal joint (IP joint). The wrist contains eight small bones (11-18) that form the bridge between the hand and the two bones of the forearm, the radius and the ulna [12].

The movement of this complex bone structure requires a complex nexus of muscles, tendons and bands innervated by a delicate nervous system. Due to limited space in the hand, particularly in the fingers, the muscles are distributed over the palm (intrinsic muscle-tendon system) and the forearm (extrinsic muscle-tendon system). The intrinsic muscles predominately accomplish fine coordination of precise finger and hand movements. They are divided into thenar (concerning the palm on the thumb side) and hypothenar (concerning the palm on the little fingers side), as well as the four lumbrical muscles and the seven interossei muscles between the five rays of the hand (Figure 1-5)

[12].

Figure 1-5 The muscles of the hand [10]. Left: dorsal view of left hand with the four dorsal interossei muscles (the volar three interossei muscles are conceiled in this view). Right: palmar view of left hand with (A) hypothenar muscles (including musculus abductor digiti minimi, flexor digiti minimi brevis and opponens digiti minimi), (B) thenar muscles (including musculus abductor pollicis brevis, flexor pollicis brevis and opponens pollicis) and the lumbrical muscles marked I,II,III, IV.

The flexion and extension of the fingers though are mainly executed by the extrinsic muscles of the forearm transmitted to the fingers via long tendons. The extrinsic extensor muscles arise from the lateral epicondyle and roughly provide extension and abduction to the fingers. The extrinsic flexor muscles arise from the medial epicondyle and roughly provide flexion and adduction to the fingers. In total, nine extrinsic flexor 11

tendons enter the hand through the carpal tunnel: The flexor palmaris longus (FPL) tendon to the thumb, and four flexors digitorum profundus (FDP) tendons and four flexor digitorum superficialis (FDS) tendons to the remaining four long fingers. The four FDS tendons lie superficial to the FDP tendons in the carpal tunnel [10]. At the level of the Meta Carpal Phalangeal (MCP joint), they pass through the FDP (also known as the Chiasm of Camper) to insert into the volar base of the middle phalanx.

The four FDP tendons lie deep in the carpal tunnel, traversing the palm and entering the volar base of the distal phalanges of the four long fingers [13].

1.2.1.3 Pulley System

In the distal hand the flexor tendons run through a sheath which is composed of a pulley system and a membranous tissue component. The membranous component forms a synovial tunnel providing a gliding surface between the tendon and its adjoining environment. It consists of two layers, a visceral layer investing the tendon and a parietal layer lining the tunnel [14].

The pulley system comprises of eight annular fibrous tissue condensations along the area between the distal palm and the DIP joint that form a fibro-osseous tube through which passes the FDP and FDS tendons (Figure 1-6) [15].

Figure 1-6 Flexor tendon synovial sheath and pulleys including the palmar aponeurosis pulley, five annular pulleys and three cruciform pulleys [16]. Each digit (except the thumb) entails five annular pulleys, also referred to as major pulleys because they accomplish the main biomechanical pulley function. Additionally, there are three smaller cruciform pulleys, referred to as minor pulleys that offer additional support and stability to the tendon sheath. Three of the five annular pulleys (A1, A3, A5) are located over the joints, while two broader pulleys (A2, A4) are located between the joints. The three thinner cruciate pulleys (C0, C1, C3) are located nearby to their respective joints [14]. The pulley system provides the framework for the high level of control to move the fingers with great precision and power. It prevents the tendons

from moving away from the joints during flexion (bowstringing), which enhances

mechanical efficiency (

Figure 1-7) [17].

Figure 1-7 Mechanics of the pulley system [17].

1.2.1.4 Flexor Tendon Blood Supply

In general tendons have limited blood supply as their oxygen consumption is 7.5 times lower compared to that of skeletal muscles [18]. While this low metabolic rate and well- developed anaerobic energy-producing ability are important to carry loads and maintain tension for long periods, on the downside, it results in slow healing after injuries [19]. 14

The tendons in the hand receive blood supply from vascular and synovial sources [20,

21]. Vascular blood supply to the flexor tendons in the hand is provided by longitudinal vessels that travel parallel to each ray into the fingers. Two digital vessels, one ulnar and one radial, run along each finger. Each tendon gets vascular perfusion from segmental branches from these digital arteries.

These branches enter the tendon sheath through so called long and short vincula, which are folds of mesotenon. For each FDS and FDP tendon, there is one short and one long vinculum. The vinculum brevis superficialis (VBS) and the vinculum brevis profundus

(VBP) are located near their respective osseus insertions. The long vinculum for the

FDS (VLS) emerges from the floor of the digital sheath of the proximal phanlanx and the long vinculum for the FDP (VLP) arises from the superficialis tendon at the level of the proximal interphelangeal joint (Figure 1-8) [22].

Figure 1-8 Vascular blood supply of FDS and FDP tendons [23]

There are areas of potential poor blood supply particularly at the volar side of the flexor tendons in zones I and II compared to richer blood supply on the dorsal sides that directly connect with the vincula [24].

Important additional nutrition but also lubrication and waste transport is provided by the synovial fluid within the flexor tendon sheath. Nutrition delivery is accomplished through diffusion in combination with a pumping mechanism (imbibition) during flexion and extension of the tendon. The fluid is pumped into the interstices through small channels in the tendon surface. There is debate about the exact contribution of each type of nutrition [e.g., 25, 26, 27] but the general importance of the synovial fluid for tendon nutrition and healing is commonly acknowledged. 16

1.2.1.5 Zones

Verdan et al. divided the forearm and the hand into five zones according to the difficulties faced during surgeries in these zones [28].

Figure 1-9 Flexor tendon zones [29].

Zone I reaches from the fingertip to the insertion of the FDS. It includes the C3, A4 and

A5 pulleys. Injuries in this zone only affect the FDP tendon, which make surgical results more predictable and adhesion formations are rare.

The focus of this present research is on zone II, historically referred to as “no man’s land” [30]. The term originated from the belief that no surgeon should attempt repairs in this zone due to its complex anatomy and problematic surgical results. Nevertheless it is a very common area for injuries due to its importance in the grabbing or catching of objects. As described earlier, within this zone the FDP and FDS change their relative position at the Camper’s chiasm under the A2 pulley. One of the main problems is adhesion formations between the bone, the tendons and the sheath, but also the limited surgical access to the tendons and the small space in general [31].

Zone III extracts over the palm, from the proximal edge of the A1 pulley to the distal edge of the carpal tunnel. Surgical procedures in this zone are less problematic because the flexor tendons are more accessible due to absence of the fibro-osseous tunnel [32].

Zone IV stretches over the carpal tunnel. Regarding repair difficulties, tendon injuries in this area also have tendencies for adhesion formation due to the compressed space conditions in the tunnel and due to the relative avascularity of tendons in this area [29].

Zone V is described as the area proximal to the carpal tunnel. Here, as well as in zone

IV one problem occurs due to the fact that injuries in this zone often lacerate multiple structures which require wide exposure for surgical procedures [33].

1.2.1.6 Tendon biology and composition

Tendons are dense longitudinal structures with a complex microarchitecture. They consist of a hierarchical arranged composition of unique tissues. Main components are collagen molecules, fibrils, fibers, fiber bundles, fascicles and tendon units. All of these components are aligned in a parallel fashion along the tendon's longitudinal axis [34,

35]. This composition makes the tendon stiff and at the same time flexible and smooth and provides the tendon with adequate strength [36].

Figure 1-10 Representative schematic of a multi-unit hierarchical structure of the tendon [36].

At the top of the tendon hierarchy is the epitenon. It is carrying blood vessels, nerves and lymphatics in its dense network of collagen strands and surrounds the entire tendon structure [37]. The inner surface of the epitenon is adjacent to the endotenon, similar to the epitenon in that it also can carries blood vessels, nerves and lymphatics but consisting of a finer, loose reticular network sheath of connective tissue and reaching deeper into the tendon [28].

Within the epitenon are primary, secondary and tertiary bundles, all of them surrounded by the endotenon. Primary and secondary fiber bundles are formed of collagen fibers and are known as fascicles [38]. Together, a group of fascicles makes up a tertiary bundle. A group of tertiary bundles forms a tendon with its outer layer surrounded by the epitenon. While a fascicle is a bundle of fibers [38], the terms ‘primary fiber bundle’, ‘secondary fiber bundle’ and ‘tertiary fiber bundle’ and ‘fascicle’ co-exist with some level of confusion [35]. The hierarchical nature of tendons and the terminology of their component characteristics is not entirely straightforward. Further down the tendon hierarchy is the collagen fiber, made of a bunch of collagen fibrils [38]. Laid end to end and longitudinally aligned, collagen fibers form the basic unit of a tendon and the collagen fibers are also the smallest component of a tendon that is visible by light microscopy [39].

At the lower end of the tendon hierarchy are the fibrils. Composed of collagen molecules, fibrils are embedded in the matrix of the tendon again in parallel alignment

[40]. With diameters varying from 10 to 100 nm, depending on sample location, age and species [36], fibrils are the smallest structural unit of the tendon.

At a cellular level, tendons are primarily composed of fibroblast cells. Loosely distributed along the collagen fibrils which together form collagen bundles (figure 1-11).

These fibroblasts synthesise the extracellular matrix (ECM). Also known as tenoblasts or tenocytes, the spindle-shaped fibroblasts make up the large majority, 90-95%, of a tendon’s overall cellular structure [36, 37, 41].

Figure 1-11 Fibroblast distribution within a collagen bundle. Collagen fibrils tightly surround the cells [42].

Chondrocytes, synovial cells and vascular cells make up the remaining cellular population of tendons. Chondrocytes are collected in the region where the tendon joins the bone, the entheses, while synovial cells are found on the tendon’s surface.

The arterioles smooth muscle cells, found within the dense network of the endotenon and epitenon and the capillary endothelial cells form the vascular cell population of a tendon. Together, chondrocytes, synovial cells and vascular cells make up 5-10% of a tendon’s overall cellular population [37].

The extracellular matrix (ECM), synthesised by fibroblasts, consists of proteoglycans, glycoproteins, water and collagens.

Proteoglycans exist in small quantities, but influence the tendon’s mechanical properties

[36]. Depending on the tensional or compressive forces on the tendon proteoglycan content varies within the tendon, for example under pulleys or at joints [43].

Alongside proteoglycans the ECM also consists of glycoproteins, including tenascin-C and fibronectin [36]. The increased synthesis of fibronectin has been associated with healing [44].

Collagen plays the most important role of the ECM and is responsible for the tendon’s strength. Part of the mechanical stability of collagen is related to the structure of its molecules. Collagen forms a right-handed superhelix in a coiled-coil structure that form

long fibrils [45]. The other part that stabilises collagen is its ability to form cross-links.

These cross-links significantly improve tensile strength of the tendon [38, 41, 46].

Three types of collagen exist in the ECM, known as Type I, Type III and Type V collagen. 95% of the collagen found in tendons is Type I collagen. It makes up 60% of the dry mass of the tendon. Type III and type V collagens account for the other 5% [36].

Type I collagen forms larger, more organised fibrils than Type III collagen. Type V collagen is responsible for regulating fibril growth and is interwoven into the core of

Type I collagen fibrils [47].

1.2.2 Flexor tendon repairs

1.2.2.1 History of tendon repairs

From the early age on Hippocrates and other ancient physicians described tendons as white cord-like structures but commonly misinterpreted them as nerves or did not recognise tendons as a distinct structure at all. Galen, who lived in the 2nd Century AD, was the first physician to understand the anatomical differences between nerves and tendons and described the different function of nerves, muscles and tendons [48].

Nevertheless, he advised surgeons not to repair this structure.

In the 11th century the Persian polymath Avicenna was probably the first physician who described tendon repairs and his teachings were subsequently adopted and modified by

European surgeons from the 14th to the 16th century [49].

In 1767, John Hunter was the first scientist who investigated tendon healing [50].

However, most investigations on tendons before the 19the century were made on

Achilles tendons and only in the early 20th century did scientists report on investigations on the more complex deep flexor tendons, with their distinctly different anatomic environment in the synovial lined flexor digital sheath [51].

Around 1920, Bunnel et al. started to define the concept of adhesion formations in the digital sheath after tendon repairs and subsequently founded the term "no man’s land"

[30] . He advocated that only under certain conditions repairs of flexor tendons in the digital sheath would produce satisfactory results. One of them was postoperative immobilisation. These conditions were further defined by Mason et al. [52] but common opinion in the first half of the 20th century was to not repair deep flexor tendons in the

“no man’s land” due to the bad results caused by high rates of infections and repair failures. Instead hand surgeons promoted the concept of secondary tendon grafting for reconstruction.

In 1967 Kleinert presented the breakthrough paper "Primary Repair of Flexor Tendons in the “no man’s land" [53]. This marked a turning point. Kleinert's work and further investigations of the deep flexor tendon biology by various investigators [54-58] helped to understand tendon vascularisation, nutrition and intrinsic healing mechanisms. The previously promoted concept that adhesion formations were necessary for tendon repair nutrition and healing went overboard. Modern concepts for minimising adhesion formations, early passive and active mobilisation of repaired tendons and the resulting need for more stable multi strand repair techniques to ensure sufficient repair strength to mobilise the tendon after surgery was established [59-66].

1.2.2.2 Mechanism of injuries

In general two types of tendon injuries have to be distinguished: Open and closed injuries. Common closed flexor tendon injuries are sports injuries like the “Jersey finger” (deep flexor tendon tear at, or close to, insertion of tendon into distal phalanx)

[67].

Less frequently closed flexor tendon disruptions are caused by chronic illnesses like rheumatoid arthritis [69] or also after other operations like an open reduction and internal fixation of a distal radius fracture with a volar plate [70].

This thesis focuses on open tendon lacerations because more frequently flexor tendons get lacerated or injured in trauma cases, often through the use of knifes, blades, saws or other machinery. These mechanisms of injury especially apply to work related injuries.

In a statistical report by “Safe Work Australia” from the year 2008 it was discovered that trauma to the hand and forearm account for about 8400 hospital admissions per year in Australia [71]. Not all of these injuries affect tendons, but with the close proximity of the tendons to the skin of the forearm, wrist, palm and fingers, tendon injuries are common in hand trauma cases [72-74].

Figure 1-13 Typical multi-level hand injury with tendon lacerations (extensor + flexor tendons of the thumb) after work trauma with lawn mower. Note: the multiple lacerations with different penetration depth in different hand zones and areas make it difficult to examine the exact extent of the injury.

Figure 1-14 Typical flexor tendon injury after work trauma with knife. Note: At index finger both tendons, superficial flexor and deep flexor tendon (FDS and FDP) are cut. Compared to middle finger with only the deep flexor tendon (FDP) cut. Hence the different postures of the two fingers.

In the UK alone more than 3200 patients present with open flexor tendon injuries at hospital emergency departments per year [75].

Of all flexor tendon injuries the laceration of a flexor tendon in zone II is not only the most common one but also the most complicated to treat [76]. A lot of times injuries to the deep flexor tendon happen in the act of grabbing a sharp object. If this happens the area of injury is often located in the dreaded zone II.

Not only that sharp injuries in this area often lacerate flexor tendons, but also the distal tendon end of the lacerated deep flexor tendon can retract into the distal finger. If this is the case, reconstruction is further complicated [77].

This area of injury was previously called “no man’s land” and up until the early 70ies the treatment of tendons in this area was considered too complicated [30]. The close proximity of the superficial and deep flexor tendons in the limited space of synovial sheath and pulleys in this area made it seem impossible to perform repairs without causing severe adhesion formations [78].

Figure 1-16 Zone II or no man's land with it’s complicated anatomy of deep flexor tendon, superficial flexor tendon, sheath and pulleys [29]

Only with modern multi strand repairs and early rehabilitation regimes the restoration of hand function after zone II tendon repairs became successful [79, 80].

The high incidence of flexor tendon injuries in this area, as well as the complicated treatment is the reason why most research and also this thesis focuses on lacerations of deep flexor tendons in zone II.

1.2.2.3 Repair techniques

Several factors influence the outcome of treatment of an injured flexor tendon. These include the repair technique, the management of the flexor sheath and pulleys, the rehabilitation regime and biological factors influencing tendon healing [65]. It is recommended to repair both, the superficial and the deep flexor tendon in zone II [81] but regarding the functional outcome for the individual finger an intact flexor digitorum profundus tendon is more important than an intact flexor digitorum superficialis tendon since the full range of motion of a finger is only facilitated through the FDP tendon. In certain cases of tendon repairs in zone II it is even indicated to resect the superficial tendon, or respectively one slip, to be able to achieve a better repair to the deep flexor tendon and to create minimal gliding resistance [82, 83]. This PhD thesis focuses primarily on repair techniques and research methods related to core repairs of deep flexor tendons in zone II.

An overview of important challenges in repairing tendons in zone II is essential for understanding developments in repair methods. The main issues to consider are gapping of repairs (distraction of tendon ends when repair is tensioned), gliding resistance, adhesion formation and finally repair failure (either by gapping more than 2mm or total disrupture of repair). It will be stressed that these factors are all interrelated.

The ideal flexor tendon repair should be easy to reproduce and add as little bulk as possible. It should possess a high tensile strength to safely allow early mobilisation, and be resistant to gap formation to facilitate healing with minimal scarring [22]. Various repair techniques have been described which differ in their ability to meet these aims.

Repair of a completely lacerated tendon involves a core suture, which is predominantly responsible for repair strength, and a peripheral suture, which is additionally important for gap resistance.

1.2.2.3.1 The core suture

Core repair strength relates to several factors including suture material [84-87] and size

[88], suture configuration [63, 89-91], and the number of suture strands crossing the repair site [92]. Yet the complexity of achieving a six or even eight-strand repair has limited practical application [60, 93]. Potential pitfalls with using more than four-strands include the technical ability of the surgeon, an increase in surgical time, tissue handling, repair bulk, and injury to tendon vascularity [22, 94].

The number of strands can be increased by using a multistrand technique (see Figure

1-17 and Figure 1-18), using a double stranded suture (Figure 1-19), or performing two or more two-strand repairs at once (Figure 1-20). This latter approach however usually means more knots and inevitably more repair bulk.

Figure 1-17 Example of a two-strand technique.

Figure 1-18 Example of a true four-strand technique.

Figure 1-19 Two-strand Kessler type repair performed with a single stranded suture (above) and with a double stranded suture (below) to create a four-strand repair [85].

Figure 1-20 : Double and Triple Kessler repair configuration [64].

Studies by Barrie et al. [60, 93] showed that a four-strand repair has greater resistance for gapping than a two-strand technique, but there was no difference in gapping when a four-strand technique was compared to an eight-strand repair. Nevertheless, this concept is likely to be technique specific and not solely dependant on the number of strands. But in general four-strand repairs are considered as a good compromise of repair stability and complexity.

When performing the repair, some tension in addition to that required to oppose tendon ends should be the aim, ideally with equal tension across all suture strands [95]. This may become increasingly difficult to achieve with increased number of strands and more complex repair configurations [96].

Increasing suture calibre increases repair strength [97-99]. However, larger suture calibre do not significantly influence gap formation [100]. 3-0 and 4-0 are the most commonly used sizes. 2-0 and larger sutures have been shown to significantly increase gliding resistance [99].

Suture material is also shown to affect the biomechanical properties of multi-strand repairs [87, 101]. Non-absorbable braided sutures are considered most practical.

Although absorbable sutures are appealing, these generally lack adequate strength or often lose their strength too early and promote more tissue reaction increasing adhesion formation [87].

Suture pullout can be minimised by optimising the holding configuration into the tendon and by taking sufficient purchase of the tendon. A core suture loop that “locks” into the tendon, hence sandwiching intervening fibres under tension, generally has better biomechanical performance over a configuration that “grasps” the tendon [62, 98].

Despite the terms "locking" and "grasping" being not very well defined, cross locks like those used in Adelaide repair have been shown to perform better than commonly used loop locks as in Kessler repair [89, 102].

There is no difference in gliding resistance between similarly designed locking and grasping techniques [103, 104].

The modified Kessler technique is the most common traditional repair to which other methods are compared. Modifications of the original repair give a good illustration of how developments in understanding have been applied to improve the strength of a repair. Chapter 2.2.1 gives an in-depth analysis of the Kessler repair method. Modern multi-strand repairs follow the same evolutionary steps to reinforce the anchoring capabilities and to improve integral geometric stability. Chapter 2.2.2 gives a more detailed analysis of current repair methods with four-strands.

The length of core suture purchase influences both, the forces required to gap the repair and ultimately break the repair [105]. The optimal purchase length is 7-10 mm; increases beyond this do not result in increased repair strength [106]. Further, the size of the locking loops has an influence on the repair stability [107, 108]. This is discussed in detail in chapter 1.3.1.

There is a decrease in strength with more than one knot [109]. The position of the knots is important as well [110]. Repairs with the knot on the outside have greater ultimate strength compared to a knot within the repair site. But external exposure of knots increases gliding resistance and can abrade the undersurface of a pulley with repetitive movement [111]. Hence it is recommended to place the knot between the tendon ends.

Despite concerns that a knot at the repair site may influence healing, studies show no difference in strength after six weeks [110].

1.2.2.3.2 The peripheral suture

The peripheral suture is also referred to as the epitenon suture, epitendinous suture, and the circumferential suture. It was once considered a “tidy up” stitch for its effect on smoothing the repair site. However, its second important role is for resisting gap formation [22, 112, 113].

In biomechanical testing, significant gapping often occurs only once the peripheral suture has given way. The details of the optimal peripheral suture remain controversial.

The simple running technique was initially introduced Kleinert et al. [114]. However, the impact of the peripheral suture was not appreciated until Silfverskiold and

Andersson, described their “cross-stitch” technique, and showed superior repair stability with this technique [115]. It was perhaps this study that established the routine use of a core and peripheral suture repair.

Various peripheral suture techniques have been described. Continuous sutures are preferred because they minimise knots, bulk and gliding resistance. The most common suture material used is a 6-0 polypropylene [111]. Kubota et al., 1996, compared the biomechanical characteristics of multiple different continuous peripheral suture techniques in human cadaver flexor digitorum profundus tendons: Simple-locking,

Simple, Lembert, Halsted, Cross-stitch and Lin-locking [116]. Increasing the number of peripheral suture strands was shown to increase ultimate strength, but gap formation was not an endpoint in this study.

Including tendon fibres in the suture pass of a circumferential suture adds significant strength over a suture pass picking up only the epitendon layer [117, 118]. Furthermore, increasing suture purchase also improves performance [119].

Also techniques with a transverse component to the suture are more resistant to pullout

[120]. Silfverskiold and Andersson, 1993 used an interlocking modification of the cross- stitch, the so-called Type A construct, which was hypothesised to further increase hold on the tendon under greater loads [115]. Dona et al., 2003 applied this principle to introduce a new technique, the interlocking horizontal mattress suture, and demonstrated its superiority to the cross-stitch configuration (Figure 1-21) [1]. Nevertheless, the conventional simple running sutures with buried knot remain most prevalent in clinical practice because of their relative simplicity and minimal gliding resistance.

Figure 1-21 Common peripheral suture techniques: (A) simple running; (B) cross-stitch; (C) Interlocking cross-stitch; and (D) interlocking horizontal mattress [1]

Lotz et al. showed the importance of equal load sharing between the core and peripheral suture to produce the maximum tensile strength [121]. If more load is carried by either suture, then this will fail first and then all the load will be transmitted to the remaining suture.

1.2.3 Biomechanical testing

The ideal approach to testing tendon reconstructions remains controversial. The variability in tendon models, testing protocols and endpoints between studies have complicated comparisons between different tendon research groups.

Static pull-to-failure testing of repaired tendons remains the most common ex vivo method. This approach is relatively simple and the usual endpoints are load to failure, load to 2mm gap formation, and stiffness. While providing important comparative information, the clinical utility of this data is limited; it fails to appreciate the effect of dynamic stress at the repair site during rehabilitation. Cyclic testing is less frequently utilised. It involves a relatively greater level complexity and the testing parameters and endpoints are more variable between investigators. Pruitt et al. was the first to cyclically load repaired flexor tendons. The clinical utility of this approach for evaluating gap formation was noted, and has been reaffirmed by multiple investigators [60, 99, 122,

123]. This method may provide a more clinically realistic assessment of repair behaviour.

1.2.4 Flexor tendon healing and rehabilitation

1.2.4.1 Healing

There are two approaches to tendon healing: (A) The biological approach on a cellular basis and (B) the clinical approach on the basis of biomechanical repair stability after a tendon repair (see 1.2.4.2 Rehabilitation). Both are interrelated, but this thesis concentrates on the latter. Figure 1-22 shows the relationship of biological healing and biomechanical stability after a tendon repair.

Tendon healing on a cellular basis occurs in three overlapping phases. In the initial inflammatory phase, erythrocytes, macrophages, monocytes and inflammatory cells enter the site of injury. Vasodilative and chemotactic factors are released angiogenisis and tenocyte proliferation are initiated and collagen type III synthesis is started. This phase only lasts for the first days, then the proliferative phase begins. During this phase, lasting about four to six weeks, the collagen type III synthesis reaches a maximum and glycosamin concentrations are high [124]. This forms the repair scaffolding for neovascularisation and stabilises the repair [79]. The last stage of tendon healing starts after four to six weeks and is known as the remodeling phase. Cellularity decreases, a change from type III to type I collagen synthesis is noticeable and collagen fibers are aligned [125]. In the following months and continuing for up to a year, the fibrous tissue further maturates to tendon-like tissue.

Despite the remodelling of the tendon tissue, the biochemical stability of the healed tendon tissue never reaches those of intact tendons [126]. Rupture forces as low as 57% of the rupture force of intact tendons are reported for repaired Achilles tendons after 12 months [127].

Figure 1-22 Illustration of biological and biomechanical healing of repaired tendon.

1.2.4.2 Rehabilitation

Immobilisation of repaired flexor tendons is associated with adhesion formation, joint stiffness, and poor functional outcomes. Early mobilisation commences within the first postoperative week and undisputedly yields superior outcomes. Multiple different protocols are used by different hand units, but none have convincingly demonstrated to be superior [111]. Recent evidence supports the use of regimes emphasising greater tendon excursion without increasing force at the repair site.

Gliding resistance of a repaired tendon decreases over the first five days as oedema settles. If the digit is not mobilised at this stage, early adhesions and joint stiffness begin to set in. It is generally held that early mobilisation should thus commence 5 days after surgery or earlier [95, 111]. In fact, there is a trend for mobilisation to now commence as soon as practicable, often on postoperative day 1 or even start first cycles on the day of operation.

The strength of a repaired flexor tendon progressively decreases during the first 2-3 weeks. After this initial phase there is an increase in repair strength [128].

Early mobilisation keeps the loss of strength in the first 3 weeks to a minimum but still repair strength decreases [129]. Early mobilisation also increases resulting total active range of motion, repair site stiffness, and ultimate strength [130, 131]. Therefore movement is perhaps the single most important factor in preventing adhesions and adverse outcomes in tendon repairs. 45

The decreasing repair stability in the first three weeks on the one hand and the need for tendon excursion and controlled tension on the repair site on the other hand are challenging tasks to overcome for the patient, the therapist and the surgeon to achieve good overall outcomes. There is controversy regarding the optimal amount of force that should be applied to the tendon during rehabilitation. The bulk of the repair, associated oedema, and joint stiffness contribute to the work of flexion that must be overcome to move the tendon [132, 133]. Postoperative passive finger exercises might not result in adequate movement of the tendon and higher forces, such as those generated by unresisted active finger movements might be required [134]. A traditional two-strand repair cannot handle active rehabilitation; however, it is tolerated by a well-constructed four-strand technique as illustrated in figure 1-23 [79, 111]. A controlled early active mobilisation protocol is now the most commonly used post-op rehabilitation regime in large hand centres.

Figure 1-23 Relative tolerance of repairs to early mobilisation. Active rehabilitation should not be used for a tendon repaired with a two-strand technique [79].

Increased motion rather than force is supposed to be the key to achieving better final outcomes in regards to adhesion formations [135]. A study in canines demonstrated that only a small amount of force and tendon excursion are required to facilitate healing and prevent adhesions, without significant benefit with forces above 5N or excursion of

1.7mm [136]. However, gliding resistance is sometimes not overcome by passive manoeuvres, hence adhesions sometimes still occur [111]. This is one explanation for the better outcomes often observed clinically with early active rehabilitation.

Nevertheless, tendon excursion can be increased without increasing the tendon force by employing dynamic movements at the wrist [137, 138]. Such synergistic motion protocols utilise wrist flexion and finger extension, alternating with wrist and metacarpophalangeal joint extension and finger interphalangeal joint flexion (figure 1-

24). Graded rehabilitation programs of passive to active therapy have been described

[139], and this might be the most common approach worldwide. Usually a patient commences on a passive mobilisation protocol, and is progressed to an active regime

[111].

1.2.5 Complications

1.2.5.1 Gapping of repair constructs

Repair site elongation/separation, referred to as “gap formation” or “gapping” is due to insufficient repair stability and is associated with increased adhesion formation and gliding resistance during tendon excursion [128]. A large gap may cause a flexor lag, catching of reconstruction in pulleys and in the end the risk of early or delayed rupture.

The main factors effecting gap formation are the repair technique and rehabilitation protocol [79]. But while mobilisation may increase gapping, the superior functional outcomes of early mobilisation outweigh the harmful effects of immobilisation. A gap above 2mm significantly increases gliding resistance [140], and is a common endpoint for repair failure used throughout the literature. The catching of repairs to the edges of peritendinous structures such as pulleys produces a clinically notable interruption in movement, it is referred to as “triggering”. The pulleys and sheath edges in Zone II can serve as trigger points as shown in Figure 1-25.

Figure 1-25: Gapped tendon catching beneath pulley [111].

It should be emphasised again that more stable repair methods that are less prone to gapping combined with early active rehabilitation produce best outcomes.

1.2.5.2 Increased gliding resistance

All tendon repairs increase bulk and hence increase gliding resistance. The work of flexion is the sum of all the resistances to gliding. Increasing suture strands and exposure of knots will increase friction. The management of the sheath, oedema at the repair site and in surrounding tissues, joint stiffness, and the effect of fibrous adhesions also affect gliding resistance [141]. 50

Depending on the repair method and the rehabilitation protocol, the gliding resistance gradually increases over the first five days after surgery and begins to settle thereafter

[142].

1.2.5.3 Adhesion formation

Adhesion formation remains the most common complication and limits ultimate range of motion [143]. Two types of adhesions are distinguished. The first type is the loose and filmy adhesions which are of little clinical significance because they interfere little with tendon excursion and function. The second type is the thick and dense fibrous adhesions that adversely affect outcomes. The latter type occurs at a frequency of about

5% and often requires secondary salvage surgery [111].

Tang et al. more specifically classified adhesions in four categories (table 1) but followed the same concept by assessing the rigidity of the formed adhesion and its length [144].

Table 1 Classification for macroscopic adhesion formation by Tang et al. [144]

Novel measures for minimising adhesions are being investigated. These include the use of both autologous and synthetic barrier methods, as well as chemical and biological mediation of the inflammatory response [145].

While some of these methods may attest to be viable, none have yet made their way into widespread clinical use. The most effective method for reducing adhesions has been the use of early postoperative mobilisation regimes [146]. This was only made possible by developments of more stable multi-strand repair techniques in recent years.

1.2.5.4 Repair failure

Generally speaking, the greater the loads exerted on the tendon, the weaker the repair method, the higher the likelihood of gapping and rupture. The rupture rate was originally considered higher with an active mobilisation protocol compared to a passive protocol [147], but this view is challenged with modern multi-strand repair techniques and controlled active rehabilitation methods producing superior outcomes [111]. The mechanism of rupture is usually either by suture pullout (prolonged gapping leading to pull out of the repair anchoring configuration out of the tendon) or by suture breakage.

Additionally, as mentioned previously, authors consider a gap of repair site larger than

2mm as repair failure [60, 79, 128, 148]. Many efforts have gone into producing repair techniques that withstand higher loads. This has allowed early active mobilisation protocols with superior clinical outcomes.

In general focus has shifted towards producing repairs with greater gap resistance, which is considered a better measure of performance and stability of a repair construct

[111].

1.3 Thesis Rationale and Aims

1.3.1 Thesis Rationale

The restoration of hand function after tendon injuries is of utmost importance. The patient's ability to master everyday living tasks is strongly dependent on the utility of healthy hands: dressing, cleaning, working, writing, creating, exploring, loving, only to mention some essential ones.

Human hands are a series of complex, delicately balanced bones, joints, muscles and tendons and are one of our most exposed and least protected organs. Therefore they are prone to get harmed. Especially the laceration of a flexor tendon is a common injury. In the UK alone more than 3200 patients present with flexor tendon injuries at hospital emergency departments per year [75]. Most patients are young working people [71], therefore not only causing primary, personal treatment and rehabilitation costs, but also causing high indirect costs for the society in the form of lost productivity. Additionally the reconstruction of a lacerated flexor tendon is technically challenging and carries high risks of complications.

Flexor tendons in the hand are embedded in a subtle system of sheaths and pulleys and any disruption of the tendons smooth gliding in this sheath causes severe limitations of hand function. The management of tendon injuries has evolved with time, and extensive research into the topic has led to fundamental improvements. Repairs of flexor tendons in the hand became more manageable and reconstructions became stronger.

But still complication rates for deep flexor tendon repairs especially in the distal palm of the hand and in the proximal finger (zone II) remain high and final functional outcomes drag far behind other common musculo-skeletal surgeries [149]. This thesis investigates current concepts of tendon repairs, recommends new modifications and subsequently introduces a completely new method to repair tendons.

1.3.2 Thesis Aims and Hypothesis

The basic aim of this study was to design and investigate new, better methods for repairing deep flexor tendons in zone II, with the ultimate goal to translate these new tendon repair concepts into clinical practice and to improve tendon repair outcomes in the human hand.

Nevertheless, at the start a laboratory research model has to be chosen to conduct the investigations. Secondly, current tendon repair concepts have to be analysed to verify possibilities for improvements and to understand how to design a new superior repair method. Finally the new repair concept has to be designed and tested. Tested firstly in an ex vivo laboratory setting, and secondly in an in vivo animal experiment.

This present thesis evolved following the just described path.

Tendon research models

Various animal tendons and also human cadaveric tendons were previously used to perform laboratory tendon experiments. Analysing the big amount of published work in this area, shows that the pig and the sheep tendons are the most used substitute for tendon repair experiments. Still there exists no clear consensus which of these two animal tendon models is better to use in laboratory studies. Therefore the first experiment of this thesis (chapter 2.1) compares the biomechanical and histological properties of sheep, pig and human deep flexor tendons. The aim was to analyse which animal model better reflects human flexor tendon properties.

Hypothesis: The sheep tendon model reflects human tendon properties better than the pig tendon model.

Investigations of current tendon repair techniques

The second block of experiments investigates current tendon repair concepts (chapter

2.2). The main point of focus in this chapter was set on geometrical changes of core repair techniques when put under tension. This is a new approach to investigate the main weakness of tendon repairs, the gapping of repairs (separating of tendon ends) after tensioning.

Hypothesis: In comparison, the gold standard in tendon repairs, the Adelaide repair is the most stable four-strand repair in regards to gapping when put under tension.

Improvements to the Adelaide repair

Chapter 2.3 further investigates the Adelaide tendon repair, the current gold standard in tendon repairs. Chapter 2.3.1 details investigations into how the Adelaide repair is best performed to achieve superior repair results. And chapter 2.3.2 describes the authors modification to the Adelaide repair to additionally improve its repair stability and gapping resistance.

Hypothesis: The stability of the Adelaide repair can be further improved by performing the suture configuration with bigger cross locks (2.3.1) and by interlocking the distal suture components (2.3.2).

A new tendon repair method: The knotless barbed suture repair

In chapter 2.4 the author introduces a completely new repair method, the knotless 3D barbed suture flexor tendon repair technique. This knotless repair concept distributes loads throughout the repair instead of relying on knotted anchor points and therefore reduces gapping and maximizes repair stability.

Hypothesis: The unknotted 3D barbed suture repair can show better repair stability and gapping resistance than the Adelaide repair in an ex vivo laboratory setting.

A new tendon repair animal model

To prove that this repair method is also applicable in an in vivo setting, an appropriate in vivo animal model had to be found. Unfortunately there are not many in vivo animal models available and most of the available animal models (rat, chicken, rabbit) provide only very small tendons. Pig and sheep cannot be used for in vivo experiments because their deep flexor tendons are too deep in the hoof for surgical access. Therefore, at the start of the in vivo chapter a new animal model for tendon repair experiments providing tendon sizes comparable to human tendons, the turkey tendon model is introduced and validated (chapter 3.1). Following the introduction of the turkey model, chapter 3.2 analyses the husbandry of the turkey and describes in detail the surgery on the turkey’s foot.

Hypothesis: The turkey tendon model not only shows similar ex vivo tendon properties, but is also a practicable animal model for in vivo surgery experiments that reflect human deep flexor tendon anatomy, physiology and biomechanics.

The knotless barbed suture repair tested in an in vivo experiment

The final experimental chapter of this thesis is chapter 3.3. This chapter depicts a comparative investigation on the in vivo performance of the author’s new unknotted barbed suture repair method in the turkey foot.

Hypothesis: In comparison to the Adelaide repair, the unknotted barbed suture tendon repair method can show improved repair stability in an in vivo setting.

CHAPTER 2: Ex vivo studies

2.1 What is the ideal animal model for ex vivo tendon repair

experiments?

2.1.1 Introduction

The functional reconstruction of injured flexor tendons in zone II is technically demanding. Many surgical techniques and refinements have been described since

Kleinert et al. proposed primary repair in the early 1970s [114]. Clinical reports and peer reviewed investigations indicate superior outcomes and more consistent results using newer techniques [22, 66, 111], advances which have been due to both lab-based and clinical research [79]. The laboratory environment caters for controlled testing conditions and clinical studies complement this data. This provides the basis for the introduction of new techniques in modern clinical practices. However, at the outset a suitable and reliable laboratory model has to be defined.

Fresh human cadaveric tendons are theoretically ideal and have been used for ex vivo tendon experiments [89, 98, 99, 120, 122, 150]. But size and quality of the specimens varies with age, gender, health status, hand dominance and other factors, including the timing and preservation method of donor cadavers. Inconsistencies between human flexor tendons may therefore influence results. There are also logistic and ethical issues regarding the use of human tendons. They are relatively expensive and laboratories require appropriate licensing permits for the use of human tissues. Various animal models have therefore been used. Non-human primate cadaveric tendons are the most representative both anatomically and physiologically [151], but similar barriers, namely cost, availability and ethics preclude their widespread use. Porcine deep flexor tendons are the most commonly used substitutes [102, 106, 108, 152-157], perhaps followed by the sheep tendon model [115, 154, 158-161]. They are easy to obtain and have a comparable size to human flexor digitorum profundus tendons. Comparative studies examining pig, sheep or human tendons have not been well documented. This is surprising considering the volume of research and literature on deep flexor tendon repairs.

This ex vivo laboratory study directly compared porcine, ovine and human deep flexor tendons. To compare the biomechanical properties we assessed the tendons' resistance to cheese-wiring. Cheese-wiring is the act of splitting a tendon longitudinally, or in this case, dragging a suture loop through the tendon longitudinally.

Cheese-wiring of a suture through a tendon is an essential factor for the occurrence of tendon repair failure; both, by gapping and suture pull out. The null hypothesis was that there was no difference between the forces needed to cheese-wire a pig, sheep or human tendon. Additional representative histology was used to describe bio-structural differences in the investigated tendon models.

2.1.2 Methods

Deep flexor tendons were harvested from adult (two year old) pig fore limbs (n=24), adult (two year old) sheep fore limbs (n=24), and fresh frozen human cadaveric hands

(n=12). Figure 2-1, 2-2 and 2-3 depict the harvesting process of the deep flexor tendons of pig, sheep and human tendons.

Figure 2-1 Harvesting of pig deep flexor tendons.

Figure 2-2 Harvesting of sheep deep flexor tendons.

Figure 2-3 Harvesting of human deep flexor tendons.

The human tendons were attained after appropriate ethical approvals (Royal North

Shore Hospital, ScienceCare (USA)) from four fresh cadaveric hands. The mean age of donors was 68.5 years (range 60-71 years) and all were male. There was neither known history of nor any apparent musculoskeletal disease. The flexor digitorum profundus tendons from the index, middle and ring fingers were used.

Figure 2-4 Harvested human deep flexor tendons.

The animal deep flexor tendons were harvested from the two middle toes from each foot for the porcine and ovine samples.

Figure 2-5 Harvested pig and sheep tendons.

2.1.2.1 Standardising site in zone II

Zone II defines the segment between the A1 pulley and the flexor digitorum superficialis tendon insertion into the middle phalanx. Dissections of five pig and five sheep flexor tendon systems were performed to examine zone II in these species (figure

2-6 A&B). Radiographs were taken from the specimen using a MX 20 radiography system (Faxitron X-Ray, Lincolnshire, IL, USA). An energy setting of 20kV and exposure time of 20 seconds was used for all images. 67

Dissections and radiographs were compared to analyse the relationship of zone II to anatomical landmarks and the phalangeal bones (figure 2-6).

Results were compared to dissections of cadaveric human hands and a standardised point 5mm distal to the A1 pulley was chosen as testing site in all specimens. This point was marked with black ink during harvest with the digits held in extension.

The mean diameters (medial to lateral) of the pig tendons, sheep tendons and the human tendons at the testing site were measured with a digital caliper (CD-6_CS Absolute

Digimatic; Mitutoyo, Tokyo, Japan). After harvest, tendons were wrapped in phosphate buffered saline soaked gauze and deep frozen at -17 degrees Celsius until the time of testing. Pig and sheep tendons underwent a single freeze-thaw cycle, but human tendons were subjected to two freeze-thaw cycles prior to testing. This however is unlikely to significantly affect the tendon properties [153, 162].

2.1.2.2 Biomechanical testing

The mechanical properties of 24 tendons of each animal species and 12 human tendons were assessed by measuring the force required to cheese-wire a suture through the tendon. A 3/0 silicone coated polyester suture (Ticron; Tyco Healthcare. Norwalk, CT) on a round-bodied taper point needle was used to take a 2mm depth bite into the lateral aspect of each tendon at the marked point in zone II. The suture loop was knotted with 5 square throws over a 10mm diameter plastic rod to create uniform loops. The rod was removed and suture ears were cut at 8mm. All tendons were prepared by the same investigator (T.P.) with the aid of x3.2 loupe magnification. Tendon dehydration was avoided by wrapping tendons in phosphate buffered saline gauze while awaiting preparation and mechanical testing.

The force to pull the loop through the tendon (split the tendon) was tested with linear uniaxial distraction using a Mach-1 micromechanical testing machine (Biosyntech,

Quebec, Canada) with a 100 Newton (N) load cell. The proximal end of the tendon was secured superiorly in a screw-tightened clamp and the suture loop fed over a fixed steel hook inferiorly (Figure 2-7).

Figure 2-7: The tendon-loop construct fixed in the testing machine.

The starting gauge length between the clamp and hook was 20mm. The tendon-loop construct was preloaded to 1.5N then distracted at a rate of 10mm/min for 10mm.

Load, displacement and time were continuously recorded at 100Hz. Data were used to generate load-displacement graphs for each tendon. The loads at start of cheese-wiring,

2mm, 5mm, and 10mm displacement from initial cheese-wiring were recorded (Figure

2-8).

Figure 2-8 Example graph of force displacement curve generated during mechanical testing.

Data are presented as means (±standard deviation) and were analysed using Statistical

Package for the Social Sciences (SPSS) version 17.0 (SPSS Inc., Chicago, IL, USA).

Analysis of variances (ANOVA), post hoc Tukey test and Pearson correlation test was used. The significance level was set at p<0.05.

2.1.2.3 Histology

Five separate tendons from the two animal species and four little finger FDP tendons from the human cadaveric hands were harvested, marked at the same standard site in zone II and prepared for routine paraffin histology after formalin fixation for 72 hours using our standard laboratory protocol. Six longitudinal and six cross sections were taken from each specimen and embedded in paraffin. After trimming, 4µm sections were cut using a hand-operated microtome (Leica RM2165, Germany) and mounted onto salinised slides for staining with Harris Haematoxylin and Eosin (H&E). Fascicle size, number, and distribution were qualitatively compared on cross sections using light microscopy (Olympus BX51, Sydney, Australia). The structural waviness, “crimp,” was qualitatively evaluated using longitudinal sections. Polarised microscopy was used to aid the evaluation of crimping patterns.

2.1.3 Results

The required loads to start cheese-wiring a suture loop through the tendon, and the recorded loads at 2mm, 5mm and 10mm are given in Figure 2-9, and represented diagrammatically in Figure 2-10.

Figure 2-9: The loads generated with pulling the suture loop through the deep flexor tendon (splitting) of human, sheep, and pig tendon at start of splitting, 2mm splitting, 5mm splitting and 10mm splitting.* indicate significant difference (p<0.05).

Figure 2-10: The loads generated with pulling the suture loop through the deep flexor tendon (splitting) of human, sheep and pig illustrated graphically.

There was no significant difference in forces at initial cheese-wiring between pig, sheep

and human tendons. There was no significant difference in forces at 2mm splitting

between sheep and human tendons but forces recorded to cheese-wire the pig tendons 74

were significantly higher . In the later stages of testing (5mm and 10mm), the forces to split sheep, and pig tendons became significantly higher than those of human tendons.

Table 2 Significances (p values) for forces needed to split tendons.

Start of splitting Human Sheep Pig Human / p=0.36 p=0.06 Sheep p=0.36 / p=0.11 Pig p=0.06 p=0.11 /

2mm splitting Human Sheep Pig Human / p=0.15 p<0.05 Sheep p=0.15 / p<0.05 Pig p<0.05 p<0.05 /

5mm splitting Human Sheep Pig Human / p<0.05 p<0.05 Sheep p<0.05 / p=0.42 Pig p<0.05 p=0.42 /

10mm splitting Human Sheep Pig Human / p<0.05 p<0.05 Sheep p<0.05 / p=0.20 Pig p<0.05 p=0.20 /

Representative histological sections at x4 magnification are demonstrated in Figure

2-11.

The pig tendons had a greater eosinophilia and more prominent nuclei than the sheep and human tendons. In comparison to the sheep and human tendons the pig tendons fascicles were more poorly segregated and seemed to be denser. Human tendons had the largest fascicle size, and these were well delineated throughout the tendon cross section.

Sheep had a smaller fascicle size, especially in the peripheral areas of the cross section.

On longitudinal sections the fibres in human tendons were difficult to define, with the tendon having a more homogenous hyaline appearance.

Figure 2-12: Greater crimping of pig fibrils in the longitudinal projections compared to human and sheep (x4 magnification).

Pig fibres possessed the greatest degree of structural waviness, "crimp", with less crimping seen in the sheep tendon (figure 2-12 and 2-13).

Figure 2-13 Greater crimping of pig fibrils in the longitudinal projections compared to human and sheep (x20 magnification).

These differences were best highlighted with polarised microscopy (Figure 2-14).

Figure 2-14: Polarised microscopy x10 magnified

Tendon widths are represented in Figure 2-15. Pig tendons were 25% wider than human tendons. Sheep tendons were 10% wider than human tendons. These differences were significant p<0.05. 79

Figure 2-15: Mean tendon widths of pig, sheep and human tendons 5mm distal to A1 pulley. Differences in width were significant (p<0.05).

2.1.4 Discussion

An important consideration when choosing an animal model is the ability to compare results with existing literature. Most authors use the pig tendon model and this might enforce the further use of this model. But only few authors have specifically investigated

the pig tendon model [152, 163]. And these few reports did not investigate the animal model in a direct comparison to human tendons or only investigated the models anatomy. Other authors refer to similar biomechanical investigations on either human or pig flexor tendons and conclude the tendons comparability [163-165]. No specific studies investigating the sheep flexor tendon model were found.

A flexor tendon reconstruction is a dynamic interface. The repair configuration adjusts as the tendon is loaded and repetitively stressed. These changes include equilibration of tension across suture strands and geometric adjustment of the repair configuration [166].

If the suture material and knot are relatively strong, further increases in load will begin to overcome the resistance of the tendon tissue causing cheese-wiring. This inevitably leads to gap formation at the repair site and eventually, if a sufficient load is reached and suture and knot are strong enough, the repair will fail by suture pull out.

Tendon resistance to cheese-wiring relates to the tendon microstructure and this is a surgically relevant indicator. Therefore, the forces needed to cheese-wire a uniform suture loop through human, pig and sheep tendons were compared.

A 2mm bite was chosen because this depth reflects a typical locking loop bite into the tendon substance used in most suture configurations, such as Kessler [167], Tsuge

[168], Cruciate [89], Cross Locked Cruciate (Adelaide) [108] or Tang [120] repair. A

2mm bite is also the recommended minimum depth for any circumferential repair [118,

119].

In this study, the force required to pull a suture loop through a pig or sheep tendon at zone two was greater than that force required in human tendons. In the initial phase of cheese-wiring these differences were not statistically significant. At 2mm distraction, the forces needed in the pig tendons became significantly higher than in the sheep and human tendon. Differences became more pronounced with greater distraction. At 5mm and 10mm distraction, pig and sheep tendons were both significantly more resistant to cheese-wiring than human tendons.

Difference in resistance to suture pullout reflects discrepancies in suture-tendon interaction during loading of the reconstruction. Although these differences may be small, the effects on gap formation may be sufficient enough to class a repair as a

“failure” in one tendon model, yet having adequate performance when tested in a different tendon model.

Results showed no significant differences between pig, sheep or human tendons in the initial phase of cheese-wiring of a single suture loop through the tendon substance, or if formulated differently, in the initial phase of gapping. When compared with human flexor tendons, however, forces in animal tendons became disproportionately higher the

further the suture advanced through the tendon. Or, the larger the gap, the disproportionately higher the resistance to gapping in the animal tendons (see figure 2-

10). This effect was more pronounced in the pig than in the sheep tendon. In a composite tendon repair, combining core and peripheral suture, gaps of 2mm are considered a failure [128]. However in biomechanical laboratory studies, tendon repairs are often tested to their final suture pull out or as a "core only" or "peripheral repair only" construct. In these cases forces and gapping are measured beyond 2mm to gain biomechanical data to compare the investigated repair methods [60, 85, 91, 98, 113,

115, 148, 169-173]. From this investigation it becomes clear that in laboratory tests chosen tendon models make a difference if repairs are stressed beyond initial gapping.

Biomechanical results regarding measured forces and gaps beyond initial gapping might not be as comparable between species as previously thought. In this study, this effect was more pronounced in pig tendons than in sheep tendons when compared to human tendons.

Histological examinations were consistent with these findings. On light microscopy, pig tendons had a denser fascicle arrangement than human and sheep tendons (figure 2-11).

The longitudinal sections of pig tendons demonstrated a more "zigzag" configuration than human and sheep tendons (2-12, 2-13, 2-14). This constellation is also called crimping correlating with helical angle changes in the collagen fibres. It is described as a biological buffer system especially in areas with alternating force vectors [174-176].

The best investigation modality to analyse these crimping patterns is polarized light microscopy [177]. Figure 2-14 illustrates the different crimping patterns between human and pig tendons. These crimps straighten under tension, but still regarding cheese- wiring, the crimping most probably causes resistance. In terms of fascicle density and crimping, sheep tendons had a more comparable organisation to the human specimens than the pig tendons. In general fascicle architecture was more longitudinal orientated in the human tendons compared to the pig and sheep tendons. It was assumed these differences influenced the results in this study, but differences in tendon structure in terms of crimping configuration and fascicle distribution requires more detailed investigations.

In the author's previous experience with both models, tendon length has never been an issue. Sheep and pig provide long enough deep flexor tendons to achieve a zone II repair with 10mm purchase and still have sufficient tendon length to securely clamp the tendons in a biomechanical testing machine. In terms of diameter, the pig tendons were slightly larger than sheep and human tendons. A larger tendon circumference in the pig model may be particularly important when comparing the effect of repair constructs.

Transverse components of the core suture might be longer, suture bites might be deeper and the peripheral circumferential repair might need more passes to complete.

Our testing setup was not influenced by tendon width, but significant differences between species were noted. Sheep tendons had a more similar tendon size to human tendons, pig tendons had the biggest mean medial to lateral diameter.

This was also discussed in a paper published by Hausmann et al., 2009 [154]. In that study the biomechanical performance of either modified Kessler core repairs or simple running circumferential repairs were tested dynamically in the deep flexor tendons of sheep, pig, calf, and humans. Repaired sheep tendons were found to perform most similarly to repaired human tendons with regard to gap formation and failure loads in both groups, "core only" and "peripheral only". Porcine tendons also performed similar to human tendons in regards to ultimate failure force in both groups. But when focusing on gapping, the "core only" pig tendon group showed significant less propensities to gapping than the sheep and human tendons. The authors mention that the larger diameter of the pig tendons was suspected to have influenced this result, but concluded it had no statistically notable influence on ultimate load nor gap formation.

In a letter to the editor regarding that study Cao et al. explain that the larger size of the pig tendons might only have an influence on test results if methods (e.g. number of runs used in the peripheral repair) are not standardised between groups [163].

They further claim that the broader surface of pig tendons allow easier placement of standardised repair configurations. This combined with the great availability of pig tendons make this tendon their preferred model.

The letter also criticised the setup of Hausmann's tendon comparison, namely that no composite tendon repair group (core suture + peripheral suture) was tested [163]. In a reply to the letter Hausmann states that in a composite repair the core repair augments the peripheral repair and points out: "Our results suggest that porcine core sutures will support the running suture earlier than human or sheep tendons." This is in accordance with this study's results, considering the fact that in Hausmann's "core only" group the final gaps were measured beyond initial gapping (up to 8mm).

Furthermore, other factors are at play in determining comparative repair performance.

Tendon age, repair site, tendon orientation, handling and preparation might all be important. Changing any of these variables may affect testing results. This study endeavoured to standardise these factors. Animals were the same age, the site of testing

(biomechanical and histological) was exactly 5mm distal to the A1 pulley, a precise

2mm bite was performed with the same suturing material at the lateral border of the tendon (3 o'clock and 9 o'clock). Care was taken that the tendons were oriented the same way and the force vector, or pull direction, was always from proximal to distal.

It is believed that regardless of the animal model chosen, the importance of consistency of samples in a study with regards to these factors cannot be overemphasised.

A limitation of this study is that it only tested a focal region of each tendon model representative of zone II, because research attention is predominantly focused on this

domain. But tendon biology and mechanical properties differ along the length of the tendon. The structural heterogeneity reflects a response to different compressive and tensional forces, especially at the joints and beneath pulleys [24, 178]. The segmental differences in tendon microanatomy mean that these findings cannot be extrapolated to other flexor tendon segments.

Further, these structural adaptations differ between human, pig, and sheep. The tendons from quadripedal animals have different functions and are subjected to greater compressive forces with “walking". Therefore it cannot be verified if the definition of zone II in a pig and sheep tendon are structurally comparable to zone II in humans. Even the two animal models, although they show very similar anatomy, might differ in the loading patterns of their tendons and for example crimp distribution in the tendon might be different between species. Finally, the tendon selection of the human group must be taken into consideration, albeit the homogenous age group (average 68.5 years ±4.4), tendons might differ between human individuals depending on age and physical constitution [179-181].

In summary both the sheep and pig tendon are good, reliable models for the ex vivo biomechanical assessment of zone II flexor tendon repairs. Pig tendons and sheep tendons have somewhat greater resistance to cheese-wiring, especially when forces are measured beyond initial start of cheese-wiring and this appears to be related to structural

differences in the tendon. This finding was more significant in the pig tendon.

Differences in tendon bio-properties should be kept in mind when interpreting and comparing biomechanical data of tendon research reports.

2.2 Biomechanical and geometrical investigations of

current flexor tendon techniques: Changes in core

suture geometry within repaired flexor tendons under

tension

2.2.1 Two strand repairs or the Kessler Dilemma

2.2.1.1 Introduction

Techniques for repairing flexor tendons in the human hand have passed through multiple stages of evolution. Clinical results are more consistent and reliable with modern multi strand reconstruction methods [64, 66, 182, 183]. Nonetheless, many surgeons still rely on a traditional two strand Kessler repair technique. Perhaps inertia is the main reason for this – surgeons are most comfortable using the techniques they were originally trained to use.

Even with research showing inferior repair results for Kessler repairs [184] and the proof that modern multi strand repairs are not necessarily more difficult to perform

[101], its popularity is unbroken.

Since its introduction by Isidor Kessler 1969, the Kessler repair went through different modifications [185]. The nowadays famous Kessler configuration, also called grasping

Kessler repair (Figure 2-16) is in fact a modification of the Kirchmayr repair from 1917

[186].

Figure 2-16: Grasping Kessler Repair. Note, the transverse strand passes deep to the longitudinal strands.

Urbaniak presented the Kessler repair in its modified form to the American Society for

Surgery of the Hand in 1973 and published the paper in 1975 [187, 188]. The main problem of this still commonly-used two strand repair technique is its inferior stability. 90

Subsequently locking configurations were introduced to circumvent this weakness. Main evolutionary steps were the introduction of the commonly named locking Kessler repair

(Figure 2-17) by Pennington 1979 [189] and the introduction of the modified

Pennington repair or modified locking Kessler repair (Figure 2-18) by Hatanaka and

Manske in 1999 [107].

Figure 2-17: Locking Kessler repair. Note, the transverse strand passes superficial to the longitudinal strands.

This study examined the geometrical changes of these three Kessler repair configurations before and after loading and presents an in depth analysis of failure mechanisms in sheep tendons.

2.2.1.2 Pilot experiments

To demonstrate the three dimensional configuration of tendon repair techniques, it was necessary to visualize the suture path in the tendon. This can be achieved by using radiographic methods. Nevertheless to achieve a good contrast of the suture in the tendon the suture needed to be contrasted to increase its radio opacity. This was tried to be achieved by incubating sutures in iodine contrast solution. In a pilot study we contrasted 3/0 silicone coated braided polyester sutures (Ticron. Tyco, Lane Cove,

NSW, Australia) by incubating the sutures for 24 hours in 240mg/ml iodine solution

(Omnipaque. GE Healthcare, Rydalmere, NSW, Australia). Sutures were used to perform Kessler tendon repairs and X-rays were taken before and after tensioning the repair construct to compare the repair geometries (figures 2-19 and 2-20)

Figure 2-19 Anterior posterior (AP) X-rays of Kessler repair configuration performed with iodine contrasted braided polyester suture. Left pre tensioning, right post tensioning.

Figure 2-20 Lateral X-rays of Kessler repair configuration performed with iodine contrasted braided polyester suture. Left pre tensioning, right post tensioning.

This technique provided a first insight into changes in repair geometries, but evaluability was limited due to contrast agent diffusing throughout the investigated specimen after tensioning.

The solution to the problem was the use of steel wire sutures that provided clear contrast and made in depth analyses of suture configurations possible (2-21 and 2-22).

Figure 2-21 AP X-rays of Kessler repair configuration performed with stainless steel wire suture. Left pre tensioning, right post tensioning.

Figure 2-22 Lateral X-rays of Kessler repair configuration performed with stainless steel wire suture. Left pre tensioning, right post tensioning. 96

2.2.1.3 Methods

30 adult sheep forelimb deep flexor tendons were harvested. Tendons were individually wrapped in saline-soaked gauze and deep-frozen at -18°C until the day of experimentation. On the day of experimentation, tendons were thawed to room temperature, transected at zone II (5mm distal to the A1 pulley using a method we previously described [108] and randomly allocated to one of three groups of n = 10.

- Group 1: "grasping Kessler repair", (Figure 2-16).

- Group 2: "locking Kessler repair", (Figure 2-17).

- Group 3: "modified locking Kessler repair", (Figure 2-18).

Only the proximal stumps of the tendons were used and a hemi-core-repair was performed (half the core repair). 3/0 stainless steel multifilament sutures (CrNi, Ethicon,

Sommerville, NJ, USA) were used. Each repair was accomplished by the same surgeon using X3.2 loupe magnification. Repair purchase length was set at 10mm. A digital caliper (Mitutoyo CD-6–inch CS; Absolute Digimatic, Tokyo, Japan) was used during the repair to keep suture configurations and lengths constant.

After repair a small titanium ligation clip (Ligaclip, Ethicon, Sommerville, NJ, USA) was applied to each suture strand at the tendon-suture borderline. This served as marker for later measurements (figure 2-23). Locking loop sizes were set at 1/3 of the tendon surface diameter [61, 107].

Figure 2-23 Illustration of measurements of "gapping" with Ligaclip markers.

Tendons were radiologically imaged in both anteroposterior and lateral projections using a MX 20 radiography system (Faxitron X-Ray, Lincolnshire, IL, USA). An energy setting of 20kV and exposure time of 20 seconds was used for all images. A 50.0mm metal marker was placed adjacent to the tendon to serve as a reference for measurements

(figure 2-24). 98

This load was held for 30 seconds and then released. Tendons were re-imaged post- tensioning in anteroposterior and lateral projections.

Pre and post tensioning images were qualitatively and quantitatively compared. Suture elongation (relative gap formation) was digitally measured using Image J software

(ImageJ software version 1.41, National Institute of Health, USA). Data are presented as means (±standard deviation) and were analysed using SPSS version 17.0 (SPSS Inc.,

Chicago, IL, USA). Analysis of variances (ANOVA), post hoc Tukey test and Pearson correlation test was used. The significance level was set at p<0.05. Sample size of n=10 for each group was based on a pre-hoc power assessment using a significance level of

5% and power of 80%.

2.2.1.4 Results

Qualitative comparison of change of suture configurations from pre to post tensioning showed differences in failure mechanisms between the grasping Kessler repairs (group

1) and both locking variants of the Kessler repair (group 2+3).

Grasping Kessler repairs (group 1) deformed according to previous reports [148, 166,

190]. On the AP pictures (figure 2-24) a significant narrowing of the tendon at the transverse suture component and a complete loss of the Kessler "pretzel" with transformation to a U-shaped configuration was observed.

Pre tensioning Post tensioning

Figure 2-24: Pre and Post tension anterior-posterior radiographs of Grasping Kessler (top), Locking Kessler (middle) and Modified Locking Kessler (bottom). Note the loss of the Kessler Configuration in the Grasping Kessler Repair post tension.

On lateral films (Figure 2-255), a change in the angle of the exposed Kessler loops relative to the longitudinal axis of the tendon was apparent. The repair configuration unfolded around the transverse component in the direction of the volar tendon surface

(anticlockwise in the below image). 100

Pre tensioning Post tensioning

Figure 2-25: Pre and Post tension lateral radiographs of Grasping Kessler (top), Locking Kessler (middle) and Modified Locking Kessler (bottom). Note the different directions of unfolding rotation of Grasping Kessler compared to Locking and Modified Locking Kessler post tension indicated with arrows.

On AP films the locking variants of the Kessler repair (group 2+3) also showed significant narrowing of the tendon at the transverse component (figure 2-24), but the typical Kessler configuration prevailed. Nevertheless a significant constriction of the

Kessler loops from pre tensioning to post tensioning was observable in this projection.

On lateral films Figure 2-255) again a change in the angle of the exposed Kessler loops relative to the longitudinal axis of the tendon was apparent, but in contrast to group 1 the repairs unfolded around the transverse component in the direction of the dorsal tendon surface (clockwise in the above picture).

101

Quantitatively no significant differences in gaps between groups were detectable (Figure

2-266).

Figure 2-26 Measured gaps in mm post tensioning.

Average gap post tensioning for the grasping Kessler repairs (group 1) was 5.69mm

(±0.69), for locking Kessler repairs (group 2) 5.08mm (±0.53) and for the modified locking Kessler repairs it was 5.01mm (±0.56). Differences were not significant

(p>0.05).

102

2.2.1.5 Discussion

Kessler type tendon repair techniques still remain the most widely applied tendon suturing method. It is a traditional technique that is simple to learn and perform and is frequently used as a double Kessler repair to achieve a four strand repair. There are several fine points to this repair method that are often discussed, but controversy remains. The confounding nomenclature system for the different Kessler repair methods and variations of numbers and placements of knots were nicely specified in a recent paper by Sebastin et al., 2013a [167]. For this present study the three most common

Kessler variants and their most common and descriptive names were used: The grasping

Kessler repair (Figure 2-16), the locking Kessler repair (Figure 2-17) and the modified locking Kessler repair (Figure 2-18).

Besides numbers and placements of knots in a Kessler repair, the main discussion focuses on the relationship of the transverse and longitudinal strands to each other.

When the transverse component passes deep to the longitudinal strands, a grasping type

Kessler repair is created (Figure 2-16). When the transverse limb passes superficial to the longitudinal strands, a locking type Kessler repair is created (Figure 2-17).

This is a clear definition of the two repair variants, but problems arise when the question is asked what repair type was actually achieved by the surgeon.

103

It is very difficult, even under controlled laboratory conditions, to know if the transverse pass was placed over or under the longitudinal strands. Figure 2-27 illustrates one of the repairs that was not used for this comparison, but clearly shows the problem.

Pre tensioning Post tensioning

104

In this case on one side a grasping loop was created and on the other side a locking loop.

The only solution to this problem is to exit the dorsal surface of the tendon and re-enter it to be sure to be deep to the transverse suture component and therefore to be sure to create a locking Kessler repair (Figure 2-178). This is the only way to securely create a locking Kessler suture configuration.

Despite the large amount of literature citing Kessler repair methods [60-62, 64, 101,

107, 150, 191] only few studies investigated the grasping and locking type Kessler repair in a direct biomechanical comparison.

Wu and Tang found no significant difference between grasping and locking Kessler groups in regards to ultimate failure strength but a trend to more gapping resistance in the locking Kessler group in their biomechanical laboratory experiment [172]. Testing these same two suture methods in a double configuration (to create a four-strand repair) showed no significant differences in ultimate failure force or gapping.

Hatanaka et al. could only find significant differences in repair stability for locking

Kessler repairs compared to grasping Kessler repairs when using large suture calibers.

Using 4/0 sutures showed no significant differences between groups [61].

105

Further, Walbeehm et al. could not find significant biomechanical benefits when comparing locking Kessler repairs with grasping Kessler repairs using 4/0 sutures [148].

Reported results are in accordance to the results found here. In this present laboratory study the locking variants of the Kessler repair showed a trend to more resistance to gapping in comparison to the grasping Kessler group, but the difference was not significant.

This study focused more on a qualitative comparison of the different Kessler repair configurations to further investigate the failure mechanisms.

Stainless steel wire sutures have been used for tendon repairs previously [87, 192-194].

Radiography of flexor tendons repaired with these steel wire sutures provides the possibility to examine the intratendinous stitch geometry of a repair. Analysing anterior- posterior (AP) and lateral radiographic projections of repaired specimen before and after loading allows detailed descriptions of geometric changes of the three dimensional suture configuration and therefore offers a unique way to precisely describe failure mechanisms.

106

After tensioning the grasping Kessler repair (group 1) AP projections not only show a significant narrowing at the transverse component of the repair, but more importantly show a transformation of the repair geometry from the typical Kessler configuration into a U-shaped loop. The transverse suture component stays relatively in place. This phenomenon was reported previously [148, 166, 190]. This doesn't apply to the locking repairs to the same degree. In our locking groups (group 2 and 3) the repair geometry from an anterior-posterior perspective remained relatively stable, but significant tightening of the loops and also narrowing at the transverse suture component was apparent. Again the transverse suture component remained relatively stable in position.

On lateral projections the main failure mechanism becomes apparent. All tested Kessler constructs, group 1, 2 and 3 rotate around the transverse suture component of the repair.

The only difference is that the grasping Kessler repairs (group 1) are forced to rotate around the transverse component in direction of the volar tendon surface (anti-clockwise in attached pictures). On the contrary, the locking variants, group 2 and 3 unfold and rotate around the transverse component in the other direction, towards the dorsal tendon surface (clockwise in attached pictures). In all cases the axis of rotation is the transverse suture component.

107

This adds a new insight to the discussion of the definition of a locking loop. Analysing geometric changes in suture configurations and following the results of this present study, the term 'locking loop' might be misleading. Any single suture loop no matter if used in a Kessler type repair or any other repair remains a loop that simply grasps tendon substance, there is no lock around a defined strand of collagen fibres. The present experiments showed that no matter if a grasping loop or a locking loop was created, under tension both unfold in the one or the other direction.

In a Kessler repair the suture configuration unfolds around the transverse component and the resulting suture elongation of the longitudinal strands causes gapping. To qualify for the term locking, the construct must not unfold, but just tighten and lock in itself like in a cross lock in a Savage repair [63, 183] or Adelaide repair [76, 93, 195].

This failure mechanism of the unwinding locking loop was already assumed by

Hotokezaka and Manske in their classic paper from 1997 [150]. They claimed that if a very strong suture like a steel wire suture is used, the collagen fibres are either cut by the suture or unwind with the suture. The results of this present study indicate that a combination of both is the case when high enough loads are applied. Importantly, these findings were also replicated when using Iodine contrast impregnated 3-0 braided polyester sutures (Ethibond Excel, Ethicon, Sommerville, NJ, USA) but with inferior image quality.

108

The limitations of a two-strand reconstruction are now well documented [64, 91, 182,

183, 196]. The double-Kessler repair represents a relatively simple method for achieving four-strand repairs. But double Kessler repairs rely on the same Kessler holding loops and are therefore prone to the same unfolding principles. This is further described in chapter 2.2.2.

This study investigated only one half of a tendon reconstruction. The tendon was statically loaded for an arbitrarily defined time. Even though it was chosen to load tendons with 35N, what is regarded as a typical load for active unresisted finger motion

[197], these conditions don’t precisely reflect in vivo loading which is better replicated using dynamic testing of a whole tendon reconstruction. The properties of steel wire also differ significantly from polyester or polypropylene sutures. Nonetheless, our methodology permitted a direct evaluation of suture configurations and their three dimensional adaptations with loading. The main aim of this study was to clarify the failure mechanisms of grasping and locking Kessler repair variants.

In summary this study reaffirms that reconstructions employing Kessler tendon repair configurations are geometrically unstable. All three tested Kessler type repairs unfold by rotations around the transverse suturing component on a three dimensional level.

Therefore the term locking loop is misleading. Furthermore it is difficult to be sure if a locking or grasping type of Kessler repair was achieved.

109

2.2.2 Multi strand repairs

2.2.2.1 Introduction

Repair strength relates to several factors including suture material [84-87] and size [88], suture configuration [63, 89-91], and the number of suture strands crossing the repair site [92]. The latter is probably the most effective way to further stabilize a tendon repair, but the complexity of achieving a six or even eight-strand repair has limited practical application. Potential pitfalls with using more than four strands include the technical ability of the surgeon, an increase in surgical time, tissue handling, repair bulk, and injury to tendon vascularity. Studies by Barrie et al. [60, 93] showed that a four- strand repair has greater resistance for gapping than a two-strand technique, but there was no difference in gapping when a four-strand technique was compared to an eight- strand repair. In general four-strand repairs are considered as a good compromise of repair stability and complexity [59], a well-constructed four-strand repair in conjunction with a peripheral suture is capable of safely tolerating early rehabilitation protocols [79,

197]. Therefore this thesis predominantly focuses on four-strand repair methods.

As in the previous study, further analysis of the weaknesses of common core suture configurations was carried out with the novel X-ray technique, but this time four-strand constructs instead of the previously described two strand repairs were investigated.

110

The reasons for differing performances of equivalent strand repairs are based primarily on the holding configurations of the suture into the tendon. Small modifications of these locking zones can cause significant differences in gap formation and ultimate strength.

Focus was on the most common current four-strand repair techniques: Double Kessler technique, Cruciate technique and Adelaide technique.

2.2.2.2 Methods

10 tendons were tested per repair method. Therefore 40 adult sheep forelimb deep flexor tendons were harvested. Tendons were individually wrapped in saline-soaked gauze and deep-frozen at -18°C until the day of experimentation. On the day of experimentation, tendons were thawed to room temperature, transected at zone II (5mm distal to the A1 pulley like previously described) and randomly allocated to one of four groups of n=10 tendons.

111

- Group 1: Grasping Double Kessler Repair (figure 2-28 above)

- Group 2: Locking Double Kessler Repair (figure 2-28 below)

Figure 2-28 Grasping (above) and locking (below) double Kessler repair techniques

112

- Group 3: "Cruciate Repair" (figure 2-29)

Figure 2-29 The four-strand cruciate repair technique [89]

113

- Group 4: "Adelaide Repair" (figure 2-30)

Figure 2-30 The cross locked cruciate or Adelaide technique [76].

Again, like the previous study, only the proximal stumps of the tendons were used and a hemi-core-repair was performed (half the core repair). 3/0 stainless steel multifilament sutures (CrNi, Ethicon, Sommerville, NJ, USA) were used. Each repair was accomplished by the same surgeon using X3.2 loupe magnification. 114

Repair purchase length was set at 10mm. A digital caliper (Mitutoyo CD-6–inch CS;

Absolute Digimatic, Tokyo, Japan) was used during surgery to keep suture configurations and lengths constant.

After repair a small titanium ligation clip (Ligaclip, Ethicon, Sommerville, NJ, USA) was applied to each suture strand at the tendon-suture borderline. This served as marker for later measurements.

Hemi-repairs were then statically tensioned using an MTS 858 Mini Bionix materials testing machine (MTS Systems Corp., Eden Prairie, MN, USA). Samples were preloaded to 1.5 N, then loaded at a distraction rate of 20 mm/min up to 35N. This load was held for 30 seconds and then released. Tendons were re-imaged post-tensioning in anteroposterior and lateral projections.

Pre and post tensioning images were qualitatively and quantitatively compared. Suture elongation (relative gap formation) was digitally measured using Image J software

(Image J 1.41o, National Institute of Health, USA).

115

Data are presented as means (±standard deviation) and were analysed using SPSS version 17.0 (SPSS Inc., Chicago, IL, USA). Analysis of variances (ANOVA), post hoc

Tukey test and Pearson correlation test was used. The significance level was set at p<0.05. Sample size of n = 10 per group was based on a pre-hoc power assessment using a significance level of 5% and power of 80%.

2.2.2.3 Results

Figure 2-31 shows example pictures of repairs of each group, before and after tensioning.

Figure 2-31 Example pictures of repairs from each tested group before and after tensioning. 116

Figure 2-32 illustrates the average measured distances of the marker clips from the tendons end for each repair group. Gap in mm after tensioning (1 = Grasping Double Kessler Repair; 2 = Locking Double Kessler Repair; 3 = Cruciate Repair; 4 = Adelaide Repair).

Group one, the grasping double Kessler repairs, least withstood the forces applied and showed most gapping (3.32mm ±0.79) (figure 2-32).

Group two, the locking variant of the double Kessler repair, could show slight improved gapping propensities (3.12mm ±0.36) but the difference was not significant (p=0.60)

(figure 2-32).

117

Group three, the cruciate repairs could show a further improvement in gapping resistance (2.61mm ±0.44) but again statistical analysis showed no significant difference to the grasping or locking Kessler variants (p≥0.05) (figure 2-32). The Adelaide repair method was the most stable repair method and could resist the applied loads with least gapping (1.71mm ±0.47). This improvement was statistically significant to all other repair methods (p<0.05) (figure 2-32).

2.2.2.4 Discussion

Radiography of flexor tendons repaired with steel wire suture allows the intratendinous and extratendinous stitch geometry to be examined. Indeed, the first radiograph was taken by Wilhelm Roentgen of his wife's hand in 1895. Steel markers have been used for monitoring flexor tendon repairs clinically by Silverskiold previously [198]; these provided a powerful method for imaging gap formation and tendon excursion. Stainless steel wire itself has been used in the past for tendon repair [199]. It has a high tensile strength and excellent knot holding capacity. However, the difficult handling properties and risk of abrasion of the lining of the tendon sheath lead this material to lose favour to newer sutures.

118

The limitations of a two-strand reconstruction are now well documented. The double-

Kessler repair represents a relatively simple method for achieving four-strands. But this technique relies on the same Kessler holding loops and is therefore prone to the same inherent instability. More importantly the problem to reliably create a locking variant of the two strand Kessler repair doubles with the four-strand Kessler repair. All four longitudinal strands have to be deep to the transverse component respectively. This is hard to achieve. Like explained in the previous chapter, the only way to reliably create a locking type double Kessler repair would be to perform two modified locking double

Kessler repairs, exiting the dorsal tendon surface distal to the transverse strand and reentering the tendon proximal to the transverse component. This seems to be very complicated and regarding the results of our previous study this procedure would not even make a big difference regarding the biomechanical stability. Furthermore, the second Kessler configuration has a smaller, less than ideal purchase length. Also of concern is the risk of “catching a suture,” because the suture passes are crossing the line of previous suture strands. The inferior biomechanical properties of the two double

Kessler repair variants are also reflected in quantitative study results. The locking variant of the double Kessler repair method could improve the gapping propensity by

6%, but the improvement was not significant. Wu and Tang came to the same conclusion. They found no significant difference between grasping and locking Kessler groups when used in a double configuration to create a four-strand repair in their biomechanical comparison [172].

119

Previously McLarney’s four-strand "Cruciate" repair, was said to achieve an ideal balance between simplicity and performance [79]. This four-strand repair was designed with loop-locks [89]. It was demonstrated that these have a tendency for loosing grip when tensioned. In biomechanical experiments the Cruciate repairs showed significant cheese-wiring tendencies with a resultant propensity for repair site gapping (figure 2-

33).

2-33 Radiographs of Cruciate repair, pre and post tensioning, left AP and right lateral projections.

In contrast, the cross-locking variants in the Adelaide repair tightened (locked) under tension and did cheese-wire much less. In the lateral projections the flattening out of the cross-lock “pyramids” was the most recognisable geometrical change (figure 2-34).

120

2-34 Radiographs of Adelaide repair, pre and post tensioning, left AP and right lateral projections.

The superior holding capacity of cross locks has been reported previously by Xie and others [103, 171]. In this biomechanical study, a significantly greater gapping resistance in the repair method incorporating cross locks (Adelaide repair) with 1.71mm displacement compared to the repair method incorporating loop locks (Cruciate repair) with 2.61mm displacement was found. This improvement of gapping resistance for the

Adelaide repair was statistically significant (p<0.05).

In conclusion, if a surgeon does choose to use a double Kessler repair, it is recommended to use a locking variant. But like previously reported this might not always be achieved. From the outside it is not possible to see if a locking or grasping variant of the Kessler configuration was achieved.

121

Additionally, it is thought that a double Kessler repair shows not only biomechanical disadvantages to modern four-strand constructs, but also requires two knots.

Furthermore, compared to the Cruciate and Adelaide repair, the level of complexity when performing a double locked Kessler repair is not less challenging. The Cruciate repair is easy to perform and shows improved gapping resistance compared to the double Kessler configurations but cannot achieve biomechanical stability like the

Adelaide repair. A cross locked Adelaide repair undergoes minimal constitutional change under tension and shows least propensity for gapping.

122

2.3 The current Gold Standard in Flexor Tendon repairs, the

“Adelaide Repair” - Improvements and Modifications

2.3.1 The influence of locking stitch size in an “Adelaide

Repair”

2.3.1.1 Introduction

The ideal flexor tendon repair should be easy to reproduce and add as little bulk as possible. It should possess a high tensile strength to safely allow early mobilisation and be resistant to gap formation to facilitate healing with minimal scarring. Various repair techniques have been described which differ in their ability to meet these aims. Four- strand techniques provide a good compromise between minimising complexity and adequate performance.

The cross-locked cruciate repair was firstly reported by Sandow and McMahon, as a four-strand modification of the Savage six-strand repair technique using two single cross-locks to grasp the tendon ends on each side [91, 195, 200] (Figure 2-35).

123

Figure 2-35: Top: Schematic drawing of the cross-locked cruciate repair technique or Adelaide repair technique. Bottom: Picture of Adelaide core repair performed in a pig tendon with stitches one to nine and recommended suture purchase marked.

124

This repair technique is also known as “Adelaide repair technique”. Current literature suggests that this cross-locked cruciate repair is not only stronger than other four-strand techniques, but also is a favourable repair in terms of gap formation and simplicity [60,

98, 102, 201, 202]. However the optimum size of the cross-locks in this specific repair technique, or indeed if the size of the locking stitch effects the repair strength or gapping behaviour, has not been investigated. The purpose of this study is to address this question in an ex vivo biomechanical comparison.

125

2.3.1.2 Materials and methods

Twenty-two deep flexor tendons were harvested from the forelimbs of adult pigs, and were randomly allocated into two repair groups of n=11 tendons each. A <5% partial laceration was used to mark the planned transection point 5mm distal to the A1 pulley at time of harvest (Figure 2-36).

126

Tendons were individually wrapped in saline soaked gauze and deep frozen at -18 degrees Celsius up until the day of experimentation.

On the day of mechanical testing, each tendon was thawed to room temperature immediately prior to transection and repair. Care was taken to minimise tensile strengthening of the tendon resulting from saline soaking [153, 203]. A sharp transverse laceration of the tendon was performed at the pre-marked site using a #22 scalpel blade.

All tendons were then repaired with a four-strand cross-locked cruciate suture technique

(Figure 2-31). The repairs were performed using 3/0 silicone coated braided polyester sutures (Ticron. Tyco, Lane Cove, NSW, Australia) and each repair was accomplished by the same surgeon (T.P.) using x3.2 loupe magnification. The cross-locks were uniformly set 10mm from the line of tendon division. A cross-lock width of 2mm was used for group 1 (n = 11) compared to a width of 4mm for group 2 (n = 11) (Figure

2-37).

127

Figure 2-37: Schematic drawing of the repairs in group one (2mm cross-locks) and group two (4mm cross-locks).

All repairs were completed with a 6-0 polypropylene (Prolene. Ethicon, Somerville, NJ,

USA) simple running peripheral suture with 16 loops and 2mm purchase. A digital caliper (Mitutoyo CD-6’’CS Absolute Digimatic. Tokyo, Japan) was used during surgery to produce exact measurements of distances and lengths.

128

Repaired tendons were mechanically tested in a uniaxial manner using a MTS 858 Mini

Bionix materials testing machine (MTS Systems Corp. Eden Prairie, MN, USA).

Tendons were gripped in grooved pneumatic clamps with a gripping pressure of 60psi.

Gauge length was standardised at 40mm. Samples were preloaded to 1.5N then tested to failure at a relative slow distraction rate of 10mm/min. Load, displacement and time were continuously recorded at 100Hz. Data were used to generate load-displacement graphs for each tendon. Additionally a digital caliper fixed at 2.0mm was placed adjacent to the repair site as a reference. The repair site was continuously filmed during testing using a high definition Sony Handycam camcorder (Sony Corp. Tokyo, Japan) to monitor gap formation. The timing of the video was coupled with the data recorded by the testing machine. This method allowed an exact visual and biomechanical correlation of 2 mm gapping.

Load to failure (measured in Newton), load to 2mm gap formation (measured in

Newton), stiffness (measured in Newton/mm), and mechanism of failure were determined for all samples. The video and mechanical data were correlated after each test to determine the load to first point of 2mm gapping. Stiffness was defined from the tangent of the linear middle third of the load displacement curve. The mechanism of failure was recorded as either suture breakage at the knot, suture breakage away from the knot, or suture pull out.

129

Data are presented as means (± standard deviation) and were analysed using SPSS v17.0

(SPSS Inc. Chicago, IL, USA). Independent t-tests assuming unequal variances were used for comparisons with p<0.05 was considered statistically significant. A sample size of n = 11 per group was based on a pre-hoc power assessment using a significance level of 5% and power of 80%.

2.3.1.3 Results

There was no pullout of the core suture noted. 21 out of 22 repairs failed by breakage of the core suture at the knot; one repair in the 2mm cross-lock group failed by breakage of the thread 1mm adjacent to the knot.

The load to failure was 71.7N (±11.0N) for group 1 (2mm cross-locks) compared to

71.1N (±11.8N) for group 2 (4mm cross-locks). Therefore no significant difference in load to failure was detected (p = 0.89, Figure 2-38).

130

Figure 2-38: Mean load to 2mm gap formation and mean load to failure by repair group.

The load to 2mm gap formation was 55N (±8.7N) for group 1 (2mm cross-locks) compared to 62.2N (±9.4N) for group 2 (4mm cross-locks). This showed a trend towards higher loads in the group with bigger locking stitches (p = 0.07, figure 2-38).

131

The stiffness was greater in repairs with 4mm locking stitches (4.1N/mm ±0.8N/mm) compared to repairs with 2mm locking stitches (4.6N/mm (±0.6N/mm). This difference, however, was not statistically significant (p = 0.33).

2.3.1.4 Discussion

According to [59], a well-constructed four-strand repair in conjunction with a peripheral suture is capable of tolerating light composite grip during the healing period. But modern “early active” rehabilitation regimes may require high repair loads of up to 50N

[197, 202] and a trend to reduced splint usage is recognisable [204]. Taking this into consideration and bearing in mind that unavoidable mishaps during the rehabilitation process might require additional repair strength lead to the conclusion that stronger repairs might still be better repairs. Although defining a number of Newton that reflects the necessary repair strength for tendon repairs is difficult and may be impossible.

Anchoring of the cross-locks in the tendon substance with the cross-locked cruciate repair in this study was strong, stronger than the suture knot. No suture pullouts were noted. Considering this, it is doubted that the measurement of final failure force is the best parameter for the assessment of tendon repair strength, particularly if the suture configuration within the tendon is not failing. The strength of any repair technique cannot exceed the ultimate tensile strength of the suture with its knot [87, 91, 205].

132

Bearing this in mind, the most important parameter becomes gap formation. Gelberman et al. have demonstrated that while a gap up to 1mm may have little effect on gliding resistance, a gap of 3mm can block motion [128]. Two millimetres is considered the threshold and is a common endpoint in comparative biomechanical studies.

Suture loop configurations can be either grasping or locking. Locking configurations show less propensity for gapping than grasping configurations [60-62, 98, 107, 150].

Recently published data suggests superior biomechanical properties with the cross- locked cruciate technique [60, 98, 102, 201]. No data regarding the influence of the size of the cross-locks in a cross-locked cruciate repair on repair strength and gap resistance was found.

Xie et al. compared simple circle locks perpendicular to the longitudinal axis of the tendon in different diameters (1mm, 2mm, 3mm) [103]. In this study an increase of repair strength was noted when increasing the locking loop size from 1mm to 2mm. No further increase of strength could be shown for the group with 3mm loops. Gap formation was significantly higher in the 1mm group compared to the 2mm and 3mm group.

133

Dona et al. investigated the effect of the size of the locking stitches in four-strand cruciate repairs [159]. This repair technique consists of the same four-strand cruciate core configuration used in the cross-locked cruciate repair, but utilises simple loop-locks instead of the more stable cross-locks. They found least gapping when the loop-locks incorporated 25% of the tendon. Loop-locks incorporating 33% or 50% of tendon gapped earlier. It is noteworthy that no peripheral tendon suture was used in either of these two studies.

Hatanaka and Maske showed a positive linear relationship between ultimate strength and cross sectional area of the locking loops in their study using a two-strand

Pennington modification of the Kessler repair. But, they also noted a higher propensity for gap formation with larger locking loops in this two-strand repair technique [107].

These reported results are not directly comparable to this study's findings since the chosen repair techniques differ vastly to the cross-locked cruciate technique used in this study. The question of whether larger or smaller locking stitches are more favourable is heavily dependent on the actual repair technique or locking configuration. Greater loop sizes in open lasso-type repair techniques (like the two-strand modified Pennington repair) allow greater tendon constriction and suture “give” when tensioned. This may result in elongation and therefore gapping.

134

This effect is not conferrable to the cross-locked cruciate repair technique. The cross- locks in a cross-locked cruciate repair lock in itself and if appropriately pre tensioned grip the tendon with little constricting effect.

A higher resistance to gap formation in the repairs with larger (4mm) cross-locks than in repairs with smaller (2mm) cross-locks was noted. The simple augmentation of the repair with larger cross-locks resulted in an increase of gap resistance of 13% (55N –

62.2N, p = 0.07).

An alternative explanation for the inferior stability of repairs with smaller cross-locks may be that smaller cross-locks inevitably grab the tendon more superficially than larger cross-locks (see shaded area in Figure 2-37). The depth of a stitch in a tendon is predefined by the curvature of the needle used and the distance of entry and exit point of the stitch. If the same suturing needle is being used, smaller cross-locks will have a more superficial anchoring character than larger cross-locks. Therefore slightly more buckling and asymmetric gapping was noted in the group with 2mm cross-locks. The posterior tendon wall gapped earlier than the anterior wall. This effect was not observed in the group with 4mm cross-locks.

135

An optimal cross-locked cruciate repair can only be achieved if all four-strands are correctly balanced and pre-tensioned with the same force. The observation that all our repairs failed at the weakest point (the knot) suggests that this was adequately achieved.

This study was a laboratory study in nature and utilised a common ex vivo animal model. While it was demonstrated that larger cross-locks in the cross-locked cruciate repair may provide better gapping behaviour, they may also have a larger effect on the tendon biology, thereby compromising healing. Despite earlier studies indicating that reduced tendon vascularisation does not necessarily interfere with the tendon healing process [57, 206-209], recent evidence suggests that surgical intervention may disturb tendon biology more than previously thought [210, 211]. Not only a reduced vascularisation at the repair site, but also a slight increase of suture exposure might interfere with therapy success.

It is important to test the complete dynamic repair construct with core and peripheral suture to allow comparability to other tendon repair studies and clinical scenarios.

Nevertheless, the reinforcement of the core repair with the peripheral suture might have diminished more significant differences in our two study groups. Furthermore, the static testing model used in this experiment may not reflect the repetitive stress of rehabilitation seen in clinical situations [155].

136

In conclusion, four-strand cross locked cruciate repairs with cross lock sizes of 2mm or

4mm provide high tensile strength and are resistant to pullout. Repairs with 4mm cross- locks tend to provide better gapping resistance or a more central load distribution than repairs with 2mm cross-locks.

137

2.3.2 The interlocking Adelaide Repair: A stronger modification

of the Adelaide Repair

2.3.2.1 Introduction

In the author's experience of the Adelaide repair, there is always one particular cross lock that tightens first and gives way for gapping (Figure 2-39).

Pre Tension Post Tension

Figure 2-39 Pictures of Adelaide repair pre and post tensioning

138

To circumvent the phenomenon and its inevitable effect on gapping and at a later stage repair failure the author modified the Adelaide repair by interlocking the distal components of the cross locks with each other (Figure 2-40)

Figure 2-40 First drawings of the modification for the Adelaide repair

This represents a novel repair method that was not described previously.

It is established that early mobilisation of reconstructed flexor tendon injuries has significantly improved functional outcomes [147, 212]. This has been enabled through improved rehabilitation protocols and also through the evolution in repair techniques that withstand gapping better than previous techniques.

139

The four-strand, cross-locked cruciate repair or Adelaide repair demonstrate a good compromise between strength, simplicity and bulk without excessive tissue handling

[60, 98, 102, 201, 202]. Previous work has demonstrated that the properties of the cross locks are integral to the overall biomechanical behavior of the repair [108]. Still, since there are no connecting components between the cross locks, gapping occurs when the weakest single cross lock tightens. It is therefore hypothesised that a simple interlocking modification to the cross locks can increase resistance to gapping. This study aims to compare the standard Adelaide repair to the modified, interlocking Adelaide repair using both cyclic and static testing protocols.

2.3.2.2 Materials and methods

30 deep flexor tendons were harvested from the fore limbs of adult sheep of similar age.

15 forelimbs were used in total, each limb containing two deep digital flexor tendons.

One tendon from each limb was randomly allocated to either the standard Adelaide or modified (interlocking) Adelaide group. As previously described all tendons were sharply transected at a standardised point immediately distal to the A1 pulley in order to represent the equivalent of a zone II flexor tendon laceration [108].

140

2.3.2.2.1 Repair techniques

Fifteen tendons in the standard group underwent Adelaide repairs performed according to previously described techniques (Figure 2-41) [200]. Tendons were repaired on the volar aspect using 3-0 braided polyester suture (Ticron; Tyco Healthcare. Norwalk, CT) for the core strands. Suture purchase was kept constant at l0mm. The cross-locks were kept at 4mm in both transverse and longitudinal dimensions. Care was taken to keep the repair technique standardized and a metric ruler was used for this aid. A single four throw (1+1+1+1) square knot was tied between tendon ends.

Figure 2-41 Adelaide repair technique

141

The remaining fifteen tendons were repaired using an interlocking modification which involved interlocking of the distal angles of the cross locks through each other (Figure

2-42).

Figure 2-42 Interlocking or modified Adelaide repair technique

Suture purchase and depth was again kept constant at 10mm as in the conventional

Adelaide repair group. Repairs were all carried out using 3.2x loupe magnification.

142

As in previous studies, after repair, tendons were immediately wrapped loosely in cotton gauze swabs moistened with phosphate buffered saline and sealed in airtight containers before being stored in a freezer. They were then allowed to fully defrost at room temperature for eight hours before mechanical testing [153].

2.3.2.2.2 Mechanical testing

Repaired tendons were mechanically tested with the MTS 858 Mini Bionix material testing machine (MTS Systems Corp. Eden Prairie, MN). Uniaxial testing was performed using a 500N load cell. Tendon ends were gripped in the same orientation between pneumatic grips with a clamping pressure of 60 pound per square inch (psi).

The testing gauge length (distance between the two grips) was standardised at 40mm. A

3N preload was maintained for 10 seconds before the cyclic loading protocol started.

Loads between 3N (minimum) and 30N (maximum) were applied for 250 cycles at 1

Hertz (Hz) (i.e. one cycle of loading and unloading per second) (Figure 2-43). These loads are reflective of the dynamic stress applied during early active rehabilitation [197].

250 cycles were chosen since a previous dynamic tendon repair study showed that 90% of the measured gap at 1000 cycles is reached by 200 cycles in a core only repair experiment [213].

143

To evaluate gap formation during this dynamic phase of loading, testing was paused at

10, 20, 30, 40, 50, 100, 150, 200 and 250 cycles under the application of a static 3N load. Digital images of the repair site were captured during each pause. The loading and imaging sequence is illustrated in Figure 2-43.

Pictures were taken using a Canon EOS 1000D digital camera with a 100mm Canon macro lens (Canon Inc, Tokyo, Japan). The camera was mounted on a stable tripod positioned at a constant distance and level to the repair site. A digital caliper (CD-6_CS

Absolute Digimatic; Mitutoyo, Tokyo, Japan) fixed at 10mm was secured in the same plane as the tendon to serve as a measurement reference in the images (Figure 2-44). 144

After 250 loading cycles, the repaired tendon was distracted to failure at a distraction rate of 20mm/min. Load, displacement, and time were continuously recorded at 100Hz and load-displacement curves were generated from the recorded data. No slippage of the tendon within the clamps was observed during testing. Care was also taken to avoid tendon dehydration by intermittent spraying with phosphate-buffered saline.

145

The primary endpoints were gap formation during dynamic loading (recorded in mm), ultimate failure load (recorded in N) and mode of final failure (suture pull out or breakage). Mean gap was calculated using the digital images taken at each cycle pause.

Therefore measurements were made in each picture at eight locations across the repair site using image analysis software (Image J 1.410; National Institutes of Health,

Bethesda, MD). Mean gap (recorded in mm) was then determined by averaging the eight measured lengths (Figure 2-5144). The caliper in the image served as 10mm reference for scaling. The number of cycles necessary to create a 2mm gap was calculated by correlating the recorded gap data with the amount of cycles at each measurement point

(logarithmic interpolation).

All data were analysed in SPSS v17.0 (SPSS Inc., Chicago, IL). Independent sample t- tests with equal variances not assumed were used for comparisons of groups. A p-value of less than 0.05 was considered statistically significant. A sample size of n = 15 per group was based on a pre-hoc power assessment using a significance level of 5% and power of 80%.

146

2.3.2.3 Results

2.3.2.3.1 Gap Formation

The average gap formation of both groups after 250 cycles was similar (3.95mm in the standard Adelaide group, 3.80mm in the modified group) and observed differences did not reach statistical significance.

The rate of gap formation was highest over the first 50 cycles. However, in the early phase of cyclic loading (up to 10 cycles) gap formation was significantly lower (p =

0.02) in the modified Adelaide group (0.72mm +/- 0.6) compared with the standard

Adelaide group (1.56mm +/- 0.49) (Figure 2-45).

Figure 2-45: Illustration of gapping propensities of repair techniques.

147

2.3.2.3.2 Final Load

The average final load to failure in the standard Adelaide group was 48.6N (+/- 5.94) compared to 51.9N (+/- 3.0) in the modified Adelaide group (Figure 2-46).

Figure 2-46: Final load at failure of repair techniques.

148

Results showed that by interlocking the distal locking components in an Adelaide repair, the final load capacity increased by 6.83%. Although a trend towards increase load to failure in the modified Adelaide group was observed, this did not achieve statistical significance (p=0.13).

2.3.2.3.3 Mechanism of failure

Static pull to failure testing caused complete repair rupture and failure at the knot in most cases. Only one suture pullout was observed in the conventional Adelaide repair group.

2.3.2.4 Discussion

The ideal technique for tendon repair should be easily reproducible, possess high tensile strength, high resistance to gap formation and ultimately allow for early active mobilisation [59]. Flexor tendon repair techniques have evolved and undergone constant modification in a bid to achieve the desired balance between strength and simplicity. Six core strand techniques impart strength but increased bulk of the repair and increased tissue handling are potentially detrimental to the tendon [210]. Two core strand techniques encounter these problems to a lesser degree but do not possess the desired durability under loads imparted by active mobilisation regimes [92].

149

Four core strand repairs appear to achieve this balance more successfully. In particular, previous studies have demonstrated the favorable biomechanical characteristics of cross- locked cruciate four-strand techniques [98, 102, 105, 201, 202] and indicate that these techniques may offer a good compromise between the desirable and undesirable aspects of tendon repair.

During the evolution of four-strand repairs, several authors have described differences in the biomechanical properties of grasping loops and locking loops [62, 102, 107, 150,

159]. These differ from the cross locks seen in the Adelaide repair and the modification described in this present study. [102], [60] and [98] Several authors demonstrated that within the context of four-strand repairs, the use of locking cross stitches confers significant benefits over repairs using looped techniques . They demonstrated that gap force and ultimate tensile strength were both improved by using a cross lock. The benefit of the cross locks is that they have less propensity to deform under load than looped techniques [61, 62, 98, 104, 107, 108, 150, 201]. Consequently, the tendon is less prone to bunching and gap formation is minimised. Previous work [108] examined the influence of altering the size of the locking cross stitch in the Adelaide repair. In contrast to simple loop locks, increasing the size of the cross locks can offer improved resistance to gap formation. Therefore subtle differences in the configuration of cross locks with this repair can have a significant impact on its biomechanical performance under load.

150

This present study intended to examine whether further modification in the form of interlocked cross stitches could confer additional benefits without unduly adding to repair time, complexity or bulk.

The modification described requires almost no departure from the established technique and neither does it add further bulk. Interlocking of the cross stitch can be simply achieved by altering the placement of the suture.

Interlocking of sutures in tendon repairs has previously been described. This principle has been utilised in the optimisation of epitendinous repair techniques (Type A

Silfverskiold cross-stitch and the interlocking horizontal mattress) with demonstrated biomechanical advantages [154, 159]. Interestingly, the unlocked type B Silfverskiold suture and the interlocking horizontal mattress are very similar in design, differing by virtue of the fact that one is interlocked and the other is not. In a comparative biomechanical comparison of these two, the interlocked version performed better [159].

The same principle has not often been considered in the context of a core suture.

Robertson and al-Qattan, 1992 described a four-strand technique consisting of looped core strands in each half of the transected tendon which interlocked with each other between the cut ends of the tendon [90]. They concluded that their repair was stronger and more resistant to gapping than the modified Kessler or Strickland techniques.

151

However, theirs is a 6-strand technique requiring more tissue handling while placement of the interlocking loop between the ends of the divided tendon can make it difficult to tension appropriately.

Dona et al., 2004 demonstrated that in a four-strand model, the addition of a simple looped epitendinous suture did not affect the load to failure, but significantly affected

2mm gap force from 14N to over 30N [159]. Therefore, a circumferential suture to avoid masking of differences in the performance of the two different core sutures was not used. The comparative improved resistance to early gap formation that was demonstrated inthis experiment did not quite achieve statistical significance but a trend to less gapping was observed. The aim was to establish whether there would be an observable difference in a repair that would mimic the clinical scenario as closely as possible. This led to the use of a sheep tendon model which in previous studies have proven to be more similar to human than porcine or bovine tendon [213].

This experiment confirms earlier observations regarding the mechanism of gap formation in the Adelaide repair [108]. As the forces within the strands of the suture equilibrate, the tendon deforms in association with tightening of a single cross lock. The resultant redundancy in the suture material leads to gap formation at the repair site. It is possible that by interlocking the cross locks, their resistance to tightening under load may be increased and thus delay gap formation during early cyclic loading.

152

This early period may well be critical to the overall biomechanical integrity of the repair

[155]. The observed resistance to early gap formation may be due to the fact that interlocking cross locks may increase the volume of tendon material incorporated into the repair.

A modest increase in the final load to failure was also demonstrated although this did not reach statistical significance in our study. This experiment measured ultimate tensile strength (UTS) following cyclic loading of the repaired tendon. UTS measured as a single load to failure may not be a reliable indicator of the biomechanical performance of tendon repairs. It has been demonstrated that the UTS of the Adelaide significantly decreases with cyclic loading and differs from the single load to failure in a previously unloaded repair [128]. This emphasises the importance of tendon gapping as a mode of failure as described by Gelberman [128]. The work of flexion in repaired tendons demonstrating gapping after cyclic loading may actually approach the actual lower UTS seen in physiological conditions. Therefore the behavior of the repair during cyclic loading may be a more accurate representation of its performance and its propensity to fail via gap formation.

153

The simple modification that we describe to the already proven technique does not appear to have any noticeable drawbacks. Although there is a theoretical possibility that the suture may eventually fail at the site of interlocking due to a tear caused by the friction of one strand directly over the other, this has not been observed in the author's experience. Therefore, it is proposed that the use of the interlocked modification of the cross locked cruciate repair may have relevant clinical implications with regard to increased early resistance to gap formation. Further work will be required to demonstrate this.

154

2.4 A new concept for flexor tendon repairs: The knotless

3D barbed suture flexor tendon repair

2.4.1 Introduction

The best approach for restoring flexor tendon function after injury has evolved with time. Primary repair with a core and peripheral suture is now standard [59, 114]. Multi strand core suture techniques are ideal, although contention regarding the ideal number of strands exists [60, 92, 93, 100, 214]. Together with the emphasis on early mobilisation [22, 112, 131, 134, 204], these represent major developments in tendon reconstruction. As previously described, four-strand techniques in general and the four- strand cross locked cruciate repair in particular (or Adelaide repair) provide a good compromise between complexity and performance [201, 215].

But there are several limitations with conventional tendon repairs. Primarily, the presence of a knot introduces a weak point into the system [216, 217]. Furthermore, conventional repairs are dependent upon the holding configuration of the suture in the tendon to prevent pullout and gapping. Therefore grasping configurations were superseded by locking techniques [61, 98, 150, 189, 216, 217] and loop-locks were superseded by cross-locks [91, 108, 195].

155

Despite these advancements, the effect of constricting elements in a conventional repair configuration on vascularity and healing of tendon remains uncertain [25, 209-211,

218].

Potential problems with conventional knotted tendon repairs:

 Time consuming

 Nidus for infection

 Ischemia through constricting loops

 Knot slippage

 Spitting of knots

 Impaired healing due to knot

 Complexity of surgical procedure

Barbed suture tenorrhaphy (tenorrhaphy being tendon reconstruction or tendon repair) offers theoretical benefits to circumvent some of these issues. It may allow better distribution of loads along the entire intratendinous suture length, rather than localising stress loading to isolated anchoring or “locking zones”.

156

Also, the constricting forces may be reduced. Both factors theoretically could lead to less gapping of the loaded repair construct.

Theoretical benefits of unknotted barbed suture tendon repairs:

 Tension is broadly distributed and not concentrated at single stress areas?

 Less traumatic constricting forces?

 More even load distribution improves gapping resistance?

 Easier, less complex surgical procedure?

 Faster?

 No knot (no knot slippage, no knot spitting, no impaired healing)

Concepts of using barbed devices to repair flexor tendons were reported in the early

1950s by Jennings et al., 1952 [219] and Bunnell, 1954 [220], but the first scientific investigation of a barbed nylon suture for tendon repairs was published in 1967 by

McKenzie [221]. Whilst McKenzie confirmed a potential role for barbed sutures, further optimisation of the concept was not pursued by his group.

157

The recent commercialisation of barbed suture materials has reignited interest in the utility of a barbed material for various surgical applications, in particular wound closure

[222, 223], but also aesthetic facial surgery [224, 225], abdominal and gastrointestinal surgery [218, 226, 227] and gynaecological applications [228]. The main advantages of barbed sutures are claimed to be an even distribution of tension across the wound, no knot, and a potentially faster repair. These facets are all ideal for flexor tendon reconstruction and support the re-investigation of barbed sutures for this purpose.

This ex vivo study compares the bio mechanical performance of a new three dimensional knotless four-strand barbed suture flexor tendon repair to the four-strand cross locked cruciate technique (Adelaide repair) in a dynamic testing model. A dynamic investigation of a barbed suture tenorrhaphy focusing on gap formation and final load performance has not yet been published.

158

2.4.2 Materials and methods

40 deep flexor tendons were harvested from sheep forelimbs. This model has been reported to best mimic human flexor tendon biophysical properties [154]. A standardised point in zone II was marked 5mm proximal to the long vinculum using our previously described procedure [108]. Tendons were wrapped in phosphate buffered saline soaked gauze, placed inside plastic specimen bags and deep frozen at -20 degrees

Celsius.

On the day of experimentation, tendons were thawed at room temperature (22 – 25 degrees Celsius) and randomly allocated to two groups of n = 15 and one additional group of n = 10 tendons. The sample size of n = 15 per group was based on an a priori power assessment using a significance level of 5% and power of 80%. The thawed tendons were cut at zone II with a size 15 scalpel and immediately repaired according to the group allocation.

Tendons in the conventional repair group underwent a four-strand cross locked cruciate core repair (from now on referred to as “Adelaide repair”, see Figure 2-47) using a 3/0 silicone coated braided polyester suture (Ticron; Syneture, Norwalk, CT).

159

Figure 2-47: Illustration of the four-strand cross locked cruciate repair technique (Adelaide Repair). Tendon purchase and cross-lock width indicated in mm.

Suture purchase was 10mm and cross-locks were 4mm wide [108]. A four throw square knot was tied between tendon ends. An epitendinous suture was not performed.

Tendons in the barbed suture repair group underwent a novel three dimensional four- strand reconstruction (from now on referred to as “barbed suture 3D repair”) (Figure

2-48) using a 3/0 unidirectional barbed (Figure 2-49) glycolic-carbonate suture (V-Loc

180TM; Covidien. Mansfield, MA).

160

Figure 2-48: Illustration of the three dimensional knotless four-strand unidirectional barbed suture repair technique (barbed suture 3D repair). Tendon purchase indicated in mm. a) First strand runs in the horizontal plane. Note, it passes through the suture loop at the end of the suture thread (see 3). b) Second strand (see 5) runs from horizontal into vertical plane (three dimensional repair). Third strand (see 7) runs in the vertical plane. c) Fourth strand stays in the vertical plane and finishes with a triple zigzag. d) Suture is cut flush with the tendon.

161

Figure 2-49: Scanning Electron Microscopy (SEM) picture of the used 3/0 unidirectional barbed glycolic-carbonate suture (V-Loc 180TM; Covidien. Mansfield, MA).

Again a suture purchase of 10mm was used, and no epitendinous suture was applied. No knot was required.

162

To complete and further analyse this comparison an additional group of 10 tendons repaired with the above described three dimensional four-strand repair method, but using a conventional 3/0 silicone coated braided polyester suture (Ticron; Syneture,

Norwalk, CT) and a 5 throw square knot between the tendon ends (from now on referred to as “conventional suture 3D repair”) was added.

As in the author's previous study repaired tendons were mechanically tested with the

MTS 858 Mini Bionix material testing machine (MTS Systems Corp. Eden Prairie,

MN). Uniaxial and cyclical testing was performed. Again testing gauge length was kept at 40mm and loads between 3 and 30 Newton were applied for 250 cycles (reflective of the dynamic stress during early active rehabilitation). Again to evaluate gap formation during this dynamic phase of loading, testing was paused at 10, 20, 30, 40, 50, 100, 150,

200 and 250 cycles under the application of a static 3N load and digital images were taken. Figure 2-50 and Figure 2-5151 show the testing sequence and the testing setup.

163

164

Again the primary endpoints were gap formation during dynamic loading (mm), ultimate failure load (N) and mode of final failure (suture pull out or breakage). An additional endpoint was added to this testing setup: "number of cycles to 2mm gap formation". The number of cycles necessary to create a 2mm gap was calculated by correlating the recorded gap data with the amount of cycles at each measurement point

(logarithmic interpolation).

165

2.4.3 Results

2.4.3.1 Gap formation

Mean gap was smaller at each cycle point for the barbed suture 3D repair group compared to the Adelaide repair group (figure 2-52). The differences in gap formation were statistically significant at 10 cycles (p<0.001), 20 cycles (p=0.007), 30 cycles

(p=0.018) and 250 cycles (p=0.031) of dynamic testing. The additional repair group using the conventional suture 3D repair showed significant less resistance to gap formation at all cycle points of dynamic testing in comparison to both other groups

(Figure 2-52).

166

Figure 2-52: Gapping in mm per load cycle.

167

2.4.3.2 Ultimate load

The ultimate loads on failure testing after 250 cycles were 61.5± 9.3N for the barbed suture 3D repair group compared to 48.6 ±5.9N for the Adelaide repair group (p =

0.002) and 47.2 ±6.3N for the conventional suture 3D repair group respectively

(p<0.001) (Figure 2-53).

There was no significant difference in ultimate failure force between the Adelaide repair group and the conventional suture 3D repair group (p = 0.623). 168

2.4.3.3 Cycles to 2 mm gap

A logarithmic regression equation was applied to the data and used to calculate the number of cycles necessary to create a 2mm gap (figure 2-54).

Figure 2-54 Example logarithmic regression calculation (black line). From the logarithmic regression for each group, the number of cycles to 2mm gap was interpolated.

169

The Adelaide repair group required 24.4 cycles (y = 0.6501ln(x) - 0.0754, R² = 0.9251) of loading to gap by 2mm. The outcome for the calculation for the barbed suture 3D group was 102.5 cycles (y = 0.6711ln(x) - 1.1073, R² = 0.9383). In contrast to the conventional suture 3D repair group that gapped by 2mm already after an average of 4.7 cycles of loading (y = 0.9791ln(x) + 0.4936, R² = 0.9759) (Figure 2-55).

Figure 2-55: Average amount of loading cycles (3N to 30N) to cause a 2mm gap.

170

2.4.3.4 Mode of final failure

In the Adelaide repair group one suture pull out was noted and fourteen repairs failed by suture breakage. Thirteen sutures broke at the knot, one suture broke adjacent to the knot. In the barbed suture 3D repair group, 5 repairs failed by suture pull out and 10 by suture breakage. In the conventional suture 3D repair group 4 pull outs were noted and 6 repairs failed by breakage of suture material at knot.

2.4.4 Discussion

This ex vivo study supports a potential role for barbed suture materials in tendon repair surgery. The barbed suture 3D reconstruction demonstrated an improvement in ultimate repair strength of 27% compared to the Adelaide technique. Additionally, a significant reduction in gap formation throughout the testing was noted in the barbed suture group.

The conventional Adelaide repairs resisted an average of 24 cycles of loading before gapping by 2mm, the barbed suture repairs resisted an average of 103 cycles of loading before gapping by 2mm. The control group with the conventional suture 3D repairs using a knot performed significant worse than the main testing groups in regards to gap formation. A 2mm gap was already noticeable after an average of 5 cycles of loading. In regards to final failure loads the conventional suture 3D repairs showed similar loads to that of the Adelaide repair group.

171

It should be noted that no circumferential stitch was applied to any of the repairs of all three groups to exaggerate the behavior of the core suture materials and methods tested in isolation. An additional circumferential suture would only have masked differences and not helped to investigate the principle differences in these repair constructs.

The Adelaide repair technique performed with a braided polyester suture was chosen as the comparator because of its common use and high regard amongst several authors for superior performance and stability in comparison to other four-strand repairs [22, 60, 98,

102, 201, 202].

Two barbed suture materials are commercially available. The Quill™ device

(Angiotech. Vancouver, BC, Canada) is a bidirectional, double armed, barbed suture material and the V-Loc™ device (Covidien. Mansfield, MA, United States) is a unidirectional, single armed, barbed suture. The literature examining the use of these materials in tendon surgery is limited (table 3).

172

Table 3 Overview of published reports on barbed suture tendon repair studies.

I -- cycles N 84N . 22.2 25N/30N 62 32N t t 2mmgap at at a a loading After103

. . 6N . ailure f 145 88N 38N / /

I 6N 5N ce . . r 20N-25N 29 61 38N 30N 7239N SON Final fo .36N .

. . . . . dynamic / ng Static static static static static static static static static +dynamic testi .

no / yes repair no no no no no no no ~

. · ! unknotted unknotted unknotted unknotted unknotted unknotted unknotted unknotted kriorted Knotted · . · 0 0 0 / / / 1/0 4/0 2/0 310 2 3/0 2 3 .

-Loc) -Loc) . (Quill) (Quill) (Quill) (Quill) 01 01 .

material ectional r Suture QO!~glyconate Qolyglyconate bidirectional bidirectional polypropylene bidirectional polypropylene monodi[ectional bidirectional bidirectional polypropylene bidi polypropylene monodirectional nylon ~(Quill) . ·

. · · four / Strands four four four two two threelsix two two/four . . . . . own own own . passes Kessler Kessler Kessler Kessler . . . simple authors authors authors Corerepair method mod. mod mod mod . .

. . 1967 Year 2011 2011 2011 2011 2009 2009 2012 .

. . . . . al eta!: al et in al. et tal. . Marrero-~etal

Publication McKenzie McClellan Parikh Buschmiu ~et ~etal. ~e .

173

Most studies use the bidirectional suture device in tendon experiments. Little data is reported on the monodirectional device. There are no studies available reporting on the biomechanical behavior of barbed suture repair constructs when put under repetitive cycling loads to investigate the gapping performance of such repair constructs.

Parikh et al. investigated the use of the Quill device for tenorrhaphy in an ex vivo analysis using a novel repair method with either three or six-strands [229]. They reported loads to failure of 36N for their three-strand barbed suture repair and 88N for their six-strand repair method (no peripheral repair was performed in any group). The six-strand variation, constructed using two Quill devices, performed significantly better than the conventional four-strand cruciate repair. The three-strand barbed suture technique, sutured with one Quill device, was not different to the conventional four- strand repair. Since a 2/0 suture and a three or six-strand method was used for the barbed repairs and 4/0 sutures and a four-strand method was used for the conventional repairs, comparison is limited. More importantly, no dynamic testing or quantitative evaluation of gap resistance was carried out.

Trocchia et al. performed a modified Kessler-Bunnell two strand tendon repair with the

2/0 Quill bidirectional suture device and compared it to a conventional modified Kessler repair with 3/0 Ethibond.

174

In their static testing scenario the two strand barbed suture repair was weaker than the conventional two strand modified Kessler repair, but they did not demonstrate a difference in 2mm gap forces between groups. No dynamic testing was carried out

[230].

Zeplin et al. investigated ultimate failure forces of a four-strand modified double Kessler repair performed with either 3/0 conventional sutures or 3/0 V-Loc barbed sutures, the same device tested in our study. The mean ultimate repair strengths published by this group were in the order of 145.6N for the barbed suture repairs and 149.5N for the conventional repairs. No significant difference between groups was demonstrated.

Nevertheless, these reported loads appear significantly higher than those reported in the literature. This may be explained by their unconventional testing methodology considering the tendons were repaired end-to-end, creating a looped testing configuration. Comparisons to this data are therefore rather complex. This study did not utilise dynamic testing nor comment on gap formation [231].

Marrero-Amadeo et al. investigated deep flexor tendon repairs using a Quill bidirectional barbed suture device in an ex vivo cadaver model.

175

Their unusual approach of testing a four-strand Tajima repair augmented with a circumferential epitendinous repair versus a four-strand modified double Kessler barbed suture repair, in contrast not augmented by a circumferential repair, showed no significant differences in groups in regards to final load or 2mm gapping. No dynamic testing was carried out [232].

McClellan et al. showed similar biomechanical performance for their knotless four- strand barbed suture repair construct in comparison to a modified four-strand Savage repair. Although they used two Quill suture devices for each barbed suture repair in their static ex vivo testing scenario, they could show that in comparison to the conventional knotted repairs the cross sectional area of the barbed suture repairs was significantly smaller after testing [233].

On balance there is modest support for McKensie’s positive conclusions regarding the use of barbed sutures in tendon repairs half a century ago. Nonetheless, many studies have reiterated the importance of gap formation for tendon repair failures [60, 128, 140,

148, 213, 234]. Many studies have also pointed out the importance of using a dynamic testing protocol to safely make a statement about gap formation [93, 99, 113, 122, 123,

155, 213]. The author could find no study applying a dynamic testing protocol to barbed suture repairs.

176

A common problem in biomechanical tendon repair studies is that authors do not specify how they access the load at the 2mm gapping point, or how they measure the 2mm gap itself. We feel confident that with our method of accessing the gap of the repair site at several time points with exact digital gap measurements it was possible to precisely analyse this endpoint.

A principal benefit of the barbed suture is that tension is broadly distributed along the intratendinous path of the suture, and not concentrated at localized, high-stress points/areas. This is an important factor for resisting gap formation. Furthermore, the author believes that merely applying a traditional repair configuration with barbed devices fails to draw the greatest utility from this material. Traditional repair techniques are designed for conventional sutures and derive their performance by the number and configuration of holding zones in the tendon [63, 100], as well as relying on knotted configurations [93, 103, 216, 217]. With the knotted conventional suture 3D repair group it could be shown that a conventional suture four-strand repair construct without locking configurations might hold final failure loads comparable to constructs with locking configurations, but in regards to gap formation the unlocked construct shows significant inferior stability. Many studies on improving the design and geometry of these locking zones have been carried out [22, 62, 63, 93, 105, 107, 108, 148, 150, 159,

195, 234, 235]. Still, these holding zones are sites of stress concentration. They tighten and are the points of cheese-wiring and constriction, which inevitably leads to gapping.

177

It is likely that a sufficient/threshold number of barbs need to be “buried” within the tendon, which can be achieved by increasing suture purchase [105, 235] or increasing the number of strands [64, 92, 93, 214]. To make comparisons consistent, the author chose to use four-strands and a common suture purchase of 10mm for all three groups.

Barbed suture repair constructs with less strands and/or less suture purchase length may produce insufficient repair stability [230, 231, 233, 236].

The author believed that repairing tendons in a three dimensional fashion using the whole body of the tendon for maximal load distribution and barb interaction with tendon fibrils helps to improve barbed suture tendon repairs. As shown with the control group this does not apply to conventional suture constructs to the same degree.

Exact tendon apposition and equal tensioning on each strand during repair is important for the repair performance [95], but difficult to achieve in practice when using conventional multi strand tendon repair methods. The author noted that tendon apposition and equal strand tensioning is more easily maintained during repair with barbed sutures since sustained apposition of tendon ends can already be achieved after the first repair pass. Although the illustration of the authors repair construct looks complicated, it essentially consists of four simple passes through the tendon body in two planes and can be performed very quickly. The placement in two planes is easily achieved by rotating the tendon 90° to the left and right.

178

In spite of the attraction of barbed material, several downsides remain. The used 3/0 barbed suture in fact has the calibre of a 2/0 material, with the strength of a 3/0 suture.

In general, exposure of barbs on the tendon surface is a concern. There is a potential for these to abrade the lining of the flexor sheath, adversely affecting gliding resistance as well as promoting adhesions. Exposure can be minimised but not entirely avoided by reentering the tendon very closely to the exit point of the suture at each pass. Perhaps another caveat is the current limitations of the available barbed material. Most devices are absorbable and primarily marketed for wound closure. The needle has a cutting edge and hence is not ideal for tenorrhaphy.

In conclusion, this study shows promising biomechanical characteristics of the barbed suture tendon repair method. Barbed sutures better distribute loads throughout the tendon substance and therefore help to produce a stable repair with improved resistance to gap formation. This might be an important benefit for the positive outcome of early active rehabilitation regimes. But further refinements in the material and in vivo studies will be necessary before its use can be appraised in a common clinical setting.

179

CHAPTER 3: In vivo studies

3.1 A new tendon model for ex vivo and in vivo experiments:

The turkey foot

3.1.1 Introduction

Flexor tendon repairs are still a subject of much scientific research. As described in chapter 2.1 animal tendons can be a good substitute for human flexor tendons in biomechanical laboratory experiments. Many ex vivo studies have been conducted using a variety of animal species. The ideal substitute would resemble the biomechanical properties of human tendons. As shown, the sheep laboratory tendon model reflects this requirement best and is the ideal animal model for ex vivo tendon repair experiments (see chapter 2.1).

But tendon healing and adhesion formation is too complex to replicate in in vitro or ex vivo laboratory studies. Only an in vivo model can show healing propensities and final functional outcome of different repair methods.

180

And only the combination of ex vivo and in vivo experimental data can prove new concepts and open the door to clinical trials. However, current in vivo animal models for tendon repairs like the chicken or rabbit model have tendons too small to perform modern multi-strand tendon repairs.

Figure 3-1 Chicken tendon model

In our laboratory an analysis of three laboratory chickens (six chicken feet) deep flexor tendons showed an average AP diameter in zone II of 1.1mm (±0.16mm) and a lateral diameter of 2.5mm (±0.17mm).

181

This tendon calibre only allows two strand repair techniques or multi-strand techniques using significantly smaller suturing material compared to human scenarios (figure 3-1).

This has been reported in the literature previously [94]. Therefore it is not an ideal model for investigating modern multi strand repair techniques. The same accounts for the rabbit tendon model.

Figure 3-2 Dissected rabbit rear paw.

The flexor tendons of the rear paw of three laboratory rabbits (total of 6 paws) were analysed. Tendons had a more round aspect than chicken tendons, but the diameter was only marginally bigger: AP 1.5mm (±0.2mm) and Lateral 2.1mm (±0.18mm).

182

Again the size of the tendon is not appropriate to perform multi strand tendon repairs without applying significant changes to the technique or suturing material as in human scenarios. Additionally the rabbits rear paw flexor tendon anatomy is not comparable to the human deep flexor apparatus in the hand (figure 3-2). The forepaw's flexor tendons are even smaller and further the anatomy regarding deep and superficial tendon arrangement, sheath, pulleys and vinculae is different.

Some studies employed the repair of rabbit achilles tendons instead. But in the author's opinion this is not comparable to the repair of a human deep flexor tendon.

The rabbit's Achilles tendon is multi fascicled, has no tendon sheath and is also different in its anatomy and biomechanical function (figure 3-3 and 3-4).

Figure 3-3 Dissected rabbit leg with flexor tendon and Achilles tendon marked. 183

Figure 3-4 Achilles tendon with gastrocnemic and soleus tendon.

Furthermore, the rabbit Achilles tendon model only provides tendon widths of about

4mm. This is barely enough tendon width to perform a modern four-strand repair.

184

And especially four-strand tendon repairs performed with a barbed suture require tendon width of at least 4-5mm diameter to guarantee sufficient stability.

Sheep and pig trotters are commonly used in ex vivo experiments and, as described (see chapter 2.1), have similar size, anatomy and biomechanical properties in comparison to human tendons, but surgical access to the deep flexor apparatus is not practicable in an in vivo setting.

Figure 3-5 Dissection of sheep trotter (left) and pig trotter (right).

The tendons are hidden deep in the hoof and surgical access would cause too much disruption of anatomical structures (figure 3-5). The operation would be difficult to perform, it would take considerable time, the animal's recovery could be complicated and the wellbeing of the animal could not be assured.

185

Theoretically it is possible to use dog flexor tendons for tendon repair experiments.

Inbred laboratory dogs provide flexor tendons in the range between 4-5mm in diameter but the comparability to human flexor tendon anatomy or physiology is limited.

Furthermore availability is a problem. Also the possibility to use dogs in in vivo research experiments is limited to a small number of licensed institutions worldwide and the use of dogs in research is still ethically questionable. The same problems account for primates. Additionally most western countries generally do not allow the use of primates for invasive in vivo experiments.

In summary there is no animal model available that can be used in ex vivo and at the same time in vivo settings that replicates human tendon size, anatomy and physiology.

Thus, such a model would be required to further investigate our introduced knotless four-strand 3D barbed suture repair method.

Therefore it became necessary to develop a new animal model for ex vivo and in vivo experiments: The Turkey Tendon Model.

186

Just like pigs or sheep, turkeys are easy to obtain, they are relatively cheap and they are uncomplicated to farm, making it an attractive model for ex vivo and in vivo studies.

Figure 3-6 Male Turkey (Meleagris Gallopavo) 187

Also in regards to size and anatomy the turkey foot is similar to a human hand (figure 3-

7 and 3-8).

Figure 3-7 Male adult turkey leg

The middle toe reflects comparable sizes to human fingers (figure 3-8).

Figure 3-8 Middle toe of an adult male turkey. Note similar size to human finger. 188

Dissection of a turkey digit reveals similar anatomical landmarks as in the human finger

(figure 3-9).

Figure 3-9 Dissected adult turkey middle toe with paired neurovascular bundles

Also the composition of flexor aparatus is comparable to the one in human fingers.

The superficial flexor tendon (FDS) forms a chiasma at the level of the proximal phlanx where the deep flexor tendon passes through. Insertion of both tendons is similar to human fingers at the middle phalanx for the FDS and the distal phalanx for the FDP tendon respectively (figure 3-10).

189

Figure 3-10 Dissected male turkey middle toe with FDS tendon, FDP tendon and chiasma tendineum

Like in the human hand the tendons run through a synovial sheath providing a gliding surface. Also the annular pulley system is constructed comparably to the human pulley system. Long and short vincula are more dominantly developed in the turkey foot but origin and insertion are similar to human deep flexor tendons (figure 3-11).

190

Figure 3-11 Dissected male turkey middle toe, short vinculum (SV), long vinculum (LV), and annular pulleys (A1-4) are marked.

191

Furthermore the bony skeleton with proximal, middle and distal phalanx and Proximal

Inter Phalangeal (PIP) joint as well as Distal Inter Phalangeal (DIP) joint shows the comparable composition as the human digit (see figure 3-12).

Figure 3-12 AP and Lateral radiographs of male turkey middle toe, clamp indicates long vinculum.

It must be noted that if the claw itself is considered as the distal phalanx, the turkey skeleton with its four phalangeal bones, shows some difference to the human digit with its 3 phalangeal bones. There are also some additional flexor tendons present for the flexion of the first proximal phalanx, nevertheless if the last three phalangeal bones are considered as digit, the similarity to a human digit is surprising.

192

The dissection of a turkey PIP joint showed comparable size and anatomy to human PIP joints (see figure 3-13). Nevertheless the bone density seems to be lower than in human phalanges and the trabecular structure shows differences to that of human bones (figure

3-14 and 3-15).

Figure 3-13 Dissected PIP joint with collateral ligament

193

Figure 3-14 Turkey PIP joint AP radiograph. Note lower bone density than in human phalanges.

Figure 3-15 Turkey PIP joint in lateral projection. Again, note lower bone density than in human phalanges.

Access to the turkey tendon is practicable and comparable to surgery in human fingers

(figure 3-16).

194

Figure 3-16 Volar (plantar) access to tendon through Brunner incisions

To see if it is possible to perform the 3D knotless barbed suture repair method from chapter 2.4 in the turkey foot an ex vivo pilot study with five turkey deep flexor tendon repairs using the 3D barbed suture repair method was conducted (figures 3-17, 3-18, 3-

19).

195

Figure 3-17 Transected deep flexor tendon in zone II

Figure 3-18 Unknotted 3D four-strand core repair with 4/0 barbed suture (technique see chapter 2.4)

196

Figure 3-19 Finished with a 6/0 Prolene simple running suture. Stable repair of deep flexor tendon , "no" bulk.

In conclusion, our first ex vivo repairs in the turkey foot showed no problems regarding surgical access, suturing methods or tendon size. The surgical procedure was directly comparable to zone II flexor tendon repairs in humans. The 4/0 barbed suture was easy to use in the turkey tendon and repairs were quick to perform.

This following study on turkey tendons evaluates the turkey model in regards to dimensions and biomechanical properties of the turkey deep flexor tendons. Results are compared to the gold standard in animal tendon models, the sheep tendon model.

197

3.1.2 Methods

25 flexor digitorum profundus tendons were harvested from the middle toe of adult turkey cadavers. Respectively 25 deep flexor tendons from adult sheep cadavers were harvested (as previously described, see chapter 2.1) [108]. All animals were obtained from local abattoirs and had the same age. The distance of the musculotendinous junction to the insertion of the tendon on the distal phalanx was measured. The site of transection in zone II was marked at a constant point 5 mm proximal to the long vinculum insertion for both species. Cross sectional diameter (widest, medio-lateral, diameter) was also measured at this point for all tendons. To determine exact measurements of distances and lengths a digital caliper (MitutoyoCD-6 –inch CS;

Absolute Digimatic,Tokyo, Japan) was used. The tendons were then individually wrapped in saline soaked gauze and deep frozen at -18 degrees Celsius until the day of experimentation. On the day of mechanical testing, each tendon was thawed to room temperature. A sharp transverse laceration of the tendon was performed at the pre marked site, using a number 22 scalpel blade. All tendons were then repaired with a 4- strand cross-locked cruciate suturing technique (Adelaide technique).

198

Figure 3-20 Cross Locked Cruciate (Adelaide) tendon repair technique.

The same surgeon (T.P.) using 3.5x loupe magnification accomplished each repair with

4-0 silicone-coated braided polyester sutures (Ticron; Tyco, Lane Cove, NSW,

Australia). The cross-locks were uniformly set 10 mm from the line of tendon division

(10mm tendon purchase). All repairs were completed with a 6-0 polypropylene

(Prolene; Ethicon, Somerville, NJ) simple running peripheral suture.

199

Figure 3-21: Left: Harvested Turkey tendon. Right: Repaired Turkey tendon

3.1.2.1 Mechanical Testing

The repaired tendons were tested in a uniaxial manner using an MTS 858 Mini Bionix materials testing machine (MTS Systems Corp., Eden Prairie, MN). Tendons were gripped in grooved pneumatic clamps with a gripping pressure of 60 psi. Gauge length was standardized at 40 mm. Samples were preloaded to 1.5 N, then tested to failure at a relative slow distraction rate of 10 mm/min. Load, displacement, and time were continuously recorded at 100 Hz. Data were used to generate load-displacement graphs for each tendon. In addition, a digital caliper fixed at 2.0 mm was placed adjacent to the repair site as a reference.

200

The repair site was continuously filmed during the mechanical test using a high- definition Sony Handycam camcorder (Sony Corp., To- kyo, Japan) to monitor gap formation.

The timing of the video was coupled with the data recorded by the testing machine. This method allowed an exact visual and biomechanical correlation of 2-mm gapping. Load to failure (N), load to 2-mm gap formation (N) and mechanism of failure were determined for all samples. The video and mechanical data were correlated after each test to determine the load to first point of 2-mm gapping. The mechanism of failure was recorded as suture breakage at the knot, suture breakage away from the knot, or suture pullout. The specimens were kept moist during testing by intermitted buffered saline spraying, to prevent desiccation.

3.1.2.2 Statistical Analyses

All analyses were carried out using SPSS version 18.0 (SPSS,Inc., Chicago, IL, USA).

Data were described as mean (± standard deviation). In order to decrease bias and increase power of the analyses we used multiple imputation (MI) chain equations to impute missing values on the variables "load to 2mm gap" (n=2) and "diameter" (n =

11) – 6% of all required values. Briefly, the imputation model for each of these variables used the information provided by all the other variables as predictors by appropriate regression models.

201

Five imputed datasets were generated and used to perform the comparisons of interest; results reported were those retrieved from pooled analyses on all five imputed datasets.

Comparisons of load to 2 mm gap, load to failure and diameter between animals were tested and the magnitude of the differences (with respective 95% confidence intervals) was estimated with multivariate regression models. Statistical significance was set at p<0.05.

3.1.3 Results

Measurements:

The mean tendon length from the most proximal muscultendinous junction to the insertion on the distal phalanx measured 102mm (range 92-130mm) for turkey tendons and 98mm (range 86-113mm) for sheep tendons. The average cross sectional medio- lateral diameter at the chosen transection site was 5.4mm (range 4.8-5.5mm) in the turkey tendons and 5.4mm (range 3.7-5.6mm) for sheep tendons.

Mechanical testing

Due to mechanical problems in the testing machine, 4 turkey tendons were not included in the analysis.

202

Therefore the total group of tested turkey tendons was n=21 and total number of sheep tendons tested was n=25. Ultimate load to failure of the tendon repairs was 71.6N

(±12.0N) in the turkey tendons and 67.3N (±11.6N) in the sheep tendons.

2mm gapping occurred at 66.0N (±15.9N) in turkey tendons and 57.6N (±19.3N) in sheep tendons. The mechanism of failure was core suture gapping followed by peripheral suture pullout and eventually core suture breakage at the knot in all specimens. When considering the full data set of 46 tendons (21 turkey and 25 sheep) the load to 2mm gap and ultimate load to failure did not differ significantly between turkey and sheep.

3.1.4 Discussion

Animal flexor tendons are common substitutes for human flexor tendons in ex vivo biomechanical experiments. The ideal tendon model would resemble the biomechanical properties of human flexor tendons, have a comparable size, be inexpensive and easy to obtain. As could be shown in chapter 2.1, sheep tendons resemble the human flexor tendons bio-properties better than pig tendons in biomechanical testing scenarios [154], although pig tendons have been favoured for years [105]. Both tendon models, the pig and the sheep are relatively easy and cheap to obtain from either animal laboratories that use these animals for other experiments, or from local abattoirs.

203

As previously discussed all other available animal tendon models, like mouse, rat, chicken, rabbit, dog or primates either provide tendons too small to perform modern multi strand tendon repairs or ethic regulations forbid to use the species. Turkey tendons, just like sheep and pig tendons are relatively easy to obtain and turkey deep flexor tendons have a similar size to human tendons.

This present study’s results show that the dimensions of the adult turkey deep flexor tendons are comparable with sheep deep flexor tendons. Recorded lengths from the proximal musculotendinous junction to the insertion at the distal phalanx were comparable between turkey and sheep. No significant difference in length between species was found. More importantly, the medio-lateral width at Zone II showed no significant difference between turkey tendons and sheep tendons. When comparing these tendon dimensions to human flexor tendon sizes, equal diameters are reported in literature (table 4).

204

Table 4 Overview of published tendon dimensions.

Species Location of Diameter mm transection (medio-lateral) Boyer 2001 [136] Human Zone II 4,5 (±0,4) – 4,7 (±0,7) FDP Havulinna 2011 Human Zone I,II,III Mean 4,5 (±0,5) [237] FDP Havulinna 2011 Pig FDP Zone I,II,III Mean 3,6 (±0,4) [237] Mao 2011 [152] Pig FDP Zone II 6.4 (±0,5) - 7.1 (±0,2) Peltz TS 2014 Sheep Zone II 5,4 (3,7-5,6) FDP Peltz TS 2014 Turkey Zone II 4,9 (4,8-5,5) FDP

Besides the analysis of comparable size, the first step to establish this new tendon model was to prove the biomechanical comparability. Therefore the turkey tendon model was tested versus the current gold standard in tendon repair animal models, the sheep tendon model. The author repaired 25 sheep deep flexor tendons and 25 turkey deep flexor tendons with a four-strand Adelaide repair technique, augmented with a simple running circumferential repair. Tendons were then tested to 2mm gap formation and complete failure.

205

In this study the biomechanical testing of the repaired turkey and sheep tendons showed a trend to higher resistance to 2mm gap formation and repair failure in the turkey tendons, but no significant difference in load to 2mm gap formation or ultimate load to failure could be proven. The mechanism of failure was identical in all samples of both groups.

Boyer et al. conducted a study in which they tested the influence of tendon width on repair strength in human cadaveric flexor tendons. In this study the authors found a direct correlation of tendon width and pull out strength when testing a simple double stranded suture loop passed through the tendon but could not find any significant differences in repair strength between different sized tendons when using a double stranded modified Kessler repair [136].

In this present study, the medio-lateral diameters of the sheep and turkey tendons were similar, no significant size differences could be seen. Also the relation between diameter and load to 2mm gapping or load to failure showed no statistical significance in sheep or turkey tendons.

206

Not much has been published on turkeys regarding this animal's use in laboratory experiments. And no study investigating the deep flexor tendon apparatus of the turkey could be found. Therefore this is the first study to investigate the exact anatomical construction of the turkey flexor apparatus and this is the first study reporting on the biomechanical comparability of turkey tendons.

Nevertheless, the turkey tendons seem to show a trend to more resistance to gapping and repair failure, even this could not be proven to be significant in our experiment. A possible explanation for this could be a higher calcification of the turkey tendons.

Landis et al. investigated the gastrocnemius tendon of turkeys [238]. Like in many avian species the gastrocnemius tendon of turkeys is calcified in its central portion. In our dissections we could also demonstrate this phenomenon (figure 3-22).

207

Figure 3-22 Dissection of a turkey gastrocnemius tendon with calcified area marked.

This tendon mineralisation is common but is predominantly limited to certain areas like the gastrocnemius tendon's central part, the proximal tendon-muscle intersection is free of calcification, the same accounts for the distal part of the tendon where it tunnels into the foot. This mineralization of the turkey gastrocnemius tendon is an efficient means of preserving elastic energy storage while providing for increased load-bearing ability required for locomotion of adult birds [239]. Other reasons for calcifications in tendons can be pathological conditions or following injury [240]. No reports commenting on calcifications in the deep flexor tendons of turkeys could be found and in this present study it was not possible to show any calcified areas in the turkey deep flexor tendons.

208

Cross sectional histology cuts showed a similar biological architecture between sheep

and turkey deep flexor tendons. Fascicles had similar sizes and cellular content. But

turkey tendons showed a slight denser and more streaky overall composition. The

epitenon and endotenon seemed to be more fibrous and dominant than in the sheep

tendons (figure 3-23 and 3-24).

Figure 3-23 H&E cross sections of sheep deep flexor tendon (left) and turkey deep flexor tendon (right) (2X magnified).

209

Figure 3-24 Comparison of human, sheep, pig and turkey tendon histology (H&E cross sections 4X magnified)

Further studies focussing on the turkey deep flexor tendon histology have to investigate if these described differences in histological appearance in the turkey tendon are related to the slight trend to higher resistance for gap formation and failure force found in our experiment for the turkey tendons.

210

In conclusion it can be postulated that the turkey foot is an excellent model to teach and train tendon surgery. Not only the general anatomy of the turkey digit but also the flexor apparatus itself shows similar composition of anatomical landmarks compared to the human flexor apparatus.

Regarding laboratory experiments, the turkey deep flexor tendon is comparable to the sheep deep flexor tendon model. Both models show similar tendon sizes and biomechanical strength of suture repair, albeit a non-significant trend to higher resistance for repair failure in the turkey tendons must be noted.

The turkey flexor tendon model is an adequate alternative animal model, it is an excellent teaching and training model and it can be used in biomechanical studies on flexor tendon repairs. Additionally the turkey tendon model is relatively easy and inexpensive to obtain.

211

3.2 Surgery on the turkey foot

3.2.1 Introduction

As discussed before, tendon healing and adhesion formation is too complex to replicate in in-vitro or ex vivo laboratory studies. Albeit the possibility to conduct tendon healing experiments in in-vitro tissue cultures [55, 241-245], only an in vivo model can show healing propensities of repaired tendons in regards to repair stability, adhesion formations and final functional outcome.

We could not find any available in vivo testing model to investigate our new knotless barbed suture tendon repair technique (see chapter 2.4). Either animal tendons were too small (mouse, rat, chicken, rabbit) or ethical limitations forbid the experimentation

(dogs, primates). Indeed most in vivo tendon repair experiments are done in the canine tendon model [128, 209, 246-255].

212

But inbred laboratory dog populations for tendon repair experiments are only available in certain centres in the United States, primarily at the Mayo Clinic in Rochester,

Minnesota and at the Washington University in St. Louis, Missouri. In most other countries including Australia there exist ethics regulations preventing the use of this model for surgical tendon repair research.

Rabbits are probably the second most common laboratory animals used in in vivo tendon experiments [256-262]. But as described before the rabbit deep flexor tendons are too small to perform modern four-strand tendon repair techniques and the Achilles tendon anatomy is not comparable to the human deep flexor tendon anatomy. Therefore this model is predominantly used in experiments that do not investigate repair methods per se but focus on healing and biology.

Same accounts for chicken [263-268] and rat [245, 269-275]. Although due to the similar anatomy of the chicken deep flexor tendon model to the human deep flexor tendon anatomy, some studies employ the chicken model for in vivo surgery experiments focusing on repair techniques [94, 276, 277]. Nevertheless the chicken deep flexor tendons are comparably small (see chapter 3.1) and repairs in mentioned studies are performed with significant smaller calibre suture material. Also no modern multi strand configurations were investigated.

213

As shown before, sheep deep flexor tendons are ideal regarding size and biomechanics but surgical access to the deep flexor tendon is unfeasible. Therefore in vivo studies on sheep tendons, either investigate non-surgical procedures [278], focus on the sheep's

Achilles tendon [279], or the sheep's rotator cuff [280-283].

Turkeys (meleagris gallopavo) are not commonly used in laboratory settings, compared to other laboratory animals. Only few articles employing the turkey model are published. And most of these publications investigate animal specific topics [284-287], focus on anthropological questions [288-290] or relate to genetic subjects [291-294].

This study investigates the possibility of using turkeys for in vivo tendon repair experiments.

3.2.2 Methods

3.2.2.1 Animal housing and welfare

Authority to conduct a turkey animal research project was approved by the Animal Care and Ethics Committee (ACEC) of the University of New South Wales (ACEC Number:

12/70A). Planning of experiments, investigations and procedures were based on previously conducted work and experimental data from our chicken model [94, 295].

214

Approval was conditioned on conducting a pilot study on five turkeys and passing two

ACEC inspections. Also for this pilot study, animals had to be monitored continuously and all adverse events had to be reported to ACEC immediately.

Animal housing and procedure theatres were located at St. George Hospital Biological

Research Centre (BRC). This facility is part of the University of New South Wales

(UNSW) Faculty of Medicine.

Turkeys are the largest domesticated gallinaceous bird. Gallinaceous meaning “heavy bodied ground feeding bird” [296]. Marchewka et al. performed a review of the social and environmental factors affecting the behaviour and welfare of turkeys in modern rearing systems. According to the authors, space availability, group size, density and lighting are most important factors affecting the welfare of turkeys. Based on these recommendations, the research project was designed to meet the best possible standards and to assure the well-being of the animals [297]. Turkeys exhibit behaviours such as wing flapping, feather ruffling, leg stretching and rapid locomotion. All these activities require significant space. The BRC facilities provided sufficient space for the turkeys to allow expression of all these comfort behaviours. A dedicated room was fitted out to accommodate a flock of male turkeys in groups of 15 animals. The room had a floor area of 55 square meters (3.7sqm per bird) and was climate controlled with day/night lighting cycles [297]. Domestic turkeys are highly social and become very distressed when isolated [298]. Therefore isolation of any bird was avoided.

215

To prevent animals from social tension related behaviours (for example fighting, pecking, mating) a pre socialized single sex (male) flock/group was sourced which had been reared together prior to delivery.

Turkeys were housed on a deep-litter substrate (straw, wood shavings). Care was taken that there was sufficient substrate available to ensure adequate dilution of birds’ faeces and urine. Saturated substrate was removed regularly and replaced with fresh substrate.

As mentioned it is important to provide sufficient quality space to allow turkeys to express a range of natural behaviours (environmental enrichment). Therefore, in addition to provide sufficient space for the animals, bales of straw were placed in the room to encourage climbing and perching. Straw bales also make the birds environment more interesting and did provide a refuge from dominant birds. Turkeys dust bathe regularly when conditions are appropriate and this appears to be socially facilitated

[298]. Therefore sand trays were provided for dust bathing of birds. Feed was scattered in pecking substrates and bales of straw to encourage foraging behaviour. Fresh water and a commercial pelleted diet was offered ad-lib. Additionally diet was supplemented with fresh vegetables and some fruit on a daily basis.

15 male 16 week old turkeys (meleagris gallopavo) were attained from a local commercial breeder (Inghams Enterprises, Liverpool, NSW, Australia). Average weight of turkeys was 25.9Kg (±3.8Kg). Animals were acclimatized for 10 days to get used to the new environment. Turkeys were housed in a large group pen as described before.

For the initial pilot study only five animals were used.

216

On the day of procedure individual animals were separated from the flock and treated with heat lamps for 30 minutes. A single bird can be caught by hand if the flock is approached slowly and deliberately. When catching by hand, approach from the rear and a quick grasp at the base of the neck with other arm around and under the body is necessary. The wings are folded into their natural resting position and gently but firmly restrained. Turkeys can be restrained for induction/injections and transport using this technique. Great care was taken to apply the correct amount of restraint to ensure that breathing was not restricted or that the restraint itself did not cause injury. Competent handling is vital for the safety of the handler as well as the bird's safety. It is possible to reduce handling stress by habituating turkeys to human contact. We attempted to achieve this intensively in the acclimatization period and continued to habituate birds to animal carers and laboratory personnel throughout the project.

3.2.2.2 General Anaesthesia

Anaesthetic procedures followed recommendations of Harrison et al. and were based on previous experiences with chicken surgeries [94, 295, 299]. Prior to induction, a intra muscular injection of 5mg/kg Zoletil (Virbac Animal Care, Milperra NSW, Australia), was administered. Zoletil is a combination of a dissociative anaesthetic agent (tiletamine hypochloride) and a tranquilizer (zolazepam hypochloride). Medication was allowed 10-

15min to take effect.

217

Thereafter induction of general anaesthesia was performed by inhalation with gaseous

2% Isoflurane (Baxter Healthcare, Toongabbie NSW, Australia). Inhalation was achieved with minimal stress using a facemask applied over the beak and nares (figure

3-25).

Figure 3-25 Inhalation mask with oxygen and 2% Isoflurane for general anesthesia

Isoflurane was the agent of choice for anaesthesia. It is safe and the blood gas coefficient is very low (1.4 at 37 degrees Celsius) which allows rapid induction and rapid recovery from anaesthesia, with less retention in the body tissues [300].

218

Once unconscious, the bird was injected intra muscularly with 10mg/kg Carprofen

(Apex Laboratories, Somersby NSW, Australia), an anti-inflammatory and analgesic drug [301]. Furthermore we administered a prophylactic antibiotic (tetracycline) dose of

110mg/kg (Oxytet-200 LA: Troy-ilium, Glendennin, NSW, Australia) in the pectoral muscle of the bird. The surgical procedures were also adapted from our previous in vivo work in chickens (Ethics approval ACEC: 08/18A) [94, 295]. After positioning the

Turkey in prone position on the operating table with thick padding and a heat mat, the bird was fixed with two straps and a warm towel/blanket. Hypothermia can cause a poor and prolonged recovery, peripheral vasoconstriction, bradycardia and, when severe, ventricular fibrillation and shock [299]. When positioning the bird, again care was taken to not interfere with the birds breathing pattern. General anaesthesia was maintained on

2% Isoflurane and Oxygen with 2L/min flow administered via an in circuit anaesthetic machine. The depth of anaesthesia was continuously monitored and the levels of isoflurane and oxygen were adjusted accordingly. Parameters monitored at every five minutes were: blood oxygen saturation, end tidal CO2, heart rate, pulse rate, respiratory rate & depth and eye reflex. If the vitals became significantly abnormal the surgery would have been terminated and the animal removed from the isoflurane, kept on oxygen, stabilized and monitored closely. This was not necessary in our experiments.

Important note: Turkeys have to be positioned prone once anesthetized since their unique breathing physiology and chest anatomy would make them not able to breathe in supine positioning [299].

219

3.2.2.3 Surgery

After positioning, the surgical site (left foot) was thoroughly scrubbed using iodine soap to minimize risk of bacterial contamination (figure 3-26).

Figure 3-26 Bird in prone position with scrubbed extremity

After scrubbing the extremity was sprayed with iodine solution and positioned on sterile drape for the beginning of the operation (figure 3-27).

220

Figure 3-27 Sterile operating field

One surgeon (TP) performed all operations under 3.5X loop magnification. All procedures were conducted in aseptic fashion following the usual clinical asepsis protocols (figure 3-28). Surgical hand wash, sterile gloves, disposable sterile gowns and drapes were used with appropriate sterile techniques. All instruments and swabs were sterilized in an autoclave or came from pre-packaged sterile bags.

221

Figure 3-28 Sterile operating set up with bird in prone position.

To control bleeding during the operation a "finger tourniquet" was applied. This was done by cutting one finger of a sterile glove and using this rubber band to build a tight sleeve around the base of the turkeys middle digit. An artery clip was used to hold the tourniquet tight and in position (figure 3-29).

222

Figure 3-29 Application of tourniquet for middle digit.

Following this a approximately 2X1cm flap was marked volarly (plantarly) over the proximal interphalangeal (PIP) joint (distal part of zone II) (figure 3-30).

223

Figure 3-30 Marking of 2X1cm flap (arrow) over PIP joint.

Skin incisions were made with a size 15 scalpel and the deep flexor tendon was dissected. For better access parts of A2 and A3 pulley were incised. Care was taken to protect both neuro vascular bundles on the lateral aspects of the digit. Also the long vinculum was preserved.

224

The flexor digitorum profundus tendon was then mobilized, held in position with two 26 gauge needles and cut with scalpel 5mm proximal to the long vinculum insertion (figure

3-31).

Figure 3-31 Dissection, fixation and transection of deep flexor tendon

Tendon was then reconstructed with a four strand cross locked cruciate (Adelaide) repair technique using 4-0 silicone coated braided polyester sutures (Ticron: Tyco, Lane Cove,

NSW, Australia) for the core repair (figure 3-32 and 3-33).

225

Figure 3-32 Cross locked cruciate tendon repair technique (Adelaide technique)

Figure 3-33 Insertion of the four strand Adelaide core repair

226

After finishing of the core repair the transfixing needles were removed and a simple running circumfernetial epitendinous repair was performed using 6-0 polypropylene sutures (Prolene: Ethicon, Somerville, NJ, USA) (figure 3-34 and 3-35).

Figure 3-34 Finished core repair (Adelaide technique)

227

Figure 3-35 Finished composite repair: Adelaide core repair + simple running epentendinous repair

After completion of the tendon reconstruction repaired tendon was repositioned into sheath and skin was closed with 3/0 resorbable braided polyglycolic-acid sutures

(Vicryl: Ethicon, Somerville, NJ, USA) (figure 3-36).

228

Figure 3-36 Skin closure with 3/0 Vicryl sutures

Each surgical procedure took between 40 and 60 minutes. After the operation the wound was covered with gauzes, cotton wool padding and an elastic bandage (figure 3-37).

229

Figure 3-37 Dressed digit after operation

Additionally for immobilization a well padded dorsal splint was applied (figure 3-38).

The splint was made of Vet-lite material, a thermoplastic moldable splinting material

(Vet-lite. Runlite S.A., Hopper Crossing, Victoria).

Figure 3-38 Padding of extremity and application of dorsal plaster splint 230

3.2.2.4 Postoperative Care

After completion of surgical procedure the turkey was moved to a recovery table on a heat mat and oxygen flow (2.0L/min) maintained until the turkey was conscious at which time the bird was returned to the recovery/holding room on deep litter with access to heat lamp, climate control, food and water. Birds were continuously monitored by technicians until satisfied all turkeys have recovered uneventfully [298, 299] (figure 3-

39).

Figure 3-39 Operated turkey in recovery after operation.

231

Thereafter birds were daily individually monitored for the first week, then on a weekly basis. They were monitored for behaviour (alert, responsive, quiet, depressed), the surgical site/wound (swelling, discharge, redness, suture integrity), ambulation/posture, appetite for feed & water, whether they have been urinating and defecating normally (no diarrhea) and signs of pain (ruffled feathers, depressed and isolating themselves, not weight bearing on operated limb). Casts remained attached for three weeks. For experimental endpoints and results see chapter 3.3.

3.2.2.5 Euthanization

Birds were euthanized after 6 weeks. Animals were euthanized using appropriate methods in a designated room in the absence of any other bird [302]. Following an injection of 10mg/kg Zoletil (Virbac Animal Care, Milperra NSW, Australia), combination of a dissociative anaesthetic agent and a tranquilizer in the pectoral muscle

(providing calmness, pain reduction and immobilization) an intra venous injection of

Pentobarbital 1ml per 2.0kg body weight (Lethabarb: Virbac Animal Care, Milperra

NSW, Australia) was administered. The animal was then confirmed to be dead by checking for a heartbeat, respiratory movements, pulse, skin colour, eye reflexes and look of glazed eyes [302]. After euthanasia carcasses were again checked for any apparent illnesses and then transported to UNSW's Surgical and Orthopaedic Research

Laboratories (SORL).

232

After another inspection of the carcass and check of "certain signs of death" [cessation of breathing, pallor mortis (pale skin colour), livor mortis (no blood flow), algor mortis

(low body temperature), rigor mortis (stiff muscles) and absence of pupil reflexes] the birds legs were detached from body and marked. Then samples were either prepared for histology or biomechanical testing (for precise description of groups, endpoints and results see chapter 3.3).

3.2.3 Results and Discussion

Tendon healing and adhesion formation is too complex to be replicated in vitro or ex vivo. The behaviour of barbed sutures in tendon repairs in an in vivo setting is unknown.

The author could not find any in vivo animal model to test our new tendon repair method using these barbed sutures (see chapter 2.4). Therefore a new in vivo animal tendon model had to be introduced and validated.

16 week old male turkey birds were chosen for experiments considering the similar anatomy to the human flexor apparatus and the similar size to human tendons.

In-vitro studies in our laboratory showed good biomechanical comparability of turkey tendons to sheep and respectively to human tendons (see chapter 3.1).

233

Female birds and their flexor tendons are significantly smaller and would not have been suitable for modern multi-strand tendon repair techniques [303].

To validate the possibility to operate on turkeys in general and to investigate the turkey in vivo tendon model in detail, as well as to meet ethics requirements, initially a pilot study with only five birds was conducted. For these first five deep flexor tendon reconstructions in a turkey deep flexor tendon we used the current gold standard in multi strand tendon repairs (Adelaide repair) to minimize the amount of unknown variables.

In vivo surgery on the turkey deep flexor apparatus is practicable and procedures are comparable to clinical tendon repair scenarios in humans. Surgeries of all five birds went without adverse events. Average operating time was 46.6 (±6.1) minutes. Average core repair took 7.9 (±1.6) minutes and average epitendinous repair took 9.2 (±1.2) minutes. Intra operative bleeding was no issue with the used tourniquet technique. Also in the postoperative period until the end of experiment, no bleeding, infection or other diseases were noted. Animals showed uncomplicated recovery after surgery and tolerated casting. Operated birds were healthy, were responsive to external stimuli and behaved like the rest of the not operated flock.

They showed typical behaviours like feather fluffing, emitting hissing and gobbling sounds [297] and birds were ambulating on operated leg without signs of lameness.

234

The outcomes of the tendon repair experiments of these first five operated birds are included in the study of the next chapter (chapter 3.3). Please also refer to that chapter for precise description of groups, endpoints and results.

3.3 The knotless barbed suture flexor tendon repair: An in

vivo comparison in the turkey foot

3.3.1 Introduction

Since the establishment of deep flexor tendon repairs in zone II by Kleinert in the late

60ies [53] techniques for repairing flexor tendons have passed through multiple stages of evolution. Clinical results are more consistent and reliable with modern multi strand reconstruction methods [64, 66, 182, 183]. But the main concept of tendon repairs remained the same. A more or less smooth suture is used to anchor the tendon in its core and knots secure the tensioned construct. Additionally a circumferential, “tidy up” repair is applied. As we could show in chapter 2.4 changes to this concept can improve repair stability. By distributing tension more evenly throughout the repair with barbed sutures and by not relying on anchoring suture configurations, the gapping resistance and the final failure force of tendon repairs could be further improved. Published data supports this conclusion [170, 229, 232]. 235

Nevertheless no scientific study investigated if these improvements also apply to a healing tendon in an in vivo experiment. This question becomes even more important considering that all of the commercially available barbed sutures are resorbable and no suitable non resorbable barbed suture is available.

The resorption of suture and barbs may have no effect on tendon repairs when tested in a laboratory ex vivo setting, but in an in vivo setting resorbable barbed suture tendon repair constructs might behave differently. This study is the first investigation testing a unknotted barbed suture tendon repair construct in an in vivo tendon repair model. The turkey model was chosen, since this model is the only model that replicates human tendon sizes, anatomy and biomechanics. All other in vivo animal models provide too small tendons to perform a four strand barbed suture tendon repair. Please refer to the previous two chapters (chapter 3.1 and chapter 3.2) for in depth information on the turkey tendon model and surgery on the turkey foot. Also for detailed information on the knotless 3D barbed suture repair technique please see chapter 2.4.

236

3.3.2 Methods

30 male 16 weeks old turkeys (meleagris gallopavo) were used for this study. Average weight of turkeys in our experiments was 25.9Kg (±3.8Kg). Animals were attained from a local commercial breeder (Inghams Enterprises, Liverpool, NSW, Australia). Due to ethics regulations (pilot study of 5 animals was condition for ethics approval) and housing capacities (15 birds at a time maximum) experiments were conducted in three separate time periods:

 First period 5 animals.

 Second period 15 animals

 Third period 10 animals.

Authority to conduct a turkey animal research project was approved by the Animal Care and Ethics Committee (ACEC) of the University of New South Wales (ACEC Number:

12/70A). Planning of experiments, investigations and procedures were based on previously conducted work and experimental data from our chicken model [94, 295].

Approval was conditioned on conducting a pilot study on five turkeys and passing two

ACEC inspections.

237

Also for this pilot study, animals had to be monitored continuously and all adverse events had to be reported to ACEC immediately. Please see previous two chapters for in depth description of animal anatomy, tendon’s biomechanical properties and housing and husbandry details. Please also see previous chapter for general description of surgical procedure.

3.3.2.1 Surgery

All 30 animals, albeit housed and operated at different time points went through the exact same procedures. Animals were acclimatized prior to surgery for 10 days to get used to the new environment. Turkeys were housed in a large group pen as described before. All animals were examined for general health, marked with electronic markers and weighed before surgery. Two groups of n=15 animals each were formed by choosing random marker numbers:

238

Group 1, Adelaide repair group (figure 3-40): A four-strand cross locked cruciate

(Adelaide) repair technique using 4-0 silicone coated braided polyester sutures (Ticron:

Tyco, Lane Cove, NSW, Australia) was used for tendons core repairs. Suture purchase was 10mm. A simple running circumfernetial repair technique using 6-0 polypropylene sutures (Prolene: Ethicon, Somerville, NJ, USA) was used for tendons epitendinous repairs. A four throw square knot was tied between tendon ends. Cross-locks in the

Adelaide repair were kept at 4mm [108].

Figure 3-40 Illustration of the four-strand cross locked cruciate repair technique (Adelaide Repair). Tendon purchase and cross-lock width indicated in mm.

239

Group 2, barbed suture repair group (figure 3-41): A four-strand 3D knotless barbed suture repair using 4-0 unidirectional barbed glycolic-carbonate (Polyglycolide-

Polytrimethylene-Carbonate) sutures (V-Loc 180TM; Covidien. Mansfield, MA) was used for tendons core repairs. Suture purchase was 10mm. A simple running circumfernetial repair technique using 6-0 polypropylene sutures (Prolene: Ethicon,

Somerville, NJ, USA) was used for tendons epitendinous repairs.

240

Figure 3-41: Illustration of our four-strand 3D knotless barbed suture repair technique. Tendon purchase indicated in mm. a) First strand runs in the horizontal plane. Note, it passes through the suture loop at the end of the suture thread (see 3). b) Second strand (see 5) runs from horizontal into vertical plane (three dimensional repair). Third strand (see 7) runs in the vertical plane. c) Fourth strand stays in the vertical plane and finishes with a triple zigzag. d) Suture is cut flush with the tendon.

241

Animals were operated in prone positioning. General anaesthesia was induced with a combination drug of a dissociative anaesthetic and a tranquilizer, 5mg/Kg Zoletil

(Virbac Animal Care, Milperra NSW, Australia). General anaesthesia was then maintained on 2% Isoflurane (Baxter Healthcare, Toongabbie NSW, Australia) and

Oxygen with 2L/min flow administered via an in circuit anaesthetic machine over a face mask. Additionally a dose of 10mg/Kg Carprofen (Apex Laboratories, Somersby NSW,

Australia) as anti-inflammatory and analgesic drug and a prophylactic dose of tetracycline of 110mg/kg (Oxytet-200 LA: Troy-ilium, Glendennin, NSW, Australia) was administered. The depth of anaesthesia was continuously monitored during surgery and the levels of isoflurane and oxygen were adjusted accordingly. Parameters monitored at every five minutes were: blood oxygen saturation, end tidal CO2, heart rate, pulse rate, respiratory rate & depth and eye reflex.

Surgical procedures were all performed by one surgeon (TP) using loop magnification, aseptic techniques and a “finger tourniquet” (please see previous chapter for detailed description of surgical procedures).

In group one, 15 animal tendons were lacerated and repaired in zone II using a conventional knotted Adelaide repair technique (figure 3-42).

242

Figure 3-42 Example image of completed repair in group 1 using a conventional knotted Adelaide repair technique tendon reconstruction. Core suture knot and peripheral suture knot buried between tendon ends.

In group two, the remaining 15 animal tendons were lacerated and repaired in zone II using our novel knotless 3D barbed suture repair technique (figure 3-43). Animals were randomly allocated to groups.

243

After surgery skin was closed with 3/0 resorbable braided polyglycolic-acid sutures

(Vicryl: Ethicon, Somerville, NJ, USA) and covered with gauze, cotton wool padding and an elastic bandage. Additionally for immobilization a well padded dorsal plaster splint was applied (figure 3-44).

244

Figure 3-44 Application of dorsal splint with padding.

After completion of procedures turkeys were moved to a recovery table with heat mat and maintained oxygen flow (2.0L/min) until the turkey was conscious at which time the bird was returned to the holding room on deep litter with access to heat lamp, climate control, food and water. Birds were continuously monitored by technicians until satisfied all turkeys have recovered uneventfully. For detailed description of husbandry please see previous chapter.

Splints were removed from the turkeys leg after three weeks. Birds were euthanized after six weeks using appropriate methods in a designated room in the absence of any other bird. Following an injection of 10mg/kg Zoletil (Virbac Animal Care, Milperra

NSW, Australia) an intra venous injection of Pentobarbital 1ml per 2.0kg body weight

(Lethabarb: Virbac Animal Care, Milperra NSW, Australia) was administered.

245

After inspection of the carcass and check of "certain signs of death" the birds legs were detached from body and marked.

3.3.2.2 Grouping of specimen

The following graph illustrates allocated groups (figure 3-45):

Figure 3-45 Grouping of animals

246

Operating times were noted for all 30 surgeries. From each main group, “Adelaide group” (n=15) and “Barbed suture group” (n=15), ten specimen were randomly chosen and prepared for biomechanical testing and the remaining five specimens of each main group were prepared for histology.

3.3.2.3 Endpoints

3.3.2.3.1 Operating time needed for core repair

Regardless of grouping, operating times for all 30 surgeries were noted. Special focus was laid on time needed to complete the core repair.

247

3.3.2.3.2 Endpoints for biomechanical testing group

From the 15 animals in each main group, “Adelaide group” and “Barbed suture group” ten specimens were randomly chosen and prepared for biomechanical testing. In both the ”biomechanical testing” groups additional endpoints were investigated:

 Range of motion of digit

 Amount of failed repairs and failure mode

 Classification of adhesion formation

 Repair bulk in mm

 Average gap in repair after 6 weeks

 Failure force during biomechanical testing

 Mode of failure during testing

248

3.3.2.3.2.1 Range of motion and repair failure

The following set up was used for range of motion testing. The turkeys leg was fixed on a work bench with a goniometer installed next to the specimen. The rotation axis of the goniometer was set at the Metacarpo Phalangeal (MP) joint of the turkeys middle toe. At the site of amputation (mid lower leg) the middle toes deep flexor tendon was dissected and clamped with a rough surgical clamp (figure 3-46).

Figure 3-46 Clamp clamped on middle toe deep flexor tendon.

10 Newton (N) load were attached to that clamp to move the middle toe (a 1kg weight was used to generate the load). Total range of motion was measured for operated leg and contralateral non-operated leg (figure 3-47).

249

Figure 3-47 Measurement of total range of motion.

Range of motion of operated toe was then expressed as percentage of the contralateral, non-operated toe range of motion. Also at this stage it became clear which tendon repairs failed since no toe movement equaled with repair failure.

3.3.2.3.2.2 Classification of adhesion formation

After range of motion measurements tendons were carefully dissected. Failed repairs were not classified for adhesion formation, only mode of failure was noted as “suture rupture” or “pull out”. The intact tendon repair specimen were classified for adhesion formations. As classification basis Tang et al.’s macroscopic adhesion evaluation criteria was used [144]. Adhesions were macroscopically evaluated on a scale from 0 to 6. 0 meaning no adhesions, 6 meaning severe adhesions (figure 3-48).

250

Figure 3-48 Adhesion classification criteria [144]

251

Example pictures of dissections and evaluation of adhesion classifications in operated turkeys:

Figure 3-49 Adhesion classification: 0 (no adhesion)

252

Figure 3-50 Adhesion classification: 2 (mild adhesion)

253

Figure 3-51 Adhesion classification: 4 (moderate adhesion)

254

Figure 3-52 Adhesion classification: 6 (advanced stage adhesion)

255

3.3.2.3.2.3 Repair bulk and remaining gap

After determination of adhesion classification for each tendon, tendons were dissected out of the tendon sheath and resected distally at the insertion into the distal phalanx and proximally 4cm proximal from the repair site. A digital caliper (Mitutoyo CD-6’’CS

Absolute Digimatic. Tokyo, Japan) was used to measure repair site bulk (latero-medial diameter) and remaining gap in mm. For the identification of remaining gap eight measurements were conducted over the repair site and the average gap was calculated

(see also chapter 2.3.2.2.2 and figure 2-39). If no gap was apparent no measurement was conducted and gap was noted as zero.

256

3.3.2.3.2.4 Measurement of final failure force of repair and mode of failure

Biomechanical testing followed the testing protocol from previous experiments. All dissected tendons were mechanically tested in a uniaxial manner using a MTS 858 Mini

Bionix materials testing machine (MTS Systems Corp. Eden Prairie, MN, USA) with a

500N load cell. Tendons were gripped in grooved pneumatic clamps with a gripping pressure of 60psi. Gauge length was standardised at 40mm. Samples were preloaded to

1.5N then tested to failure at a relative slow distraction rate of 10mm/min. Load, displacement and time were continuously recorded at 100Hz. Data were used to generate load-displacement graphs for each tendon. Load to failure (measured in

Newton) and mechanism of failure were determined for all samples. The mechanism of failure was recorded as either suture breakage or suture pull out (figure 3-53).

257

Figure 3-53 Dissected tendon mounted between hydraulic clamps and tensioned in uniaxial manner.

258

3.3.2.3.3 Histology group endpoints

From the 15 animals in each main group, “Adelaide group” and “Barbed suture group” ten specimens were randomly chosen and prepared for biomechanical testing (see

3.3.2.3.2). The remaining five specimen in each group were prepared for routine paraffin histology. The tendons were carefully dissected out of the tendon sheath and resected 1.5cm proximal and 1.5cm distal from repair site. Tendon specimen were then mounted in a histology cassette and prepared for routine paraffin histology (figure 3-54).

259

Figure 3-54 Tendon specimen prepared for histology

After formalin fixation for 72 hours specimen were embedded in paraffin using a tissue processing machine (Shandon Excelsior ES, Thermo Fisher Scientific, Kalamazoo,

USA). Then 4µm longitudinal sections were cut of the paraffin blocks using a hand- operated microtome (Leica RM2165, Germany) and mounted onto salinised slides.

Slides were stained with Harris Haematoxylin and Eosin (H&E).

260

Fascicle distribution, cell population, scarring and signs for healing were qualitatively compared using light microscopy (Olympus BX51, Sydney, Australia). Assessments were made in a blinded fashion.

3.3.3 Results

As reported in previous chapter, in vivo surgery on the turkey deep flexor apparatus is practicable and procedures are comparable to tendon repair scenarios in humans.

Surgeries of all birds went without adverse events. All birds survived operation and post surgery period. Intra operative bleeding was no issue with the used tourniquet technique.

Also in the postoperative period, no bleeding, infection or other diseases were noted.

Animals showed uncomplicated recovery after surgery and all animals tolerated the casting. Birds were weight bearing on operated leg without signs of lameness. Casts remained on the leg for 3 weeks. It was noted that at the time of removal of the casts after 3 weeks all splints were broken at the ankle joint (figure 3-55).

261

Figure 3-55 Broken splint after removal.

Nevertheless this was not apparent from the outside and all casts remained for 3 weeks.

After 6 weeks birds were euthanized and depending on group affiliation prepared for either histological processing or biomechanical testing.

262

3.3.3.1 Operating time for core repair

Significant differences of time measured for performing the core repair was noted between groups. The average time needed to complete an Adelaide core repair (ADL) was 7min 41s (±41s). The average time needed to complete a 3D barbed suture repair

(BRB) was 6min 32s (±1min 12s). The difference in time was significant (p<0.05)

(figure 3-56).

Figure 3-56 Operating time of core repair per group

263

3.3.3.2 Repair failures per group after six weeks and failure

mechanism

As described in 3.3.2.3.2.1, failed repairs were defined by no movement of toe at range of motion testing. There were significant more repair failures notable in the barbed suture group compared with the Adelaide group. In the barbed suture group four of ten repairs were not intact after six weeks (40%) in the barbed suture group two of ten repairs were not intact.

Figure 3-57 Number of non intact tendon repairs after 6 weeks

264

In the Adelaide repair group one repair failed by pull out of suture construct and one by suture rupture. In the barbed suture group all four repairs failed by suture pull out.

3.3.3.3 Range of motion per group after six weeks

In the intact specimen a significant difference in range of motion could be noted. In the

Adelaide group the average range of motion of the operated toe was 58% (±16%) of the contra lateral non operated toe. In the barbed suture group average range of motion of the operated toe was 82% (±13%) of the contra lateral non operated toe. The difference between groups was significant (p<0.05) (figure 3-58).

Figure 3-58 Toe range of motion in % of contra lateral toe range of motion

265

3.3.3.4 Classification of adhesion formation

After range of motion testing tendons were dissected and adhesion classification was determined. There was a trend to more adhesion formations in the Adelaide repair group notable, but the difference was not significant (p=0.27). The average adhesion classification in the Adelaide group was 4.75 (±1.6) compared to 3.67 (±1.9) in the barbed suture group (figure 3-59).

Figure 3-59 Adhesion formation classification for groups: 0 no adhesions, 6 severe adhesions (classification from Tang et al. [144])

266

3.3.3.5 Repair bulk in mm

There was a significant difference in bulk of repair detected between groups. Repairs in the Adelaide group had a medio lateral diameter of 5.6mm (±1mm) compared to 4.8mm

(±0.66mm) in the barbed suture group. The difference was significant (p<0.05) (figure

3-60).

Figure 3-60 Average bulk in mm for groups

267

3.3.3.6 Average gap in repair

Measured repair gaps between groups showed significant differences. The average repair gap after six weeks in the Adelaide group was 0.63mm (±1.06mm), in the barbed suture group 1.17mm (±1.16mm). The difference was significant (p<0.05) (figure 3-61).

Figure 3-61 Average remaining gap in mm after 6 weeks

268

3.3.3.7 Failure force during biomechanical testing

Biomechanical testing of the repaired tendons revealed significant difference in failure forces measured. The average Adelaide tendon repair resisted 53.42N (±21.73N) before disrupture. The average barbed suture repair resisted 29.74N (±17.12N). The difference in final failure forces between groups was statistically significant (p<0.05) (figure 3-62).

Figure 3-62 Average final failure forces in Newton for each group

269

3.3.3.8 Mode of failure during testing

Of the eight tested tendon repairs in the Adelaide group six failed by suture rupture and two by suture pull out. In the barbed suture group all 6 tendon repairs failed by suture pull out.

3.3.3.9 Histology

Histological Overview of turkey deep flexor tendon in situ (figure 3-63).

Figure 3-63 Longitudinal cut through middle toe of turkey tendon at level of PIP joint ( S=skin, B=bone, T=tendon) 270

All ten histology samples (five from each group) were investigated separately from the main biomechanical testing group. From the five histology samples in the Adelaide group two repairs were disrupted, in the barbed suture group three repairs were disrupted (table 5).

Table 5 Overview of amount of failed repairs in histology group after six weeks.

Barbed suture Adelaide suture group group Intact after 6 weeks 2 3 Disrupted after 6 weeks 3 2

Both Adelaide repairs in the histology group did fail by suture pullout, meaning suture and knot were intact, but anchoring cross locks were pulled out of the tendon substance

(figure 3-64).

271

Figure 3-64 Failed repair in the Adelaide group. Note, suture is intact, but anchoring cross locks got pulled out of tendon substance.

This also became apparent in the histology slides. Suture loops were pulled through the tendon and caused disruption of repair construct. Also note reactive inflammatory cell reaction (figure 3-65).

272

Figure 3-65 Histology cut from the same specimen. Clearly visible the pull out path. Also inflammatory cell infiltration is apparent.

273

All three disrupted barbed suture repairs in the histology group failed by suture pull out, but in this case meaning sutures got pulled through the tendon due to lost barb stability

(figure 3-66).

Figure 3-66 Disrupted barbed suture repair. Note suture is intact, but pulled out of the tendon.

274

Histology of failed barbed suture repair showed suture path with disrupted walls.

Disruption of collagen was noticeable at area of barb-tendon interface (figure 3-67).

Figure 3-67 Pull out path of barbed suture. Note disruption of collagen tissue due to barb interaction.

275

In the intact five tendon repairs (two barbed suture repairs and three Adelaide repairs)

classic features of tendon healing were recognizable. All histology slides of intact

repairs showed a bridging area with unorganized, scar like fibrous tissue. Collagen

fibres in the repair site were disorientated and infiltrated with inflammatory cells.

Mainly granulocytes and lymphocytes were present. The epitenon showed increased cell

proliferation at the repair site. In general epitenon in healed repairs was without

separation from tendon surface and in continuity. Histologically, no significant

differences between repair groups regarding healing of tendon tissue was recognizable

in specimens. Table 6 gives an overview of observed histological features regarding

cellularity, scarring, vascularity and suture tissue reaction in each sample:

Table 6 Histological features of repairs after six weeks.

Failures Cellularity Repair scar Vascularity Suture tissue reaction (0-3) (0-3) (0-3) (0-3) Adelaide 1 X 1 2 2 0 Adelaide 2 intact 2 2 2 1 Adelaide 3 intact 2 1 3 2 Adelaide 4 X 0 3 1 1 Adelaide 5 intact 2 2 1 0 Average Adelaide 2 1.4 2 1.8 0.8 Barbed 1 intact 2 1 2 2 Barbed 2 X 3 1 2 2 Barbed 3 intact 2 2 3 1 Barbed 4 X 1 2 1 2 Barbed 5 X 2 2 2 1 Average Barbed 3 2 1.6 2 1.6

276

277

Figure 3-69 H&E slide of same specimen of healed Adelaide repair in higher magnification. Healing tendon demonstrating no gap and aligning collagen fibres. Also note bridging tissue with inflammatory cells such as lymphocytes and granulocytes.

278

279

Figure 3-71 Stump of failed repair with cellular fibrous tissue encapsulating the collagen tendon tissue. Left native H&E, right polarised H&E.

Figure 3-72 Moderate tissue reaction around resorbable (glycolic-carbonate) barbed suture. Left, native H&E, right polarised H&E.

280

Figure 3-73 Minimal tissue reaction around non resorbable (polyester) suture in Adelaide group.

281

3.3.4 Discussion

The strength of a conventionally knotted tendon repair depends roughly on three factors:

 amount of strands crossing the repair site [92, 93]

 amount and quality of locking configurations [63, 89-91]

 suture material [84-87] and size [88]

Four-strand techniques provide a good compromise between minimising complexity and adequate performance [79, 197]. Using more than four strands interferes with surgical complexity, tissue handling and repair bulk [22, 94]. Regarding suture material and size, possibilities to improve repair stability are limited. And also the possibilities to further improve the quality of locking configurations in tendon repairs is probably exhausted with the use of cross locks. To significantly further advance tendon repair stability, especially in regards to gapping, a entirely new tendon repair concepts had to evolve.

Barbed sutures anchor the tendon without locking configurations or knots and distribute loads through the entire repair. In ex vivo experiments these repairs could show superior gapping resistance and higher final failure forces in comparison to conventional knotted repairs [170, 229, 232]. All of these studies were ex vivo studies. It is not proven if these superior tendon repair qualities also can be achieved in an in vivo experiment. 282

Therefore this study tested the previously introduced 3D barbed suture tendon repair

(see chapter 2.4) in an in vivo experiment.

Two barbed suture materials are currently commercially available. The Quill™ device

(Angiotech. Vancouver, BC, Canada) is a bidirectional, double armed, barbed suture material (figure 3-74 and 3-75).

Figure 3-74 Quill barbed suture device. Note the single angle cut of barbs.

Figure 3-75 Quill suture device. Note, double armed and barb orientation changes in the middle of the suture.

283

And the V-Loc™ device (Covidien. Mansfield, MA, United States) is a unidirectional, single armed, barbed suture with a little loop at the end (figure 3-76 and 3-77).

Figure 3-76 V-Loc barbed suture device. Note the double angle cut of barbs.

Figure 3-77 V-Loc barbed suture device. Note, the loop at the end of the suture.

284

Both devices are available in non resorbable material and two qualities of resorbable material, fast resorbing and slow resorbing. Unfortunately the non resorbable devices from both manufacturers are not available in small enough sizes to be used in tendon repairs. Probably currently available manufacturing techniques limit the production of smaller size non absorbable barbed sutures. Therefore the only possibility for using these materials in tendon repairs is to use the slow absorbing variants.

Resorbable suture materials have been used in tendon repairs previously [257, 304-306].

In a study by Viinikainen et al., conventional tendon repairs using absorbable poly-L/D- lactide sutures showed better long term results in regards to repair site collagen quality if compared to the same repairs using non absorbable sutures. Especially the persisting knot between the healing tendon ends in the non resorbable suture group caused disorientated collagen alignments in the repair site [257].

But no reports on investigations regarding the performance of resorbable barbed sutures in unknotted tendon repairs are available yet.

The described 3D unknotted barbed suture repair is performed with an unidirectional barbed suture and for this present experiment we chose the slow resorbing V-Loc 180 suture. The number 180 indicates the sutures lifespan meaning after 180 days a complete loss of tensile strength through degradation secondary to hydrolysis can be expected.

285

But significant tensile strength is already lost after much shorter time spans. The manufacturer specifies that after 7 days 80% tensile strength, after 14 days 75% tensile strength and after 21 days 65% tensile strength is remaining in the suture [307]. But the manufacturers indication of the degradation profile in regards to decreasing tensile strength is misleading. The main holding capacity of repairs using barbed sutures relies on the stability and integrity of the barbs since no knot is securing the construct.

Degradation processes not only affect the overall tensile strength of the suture body, but also the barbs. This has negative consequences in a unknotted tendon repair. Before the tensile strength of the actual suture reaches low values the barbs in the suture lose strength. This weakens the suture tendon interface and sutures pull out. To further investigate this, V-Loc 180 sutures were incubated for 3 weeks in saline and subsequently scanning electron microscopy pictures of the suture were produced. In these pictures the degradation of the barbs is clearly visible (figure 3-78).

286

Figure 3-78 SEM pictures of V-Loc 180 suture after incubation in saline for 3 weeks.

It was assumed that this phenomena was a reason for the bad outcome in our barbed suture group. In comparison to the conventional Adelaide tendon repair group, the barbed suture repair group showed significantly more failed tendon repairs after 6 weeks

(40% repair failures versus 20% repair failures). All four failures in the barbed suture group were caused due to suture pull out. In the Adelaide group one failure was by pull out and one by suture rupture. This result was also reflected in the separately analysed histology group. Also here the amount of tendons that failed was bigger in the barbed suture group than in the conventional Adelaide group (three of five versus two of five).

287

Another indicator for the problematic holding capacity of the resorbable barbed sutures was the measured gap in the repair site after 6 weeks. Average gap measured in the remaining intact eight Adelaide repairs was 0.63mm compared to 1.17mm in the remaining six intact barbed suture repairs (p<0.05).

Also the average measured final failure force in the remaining repairs was significantly lower in the barbed suture repair group compared to the Adelaide repair group (29.74N versus 53.42N, p<0.05).

High failure rates in in vivo tendon repair experiments are common and have been reported previously. Su et al. reported of 9 failures in 16 operated dogs [250]. Also Zhao et al. report of high failure rates in their dog experiments if weight bearing and mobilisation of operated extremity are not controlled [248, 249]. This seems to be the main problem in large animal in vivo tendon repair experiments. The weight of turkeys and laboratory dogs is comparable. Average weight of turkeys in our experiments was

25.9Kg (±3.8Kg). Reported weights of dogs are between 20 and 30 kg [248, 252, 254,

255]. The only difference of course is, that turkeys distribute this weight through two legs instead of four. The fact that three Adelaide repairs failed by suture pull out (one in the general biomechanic group and two in the histology group) supports the high load theory as well. Normally, Adelaide repairs resist high loads before the stable cross lock configurations pull out of the tendon substance.

288

So either the tendon substance weakened or the load on the repair was considerably high. Also the fact that in our experiment all cast were broken after 3 weeks indicate that high forces were transduced into the foot and toes during the early healing phase of operated tendons. This might have contributed to the high failure rates.

In this present study repaired tendons were investigated after 6 weeks. This is a common endpoint in in vivo tendon repair experiments [247, 248, 250, 255, 257, 261, 268].

Nevertheless the strength of a repaired flexor tendon progressively decreases during the first 2-3 weeks [128]. Only after this initial phase an increase in repair strength can be expected. It is assumed that already in this critical early phase of tendon healing high loads were affecting the repair sites due to the full weight bearing of the animals and the broken casts.

The post repair management was immobilization for three weeks. This was an attempt to keep influencing factors to a minimum and an attempt to protect repairs from high loads for the critical first 3 weeks of tendon healing. For this first study on the use of resorbable barbed sutures in an in vivo tendon repair experiment focus was primarily laid on the comparison of the two repair techniques.

289

The author is aware that a modern active or passive rehabilitation protocol would have been the appropriate post op management in a clinical scenario [79, 129, 249, 253], but applying this to our animal experiment would have introduced various factors that would have been difficult to control.

Gliding resistance of a tendon repair in the tendons sheath is an important factor for causing adhesion formations [253]. Studies show that repair configurations with more bulk or suture exposure on the tendons surface cause more adhesion formations.

Adhesion formations in a tendon repair limit total range of motion of the digit and therefore cause suboptimal final functional outcomes [66, 142, 308, 309]. This could also be observed in this present study. The significant smaller repair bulk in the barbed suture group (4.81mm versus 5.56mm, p<0.05) correlates with a trend to less adhesion formations and a significant better range of motion of the operated digit. Total range of motion in the barbed suture group reached 81.7% of the contra lateral non operated toe range of motion after 6 weeks compared to only 57.5% in the Adelaide group.

Nevertheless it is not clear if the repair bulk alone was responsible for this result or if the suture exposure on the tendons surface caused this. The barbed suture repairs show only minimal suture exposure on the tendons surface compared to the relatively prominent cross locks used in the Adelaide tendon repairs (figure 3-79).

290

Figure 3-79 Adelaide tendon repair (left) and barbed suture repair (right). Note difference in repair bulk and suture exposure on tendon surface.

That friction rather than pure bulk is more likely to cause adhesion formations was also the conclusion by Sanders et al. and also Momose et al. from their experiments [308,

310].

In conclusion it is difficult to deduct all influencing factors in an in vivo experiment.

Especially if it is a novel in vivo model. The main focus was laid on the comparison of a new knotless tendon repair technique (knotless 3D barbed suture repair) versus a common and well approved knotted repair technique (Adelaide repair). Nevertheless the turkey model is a “high load” tendon repair model and it is difficult to control forces affecting the operated digit. Turkeys seem to have little pain reception and start weight bearing on the operated foot right after surgery.

291

Also the heavy weight of turkeys and the fact that this weight is only distributed between two legs complicates the post op protection of the operated tendon, hence the broken casts. Nevertheless this was a comparative study and conditions were similar for both groups.

Barbed suture tendon repairs performed with resorbable barbed sutures seem to be weaker than conventional knotted tendon repairs using non resorbable sutures. Double the amount of failed repairs, bigger gaps in the remaining intact repairs and significant lower final failure forces after 6 weeks indicate this. Nevertheless the remaining intact barbed suture repairs could show better functional outcomes especially in regards to adhesion formations and total range of motion. It would be interesting if the weaknesses of the barbed suture group could be minimized by longer lasting barbed sutures. One weakness of our study is the small amount of samples and the little experience with the in vivo turkey model. Also the use of a non resorbable suture in the conventional knotted repair group versus a resorbable suture in the barbed suture group could be disputed. Reason for the use of the non resorbable sutures in the Adelaide tendon repair group was the aim to test the new repair against the gold standard in tendon repairs. And as reported, the Adelaide technique performed with an non resorbable suture is considered the gold standard. Future experiments have to show how the knotless 3D barbed suture tendon repair method can be improved to further advance deep flexor tendon repairs.

292

CHAPTER 4: Conclusions, limitations and future

directions

The title of this thesis reads “Deep Flexor Tendon Repairs”. This is a broad topic and it is difficult to discuss all aspects of this wide field without leaving the limits of a PhD.

The author's background of plastic surgery and hand surgery goes some way in narrowing down the subject range. Likewise, the institution in which the research is undertaken plays an important role in the breadth of investigation possibilities. In this case, a biomechanical laboratory based in a university hospital served as institution for all experiments.

This thesis therefore focuses on the main questions of tendon repair research from a surgical and biomechanical point of view: the investigations of surgical tendon repair concepts and investigations on related laboratory models for examining these concepts.

The deep flexor tendon was chosen due to its complex anatomy (hence difficult surgery) and its pattern and frequency of injury. The same reasons accounted for the choice of the area of repair, zone II, also famously known by its alternative name, “no man’s land”.

293

To create the basis for a coherent thesis it is important to follow a single, clear path of investigation. One idea must build on the other and only then a stable framework can be achieved. Applying the rule to this thesis, one project followed the other, as anticipated.

Each experiment informed and shaped the next, deepening the knowledge base of the overall topic. At times, however, it was not until the second or third experiment that the real explanations for the outcome of the first experiment became clear. It is likely, of course, that this phenomenon not only applies to the small perspective of this PhD research, but also to the big picture of research in general.

Nevertheless, in the process of analysing the research experiments building up to this

PhD thesis it became clear that some new insights into the topic of tendon repair research were reached.

294

The main results of this thesis are as follows:

 In regards to comparability to human tendons, sheep tendons are better tendon surrogates as pig tendons if used in ex vivo laboratory experiments.  When focusing on gapping resistance, “locking loop” repair configurations for tendon repairs are not substantially different to “grasping loop” configurations, and only “cross-locks”, as used in the Adelaide repair technique, deserve the adjective description “locking”.  The current gold standard of tendon repairs, the Adelaide repair, produces better repair stability if performed with larger cross locks.  The author's interlocking modification of the Adelaide repair can further improve the Adelaide repair's stability.  In an ex vivo setting, the author's new tendon repair concept, the knotless 3D barbed suture tendon repair, produces superior repair stability than the Adelaide repair.  The turkey tendon model is the first tendon model that replicates human anatomy and tendon sizes and can be used in ex vivo as well as in vivo tendon repair experiments.  In an in vivo tendon repair scenario, the use of the knotless 3D barbed suture tendon repair with resorbable barbed sutures produces inferior repair stability compared to the Adelaide repair, but improves functional outcomes. Nevertheless resorbable barbed sutures should only be used with great caution in clinical tendon repair cases.

295

This PhD thesis starts with an experiment investigating which laboratory tendon model best reflects human tendon biomechanics. Pig and sheep tendons are widely used in ex vivo experiments all over the world. However, no clear consensus did exist as to which animal model is better to use. Before embarking on tendon repair research, it made sense to thoroughly rethink this question and isolate the most suitable model for future experiments. This formed the basis of the author's first experiment, discussed in chapter

2.1. Following the results of this experiment, sheep tendons were found to be more suitable for ex vivo tendon repair experiments than the tendons of pigs. The sheep tendons' comparability to human tendons, especially in biomechanical experiments investigating gapping properties, was superior.

Namely the investigation of repair constructs with regards to gapping was the aim of the following experiments. The author initially focused on current tendon repair concepts and analysis of repair stability. However, a new approach to this question was taken. Rather than focusing purely on which suture configuration best resists gapping, the aim of the investigation instead focused on the question of what exact mechanism actually leads to the gapping. A new concept of investigating tendon repair failures was therefore designed. The solution was a combination of wire sutures and X-ray. This combination of suturing material and investigative tool made it possible to literally see what is happening when a tendon repair configuration fails.

296

This gave a new insight into long discussed questions such as the exact geometrical failure characteristics of a “locking loop” and a “grasping loop”.

Further, these experiments reaffirmed the established fact that the Adelaide repair is the best four-strand tendon repair with regards to final failure force and gapping. The reason for its superior stability was, however, not entirely known previously. During the course of this experiment, it became clear that the cross locks used in the Adelaide repair technique were the main factors contributing to its stability.

This led to the next two experiments. Their aim was to determine how to modify the

Adelaide repair to further improve its stability. The first experiment addressed the simple question of whether the size of cross locks in an Adelaide repair would influence outcomes; the second whether an interlocking modification would improve the Adelaide repair's already strong gapping resistance. These investigations led to the claim that bigger cross locks further stabilise Adelaide repairs and enable loads to be more centrally distributed. Furthermore, the interlocking of the cross locks improves the gapping resistance of the Adelaide repair, at least in the very beginning of cyclic loading.

297

Again, all of these experiments pointed to the next question: How can a tendon repair be designed to further improve overall repair stability? It was clear that the author's main effort had to be focused on gapping, since it was already acknowledged that it was not stronger repairs that were necessary but repairs that gap less. Applying the knowledge of the previous studies, it was also clear that a substantial problem was the anchoring of the sutures in the tendon. These locking configurations were, like their name indicates, locking or anchoring in certain areas in the tendon. These areas naturally became stress zones and, inevitably, failure was occurring in those sites. The aim, therefore, became to circumvent these stress areas and attempt to distribute loads throughout the repair in its entirety. Applying the author's experience of barbed sutures in wound closures in other fields of plastic surgery, the idea arose to use these sutures for tendon repairs. Very little in the way of research in the field of using barbed suture for tendon repair had been published, but initial tests in the laboratory were promising. The actual repair was designed to bury as many barbs as possible in as much tendon substance as possible.

This would guarantee sufficient hold in the tendon as well as good distribution of stress areas throughout the entire repair even without using any knot. Thus, the 3D configuration was born. The study showed that in an ex vivo setting, this unknotted repair method can produce superior holding and gapping capacities than the current gold standard in tendon repairs, the Adelaide repair. But this new repairs’ in vivo properties were still unknown.

298

Therefore the next step was to prove if this tendon repair technique had benefits in an in vivo setting as well. Although one problem already arose in the planning phase of this study: there were no in vivo animal models available to investigate this new repair method since tendon sizes of the usual research animals were too small. The solution was the introduction of the turkey tendon model, a new in vivo animal model for tendon repair experiments that replicates not only human deep flexor tendon anatomy and tendon sizes but is also comparable in biomechanical properties and can be used in ex vivo as well as in vivo settings. But another problem became apparent: only resorbable barbed sutures are available in sizes suitable for tendon repairs. Albeit resorbable sutures are used in conventional knotted tendon repair techniques, the relative early onset of barb resorbtion in the unknotted barbed suture tendon repairs had critical consequences. With the weakening of the barb structure, the integrity of the repair diminished too quickly, consequently repairs lost their holding capacity and failed.

Nevertheless, in the cases that did not fail, the repair could show better final functional outcomes. Following this investigation, the use of resorbable barbed sutures in tendon repairs in in vivo settings can only be advised with great caution.

This brings us to the limitations of this thesis. Most of the experiments were ex vivo laboratory studies and translation of laboratory experiments into clinical scenarios is difficult. Nevertheless, better ex vivo results mostly mean better in vivo results.

299

The advances in tendon surgery in recent years were, on the whole, driven by repair configurations that were designed in ex vivo experiments. Modern multi-strand repairs such as the Adelaide repair can withstand higher loads without gapping or failing. When combined with modern early active rehabilitation concepts, these repairs improved tendon repair outcomes in clinical scenarios significantly. In these cases, the interpolation of ex vivo results into clinical scenarios was correct and advocated as such.

In the case of the knotless tendon repair using resorbable barbed sutures, however, this was more difficult. The principal superior concept of repairing a tendon without a knot and relying on barbs to hold the tendon fell apart when the barbs lost their strength in the course of early healing. Nevertheless, this is an important result and must be translated into clinical scenarios. Consequently as follows: Resorbable barbed sutures might not maintain tendon repair stability for long enough to facilitate tendon healing in the early phase of rehabilitation, therefore barbed sutures must only be used with great care if applied to tendon repairs in clinical scenarios.

Finally this at the same time informs the authors commentary on future directions. It is highly probable that with non resorbable barbed sutures or sutures with barbs that do not lose their strength as quickly, it is possible to further improve deep flexor tendon repairs, both in laboratory and clinical settings.

300

To prove this, appropriate barbed sutures must firstly be designed, followed by rigorous testing of repair concepts using these new sutures. Testing by following the same path as in this PhD, first in ex vivo and then in in vivo experiments.

A constantly deepening knowledge base combined with research driven advances in technology will, one step after another, generate further improvements in this challenging field of surgery.

301

CHAPTER 5: References

1. Dona, E., et al., Biomechanical properties of four circumferential flexor tendon suture techniques. J Hand Surg Am, 2003. 28(5): p. 824-31. 2. Benjamin, M. and J. Ralphs, Tendons and ligaments-an overview. Histology and histopathology, 1997. 12(4): p. 1135-1144. 3. Thorpe, C.T., et al., Specialization of tendon mechanical properties results from interfascicular differences. Journal of The Royal Society Interface, 2012. 9(76): p. 3108-3117. 4. Ward, S.R., et al., High Stiffness of Human Digital Flexor Tendons Is. J Neurophysiol, 2006. 96: p. 2815-2818. 5. Lichtwark, G.A. and A. Wilson, In vivo mechanical properties of the human Achilles tendon during one-legged hopping. Journal of Experimental Biology, 2005. 208(24): p. 4715-4725. 6. Maganaris, C.N. and J.P. Paul, In vivo human tendinous tissue stretch upon maximum muscle force generation. Journal of biomechanics, 2000. 33(11): p. 1453-1459. 7. Moran, C.A., Anatomy of the Hand. Physical therapy, 1989. 69(12): p. 1007- 1013. 8. Penfield, W. and T. Rasmussen, The cerebral cortex of man; a clinical study of localization of function. 1950. 9. Autismindex. Motor And Sensory Homunculus. [Website] 2013; Available from: http://www.autismindex.com/Therapies/Therapy_Key_Word_Site_Map/sensory/ motor_sensory_homunculus.html. 10. Grey, H., Gray's Anatomy: The Anatomical Basis of Clinical Practice 40 ed. 2008, Churchill-Livingstone: Elsevier. 11. Veillette, C. Radiographic Anatomy of Adult Hand. 2012 [cited 2013 December, 14]; Available from: http://www.orthopaedicsone.com/x/nABeAQ. 12. Kaplan, E., Functional and surgical anatomy of the hand. Plastic and reconstructive surgery, 1968. 41(2): p. 178. 13. Idler, R., Anatomy and biomechanics of the digital flexor tendons. Hand clinics, 1985. 1(1): p. 3-11.

302

14. Doyle, J.R., Anatomy of the finger flexor tendon sheath and pulley system. J Hand Surg [Am], 1988. 13(4): p. 473-84. 15. Lin, G.-T., et al., Functional anatomy of the human digital flexor pulley system. The Journal of hand surgery, 1989. 14(6): p. 949-956. 16. Doyle, J.R., Anatomy of the finger flexor tendon sheath and pulley system. The Journal of Hand Surgery, 1988. 13(4): p. 473-484. 17. Neumann, D.A., Kinesiology of the musculoskeletal system. 2002: Mosby/Elsevier. 18. Vailas, A., et al., Physical activity and hypophysectomy on the aerobic capacity of ligaments and tendons. Journal of Applied physiology, 1978. 44(4): p. 542- 546. 19. Williams, J.G., Achilles tendon lesions in sport. Sports Med, 1986. 3(2): p. 114- 35. 20. Azar, C.A., J.E. Culver, and E.J. Fleegler, Blood supply of the flexor pollicis longus tendon. The Journal of hand surgery, 1983. 8(4): p. 471-475. 21. Lundborg, G., R. Myrhage, and B. Rydevik, The vascularization of human flexor tendons within the digital synovial sheath region—structural and functional aspects. The Journal of Hand Surgery, 1977. 2(6): p. 417-427. 22. Strickland, J.W., Development of flexor tendon surgery: twenty-five years of progress. J Hand Surg [Am], 2000. 25(2): p. 214-35. 23. Johnson, D.H. and R.A. Pedowitz, Practical orthopaedic sports medicine & arthroscopy [electronic resource]. 2007: Lippincott Williams & Wilkins. 24. Zhang, Z.Z., et al., Blood supply of the flexor digital tendon in the hand and its clinical significance. Surg Radiol Anat, 1990. 12(2): p. 113-7. 25. Manske, P.R. and P.A. Lesker, Flexor tendon nutrition. Hand clinics, 1985. 1(1): p. 13-24. 26. Peterson, W.W., P.R. Manske, and P.A. Lesker, The effect of flexor sheath integrity on nutrient uptake by primate flexor tendons. The Journal of hand surgery, 1986. 11(3): p. 413-416. 27. Gelberman, R., Flexor tendon physiology: tendon nutrition and cellular activity in injury and repair. Instructional course lectures, 1984. 34: p. 351-360. 28. Verdan, C.E., Primary repair of flexor tendons. J Bone Joint Surg Am, 1960. 42- a: p. 647-57. 29. Kleinert, H., S. Schepel, and T. Gill, Flexor tendon injuries. The Surgical clinics of North America, 1981. 61(2): p. 267. 30. Bunnell, S., Surgery of the hand. 2nd ed. 1948, Philadelphia: JB Lippincott. 31. Strickland, J.W., Results of flexor tendon surgery in zone II. Hand clinics, 1985. 1(1): p. 167-179. 32. Naam, N.H., Intratendinous rupture of the flexor digitorum profundus tendon in zones II and III. The Journal of hand surgery, 1995. 20(3): p. 478-483. 33. Stefanich, R.J., et al., Flexor tendon lacerations in zone V. The Journal of hand surgery, 1992. 17(2): p. 284-291. 303

34. Wang, J.H.-C., Mechanobiology of tendon. Journal of biomechanics, 2006. 39(9): p. 1563-1582. 35. Benjamin, M., E. Kaiser, and S. Milz, Structure‐function relationships in tendons: a review. Journal of anatomy, 2008. 212(3): p. 211-228. 36. Wang, J.H.C., Mechanobiology of tendon. Journal of Biomechanics, 2006. 39(9): p. 1563-1582. 37. Kannus, P., Structure of the tendon connective tissue. Scand J Med Sci Sports, 2000. 10(6): p. 312-20. 38. Woo, S.L.-Y., et al., Structure and Function of Tendon and Ligaments, in Basic Orthopaedic Biomechanics, V.C. Mow and W.C. Hayes, Editors. 1997, Lippincott-Raven Publishers: Philadelphia. p. 209-252. 39. Curwin, S.L., Biomechanics of Tendon and the Effects of Immobilization. Foot and Ankle Clinics, 1997. 2(3): p. 371-390. 40. Ker, R.F., Mechanics of tendon, from an engineering perspective. International Journal of Fatigue, 2007. 29(6): p. 1001-1009. 41. Woo, S.L.Y., et al., Anatomy, biology and biomechanics of tendon and ligament, in Orthopaedic Basic Science: Biology and Biomechanics of the Musculoskeletal System, J.A. Buckwalter, T.A. Einhorn, and S.M. Simon, Editors. 2000, American Academy of Orthopaedic Surgeons Rosemont, Ill. p. 581-616. 42. Louie, L., I. Yannas, and M. Spector, Tissue Engineered Tendon, in Frontiers in Tissue Engineering, J.C. Patrick, A. Mikos, and L. McIntire, Editors. 1998, Elsevier Sciences Ltd.: New York. p. 413-442. 43. Berenson, M.C., et al., Proteoglycans of human rotator cuff tendons. Journal of Orthopaedic Research, 1996. 14(4): p. 518-525. 44. Williams, I.F., K.G. McCullagh, and I.A. Silver, The Distribution of Types I and III Collagen and Fibronectin in the Healing Equine Tendon. Connective Tissue Research, 1984. 12(3-4): p. 211-227. 45. Derwin, K.A., et al., Proteoglycans and glycosaminoglycan fine structure in the mouse tail tendon fascicle. Journal of Orthopaedic Research, 2001. 19(2): p. 269-277. 46. Silver, F.H., J.W. Freeman, and G. Bradica, Structure and Function of Ligaments, Tendons, and Joint Capsule, in Repair and Regeneration of Ligaments, Tendons, and Joint Capsule, R.W. Walsh, Editor. 2005, Humana Press Inc.: Totowa. 47. Birk, D.E., et al., Collagen fibrillogenesis in vitro: interaction of types I and V collagen regulates fibril diameter. Journal of Cell Science 1990. 95(4): p. 649- 657. 48. Siegel, R.E., Galen on psychology, psychopathology, and function and diseases of the nervous system. 1973: Karger Basel. 49. Gratz, C., The history of tendon suture. Medical Journal and Record, 1928. 127: p. 156-157.

304

50. Mason, M.L. and C.G. Shearon, THE PROCESS OF TENDON REPAIRAN EXPERIMENTAL STUDY OF TENDON SUTURE AND TENDON GRAFT. Archives of Surgery, 1932. 25(4): p. 615-692. 51. Manske, P.R., History of flexor tendon repair. Hand Clinics, 2005. 21(2): p. 123- 127. 52. Mason, M., Primary and secondary tendon suture. A discussion of the significance of technique in tendon surgery. Surg Gynecol Obstet, 1940. 70: p. 392-402. 53. Kleinert, H., et al., Primary repair of lacerated flexor tendons in “no man’s land”. J Bone Joint Surg A, 1967. 49: p. 577. 54. Chow, S., G. Hooper, and C. Chan, The healing of freeze-dried rabbit flexor tendon in a synovial fluid environment. The Hand, 1983. 15(2): p. 136-142. 55. Gelberman, R.H., et al., Flexor tendon repair in vitro: a comparative histologic study of the rabbit, chicken, dog, and monkey. Journal of orthopaedic research, 1984. 2(1): p. 39-48. 56. Katsumi, M. and T. Tajima, Experimental investigation of healing process of tendons with or without synovial coverage in or outside of synovial cavity. J Niigata Med Assoc, 1981. 95: p. 532-567. 57. Lundborg, G. and F. Rank, Experimental studies on cellular mechanisms involved in healing of animal and human flexor tendon in synovial environment. Hand, 1980. 12(1): p. 3-11. 58. Potenza, A.D. and M.C. Herte, The synovial cavity as a “tissue culture in situ”— Science or nonsense? The Journal of Hand Surgery, 1982. 7(2): p. 196-199. 59. Strickland, J.W., Flexor Tendon Injuries: I. Foundations of Treatment. J Am Acad Orthop Surg, 1995. 3(1): p. 44-54. 60. Barrie, K.A., et al., The role of multiple strands and locking sutures on gap formation of flexor tendon repairs during cyclical loading. J Hand Surg [Am], 2000. 25(4): p. 714-20. 61. Hatanaka, H. and P.R. Manske, Effect of suture size on locking and grasping flexor tendon repair techniques. Clin Orthop Relat Res, 2000. 375(375): p. 267- 74. 62. Wada, A., et al., The mechanical properties of locking and grasping suture loop configurations in four-strand core suture techniques. J Hand Surg [Br], 2000. 25(6): p. 548-51. 63. Savage, R. and G. Risitano, Flexor tendon repair using a "six strand" method of repair and early active mobilisation. J Hand Surg Br, 1989. 14(4): p. 396-9. 64. Viinikainen, A., H. Goransson, and J. Ryhanen, Primary flexor tendon repair techniques. Scand J Surg, 2008. 97(4): p. 333-40. 65. Tonkin, M.A., Primary flexor tendon repair: surgical techniques based on the anatomy and biology of the flexor tendon system. World J Surg, 1991. 15(4): p. 452-7.

305

66. Tang, J.B., Clinical outcomes associated with flexor tendon repair. Hand Clin, 2005. 21(2): p. 199-210. 67. Goodson, A., et al., Current management of Jersey finger in rugby players: case series and literature review. Hand Surgery, 2010. 15(02): p. 103-107. 68. Rouzier. Jersey Finger. 2012; Available from: http://www.summitmedicalgroup.com/library/adult_health/sma_jersey_finger/. 69. Prosser, G., G. Sterne, and J.D. Nancarrow, Intratendinous rupture of flexor digitorum profundus caused by non-specific synovitis. British journal of plastic surgery, 2002. 55(1): p. 77-79. 70. Kitay, A., et al., Volar Plate Position and Flexor Tendon Rupture Following Distal Radius Fracture Fixation. The Journal of hand surgery, 2013. 38(6): p. 1091-1096. 71. Work Related Hand and Wrist Injuries in Australia. 2008, Safe Work Australia. 72. Injury Statistic of Traumatic Injuries. 2008, Queensland Employee Injury Database. 73. Clayton, R.A. and C.M. Court-Brown, The epidemiology of musculoskeletal tendinous and ligamentous injuries. Injury, 2008. 39(12): p. 1338-1344. 74. Griffin, M., et al., An overview of the management of flexor tendon injuries. Open Orthopaedics Journal, 2012. 6. 75. UK Department of Health. Hospital Episode Statistics. 2009 [cited 2014 January, 4]; Available from: http://www.hscic.gov.uk/hes. 76. Sandow, M.J. and M. McMahon, Active mobilisation following single cross grasp four-strand flexor tenorrhaphy (Adelaide repair). J Hand Surg Eur Vol, 2011. 36(6): p. 467-75. 77. Mittal, R.K., R. Garg, and A. Gupta, Unexpected retraction of distal cut end of flexor pollicis longus tendon. Journal of emergencies, trauma, and shock, 2009. 2(2): p. 145-146. 78. Walbeehm, E.T. and D.A. McGrouther, An anatomical study of the mechanical interactions of flexor digitorum superficialis and profundus and the flexor tendon sheath in zone 2. J Hand Surg Br, 1995. 20(3): p. 269-80. 79. Strickland, J.W., The scientific basis for advances in flexor tendon surgery. J Hand Ther, 2005. 18(2): p. 94-110; quiz 111. 80. Newmeyer III, W.L. and P.R. Manske, No man’s land revisited: the primary flexor tendon repair controversy. The Journal of hand surgery, 2004. 29(1): p. 1- 5. 81. Grobbelaar, A. and D. Hudson, Flexor tendon injuries in children. The Journal of Hand Surgery: British & European Volume, 1994. 19(6): p. 696-698. 82. Zhao, C., et al., Resection of the flexor digitorum superficialis reduces gliding resistance after zone II flexor digitorum profundus repair in vitro. The Journal of hand surgery, 2002. 27(2): p. 316-321.

306

83. Hwang, M.D., S. Pettrone, and T.E. Trumble, Work of flexion related to different suture materials after flexor digitorum profundus and flexor digitorum superficialis tendon repair in zone II: a biomechanical study. The Journal of hand surgery, 2009. 34(4): p. 700-704. 84. Viinikainen, A.K., et al., Bioabsorbable poly-L/D-lactide (PLDLA) 96/4 triple- stranded bound suture in the modified Kessler repair: an ex vivo static and cyclic tensile testing study in a porcine extensor tendon model. J Mater Sci Mater Med, 2009. 20(9): p. 1963-9. 85. Brockardt, C.J., et al., Evaluation of simple and looped suture and new material for flexor tendon repair. J Hand Surg Eur Vol, 2009. 34(3): p. 329-32. 86. Karjalainen, T., et al., Nickel-titanium wire as a flexor tendon suture material: an ex vivo study. J Hand Surg Eur Vol. 35(6): p. 469-74. 87. Trail, I.A., E.S. Powell, and J. Noble, An evaluation of suture materials used in tendon surgery. J Hand Surg [Br], 1989. 14(4): p. 422-7. 88. Scherman, P., et al., Cross-sectional area and strength differences of fiberwire, prolene, and ticron sutures. J Hand Surg Am. 35(5): p. 780-4. 89. McLarney, E., H. Hoffman, and S.W. Wolfe, Biomechanical analysis of the cruciate four-strand flexor tendon repair. J Hand Surg [Am], 1999. 24(2): p. 295-301. 90. Robertson, G.A. and M.M. al-Qattan, A biomechanical analysis of a new interlock suture technique for flexor tendon repair. J Hand Surg [Br], 1992. 17(1): p. 92-3. 91. Savage, R., In vitro studies of a new method of flexor tendon repair. J Hand Surg [Br], 1985. 10(2): p. 135-41. 92. Thurman, R.T., et al., Two-, four-, and six-strand zone II flexor tendon repairs: an in situ biomechanical comparison using a cadaver model. J Hand Surg [Am], 1998. 23(2): p. 261-5. 93. Barrie, K.A., et al., A biomechanical comparison of multistrand flexor tendon repairs using an in situ testing model. J Hand Surg Am, 2000. 25(3): p. 499-506. 94. Strick, M.J., et al., Adhesion formation after flexor tendon repair: a histologic and biomechanical comparison of 2- and 4-strand repairs in a chicken model1 1 No benefits in any form have been received or will be received by a commercial party related directly or indirectly to the subject of this article. The Journal of hand surgery, 2004. 29(1): p. 15-21. 95. Tang, J.B., Indications, methods, postoperative motion and outcome evaluation of primary flexor tendon repairs in Zone 2. J Hand Surg Eur Vol, 2007. 32(2): p. 118-29. 96. Healy, C., et al., Practice patterns in flexor tendon repair. Irish Journal of Medical Science, 2007. 176(1): p. 41-44. 97. Taras, J.S., et al., Evaluation of suture caliber in flexor tendon repair. The Journal of hand surgery, 2001. 26(6): p. 1100-1104.

307

98. Barrie, K.A., et al., Effect of suture locking and suture caliber on fatigue strength of flexor tendon repairs. J Hand Surg [Am], 2001. 26(2): p. 340-6. 99. Alavanja, G., E. Dailey, and D.P. Mass, Repair of zone II flexor digitorum profundus lacerations using varying suture sizes: a comparative biomechanical study. J Hand Surg [Am], 2005. 30(3): p. 448-54. 100. Viinikainen, A., et al., A comparative analysis of the biomechanical behaviour of five flexor tendon core sutures. J Hand Surg Br, 2004. 29(6): p. 536-43. 101. Lawrence, T., et al., An assessment of the tensile properties and technical difficulties of two-and four-strand flexor tendon repairs. Journal of Hand Surgery (British and European Volume), 2005. 30(3): p. 294-297. 102. Croog, A., et al., Comparative biomechanic performances of locked cruciate four-strand flexor tendon repairs in an ex vivo porcine model. J Hand Surg [Am], 2007. 32(2): p. 225-32. 103. Xie, R.G., et al., Effects of locking area on strength of 2- and 4-strand locking tendon repairs. J Hand Surg [Am], 2005. 30(3): p. 455-60. 104. Tanaka, T., et al., Gliding characteristics and gap formation for locking and grasping tendon repairs: a biomechanical study in a human cadaver model. J Hand Surg Am, 2004. 29(1): p. 6-14. 105. Cao, Y., et al., Influence of core suture purchase length on strength of four- strand tendon repairs. J Hand Surg Am, 2006. 31(1): p. 107-12. 106. Tang, J.B., et al., Core suture purchase affects strength of tendon repairs. J Hand Surg Am, 2005. 30(6): p. 1262-6. 107. Hatanaka, H. and P.R. Manske, Effect of the cross-sectional area of locking loops in flexor tendon repair. J Hand Surg [Am], 1999. 24(4): p. 751-60. 108. Peltz, T.S., et al., Influence of locking stitch size in a four-strand cross-locked cruciate flexor tendon repair. J Hand Surg Am, 2011. 36(3): p. 450-5. 109. Aoki, M., et al., Effect of suture knots on tensile strength of repaired canine flexor tendons. J Hand Surg Br, 1995. 20(1): p. 72-5. 110. Pruitt, D.L., M. Aoki, and P.R. Manske, Effect of suture knot location on tensile strength after flexor tendon repair. J Hand Surg Am, 1996. 21(6): p. 969-73. 111. Amadio, P., et al., IFSSH Flexor Tendon Committee report. J Hand Surg Br, 2005. 30(1): p. 100-16. 112. Aoki, M., et al., Biomechanical and histologic characteristics of canine flexor tendon repair using early postoperative mobilization. J Hand Surg Am, 1997. 22(1): p. 107-14. 113. Pruitt, D.L., P.R. Manske, and B. Fink, Cyclic stress analysis of flexor tendon repair. J Hand Surg Am, 1991. 16(4): p. 701-7. 114. Kleinert, H.E., et al., Primary repair of flexor tendons. Orthop Clin North Am, 1973. 4(4): p. 865-76. 115. Silfverskiold, K.L. and C.H. Andersson, Two new methods of tendon repair: an in vitro evaluation of tensile strength and gap formation. J Hand Surg Am, 1993. 18(1): p. 58-65. 308

116. Kubota, H., et al., Mechanical properties of various circumferential tendon suture techniques. J Hand Surg Br, 1996. 21(4): p. 474-80. 117. Mashadi, Z. and A. Amis, Strength of the suture in the epitenon and within the tendon fibres: development of stronger peripheral suture technique. The Journal of Hand Surgery: British & European Volume, 1992. 17(2): p. 172-175. 118. Diao, E., et al., Effect of peripheral suture depth on strength of tendon repairs. The Journal of hand surgery, 1996. 21(2): p. 234-239. 119. Merrell, G.A., et al., The effect of increased peripheral suture purchase on the strength of flexor tendon repairs. The Journal of hand surgery, 2003. 28(3): p. 464-468. 120. Tang, J.B., et al., Biomechanical evaluation of flexor tendon repair techniques. Clin Orthop Relat Res, 2001(386): p. 252-9. 121. Lotz, J.C., J.S. Hariharan, and E. Diao, Analytic model to predict the strength of tendon repairs. Journal of orthopaedic research, 1998. 16(4): p. 399-405. 122. Matheson, G., et al., Comparison of zone II flexor tendon repairs using an in vitro linear cyclic testing protocol. Clin Biomech (Bristol, Avon), 2005. 20(7): p. 718-22. 123. Sanders, D.W., et al., Cyclic testing of flexor tendon repairs: an in vitro biomechanical study. J Hand Surg Am, 1997. 22(6): p. 1004-10. 124. Oakes, B.W., Tissue healing and repair: tendons and ligaments. Rehabilitation of sports injuries: scientific basis, 2003: p. 28-98. 125. Abrahamsson, S.O., L.S. Lohmander, and G. Lundborg, Dehydration inhibits matrix synthesis and cell proliferation. An in vitro study of rabbit flexor tendons. Acta Orthop Scand, 1991. 62(2): p. 159-62. 126. Sharma, P. and N. Maffulli, Biology of tendon injury: healing, modeling and remodeling. Journal of Musculoskeletal and Neuronal Interactions, 2006. 6(2): p. 181. 127. Bruns, J., et al., Achilles tendon rupture: experimental results on spontaneous repair in a sheep-model. Knee Surgery, Sports Traumatology, Arthroscopy, 2000. 8(6): p. 364-369. 128. Gelberman, R.H., et al., The effect of gap formation at the repair site on the strength and excursion of intrasynovial flexor tendons. An experimental study on the early stages of tendon-healing in dogs. J Bone Joint Surg Am, 1999. 81(7): p. 975-82. 129. Wada, A., et al., Comparison of postoperative early active mobilization and immobilization in vivo utilising a four-strand flexor tendon repair. Journal of Hand Surgery (British and European Volume), 2001. 26(4): p. 301-306. 130. Feehan, L.M. and J.G. Beauchene, Early tensile properties of healing chicken flexor tendons: early controlled passive motion versus postoperative immobilization. The Journal of hand surgery, 1990. 15(1): p. 63-68.

309

131. Strickland, J.W. and S.V. Glogovac, Digital function following flexor tendon repair in Zone II: A comparison of immobilization and controlled passive motion techniques. J Hand Surg Am, 1980. 5(6): p. 537-43. 132. Aoki, M., et al., Work of flexion after flexor tendon repair according to the placement of sutures. Clinical orthopaedics and related research, 1995. 320: p. 205-210. 133. Tanaka, T., et al., Gliding resistance versus work of flexion—two methods to assess flexor tendon repair. Journal of orthopaedic research, 2003. 21(5): p. 813- 818. 134. Evans, R.B. and D.E. Thompson, The application of force to the healing tendon. J Hand Ther, 1993. 6(4): p. 266-84. 135. Howell, J.W. and F. Peck, Rehabilitation of flexor and extensor tendon injuries in the hand: Current updates. Injury, 2013. 44(3): p. 397-402. 136. Boyer, M.I., et al., The influence of cross-sectional area on the tensile properties of flexor tendons. J Hand Surg Am, 2001. 26(5): p. 828-32. 137. Lieber, R., et al., Relationship between joint motion and flexor tendon force in the canine forelimb. The Journal of hand surgery, 1996. 21(6): p. 957-962. 138. Lieber, R., et al., Wrist and digital joint motion produce unique flexor tendon force and excursion in the canine forelimb. Journal of biomechanics, 1999. 32(2): p. 175-181. 139. Groth, G.N., Pyramid of progressive force exercises to the injured flexor tendon. Journal of Hand Therapy, 2004. 17(1): p. 31-42. 140. Zhao, C., et al., Effect of gap size on gliding resistance after flexor tendon repair. J Bone Joint Surg Am, 2004. 86-A(11): p. 2482-8. 141. Cao, Y. and J. Tang, Biomechanical evaluation of a four-strand modification of the Tang method of tendon repair. Journal of Hand Surgery (British and European Volume), 2005. 30(4): p. 374-378. 142. Xie, R.G., et al., The gliding force and work of flexion in the early days after primary repair of lacerated flexor tendons: an experimental study. J Hand Surg Eur Vol, 2008. 33(2): p. 192-6. 143. Lilly, S.I. and T.M. Messer, Complications after treatment of flexor tendon injuries. Journal of the American Academy of Orthopaedic Surgeons, 2006. 14(7): p. 387-396. 144. Tang, J.B., D. Shi, and Q.G. Zhang, Biomechanical and histologic evaluation of tendon sheath management. The Journal of hand surgery, 1996. 21(5): p. 900- 908. 145. Beredjiklian, P.K., Biologic aspects of flexor tendon laceration and repair. The Journal of Bone & Joint Surgery, 2003. 85(3): p. 539-550. 146. Elliot, D., et al., IFSSH flexor tendon committee report 2007. Journal of Hand Surgery (European Volume), 2007. 32(3): p. 346-356.

310

147. Elliot, D., et al., The rupture rate of acute flexor tendon repairs mobilized by the controlled active motion regimen. The Journal of Hand Surgery: British & European Volume, 1994. 19(5): p. 607-612. 148. Walbeehm, E.T., et al., Influence of core suture geometry on tendon deformation and gap formation in porcine flexor tendons. J Hand Surg Eur Vol, 2009. 34(2): p. 190-5. 149. Dy, C.J., et al., Complications after flexor tendon repair: a systematic review and meta-analysis. The Journal of hand surgery, 2012. 37(3): p. 543-551. e1. 150. Hotokezaka, S. and P.R. Manske, Differences between locking loops and grasping loops: effects on 2-strand core suture. J Hand Surg [Am], 1997. 22(6): p. 995-1003. 151. Manske, P.R. and P.A. Lesker, Comparative nutrient pathways to the flexor profundus tendons in Zone II of various experimental animals. J Surg Res, 1983. 34(1): p. 83-93. 152. Mao, W.F., et al., A study of the anatomy and repair strengths of porcine flexor and extensor tendons: are they appropriate experimental models? J Hand Surg Eur Vol, 2011. 36(8): p. 663-9. 153. Hirpara, K.M., P.J. Sullivan, and M.E. O'Sullivan, The effects of freezing on the tensile properties of repaired porcine flexor tendon. J Hand Surg [Am], 2008. 33(3): p. 353-8. 154. Hausmann, J.T., et al., Biomechanical comparison of modified Kessler and running suture repair in 3 different animal tendons and in human flexor tendons. J Hand Surg [Am], 2009. 34(1): p. 93-101. 155. Gibbons, C.E., D. Thompson, and M.J. Sandow, Flexor tenorrhaphy tensile strength: reduction by cyclic loading : in vitro and ex vivo porcine study. Hand (N Y), 2009. 4(2): p. 113-8. 156. Wright, T.C., et al., A cost-effective training tool for flexor tendon repair: pig's trotters. J Plast Reconstr Aesthet Surg, 2006. 59(1): p. 107-8. 157. Smith, A.M., et al., The porcine forelimb as a model for human flexor tendon surgery. J Hand Surg Br, 2005. 30(3): p. 307-9. 158. Haddad, R., et al., A biomechanical assessment of repair versus nonrepair of sheep flexor tendons lacerated to 75 percent. J Hand Surg Am, 2010. 35(4): p. 546-51. 159. Dona, E., M.P. Gianoutsos, and W.R. Walsh, Optimizing biomechanical performance of the 4-strand cruciate flexor tendon repair. J Hand Surg [Am], 2004. 29(4): p. 571-80. 160. Tan, S., D. Power, and V. Rajaratnam, Animal models for training in tendon surgery: sheep's forelimbs. J Hand Surg Eur Vol, 2010. 35(2): p. 152-3. 161. Dogramaci, Y., et al., Does strand configuration and number of purchase points affect the biomechanical behavior of a tendon repair? A biomechanical evaluation using different kessler methods of flexor tendon repair. Hand (N Y), 2008. 3(3): p. 266-70. 311

162. Huang, H., et al., Effects of repetitive multiple freeze-thaw cycles on the biomechanical properties of human flexor digitorum superficialis and flexor pollicis longus tendons. Clin Biomech (Bristol, Avon), 2011. 26(4): p. 419-23. 163. Cao, Y., et al., Models for tendon biomechanical investigations. The Journal of hand surgery, 2009. 34(6): p. 1160-1161. 164. Wang, B., R. Xie, and J. Tang, Biomechanical analysis of a modification of Tang method of tendon repair. Journal of Hand Surgery (British and European Volume), 2003. 28(4): p. 347-350. 165. Xie, R.G., et al., Biomechanical studies of 3 different 6-strand flexor tendon repair techniques. The Journal of hand surgery, 2002. 27(4): p. 621-627. 166. Peltz, T.S., R. Haddad, and W.R. Walsh, The change in three-dimensional geometry of the Kessler flexor tendon repair under tension: a qualitative assessment using radiographs. J Hand Surg Eur Vol, 2010. 35(8): p. 676-7. 167. Sebastin, S.J., et al., History and evolution of the Kessler repair. J Hand Surg Am, 2013. 38(3): p. 552-61. 168. Labana, N., et al., A biomechanical analysis of the modified Tsuge suture technique for repair of flexor tendon lacerations. J Hand Surg Br, 2001. 26(4): p. 297-300. 169. Singer, G., et al., Use of the Taguchi Method for Biomechanical Comparison of Flexor-Tendon-Repair Techniques to Allow Immediate Active Flexion. A New Method of Analysis and Optimization of Technique to Improve the Quality of the Repair*ÔÇá. The Journal of Bone & Joint Surgery, 1998. 80(10): p. 1498-1506. 170. Peltz, T., et al., Performance of a knotless four-strand flexor tendon repair with a unidirectional barbed suture device: a dynamic ex vivo comparison. Journal of Hand Surgery (European Volume), 2013. 171. Xie, R.G. and J.B. Tang, Investigation of locking configurations for tendon repair. J Hand Surg [Am], 2005. 30(3): p. 461-5. 172. Wu, Y.F. and J.B. Tang, How much does a Pennington lock add to strength of a tendon repair? J Hand Surg Eur Vol, 2011. 36(6): p. 476-84. 173. Kim, H.M., et al., Technical and biological modifications for enhanced flexor tendon repair. The Journal of hand surgery, 2010. 35(6): p. 1031-1037. 174. Franchi, M., et al., Different crimp patterns in collagen fibrils relate to the subfibrillar arrangement. Connective tissue research, 2008. 49(2): p. 85-91. 175. Franchi, M., et al., Tendon and ligament fibrillar crimps give rise to leftÔÇÉhanded helices of collagen fibrils in both planar and helical crimps. Journal of Anatomy, 2010. 216(3): p. 301-309. 176. James, R., et al., Tendon: biology, biomechanics, repair, growth factors, and evolving treatment options. The Journal of hand surgery, 2008. 33(1): p. 102- 112. 177. Vidal, B.d.C., Image analysis of tendon helical superstructure using interference and polarized light microscopy. Micron, 2003. 34(8): p. 423-432.

312

178. Lundborg, G., R. Myrhage, and B. Rydevik, The vascularization of human flexor tendons within the digital synovial sheath regionÔÇöstructural and functional aspects. The Journal of hand surgery, 1977. 2(6): p. 417-427. 179. Dressler, M., et al., A potential mechanism for age-related declines in patellar tendon biomechanics. Journal of orthopaedic research, 2002. 20(6): p. 1315- 1322. 180. Pollock, C.M. and R.E. Shadwick, Relationship between body mass and biomechanical properties of limb tendons in adult mammals. American Journal of Physiology-Regulatory, Integrative and Comparative Physiology, 1994. 266(3): p. R1016-R1021. 181. Rosager, S., et al., LoadÔÇÉdisplacement properties of the human triceps surae aponeurosis and tendon in runners and nonÔÇÉrunners. Scandinavian journal of medicine & science in sports, 2002. 12(2): p. 90-98. 182. Wu, Y. and J. Tang, Recent developments in flexor tendon repair techniques and factors influencing strength of the tendon repair. Journal of Hand Surgery (European Volume), 2013. 183. Savage, R., The search for the ideal tendon repair in zone 2: strand number, anchor points and suture thickness. Journal of Hand Surgery (European Volume), 2013: p. 1753193413508699. 184. Peck, F., et al., A comparative study of two methods of controlled mobilization of flexor tendon repairs in zone 2. The Journal of Hand Surgery: British & European Volume, 1998. 23(1): p. 41-45. 185. Kessler, I. and F. Nissim, Primary repair without immobilization of flexor tendon division within the digital sheath: an experimental and clinical study. Acta Orthopaedica, 1969. 40(5): p. 587-601. 186. Kirchmayr, L., Zur technik der sehnennaht. Zentralbl Chir, 1917. 44: p. 906-907. 187. Urbaniak, J., J. Cahill, and R. Mortenson. Tendon suturing methods: analysis of tensile strengths. in American Academy of Orthopaedic Surgeons symposium on tendon surgery in the hand. 1975. 188. Sebastin, S.J., et al., History and Evolution of the Kessler Repair. The Journal of hand surgery, 2013. 189. Pennington, D.G., The locking loop tendon suture. Plastic and reconstructive surgery, 1979. 63(5): p. 648-652. 190. Karjalainen, T., et al., An analysis of the pull-out strength of 6 suture loop configurations in flexor tendons. The Journal of hand surgery, 2012. 37(2): p. 217-223. 191. Waitayawinyu, T., et al., Comparative biomechanic study of flexor tendon repair using FiberWire. J Hand Surg [Am], 2008. 33(5): p. 701-8. 192. Ketchum, L.D., Suture materials and suture techniques used in tendon repair. Hand clinics, 1985. 1(1): p. 43-53. 193. Srugi, S. and J.E. Adamson, A comparative study of tendon suture materials in dogs. Plastic and Reconstructive Surgery, 1972. 50(1): p. 31-35. 313

194. McDonald, E., et al., Comparison of a multifilament stainless steel suture with FiberWire for flexor tendon repairs -- an in vitro biomechanical study. J Hand Surg Eur Vol, 2012. 195. Sandow MJ, M.M., Single-cross grasp six-strand repair for acute flexor tenorrhaphy: Modified Savage technique. Atlas Hand Clin 1996. 1: p. 41-64. 196. Wade, P., I. Muir, and L. Hutcheon, Primary flexor tendon repair: the mechanical limitations of the modified Kessler technique. The Journal of Hand Surgery: British & European Volume, 1986. 11(1): p. 71-76. 197. Schuind, F., et al., Flexor tendon forces: in vivo measurements. J Hand Surg Am, 1992. 17(2): p. 291-8. 198. Silfverskiöld, K., E. May, and A. Törnvall, Gap formation during controlled motion after flexor tendon repair in zone II: a prospective clinical study. The Journal of hand surgery, 1992. 17(3): p. 539-546. 199. Lawrence, T.M. and T.R. Davis, A biomechanical analysis of suture materials and their influence on a four-strand flexor tendon repair. The Journal of hand surgery, 2005. 30(4): p. 836-841. 200. Sandow M, K.S., Flexor Tendon Injuries, in Rehabilitation of the Hand and Upper Limb., C.W. Prosser R, eds., Editor. 2003, Butterworth Heinemann: Edinburgh. p. 46-52. 201. Angeles, J.G., H. Heminger, and D.P. Mass, Comparative biomechanical performances of 4-strand core suture repairs for zone II flexor tendon lacerations. J Hand Surg [Am], 2002. 27(3): p. 508-17. 202. Hirpara, K.M., et al., A biomechanical analysis of multistrand repairs with the Silfverskiold peripheral cross-stitch. J Bone Joint Surg Br, 2007. 89(10): p. 1396-401. 203. Ng, B.H., et al., The changes in the tensile properties of tendons after freeze storage in saline solution. Proc Inst Mech Eng [H], 2005. 219(6): p. 387-92. 204. Kitis, P.T., N. Buker, and I.G. Kara, Comparison of two methods of controlled mobilisation of repaired flexor tendons in zone 2. Scand J Plast Reconstr Surg Hand Surg, 2009. 43(3): p. 160-5. 205. Herrmann, J.B., Tensile strength and knot security of surgical suture materials. Am Surg, 1971. 37(4): p. 209-17. 206. Abrahamsson, S.O., G. Lundborg, and L.S. Lohmander, Tendon healing in vivo. An experimental model. Scand J Plast Reconstr Surg Hand Surg, 1989. 23(3): p. 199-205. 207. Pennington, D.G., The locking loop tendon suture. Plast Reconstr Surg, 1979. 63(5): p. 648-52. 208. Manske, P.R., et al., Intrinsic restoration of the flexor tendon surface in the nonhuman primate. J Hand Surg Am, 1985. 10(5): p. 632-7. 209. Gelberman, R.H., V. Khabie, and C.J. Cahill, The revascularization of healing flexor tendons in the digital sheath. A vascular injection study in dogs. J Bone Joint Surg Am, 1991. 73(6): p. 868-81. 314

210. Wong, J.K., et al., The cellular effect of a single interrupted suture on tendon. J Hand Surg Br, 2006. 31(4): p. 358-67. 211. Wong, J.K., et al., The cell biology of suturing tendons. Matrix Biol, 2010. 2010: p. 21. 212. Small, J.O., M.D. Brennen, and J. Colville, Early active mobilisation following flexor tendon repair in zone 2. J Hand Surg [Br], 1989. 14(4): p. 383-91. 213. Haddad, R., et al., The relationship between gap formation and grip-to-grip displacement during cyclic testing of repaired flexor tendons. J Biomech, 2010. 43(14): p. 2835-8. 214. Gill, R.S., et al., A comparative analysis of the six-strand double-loop flexor tendon repair and three other techniques: a human cadaveric study. J Hand Surg Am, 1999. 24(6): p. 1315-22. 215. Vigler, M., et al., Biomechanical study of cross-locked cruciate versus Strickland flexor tendon repair. J Hand Surg Am, 2008. 33(10): p. 1826-33. 216. Momose, T., et al., The effect of knot location, suture material, and suture size on the gliding resistance of flexor tendons. J Biomed Mater Res, 2000. 53(6): p. 806-11. 217. Rees, L., et al., Comparison of 1- and 2-knot, 4-strand, double-modified kessler tendon repairs in a porcine model. J Hand Surg Am, 2009. 34(4): p. 705-9. 218. Patri, P., et al., V-loc, a new wound closure device for peritoneal closure--is it safe? A comparative study of different peritoneal closure systems. Surg Innov, 2011. 18(2): p. 145-9. 219. Jennings, E.R., et al., A new technique in primary tendon repair. Surg Gynecol Obstet, 1952. 95(5): p. 597-600. 220. Bunnell, S., Gig pull-out suture for tendons. J Bone Joint Surg Am, 1954. 36- A(4): p. 850-1. 221. McKenzie, A.R., An experimental multiple barbed suture for the long flexor tendons of the palm and fingers. Preliminary report. J Bone Joint Surg Br, 1967. 49(3): p. 440-7. 222. Sulamanidze, M., Evaluation of a novel technique for wound closure using a barbed suture. Plast Reconstr Surg, 2007. 120(1): p. 349-50; author reply 350. 223. Zaruby, J., et al., An in vivo comparison of barbed suture devices and conventional monofilament sutures for cosmetic skin closure: biomechanical wound strength and histology. Aesthet Surg J, 2011. 31(2): p. 232-40. 224. Paul, M.D., Barbed sutures for aesthetic facial plastic surgery: indications and techniques. Clin Plast Surg, 2008. 35(3): p. 451-61. 225. Villa, M.T., et al., Barbed sutures: a review of the literature. Plast Reconstr Surg, 2008. 121(3): p. 102e-108e. 226. Demyttenaere, S.V., et al., Barbed suture for gastrointestinal closure: a randomized control trial. Surg Innov, 2009. 16(3): p. 237-42.

315

227. Rosen, A.D., Use of absorbable running barbed suture and progressive tension technique in abdominoplasty: a novel approach. Plast Reconstr Surg, 2010. 125(3): p. 1024-7. 228. Greenberg, J.A., The use of barbed sutures in obstetrics and gynecology. Rev Obstet Gynecol, 2010. 3(3): p. 82-91. 229. Parikh, P.M., S.P. Davison, and J.P. Higgins, Barbed suture tenorrhaphy: an ex vivo biomechanical analysis. Plast Reconstr Surg, 2009. 124(5): p. 1551-8. 230. Trocchia, A.M., H.N. Aho, and G. Sobol, A re-exploration of the use of barbed sutures in flexor tendon repairs. Orthopedics, 2009. 32(10). 231. Zeplin, P.H., et al., Biomechanical evaluation of flexor tendon repair using barbed suture material: a comparative ex vivo study. J Hand Surg Am, 2011. 36(3): p. 446-9. 232. Marrero-Amadeo, I.C., et al., Flexor tendon repair with a knotless barbed suture: a comparative biomechanical study. J Hand Surg Am, 2011. 36(7): p. 1204-8. 233. McClellan, W.T., et al., A knotless flexor tendon repair technique using a bidirectional barbed suture: an ex vivo comparison of three methods. Plast Reconstr Surg, 2011. 128(4): p. 322e-327e. 234. Wu, Y.F., et al., Biomechanical comparisons of four-strand tendon repairs with double-stranded sutures: effects of different locks and suture geometry. J Hand Surg Eur Vol, 2011. 36(1): p. 34-9. 235. Lee, S.K., et al., The effects of core suture purchase on the biomechanical characteristics of a multistrand locking flexor tendon repair: a cadaveric study. J Hand Surg Am, 2010. 35(7): p. 1165-71. 236. Buschmann, J., et al., Small hook thread (Quill) and soft felt internal splint to increase the primary repair strength of lacerated rabbit Achilles tendons: biomechanical analysis and considerations for hand surgery. Clin Biomech (Bristol, Avon), 2011. 26(6): p. 626-31. 237. Havulinna, J., et al., Comparison of modified Kessler tendon suture at different levels in the human flexor digitorum profundus tendon and porcine flexors and porcine extensors: an experimental biomechanical study. Journal of Hand Surgery (European Volume), 2011. 36(8): p. 670-676. 238. Landis, W.J. and F.H. Silver, The structure and function of normally mineralizing avian tendons. Comparative Biochemistry and Physiology Part A: Molecular & Integrative Physiology, 2002. 133(4): p. 1135-1157. 239. Silver, F.H., et al., The role of mineral in the storage of elastic energy in turkey tendons. Biomacromolecules, 2000. 1(2): p. 180-185. 240. Landis, W., The strength of a calcified tissue depends in part on the molecular structure and organization of its constituent mineral crystals in their organic matrix. Bone, 1995. 16(5): p. 533-544. 241. Kew, S.J., et al., Synthetic collagen fascicles for the regeneration of tendon tissue. Acta Biomaterialia, 2012. 8(10): p. 3723-3731. 316

242. Holladay, C., et al., Preferential tendon stem cell response to growth factor supplementation. Journal of tissue engineering and regenerative medicine, 2014. 243. Oryan, A., A. Moshiri, and A. Meimandi-Parizi, In vitro characterization of a novel tissue engineered based hybridized nano and micro structured collagen implant and its in vivo role on tenoinduction, tenoconduction, tenogenesis and tenointegration. Journal of Materials Science: Materials in Medicine, 2013: p. 1- 25. 244. Müller, S., et al., Tendon healing: an overview of physiology, biology, and pathology of tendon healing and systematic review of state of the art in tendon bioengineering. Knee Surgery, Sports Traumatology, Arthroscopy, 2013: p. 1-9. 245. Tan, Q., P.P.Y. Lui, and Y.W. Lee, In vivo identity of tendon stem cells and the roles of stem cells in tendon healing. Stem cells and development, 2013. 22(23): p. 3128-3140. 246. Zhao, C., et al., The Effects of Bio-Lubricating Molecules on Flexor Tendon Reconstruction in A Canine Allograft Model In Vivo. Plastic and reconstructive surgery, 2014. 247. Zhao, C., et al., Effects of a lubricin-containing compound on the results of flexor tendon repair in a canine model in vivo. The Journal of Bone & Joint Surgery, 2010. 92(6): p. 1453-1461. 248. Zhao, C., et al., An analysis of factors associated with failure of tendon repair in the canine model. The Journal of hand surgery, 2007. 32(4): p. 518-525. 249. Zhao, C., et al., Short-term assessment of optimal timing for postoperative rehabilitation after flexor digitorum profundus tendon repair in a canine model. Journal of Hand Therapy, 2005. 18(3): p. 322-329. 250. Su, B., et al., Gross and histological analysis of healing after dog flexor tendon repair with the Teno Fix™ device. The Journal of Hand Surgery: British & European Volume, 2006. 31(5): p. 524-529. 251. Baker, A., et al., Homotypic variation of canine flexor tendons: implications for the design of experimental studies in animal models. Journal of biomechanics, 2004. 37(7): p. 959-968. 252. Hatanaka, H., J. Zhang, and P.R. Manske, An in vivo study of locking and grasping techniques using a passive mobilization protocol in experimental animals. J Hand Surg [Am], 2000. 25(2): p. 260-9. 253. Gelberman, R.H., et al., Flexor tendon repair. Journal of orthopaedic research, 1986. 4(1): p. 119-128. 254. Gelberman, R.H., et al., The Early Effects of Sustained Platelet-Derived Growth Factor Administration on the Functional and Structural Properties of Repaired Intrasynovial Flexor Tendons: An In Vivo Biomechanic Study at 3 Weeks in Canines. The Journal of hand surgery, 2007. 32(3): p. 373-379. 255. Fufa, D.T., et al., The Effect of Core and Epitendinous Suture Modifications on Repair of Intrasynovial Flexor Tendons in an In Vivo Canine Model. The Journal of hand surgery, 2012. 37(12): p. 2526-2531. 317

256. Moshiri, A., A. Oryan, and A. Meimandi-Parizi, Role of Tissue-Engineered Artificial Tendon in Healing of a Large Achilles Tendon Defect Model in Rabbits. Journal of the American College of Surgeons. 217(3): p. 421-441.e8. 257. Viinikainen, A., et al., Flexor tendon healing within the tendon sheath using bioabsorbable poly-l/d-lactide 96/4 suture. A histological in vivo study with rabbits. Journal of Materials Science: Materials in Medicine, 2014: p. 1-7. 258. Xia Sr, C., et al., Tendon healing in vivo and in vitro: neutralizing antibody to TGF-β improves range of motion after flexor tendon repair. Orthopedics, 2010. 33(11): p. 809. 259. Oryan, A., et al., Repeated administration of exogenous Sodium-hyaluronate improved tendon healing in an in vivo transection model. Journal of Tissue Viability, 2012. 21(3): p. 88-102. 260. Oryan, A., A. Moshiri, and P.A. Meimandi, Alcoholic extract of tarantula cubensis improves sharp ruptured tendon healing after primary repair in rabbits. American journal of orthopedics (Belle Mead, NJ), 2012. 41(12): p. 554- 560. 261. Sato, D., et al., Effect of Platelet-Rich Plasma With Fibrin Matrix on Healing of Intrasynovial Flexor Tendons. The Journal of hand surgery, 2012. 37(7): p. 1356-1363. 262. Håkansson, J., et al., Effect of Lactoferrin Peptide (PXL01) on Rabbit Digit Mobility After Flexor Tendon Repair. The Journal of hand surgery, 2012. 37(12): p. 2519-2525. 263. Chen, C.H., et al., Tendon Healing In Vivo: Gene Expression and Production of Multiple Growth Factors in Early Tendon Healing Period. The Journal of hand surgery, 2008. 33(10): p. 1834-1842. 264. Xu, L., et al., In vivo engineering of a functional tendon sheath in a hen model. Biomaterials, 2010. 31(14): p. 3894-3902. 265. Chen, C.H., et al., Effectiveness of MicroRNA in Down-Regulation of TGF-β Gene Expression in Digital Flexor Tendons of Chickens: In Vitro and In Vivo Study. The Journal of hand surgery, 2009. 34(10): p. 1777-1784.e1. 266. Tang, J.B., et al., Adeno-Associated Virus-2-Mediated bFGF Gene Transfer to Digital Flexor Tendons Significantly Increases Healing StrengthAn in Vivo Study. The Journal of Bone & Joint Surgery, 2008. 90(5): p. 1078-1089. 267. Cao, Y., et al., Bridging tendon defects using autologous tenocyte engineered tendon in a hen model. Plastic and reconstructive surgery, 2002. 110(5): p. 1280- 1289. 268. Siddiqi, N., Y. Hamada, and A. Noryia, The healing of flexor tendons in chickens. International orthopaedics, 1992. 16(4): p. 363-368. 269. Webb, W.R., et al., The application of poly (3-hydroxybutyrate-co-3- hydroxyhexanoate) scaffolds for tendon repair in the rat model. Biomaterials, 2013. 34(28): p. 6683-6694.

318

270. Liang, J.-I., et al., The effect of tenocyte/hyaluronic acid therapy on the early recovery of healing Achilles tendon in rats. Journal of Materials Science: Materials in Medicine, 2014. 25(1): p. 217-227. 271. Kessler, M.W., et al., Enhancement of Achilles tendon repair mediated by matrix metalloproteinase inhibition via systemic administration of doxycycline. Journal of orthopaedic research, 2013. 272. Fung, D.T., et al., Early response to tendon fatigue damage accumulation in a novel in vivo model. Journal of biomechanics, 2010. 43(2): p. 274-279. 273. Hapa, O., et al., The effect of ethanol intake on tendon healing: a histological and biomechanical study in a rat model. Archives of orthopaedic and trauma surgery, 2009. 129(12): p. 1721-1726. 274. Rickert, M., et al., Adenovirus-mediated gene transfer of growth and differentiation factor-5 into tenocytes and the healing rat Achilles tendon. Connective tissue research, 2005. 46(4-5): p. 175-183. 275. Chan, B.P., et al., Photochemical repair of Achilles tendon rupture in a rat model 1. Journal of Surgical Research, 2005. 124(2): p. 274-279. 276. Atik, B., et al., A new method in tendon repair: angular technique of interlocking (ATIK). Annals of plastic surgery, 2008. 60(3): p. 251-253. 277. Xu, Y. and J.B. Tang, Effects of superficialis tendon repairs on lacerated profundus tendons within or proximal to the A2 pulley: an in vivo study in chickens1 1 No benefits in any form have been received will be received from a commercial party related directly or indirectly to the subject of this article. The Journal of hand surgery, 2003. 28(6): p. 994-1001. 278. Martinello, T., et al., Effects of in vivo applications of peripheral blood‐derived mesenchymal stromal cells (PB‐MSCs) and platlet‐rich plasma (PRP) on experimentally injured deep digital flexor tendons of sheep. Journal of orthopaedic research, 2013. 31(2): p. 306-314. 279. Fernández-Sarmiento, J.A., et al., Histological study of the influence of plasma rich in growth factors (PRGF) on the healing of divided Achilles tendons in sheep. The Journal of Bone & Joint Surgery, 2013. 95(3): p. 246-255. 280. Dickerson, D.A., et al., In vitro and in vivo evaluation of orthopedic interface repair using a tissue scaffold with a continuous hard tissue-soft tissue transition. Journal of orthopaedic surgery and research, 2013. 8(1): p. 18. 281. Liem, D., et al., In vivo blood flow after rotator cuff reconstruction in a sheep model: comparison of single versus double row. Knee Surgery, Sports Traumatology, Arthroscopy, 2013: p. 1-8. 282. Baums, M.H., et al., Biomechanical and Magnetic Resonance Imaging Evaluation of a Single- and Double-Row Rotator Cuff Repair in an In Vivo Sheep Model. Arthroscopy : the journal of arthroscopic & related surgery : official publication of the Arthroscopy Association of North America and the International Arthroscopy Association, 2012. 28(6): p. 769-777.

319

283. Spang, J.T., et al., A Biomechanical Comparison of 2 Transosseous-Equivalent Double-Row Rotator Cuff Repair Techniques Using Bioabsorbable Anchors: Cyclic Loading and Failure Behavior. Arthroscopy : the journal of arthroscopic & related surgery : official publication of the Arthroscopy Association of North America and the International Arthroscopy Association, 2009. 25(8): p. 872-879. 284. Werner, S.J., et al., Functional significance of ultraviolet feeding cues in wild turkeys. Physiology & behavior, 2014. 123: p. 162-167. 285. Willems, O.W., S.P. Miller, and B.J. Wood, Assessment of residual body weight gain and residual intake and body weight gain as feed efficiency traits in the turkey (Meleagris gallopavo). Genetics Selection Evolution, 2013. 45(1): p. 26. 286. Case, L., B. Wood, and S. Miller, The investigation of ultrasound technology to measure breast muscle depth as a correlated trait to breast meat yield in turkey (Meleagris gallopavo). Journal of animal science, 2012. 90(10): p. 3410-3417. 287. Maxwell, E.E., Comparative embryonic development of the skeleton of the domestic turkey ( Meleagris gallopavo) and other galliform birds. Zoology, 2008. 111(3): p. 242-257. 288. Thornton, E.K., et al., Earliest Mexican Turkeys (Meleagris gallopavo) in the Maya Region: Implications for Pre-Hispanic Animal Trade and the Timing of Turkey Domestication. PloS one, 2012. 7(8): p. e42630. 289. Zhang, Y., et al., A comparative physical map reveals the pattern of chromosomal evolution between the turkey (Meleagris gallopavo) and chicken (Gallus gallus) genomes. BMC genomics, 2011. 12(1): p. 447. 290. Speller, C.F., et al., Ancient mitochondrial DNA analysis reveals complexity of indigenous North American turkey domestication. Proceedings of the National Academy of Sciences, 2010. 107(7): p. 2807-2812. 291. Dalloul, R.A., et al., Next-generation sequencing strategies for characterizing the turkey genome. Poultry Science, 2014. 93(2): p. 479-484. 292. Case, L.A., B.J. Wood, and S.P. Miller, The genetic parameters of feed efficiency and its component traits in the turkey (Meleagris gallopavo). Genet Sel Evol, 2012. 44(2). 293. Aslam, M.L., et al., Whole genome SNP discovery and analysis of genetic diversity in Turkey (Meleagris gallopavo). BMC genomics, 2012. 13(1): p. 391. 294. Dalloul, R.A., et al., Multi-platform next-generation sequencing of the domestic turkey (Meleagris gallopavo): genome assembly and analysis. PLoS biology, 2010. 8(9): p. e1000475. 295. Strick, M.J., et al., Adhesion formation after flexor tendon repair: comparison of two-and four-strand repair without epitendinous suture. Hand Surgery, 2005. 10(02n03): p. 193-197. 296. Porter, W., Family Meleagrididae (Turkeys). Handbook of the Birds of the World, 1994. 2: p. 364-375.

320

297. Marchewka, J., et al., Review of the social and environmental factors affecting the behavior and welfare of turkeys (Meleagris gallopavo). Poultry Science, 2013. 92(6): p. 1467-1473. 298. Hawkins, P., et al., Laboratory birds: refinements in husbandry and procedures. Laboratory Animals, 2001. 35(Suppl 1): p. 1-163. 299. Harrison, G.J. and L.R. Harrison, Clinical avian medicine and surgery including aviculture. 1986: WB Saunders Co. 300. Muir WW, H.L., Pharmacology of inhalation anes- thetic drugs. , in Handbook of Veterinary Anesthesia., H.L. Muir WW, Editor. 2000, Mosby: St. Louis. p. 164-181. 301. Lierz, M. and R. Korbel, Anesthesia and analgesia in birds. Journal of Exotic Pet Medicine, 2012. 21(1): p. 44-58. 302. Close, B., et al., Recommendations for euthanasia of experimental animals: Part 1. Laboratory Animals, 1996. 30(4): p. 293-316. 303. Getty, R., Sisson and Grossman's The anatomy of the domestic animals. Volumes 1 and 2. Sisson and Grossman's The anatomy of the domestic animals. Volumes 1 and 2., 1975(5th edition). 304. Bruck, J. and R. Schlögel, [Experiences with a resorbable suture (PDS) in tendon sutures]. Handchirurgie, Mikrochirurgie, plastische Chirurgie: Organ der Deutschsprachigen Arbeitsgemeinschaft fur Handchirurgie: Organ der Deutschsprachigen Arbeitsgemeinschaft fur Mikrochirurgie der Peripheren Nerven und Gefasse: Organ der V.. 1985. 17(4): p. 238-240. 305. Kang, H.J., et al., Flexor tenorrhaphy using absorbable suture materials. Archives of plastic surgery, 2012. 39(4): p. 397-403. 306. Jann, H., et al., A comparison of nylon, polybutester, and polyglyconate suture materials for long digital flexor tenorrhaphy in chickens. Veterinary Surgery, 1992. 21(3): p. 234-237. 307. Covidien, Covidien V-Loc product catalog. 2013, Covidien. 308. Momose, T., et al., Suture techniques with high breaking strength and low gliding resistance: experiments in the dog flexor digitorum profundus tendon. Acta Orthopaedica, 2001. 72(6): p. 635-641. 309. Zhao, C., et al., The effect of suture technique on adhesion formation after flexor tendon repair for partial lacerations in a canine model. Journal of Trauma- Injury, Infection, and Critical Care, 2001. 51(5): p. 917-921. 310. Sanders, D.W., et al., The effect of flexor tendon repair bulk on tendon gliding during simulated active motion: An in vitro comparison of two-strand and six-strand techniques. The Journal of hand surgery, 2001. 26(5): p. 833-840.

321