Original Article
A miniature tension sensor to measure surgical suture tension of deformable musculoskeletal tissues during joint motion
Proc IMechE Part H: J Engineering in Medicine 2014, Vol. 228(2) 140–148 Ó IMechE 2014 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0954411913518317 pih.sagepub.com
Yoshimori Kiriyama1, Hideo Matsumoto2, Yoshiaki Toyama3 and Takeo Nagura3
Abstract The aim of this study was to develop a new suture tension sensor for musculoskeletal soft tissue that shows deformation or movements. The suture tension sensor was 10 mm in size, which was small enough to avoid conflicting with the adjacent sensor. Furthermore, the sensor had good linearity up to a tension of 50 N, which is equivalent to the breaking strength of a size 1 absorbable suture defined by the United States Pharmacopeia. The design and mechanism were analyzed using a finite element model prior to developing the actual sensor. Based on the analysis, adequate material was selected, and the output linearity was confirmed and compared with the simulated result. To evaluate practical application, the incision of the skin and capsule were sutured during simulated total knee arthroplasty. When conventional surgery and minimally invasive surgery were performed, suture tensions were compared. In minimally invasive surgery, the distal portion of the knee was dissected, and the proximal portion of the knee was dissected additionally in conventional surgery. In the skin suturing, the maximum tension was 4.4 N, and this tension was independent of the sensor location. In contrast, the sensor suturing the capsule in the distal portion had a tension of 4.4 N in minimally invasive surgery, while the proximal sensor had a tension of 44 N in conventional surgery. The suture tensions increased nonlinearly and were dependent on the knee flexion angle. Furthermore, the tension changes showed hysteresis. This miniature tension sensor may help establish the optimal suturing method with adequate tension to ensure wound healing and early recovery.
Keywords Suture tension, miniature sensor, soft tissue deformation
Date received: 16 April 2013; accepted: 3 December 2013
Introduction Suturing soft tissues, such as the tendons, capsules, ligaments, muscles, nerves, and skin, is an essential component of musculoskeletal surgeries. Achieving optimal suture tension is critical for successful surgery because under-tension is one of the important factors to result in a bad apposition or a scar,1 while we undergo over-tension resulting in cutout of the surgical wound. However, there is a lack of the quantitative criteria to guide suture tension since tension is determined manually depending on surgeon’s technique and experience.2,3 Various suture techniques using suture guides and in situ tension measurement have been utilized in arthroscopic surgeries,4,5 but direct measurement of suture
tension is technically difficult because of limited working space and the small size of the target tissue in arthroscopic surgeries. Thus, measurements have been performed using an extra-articular approach with large transducers.6,7 Additionally, most of the target soft tissues in the orthopedic field undergo physiologic deformation and movement that occurs with joint
1
Department of Clinical Biomechanics, Keio University, Tokyo, Japan Institute for Integrated Sports Medicine, Keio University, Tokyo, Japan 3 Department of Orthopaedic Surgery, Keio University, Tokyo, Japan 2
Corresponding author: Yoshimori Kiriyama, Department of Clinical Biomechanics, Keio University, 35 Shinanomachi, Shinjuku, Tokyo 160-8582, Japan. Email: [email protected]
Downloaded from pih.sagepub.com at UNIV OF PENNSYLVANIA on June 14, 2015
Kiriyama et al.
141
movement. Therefore, characteristics of a useful tension sensor include small size, to enable it to be placed around the soft tissues and to avoid conflicting with adjacent tissues or sensors, as well as large measurement range of the tension. Several types of small sensors have been developed to measure static tension (i.e. initial tension) during surgery.8–11 These sensors are intended for use within soft tissues that are stationary and nondeformable, and the size and design of the sensors are not particularly amenable to use within musculoskeletal tissues. Thus, a new, smaller tension sensor with an acceptable tension range is needed to measure continuous changes in suture tension in deformable soft tissues and to provide quantitative biomechanical data regarding surgical sutures. Such a sensor could also be useful in obtaining the basic characterization of soft tissues, in developing new criteria for surgical techniques and in assessing the mechanical strength of surgical sutures. For these requirements, a new type of sensor was developed, and preliminary data were shown.12 The objective of this article was to show the detail of a miniature tension sensor to measure suture tension that is potentially modulated by joint movement. To evaluate the practical utility of the sensor, we investigated actual suture tension during joint movements in a cadaver.
Material and methods Design of miniature tension sensor The optimal tension sensor must (1) be small enough to avoid interfering with the surgical procedure or with physiologic deformation of soft tissues, (2) be able to adequately measure suture tension, and (3) have acceptable accuracy and range to measure surgical suture tension. In this study, we assumed that a maximum sensor size of 10 mm satisfied size requirements during surgery. To select adequate material, finite element (FE) analysis was conducted. An adequate range of tension measurement was designated as 0–50 N, which is the breaking strength of a size 1 absorbable suture defined by United States Pharmacopeia (USP). Thus, a sensor was designed as shown in Figure 1. In the design of a miniature tension sensor, the length is 410 mm, and the height is 4 mm maximally. Considering the possibility and ease to manufacture
requirements in addition to satisfying the requirements, we designed the sensor that is constructed with combination of simple shapes. A suture passes over the center arc via the end holes of both sides. The center arc had a radius of 2.0 mm, and a strain gage was attached at the concave side. In this study, a self-temperature compensated strain gage (KYOWA, Japan) was used for the sensor. The unique feature of the sensor was the design to measure tension, and the gage measures strain dependent on the tension when tension was applied to a suture and when the suture generates forces to bend the arc.
FE analysis Strain distribution and magnitude of the sensor model were evaluated to confirm the range and the linearity of the output from the sensor. A quarter symmetric model was constructed, as shown in Figure 2(a). To estimate the strain generated by suture tension, finite element models (FEMs) consisting of the sensor, suture, and the soft ground were constructed, as shown in Figure 2(b). The sensor, the suture thread, and the soft ground meshes had approximately 14,300, 800, and 5200 hexahedral elements, respectively. The base of the soft ground was constrained in three translations (Figure 2(c)). Since the quarter symmetric model was used in this study, all the nodes on the symmetric planes constrained to stay on the symmetric plane (Figure 2(c) and (d)). The interactions between the sensor and the soft ground and between the sensor and the thread were assumed to be frictionless. In initial condition, the sensor was placed on the top surface of the sensor. Also, the suture thread was set to touch the top round of the suture sensor. To consider the length and width of a strain gage for the sensor, five elements along the bottom arc and two elements along y-axis from the symmetrical center of the bottom arc elements were selected as sensing elements. We set the local coordinate system in which an axis corresponds to the tangential orientation of the bottom arc. The strain along the tangential orientation of the arc was calculated on the nodes that were on the surface of the sensing elements. The collected strain was averaged over the sensing elements, and the value was used as a representative strain. For the sensor, three different materials were prepared: aluminum (AL, E = 70 GPa, y = 0.3),
Tension sensor
4
Suture
4
10
tension
Figure 1. Sensor design of a miniature suture tension sensor (left top: side view; left bottom: top view; middle: sensor size compared with finger tip; and right: suture line on the sensor). The unit is millimeter.
Downloaded from pih.sagepub.com at UNIV OF PENNSYLVANIA on June 14, 2015
142
Proc IMechE Part H: J Engineering in Medicine 228(2)
Figure 2. Finite element model (FEM) of the sensor was constructed quarter symmetrically. (a) A quarter symmetric model was used. The left bottom of the shadow part is the symmetrical center, and here strain is measured. (b) Mesh model in this study. Sensing elements that consisted of 5 3 2 elements at the bottom of the sensor arc were used to collect strain data. (c and d) Triangle with rollers allows all the nodes on the quarter symmetric planes to move on the planes only. (e) Local coordinate systems in which an axis is in tangential direction of the bottom arc are set to evaluate strain.
high-tensile aluminum alloy (HTAA, 100 GPa, y = 0.3), and carbon steel for machine parts (CSMP, 210 GPa, y = 0.3). Suture material was nylon (3 GPa, y = 0.3). The soft ground as the soft tissue was assumed to be much softer than the sensor and the thread, and the material property similar to silicon (E = 10 MPa, y = 0.3) was given to the soft ground. In AL, 34 MPa as 0.2% proof strength was used to evaluate the yield, while HTAA and CSMP used 275 and 390 MPa, respectively, as yield point. Also, in this model, perfectly plastic behavior was assumed. We used the information as standard data proposed by Japanese Industrial Standards (JIS). The stepwise tension from 0.0 to 50.0 N at an interval of 5.0 N was applied to the end of the suture. A commercial FE solver (Abaqus 6.10; SIMULIA, USA) was used, and the analysis was performed nonlinearly and statically.
Output characteristic for linearity, repeatability, and reproducibility Based on the analysis, adequate material for the sensor was selected for the construction. Then, we evaluated the output from the sensors using a loading apparatus,
as shown in Figure 3. Sensor was put on a silicon sheet sustained by a metal base plate large enough. The silicon plate has two holes penetrated previous to pass a surgical suture and to reduce friction as less as possible. The ends of a suture were fixed or applied by a load. The output from the sensor was measured three times to compare between actual and simulated output for validation of the material selection. Static tension was measured, and stepwise tension from 0.0 to 50.0 N at an interval of 10.0 N was applied. In this measurement, system was reset once when the sensor was changed at initial time. This loading test also means evaluation of the linearity and repeatability of the sensor. Obtained data were used for calibration of the sensor. Then, to evaluate the process reproducibility of the sensor, arbitrary three sensors were picked up, and the output strain from the sensors was compared.
Practical evaluation of the sensor using a cadaver model For practical evaluation of the sensor, we simulated total knee arthroplasty (TKA), that is, we cut the skin
Downloaded from pih.sagepub.com at UNIV OF PENNSYLVANIA on June 14, 2015
Kiriyama et al.
143
5 mm (Height)
100 mm (Width and Depth) Tension sensor Silicon sheet
3.
Metal base plate Guide pulley Suture line
4. Applied load (Weight) Fixed end
mm was made along the medial edge of the rectus femoris muscle. An additional four sensors were set at the surgical sutures. MIS-capsule (Figure 4(c)): Following removal of all sensors, a capsule incision was made only to the distal part of the medial knee capsule. Then, four sensors were set at the surgical sutures (PDS 2-0; ETHICON). CS-capsule (Figure 4(d)): An extended incision of 50 mm was made proximally between the vastus medialis and the rectus femoris muscle. An additional four sensors were set at the surgical sutures.
(a) Suture line Suture sensor Pre-set hole
Silicon sheet
Metal base plate 4 mm
(b) Figure 3. Loading apparatus for evaluation of sensors: (a) outline of the apparatus and (b) the detail of a sensor and the peripheral condition.
and the joint capsule with the same procedure of TKA and made a suture of them, and compared two kinds of surgical techniques for TKA: a conventional surgery (CS) and a minimally invasive surgery (MIS). The miniature sensors were placed in an unfrozen fresh cadaver (69 years old). Our specimen was kept in a cold storage at 5 °C. Before the 5 h of the experiment, the specimen was out of the storage and stayed in a room at 12 °C. Before measurement, manual knee flexion–extension was repeated five times to release joint contracture. In our experiments, we set sensors on a specimen and stayed in 15 min approximately previous to the measurement because we needed to prepare other equipment. Therefore, sensor should be almost same temperature with the specimen. A senior orthopedic surgeon made an incision and performed suturing to the knee skin and joint capsule, similar to a typical surgical procedure. Sensors were then set as follows:
In the four procedures, D1–D4 are the distal sensors and a number is counted up from the knee to the distal. MIS-skin and MIS-capsule used the sensors. On the other hand, P1–P4 are the proximal sensors and a number is counted up from the knee to the proximal side. CS-skin and CS-capsule used P1–P4 additional to the sensors from D1 to D4. In these simulated surgeries, the skin and the capsule were dissected, but an artificial knee was not replaced. During suture tension measurement, the knee joint angle was simultaneously recorded by a potentiometer, as shown in Figure 5. The center of a rotatory potentiometer connecting two rods was set lateral to the knee joint center at light knee flexion. The angle between two rods expressed the rotation angle, and the rods were fixed with a slide block on a fixation belt holding the thigh and the shank. The rotatory potentiometer tended to move away from the knee joint center because the femur shows a rollback and a fixation belt slipped accompanying the knee flexion. To validate the rotatory potentiometer, the knee angles from the rotatory potentiometer and an opt-electronic motion capture system (ProReflex; Qualisys, Sweden) were compared. To a healthy subject with the rotatory potentiometer, reflective markers were placed on the bony landmarks and then the passive knee motion was measured with both the systems simultaneously. According to the comparison, the measurement error of the rotatory potentiometer was 3° in quasi-static motions. Passive knee flexion–extension was repeated five times prior to data collection for each of the four cases, and the suture tension and the joint angle in the final movements were evaluated. The above cadaver experiments were approved by an institutional review board.
Results FE analysis
1.
2.
MIS-skin (Figure 4(a)): A 70-mm incision was made from the superior edge of the patella to 20 mm distal to the knee joint line. Four sensors were set at the surgical sutures (PDS 4-0; ETHICON, Japan) at a regular interval. CS-skin (Figure 4(b)): After MIS-skin measurement, an additional proximal skin incision of 70
In all models, regardless of the type of material, strain concentration was observed on the concave side of the sensor arc. There were differences in the strain–load relationship when comparing the three kinds of materials (Figure 6(a)–(c)). The CSMP only showed linear output against the applied tensions and was selected as the best material for the sensor.
Downloaded from pih.sagepub.com at UNIV OF PENNSYLVANIA on June 14, 2015
144
Proc IMechE Part H: J Engineering in Medicine 228(2)
Distal sensor
Proximal sensor
{
D1 D2 D3 D4
(a)
{
P4 P3 P2 P1
P4 P3 P2 P1
D1 D2 D3 D4
(b)
(c)
(d)
Figure 4. (a–d) Sensor configurations. The right cadaveric knee was used. White boxes are the sensor placements. (a and c) Minimally invasive surgery (MIS) and (b and d) conventional surgery (CS) used four and eight sensors, respectively. (a) MIS-skin, (b) CS-skin, (c) MIS-capsule, and (d) CS-capsule.
Rotatory potentiometer Fixation belt
Proximal
Slide block
Distal
Signal lines
Figure 5. Knee angle measurement using a potentiometer.
Output of the sensor
Strain (103 με)
As shown in Figure 7(a), with stepwise application of 10-N loads, the sensor composed of CSMP showed linear output similar to the FE simulations. The output was in good agreement with calculated results, with a Pearson’s correlation coefficient of 0.99 in each of the three experiments (Figure 7(b)). This means that the sensor has good linearity and the repeatability of the output as the previous simulation analyzed by FEM. Figure 7(c) shows the comparison and the reproducibility among arbitrary three sensors. Each sensor has a good linearity and the difference of the output strain is very small as well. All eight sensors used in the
simulated surgeries show the same results. Therefore, the simple suture sensor is reproducible to process. This good reproducibility comes from the sensor design constructed with a combination of simple shapes. These results were not changed with two different kinds of surgical sutures, which were used in this study. As extra information, when a load over 60 N was applied to the suture using a loading apparatus, as shown in Figure 3, the sensor measured the load with the error ranged between + 0.1 N and 20.1 N and did not show a permanent offset. The estimated error by temperature was 0.06 N at the maximum between the 0 °C and 100 °C because the strain gage used in this study is a selftemperature compensated strain gage and has a characteristic of 1.8 micro-strain/°C at the maximum between 0 °C and 100 °C. When a suture was cut and a knot was made on the top of the sensor, the repetition of making a knot did not change and affect the sensor setting and experimental condition.
Evaluation of surgical suture tension in a cadaver model Although the surgeon had to remove surgical needles from a suture to pass the suture through the holes to set the sensors, that was carried out smoothly. Suture tensions of the proximal and distal four sensors in the skin and capsule in MIS and CS models (no proximal
8.0
5.0
1.5
4.0
2.5
0.7
0 0 10 20 30 40 50 Load (N)
(a)
0 0 10 20 30 40 50 Load (N)
0 0 10 20 30 40 50 Load (N)
(b)
(c)
Figure 6. Three kinds of different materials were calculated with FEM: (a) AL, (b) HTAA, and (c) CSMP. AL: aluminum; HTAA: high-tensile aluminum alloy; CSMP: carbon steel for machine parts.
Downloaded from pih.sagepub.com at UNIV OF PENNSYLVANIA on June 14, 2015
Kiriyama et al.
145
1st 2nd 3rd
Strain
0.8
1.6
1.6
R=0.98
0.8
0.8
0 0 10 20 30 40 50 Load (N)
0 0 0.8 1.6 Simulated strain
(a)
(b)
Strain
Actual strain
1.6
sensor 1 sensor 2 sensor 3
1st 2nd 3rd
0 0 10 20 30 40 50 Load (N)
(c)
Figure 7. (a) Repeatability of the suture sensor constructed by CSMP, (b) comparison between simulated and actual outputs, and (c) reproducibility of the sensor.
Suture tension (N)
4.8
4.8
D1 D2 D3 D4
3.2
4.8
D1 D2 D3 D4
3.2
3.2
1.6
1.6
1.6
0.0
0.0 0
50
100
150
0.0 0
Knee flexion angle (deg.)
50
150
100
Knee flexion angle (deg.)
150
P1 P2 P3 P4
32.0
16.0
0.0
0.0
150
48.0
1.6
1.6
100
(c)
D1 D2 D3 D4
3.2
50
50
Knee flexion angle (deg.)
4.8 D1 D2 D3 D4
0
0
(b)
4.8 Suture tension (N)
100
Knee flexion angle (deg.)
(a)
3.2
P1 P2 P3 P4
0.0 0
50
100
150
Knee flexion angle (deg.)
(e)
(d)
0
50
100
150
Knee flexion angle (deg.)
(f)
Figure 8. (a–f) Tension changes during passive knee motions. D1-4 and P1-4 are sensor’s places corresponding to Figure 4. Note that the suture tension scale in the CS-capsule (proximal) models is 10 times greater than those of the skin models: (a) MIS-skin, (b) CS-skin (distal), (c) CS-skin (proximal), (d) MIS-capsule, (e) CS-capsule (distal), and (f) CS-capsule (proximal).
sensors were set in the MIS models) are shown in Figure 8(a)–(f). All sensors showed nonlinear relationships between the knee angle and suture tension in the skin and capsule. The suture tension in the skin increased with the knee angle and reached maximum tension when the knee was in maximum flexion (;150°). MIS-skin (Figure 8(a)), CS-skin (distal (Figure 8(b)) and proximal (Figure 8(c))), MIS-capsule (Figure 8(d)), and CScapsule (distal (Figure 8(e))) showed the same peak tension (4.4 N).
The suture tension in the proximal capsule (CS-capsule, Figure 8(f)) showed a marked increase along with an increase in the knee flexion angle, and the maximum tension was 44 N. Higher tension was observed in the sensors closer to the patellar.
Discussion Various types of small sensors have been designed.8–11 Bassini et al.8 developed a lamina-type small sensor, but this sensor must be rounded along the target tissue.
Downloaded from pih.sagepub.com at UNIV OF PENNSYLVANIA on June 14, 2015
146
Proc IMechE Part H: J Engineering in Medicine 228(2)
Schubert et al.9,10 proposed a tyrolean tensiometer to pick up a thread with a U-shaped hook. Cummings11 developed a hands-free miniature suture tensiometer specific to laparoscopic surgery. These sensors are available for measurement of stationary and nondeformable soft tissues, and they cannot be applied in musculoskeletal tissues. In contrast to the previous miniature sensors that were used for stationary and nondeformable tissues, our device was small and designed for deformable and movable joints. Furthermore, the present sensor was able to measure suture tension in other soft tissues. Although our sensor is similar to buckle type,13,14 the relative width of our sensor to a target tissue is wider than previous sensors because suture thread is very thin. Therefore, the bending load applied to the sensor by a suture thread was not complicated. As a result, suture thread does not interfere with the sensing mechanism and was possible to measure suture tension, even though our sensor is a single-axis sensor. This study involved the sequential steps of designing a sensor, analyzing the output, and developing a prototype sensor. As a result, adequate material was selected, and the linearity of the output, repeatability, and reproducibility were evaluated. The linearity of the output was one of the most important factors because linear output assured that the sensor worked within the range of the expected elastic deformation. In this study, CSMP that is commonly used for industrial field was selected. On the other hand, when a harder material had been selected, the material might have allowed making a smaller sensor with linearity. However, the use of a harder material might also increase the complexity of sensor fabrication and may reduce the sensitivity of tension measurements. To avoid these problems, we selected adequate material based on the simulated output from FE analysis prior to actual construction. Such a strategy using FE analysis prior to construction of a new product is the standard used in general industry. In our cadaveric experiments, the suture tensions changed depending on the soft tissue material. Maximum suture tension in the skin was 4.4 N. The skin did not interfere with joint motion, and it was loose and wrinkled when the knee joint extended. Therefore, the tension of the skin was not overly large, even when the knee joint was flexed. Compared with the skin, the capsule showed a larger tension in the proximal location of the sensor than when the sensor was located distally. The proximal sensor had 10 times larger tension when compared with the distal sensor, while the distal sensor showed a smaller tension similar to that of the skin. The joint capsule is also an envelope surrounding the joint, but the outer layer is a fibrous membrane. The membrane has a thin structure and high tensile strength because the capsule is supported by ligaments surrounding the knee joint.15 In particular, the medial patellofemoral ligament reinforces the structure and the stiffness of the capsule.16 Hence, the
high tension (44 N) in the proximal capsule arises from the high stiffness of the medioproximal capsule because it was assumed that the area was stretched and pulled the suture tension with the knee flexion. Suture tension changed nonlinearly during knee flexion in all conditions in this study. Moreover, when comparing MIS-skin and CS-skin, the suture tension of the CS-skin had a sharp change. In CS-skin, since even distal sensor shows a sharp change, the sharpness could come from the presence of incision and dependence on the incision length. Assuming that the tissues had homogeneous material properties, the sharp change in the suture tension means that the proximal skin was stretched more rapidly in the late flexion phase. In the proximal capsule, the tension increased steeply, similar to the CS-skin. This is because the superior area of the medial patellar surface pushes and then stretches the capsule directly in the late flexion (near 150°). Hence, the proximal sensor showed large and rapid changes in tension. Actually, the largest tensions in the CS-capsule were measured with the sensor placed on the patella. In this study, changes in tension can be described as a hysteresis loop. The skin and the capsule are viscoelastic materials, and hysteresis might arise from structure changes of the soft tissues because the skin and capsule of the knee might be relaxed and wrinkled around the joint. Funk et al.17 showed that the capsular volume decreased with knee flexion, which reflects the degree of capsular or skin structural looseness. These structural alternations likely contribute to the hysteresis loop. In practical procedure, the surgeon made a knot on the sensor smoothly, although the surgical needle had to be removed. Larger hole size would allow surgeon to make a knot more easily. Also, the sensor needed 6 mm of free unknotted span, which is a distance between both of the hole centers. Hence, the requirement of 6 mm did not necessarily simulate the actual surgical condition, and the suture tension might be different from true tension. In the orthopedic field, MIS is getting a standard technique, and appropriate suture tension and surgical techniques are needed to ensure wound healing for early recovery. The size of the sensor is 410 mm, which is clinically acceptable to minimize damage to the soft tissues. Indeed, this sensor is suitable for use even within a small incision or the small working space of MIS. This study has several limitations. One limitation was the method used to detach the sensor from the suture. Specifically, the surgeon must cut the thread and remove the knot to detach the sensor because the sensor had to be set under a suture knot. This cutting process restricts to use the sensor in situ or during healing condition potentially. Second, only one specimen was used in this study. To reproduce the natural condition of the soft tissues, we used unfrozen fresh cadavers. Moreover, to keep a structural continuity of the skin, we prepared the whole lower extremity with the
Downloaded from pih.sagepub.com at UNIV OF PENNSYLVANIA on June 14, 2015
Kiriyama et al.
147
intact hip and ankle joints. The purpose of this study, however, was to develop a new miniature tension sensor and to evaluate the reliability biomechanically. Thus, the results from this specimen were likely appropriate to yield reliable data, and the low number of specimens does not obviate the demonstrated utility of the sensor. Third, our specimen was elderly, and thus, the material properties of this specimen might be different from normal adult tissues. Regardless, TKA is often performed in individuals of advanced age, and thus, we believe that the age of the specimen was relevant to an important clinical population. Regardless of the limitations, the usefulness of the suture tension sensor could indicate plans to investigate the usage of this sensor in other application in a future work. In the knee joint capsule, the structure and intraarticular pressure have been studied from the anatomical16 and pathological point of view.17–20 The anatomical structure is complicated;16 thus, biomechanical approaches to characterize mechanical properties have been limited to the investigation of the relationship between intra-articular pressure and the capsular volume.17–20 Our miniature sensor can evaluate a part of load transmitted in soft tissues via suture tension. For example, when sensors are placed on an intact tissue and sutured there, the sensor measures a part of load sharing between the tissue and the device because the deformed soft tissue changed suture tension. This helps us to understand the distribution of the tensile load of soft tissues. Hence, this miniature tension sensor represents a novel approach to measure forces applied to soft tissues. Orthopedic surgeries on tendons are often necessary, and in situ measurement of the surgical suture is required to obtain optimal initial strength. In many cases, the surgeon determines the suture technique and the number and type of suture thread and knots; thus, suture tension and the material affect initial strength of the suture and those rely on the surgeons themselves. Since the sensor needs to be removed when our suture tension sensor is used, surgeon does not obtain actual final tension. However, surgeons are possible to reproduce suture tension by their hand feeling; they can quantify the suturing tension using our sensors. Although this is indirect quantification, the present miniature sensor could provide basic biomechanical information to help surgeons to select suturing methods for tendon surgeries, such as tendon suture and transfer, primary repair of the Achilles tendon,21 and rotator cuff repairs.22 Hence, regardless of some limitations, this miniature tension sensor could provide basic biomechanical information and has a potential to elucidate a biomechanical function of the soft tissues resulting in improvement of surgical outcomes.
Conclusion We developed a miniature sensor suitable for use during space-constrained surgery that was able to continuously
measure the tension of deformable soft musculoskeletal tissues. This miniature tension sensor allows quantitative measurement of tension within soft tissues and therefore should provide basic biomechanical information to improve patient outcomes. Acknowledgements The cadavers used in our study were donated to the Clinical Anatomy Laboratory, Keio University School of Medicine with the consent of the families. The authors are grateful to Sadakazu Aiso, MD, PhD, Nobuaki Imanishi, MD, PhD, and Naoto Hirakata, MD, PhD, for approving the use of the laboratory. Declaration of conflicting interests The funding source did not influence the study design, collection, analysis, or interpretation of data. Ethical approval This study was approved by a Review Board of Keio University, School of Medicine. Funding A portion of this study was sponsored by ETHICON in Japan. References 1. Ogawa R, Akaishi S, Huang C, et al. Clinical applications of basic research that shows reducing skin tension could prevent and treat abnormal scarring: the importance of fascial/subcutaneous tensile reduction sutures and flap surgery for keloid and hypertrophic scar reconstruction. J Nippon Med Sch 2011; 78: 68–76. 2. Bisson LJ, Sobel AD and Godfrey D.Effects of using a surgical clamp to hold tension while tying knots with commonly used orthopedic sutures. Knee Surg Sports Traumatol Arthrosc. 2012; 20: 1673–1680. 3. Karahan M, Akgun U, Turkoglu A, et al. Pretzel knot compared with standard suture knots. Knee Surg Sports Traumatol Arthrosc. 2012; 20: 2302–2306. 4. Markolf KL, Willems MJ, Jackson SR, et al. In situ calibration of miniature sensors implanted into the anterior cruciate ligament, part II: force probe measurements. J Orthop Res 1998; 16: 464–471. 5. Meyer DC, Jacob HAC, Nyffeler RW, et al. In vivo tendon force measurement of 2-week duration in sheep. J Biomech 2004; 37: 135–140. 6. Kocabey Y, Taser O, Nyland J, et al. Horizontal suture placement influences meniscal repair fixation strength. Knee Surg Sports Traumatol Arthrosc. 2013; 21: 615–619. 7. Staerke C, Brettschneider O, Grobel KH, et al. Tensile forces on sutures in the human lateral knee meniscus. Knee Surg Sports Traumatol Arthrosc 2009; 17: 1354–1359. 8. Bassini R, Cantone MC, Gambarini G, et al. Method for measuring suture tension in surgery. Med Biol Eng Comput 1988; 26: 451–454. 9. Schubert HM, Moser TM, Buchegger JW, et al. Tyrolean tensiometer: a new instrument for easy
Downloaded from pih.sagepub.com at UNIV OF PENNSYLVANIA on June 14, 2015
148
10.
11. 12.
13.
14.
15.
16.
Proc IMechE Part H: J Engineering in Medicine 228(2) intraoperative tension measurement before nerve coaptation. J Trauma 2006; 61: 760–763. Schubert HM, Hohlrieder M, Buchegger JW, et al. Tyrolean tensiometer: a new instrument for easy intraoperative tension measurement before vascular anastomosis. J Plast Reconstr Aesthet Surg 2007; 60: 311–315. Cummings JF.A miniature suture tensiometer for laparoscopic applications. J Invest Surg 2000; 13: 253–258. Kiriyama Y, Matsumoto H, Imanishi N, et al. Development of a miniature tension sensor to monitor suture tension during the joint motion. J Biomech 2008; 41(Suppl. 1): S503. Keegan KG, Baker GJ, Boero MJ, et al. Evaluation of support bandaging during measurement of proximal sesamoidean ligament strain in horses by use of a mercury strain gauge. Am J Vet Res 1992; 53(7): 1203–1208. Roberts CS, Cummings JF, Grood ES, et al. In-vivo measurement of human anterior cruciate ligament forces during knee extension exercises. Trans Orthop Res Soc 1994; 19: 84. Merican AM, Sanghavi S, Iranpour F, et al. The structural properties of the lateral retinaculum and capsular complex of the knee. J Biomech 2009; 42: 2323–2329. Mochizuki T, Nimura A, Tateishi T, et al. Anatomic study of the attachment of the medial patellofemoral
17.
18.
19.
20.
21.
22.
ligament and its characteristic relationships to the vastus intermedius. Knee Surg Sports Traumatol Arthrosc. 2013; 21: 305–310. Funk DA, Noyes FR, Grood ES, et al. Effect of flexion angle on the pressure-volume of the human knee. Arthroscopy 1991; 71(1): 86–90. Geborek P, Moritz U and Wollheim FA.Joint capsular stiffness in knee arthritis. Relationship to intraarticular volume, hydrostatic pressures, and extensor muscle function. J Rheumatol 1989; 16: 1351–1358. Jawed S, Gaffney K and Blake DR.Intra-articular pressure profile of the knee joint in a spectrum of inflammatory arthropathies. Ann Rheum Dis 1997; 56: 686–689. Jayson MIV and Dixon ASJ. Intra-articular pressure in rheumatoid arthritis of the knee. I. Pressure changes during passive joint distension. Ann Rheum Dis 1970; 29: 401–408. Anderl W, Heuberer PR, Laky B, et al. Superiority of bridging techniques with medial fixation on initial strength. Knee Surg Sports Traumatol Arthrosc. 2012; 20: 2559–2566. Garofalo R, Castagna A, Borroni M, et al. Arthroscopic transosseous (anchorless) rotator cuff repair. Knee Surg Sports Traumatol Arthrosc. 2012; 20: 1031–1036.
Downloaded from pih.sagepub.com at UNIV OF PENNSYLVANIA on June 14, 2015
A miniature tension sensor to measure surgical suture tension of deformable musculoskeletal tissues during joint motion
Proc IMechE Part H: J Engineering in Medicine 2014, Vol. 228(2) 140–148 Ó IMechE 2014 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0954411913518317 pih.sagepub.com
Yoshimori Kiriyama1, Hideo Matsumoto2, Yoshiaki Toyama3 and Takeo Nagura3
Abstract The aim of this study was to develop a new suture tension sensor for musculoskeletal soft tissue that shows deformation or movements. The suture tension sensor was 10 mm in size, which was small enough to avoid conflicting with the adjacent sensor. Furthermore, the sensor had good linearity up to a tension of 50 N, which is equivalent to the breaking strength of a size 1 absorbable suture defined by the United States Pharmacopeia. The design and mechanism were analyzed using a finite element model prior to developing the actual sensor. Based on the analysis, adequate material was selected, and the output linearity was confirmed and compared with the simulated result. To evaluate practical application, the incision of the skin and capsule were sutured during simulated total knee arthroplasty. When conventional surgery and minimally invasive surgery were performed, suture tensions were compared. In minimally invasive surgery, the distal portion of the knee was dissected, and the proximal portion of the knee was dissected additionally in conventional surgery. In the skin suturing, the maximum tension was 4.4 N, and this tension was independent of the sensor location. In contrast, the sensor suturing the capsule in the distal portion had a tension of 4.4 N in minimally invasive surgery, while the proximal sensor had a tension of 44 N in conventional surgery. The suture tensions increased nonlinearly and were dependent on the knee flexion angle. Furthermore, the tension changes showed hysteresis. This miniature tension sensor may help establish the optimal suturing method with adequate tension to ensure wound healing and early recovery.
Keywords Suture tension, miniature sensor, soft tissue deformation
Date received: 16 April 2013; accepted: 3 December 2013
Introduction Suturing soft tissues, such as the tendons, capsules, ligaments, muscles, nerves, and skin, is an essential component of musculoskeletal surgeries. Achieving optimal suture tension is critical for successful surgery because under-tension is one of the important factors to result in a bad apposition or a scar,1 while we undergo over-tension resulting in cutout of the surgical wound. However, there is a lack of the quantitative criteria to guide suture tension since tension is determined manually depending on surgeon’s technique and experience.2,3 Various suture techniques using suture guides and in situ tension measurement have been utilized in arthroscopic surgeries,4,5 but direct measurement of suture
tension is technically difficult because of limited working space and the small size of the target tissue in arthroscopic surgeries. Thus, measurements have been performed using an extra-articular approach with large transducers.6,7 Additionally, most of the target soft tissues in the orthopedic field undergo physiologic deformation and movement that occurs with joint
1
Department of Clinical Biomechanics, Keio University, Tokyo, Japan Institute for Integrated Sports Medicine, Keio University, Tokyo, Japan 3 Department of Orthopaedic Surgery, Keio University, Tokyo, Japan 2
Corresponding author: Yoshimori Kiriyama, Department of Clinical Biomechanics, Keio University, 35 Shinanomachi, Shinjuku, Tokyo 160-8582, Japan. Email: [email protected]
Downloaded from pih.sagepub.com at UNIV OF PENNSYLVANIA on June 14, 2015
Kiriyama et al.
141
movement. Therefore, characteristics of a useful tension sensor include small size, to enable it to be placed around the soft tissues and to avoid conflicting with adjacent tissues or sensors, as well as large measurement range of the tension. Several types of small sensors have been developed to measure static tension (i.e. initial tension) during surgery.8–11 These sensors are intended for use within soft tissues that are stationary and nondeformable, and the size and design of the sensors are not particularly amenable to use within musculoskeletal tissues. Thus, a new, smaller tension sensor with an acceptable tension range is needed to measure continuous changes in suture tension in deformable soft tissues and to provide quantitative biomechanical data regarding surgical sutures. Such a sensor could also be useful in obtaining the basic characterization of soft tissues, in developing new criteria for surgical techniques and in assessing the mechanical strength of surgical sutures. For these requirements, a new type of sensor was developed, and preliminary data were shown.12 The objective of this article was to show the detail of a miniature tension sensor to measure suture tension that is potentially modulated by joint movement. To evaluate the practical utility of the sensor, we investigated actual suture tension during joint movements in a cadaver.
Material and methods Design of miniature tension sensor The optimal tension sensor must (1) be small enough to avoid interfering with the surgical procedure or with physiologic deformation of soft tissues, (2) be able to adequately measure suture tension, and (3) have acceptable accuracy and range to measure surgical suture tension. In this study, we assumed that a maximum sensor size of 10 mm satisfied size requirements during surgery. To select adequate material, finite element (FE) analysis was conducted. An adequate range of tension measurement was designated as 0–50 N, which is the breaking strength of a size 1 absorbable suture defined by United States Pharmacopeia (USP). Thus, a sensor was designed as shown in Figure 1. In the design of a miniature tension sensor, the length is 410 mm, and the height is 4 mm maximally. Considering the possibility and ease to manufacture
requirements in addition to satisfying the requirements, we designed the sensor that is constructed with combination of simple shapes. A suture passes over the center arc via the end holes of both sides. The center arc had a radius of 2.0 mm, and a strain gage was attached at the concave side. In this study, a self-temperature compensated strain gage (KYOWA, Japan) was used for the sensor. The unique feature of the sensor was the design to measure tension, and the gage measures strain dependent on the tension when tension was applied to a suture and when the suture generates forces to bend the arc.
FE analysis Strain distribution and magnitude of the sensor model were evaluated to confirm the range and the linearity of the output from the sensor. A quarter symmetric model was constructed, as shown in Figure 2(a). To estimate the strain generated by suture tension, finite element models (FEMs) consisting of the sensor, suture, and the soft ground were constructed, as shown in Figure 2(b). The sensor, the suture thread, and the soft ground meshes had approximately 14,300, 800, and 5200 hexahedral elements, respectively. The base of the soft ground was constrained in three translations (Figure 2(c)). Since the quarter symmetric model was used in this study, all the nodes on the symmetric planes constrained to stay on the symmetric plane (Figure 2(c) and (d)). The interactions between the sensor and the soft ground and between the sensor and the thread were assumed to be frictionless. In initial condition, the sensor was placed on the top surface of the sensor. Also, the suture thread was set to touch the top round of the suture sensor. To consider the length and width of a strain gage for the sensor, five elements along the bottom arc and two elements along y-axis from the symmetrical center of the bottom arc elements were selected as sensing elements. We set the local coordinate system in which an axis corresponds to the tangential orientation of the bottom arc. The strain along the tangential orientation of the arc was calculated on the nodes that were on the surface of the sensing elements. The collected strain was averaged over the sensing elements, and the value was used as a representative strain. For the sensor, three different materials were prepared: aluminum (AL, E = 70 GPa, y = 0.3),
Tension sensor
4
Suture
4
10
tension
Figure 1. Sensor design of a miniature suture tension sensor (left top: side view; left bottom: top view; middle: sensor size compared with finger tip; and right: suture line on the sensor). The unit is millimeter.
Downloaded from pih.sagepub.com at UNIV OF PENNSYLVANIA on June 14, 2015
142
Proc IMechE Part H: J Engineering in Medicine 228(2)
Figure 2. Finite element model (FEM) of the sensor was constructed quarter symmetrically. (a) A quarter symmetric model was used. The left bottom of the shadow part is the symmetrical center, and here strain is measured. (b) Mesh model in this study. Sensing elements that consisted of 5 3 2 elements at the bottom of the sensor arc were used to collect strain data. (c and d) Triangle with rollers allows all the nodes on the quarter symmetric planes to move on the planes only. (e) Local coordinate systems in which an axis is in tangential direction of the bottom arc are set to evaluate strain.
high-tensile aluminum alloy (HTAA, 100 GPa, y = 0.3), and carbon steel for machine parts (CSMP, 210 GPa, y = 0.3). Suture material was nylon (3 GPa, y = 0.3). The soft ground as the soft tissue was assumed to be much softer than the sensor and the thread, and the material property similar to silicon (E = 10 MPa, y = 0.3) was given to the soft ground. In AL, 34 MPa as 0.2% proof strength was used to evaluate the yield, while HTAA and CSMP used 275 and 390 MPa, respectively, as yield point. Also, in this model, perfectly plastic behavior was assumed. We used the information as standard data proposed by Japanese Industrial Standards (JIS). The stepwise tension from 0.0 to 50.0 N at an interval of 5.0 N was applied to the end of the suture. A commercial FE solver (Abaqus 6.10; SIMULIA, USA) was used, and the analysis was performed nonlinearly and statically.
Output characteristic for linearity, repeatability, and reproducibility Based on the analysis, adequate material for the sensor was selected for the construction. Then, we evaluated the output from the sensors using a loading apparatus,
as shown in Figure 3. Sensor was put on a silicon sheet sustained by a metal base plate large enough. The silicon plate has two holes penetrated previous to pass a surgical suture and to reduce friction as less as possible. The ends of a suture were fixed or applied by a load. The output from the sensor was measured three times to compare between actual and simulated output for validation of the material selection. Static tension was measured, and stepwise tension from 0.0 to 50.0 N at an interval of 10.0 N was applied. In this measurement, system was reset once when the sensor was changed at initial time. This loading test also means evaluation of the linearity and repeatability of the sensor. Obtained data were used for calibration of the sensor. Then, to evaluate the process reproducibility of the sensor, arbitrary three sensors were picked up, and the output strain from the sensors was compared.
Practical evaluation of the sensor using a cadaver model For practical evaluation of the sensor, we simulated total knee arthroplasty (TKA), that is, we cut the skin
Downloaded from pih.sagepub.com at UNIV OF PENNSYLVANIA on June 14, 2015
Kiriyama et al.
143
5 mm (Height)
100 mm (Width and Depth) Tension sensor Silicon sheet
3.
Metal base plate Guide pulley Suture line
4. Applied load (Weight) Fixed end
mm was made along the medial edge of the rectus femoris muscle. An additional four sensors were set at the surgical sutures. MIS-capsule (Figure 4(c)): Following removal of all sensors, a capsule incision was made only to the distal part of the medial knee capsule. Then, four sensors were set at the surgical sutures (PDS 2-0; ETHICON). CS-capsule (Figure 4(d)): An extended incision of 50 mm was made proximally between the vastus medialis and the rectus femoris muscle. An additional four sensors were set at the surgical sutures.
(a) Suture line Suture sensor Pre-set hole
Silicon sheet
Metal base plate 4 mm
(b) Figure 3. Loading apparatus for evaluation of sensors: (a) outline of the apparatus and (b) the detail of a sensor and the peripheral condition.
and the joint capsule with the same procedure of TKA and made a suture of them, and compared two kinds of surgical techniques for TKA: a conventional surgery (CS) and a minimally invasive surgery (MIS). The miniature sensors were placed in an unfrozen fresh cadaver (69 years old). Our specimen was kept in a cold storage at 5 °C. Before the 5 h of the experiment, the specimen was out of the storage and stayed in a room at 12 °C. Before measurement, manual knee flexion–extension was repeated five times to release joint contracture. In our experiments, we set sensors on a specimen and stayed in 15 min approximately previous to the measurement because we needed to prepare other equipment. Therefore, sensor should be almost same temperature with the specimen. A senior orthopedic surgeon made an incision and performed suturing to the knee skin and joint capsule, similar to a typical surgical procedure. Sensors were then set as follows:
In the four procedures, D1–D4 are the distal sensors and a number is counted up from the knee to the distal. MIS-skin and MIS-capsule used the sensors. On the other hand, P1–P4 are the proximal sensors and a number is counted up from the knee to the proximal side. CS-skin and CS-capsule used P1–P4 additional to the sensors from D1 to D4. In these simulated surgeries, the skin and the capsule were dissected, but an artificial knee was not replaced. During suture tension measurement, the knee joint angle was simultaneously recorded by a potentiometer, as shown in Figure 5. The center of a rotatory potentiometer connecting two rods was set lateral to the knee joint center at light knee flexion. The angle between two rods expressed the rotation angle, and the rods were fixed with a slide block on a fixation belt holding the thigh and the shank. The rotatory potentiometer tended to move away from the knee joint center because the femur shows a rollback and a fixation belt slipped accompanying the knee flexion. To validate the rotatory potentiometer, the knee angles from the rotatory potentiometer and an opt-electronic motion capture system (ProReflex; Qualisys, Sweden) were compared. To a healthy subject with the rotatory potentiometer, reflective markers were placed on the bony landmarks and then the passive knee motion was measured with both the systems simultaneously. According to the comparison, the measurement error of the rotatory potentiometer was 3° in quasi-static motions. Passive knee flexion–extension was repeated five times prior to data collection for each of the four cases, and the suture tension and the joint angle in the final movements were evaluated. The above cadaver experiments were approved by an institutional review board.
Results FE analysis
1.
2.
MIS-skin (Figure 4(a)): A 70-mm incision was made from the superior edge of the patella to 20 mm distal to the knee joint line. Four sensors were set at the surgical sutures (PDS 4-0; ETHICON, Japan) at a regular interval. CS-skin (Figure 4(b)): After MIS-skin measurement, an additional proximal skin incision of 70
In all models, regardless of the type of material, strain concentration was observed on the concave side of the sensor arc. There were differences in the strain–load relationship when comparing the three kinds of materials (Figure 6(a)–(c)). The CSMP only showed linear output against the applied tensions and was selected as the best material for the sensor.
Downloaded from pih.sagepub.com at UNIV OF PENNSYLVANIA on June 14, 2015
144
Proc IMechE Part H: J Engineering in Medicine 228(2)
Distal sensor
Proximal sensor
{
D1 D2 D3 D4
(a)
{
P4 P3 P2 P1
P4 P3 P2 P1
D1 D2 D3 D4
(b)
(c)
(d)
Figure 4. (a–d) Sensor configurations. The right cadaveric knee was used. White boxes are the sensor placements. (a and c) Minimally invasive surgery (MIS) and (b and d) conventional surgery (CS) used four and eight sensors, respectively. (a) MIS-skin, (b) CS-skin, (c) MIS-capsule, and (d) CS-capsule.
Rotatory potentiometer Fixation belt
Proximal
Slide block
Distal
Signal lines
Figure 5. Knee angle measurement using a potentiometer.
Output of the sensor
Strain (103 με)
As shown in Figure 7(a), with stepwise application of 10-N loads, the sensor composed of CSMP showed linear output similar to the FE simulations. The output was in good agreement with calculated results, with a Pearson’s correlation coefficient of 0.99 in each of the three experiments (Figure 7(b)). This means that the sensor has good linearity and the repeatability of the output as the previous simulation analyzed by FEM. Figure 7(c) shows the comparison and the reproducibility among arbitrary three sensors. Each sensor has a good linearity and the difference of the output strain is very small as well. All eight sensors used in the
simulated surgeries show the same results. Therefore, the simple suture sensor is reproducible to process. This good reproducibility comes from the sensor design constructed with a combination of simple shapes. These results were not changed with two different kinds of surgical sutures, which were used in this study. As extra information, when a load over 60 N was applied to the suture using a loading apparatus, as shown in Figure 3, the sensor measured the load with the error ranged between + 0.1 N and 20.1 N and did not show a permanent offset. The estimated error by temperature was 0.06 N at the maximum between the 0 °C and 100 °C because the strain gage used in this study is a selftemperature compensated strain gage and has a characteristic of 1.8 micro-strain/°C at the maximum between 0 °C and 100 °C. When a suture was cut and a knot was made on the top of the sensor, the repetition of making a knot did not change and affect the sensor setting and experimental condition.
Evaluation of surgical suture tension in a cadaver model Although the surgeon had to remove surgical needles from a suture to pass the suture through the holes to set the sensors, that was carried out smoothly. Suture tensions of the proximal and distal four sensors in the skin and capsule in MIS and CS models (no proximal
8.0
5.0
1.5
4.0
2.5
0.7
0 0 10 20 30 40 50 Load (N)
(a)
0 0 10 20 30 40 50 Load (N)
0 0 10 20 30 40 50 Load (N)
(b)
(c)
Figure 6. Three kinds of different materials were calculated with FEM: (a) AL, (b) HTAA, and (c) CSMP. AL: aluminum; HTAA: high-tensile aluminum alloy; CSMP: carbon steel for machine parts.
Downloaded from pih.sagepub.com at UNIV OF PENNSYLVANIA on June 14, 2015
Kiriyama et al.
145
1st 2nd 3rd
Strain
0.8
1.6
1.6
R=0.98
0.8
0.8
0 0 10 20 30 40 50 Load (N)
0 0 0.8 1.6 Simulated strain
(a)
(b)
Strain
Actual strain
1.6
sensor 1 sensor 2 sensor 3
1st 2nd 3rd
0 0 10 20 30 40 50 Load (N)
(c)
Figure 7. (a) Repeatability of the suture sensor constructed by CSMP, (b) comparison between simulated and actual outputs, and (c) reproducibility of the sensor.
Suture tension (N)
4.8
4.8
D1 D2 D3 D4
3.2
4.8
D1 D2 D3 D4
3.2
3.2
1.6
1.6
1.6
0.0
0.0 0
50
100
150
0.0 0
Knee flexion angle (deg.)
50
150
100
Knee flexion angle (deg.)
150
P1 P2 P3 P4
32.0
16.0
0.0
0.0
150
48.0
1.6
1.6
100
(c)
D1 D2 D3 D4
3.2
50
50
Knee flexion angle (deg.)
4.8 D1 D2 D3 D4
0
0
(b)
4.8 Suture tension (N)
100
Knee flexion angle (deg.)
(a)
3.2
P1 P2 P3 P4
0.0 0
50
100
150
Knee flexion angle (deg.)
(e)
(d)
0
50
100
150
Knee flexion angle (deg.)
(f)
Figure 8. (a–f) Tension changes during passive knee motions. D1-4 and P1-4 are sensor’s places corresponding to Figure 4. Note that the suture tension scale in the CS-capsule (proximal) models is 10 times greater than those of the skin models: (a) MIS-skin, (b) CS-skin (distal), (c) CS-skin (proximal), (d) MIS-capsule, (e) CS-capsule (distal), and (f) CS-capsule (proximal).
sensors were set in the MIS models) are shown in Figure 8(a)–(f). All sensors showed nonlinear relationships between the knee angle and suture tension in the skin and capsule. The suture tension in the skin increased with the knee angle and reached maximum tension when the knee was in maximum flexion (;150°). MIS-skin (Figure 8(a)), CS-skin (distal (Figure 8(b)) and proximal (Figure 8(c))), MIS-capsule (Figure 8(d)), and CScapsule (distal (Figure 8(e))) showed the same peak tension (4.4 N).
The suture tension in the proximal capsule (CS-capsule, Figure 8(f)) showed a marked increase along with an increase in the knee flexion angle, and the maximum tension was 44 N. Higher tension was observed in the sensors closer to the patellar.
Discussion Various types of small sensors have been designed.8–11 Bassini et al.8 developed a lamina-type small sensor, but this sensor must be rounded along the target tissue.
Downloaded from pih.sagepub.com at UNIV OF PENNSYLVANIA on June 14, 2015
146
Proc IMechE Part H: J Engineering in Medicine 228(2)
Schubert et al.9,10 proposed a tyrolean tensiometer to pick up a thread with a U-shaped hook. Cummings11 developed a hands-free miniature suture tensiometer specific to laparoscopic surgery. These sensors are available for measurement of stationary and nondeformable soft tissues, and they cannot be applied in musculoskeletal tissues. In contrast to the previous miniature sensors that were used for stationary and nondeformable tissues, our device was small and designed for deformable and movable joints. Furthermore, the present sensor was able to measure suture tension in other soft tissues. Although our sensor is similar to buckle type,13,14 the relative width of our sensor to a target tissue is wider than previous sensors because suture thread is very thin. Therefore, the bending load applied to the sensor by a suture thread was not complicated. As a result, suture thread does not interfere with the sensing mechanism and was possible to measure suture tension, even though our sensor is a single-axis sensor. This study involved the sequential steps of designing a sensor, analyzing the output, and developing a prototype sensor. As a result, adequate material was selected, and the linearity of the output, repeatability, and reproducibility were evaluated. The linearity of the output was one of the most important factors because linear output assured that the sensor worked within the range of the expected elastic deformation. In this study, CSMP that is commonly used for industrial field was selected. On the other hand, when a harder material had been selected, the material might have allowed making a smaller sensor with linearity. However, the use of a harder material might also increase the complexity of sensor fabrication and may reduce the sensitivity of tension measurements. To avoid these problems, we selected adequate material based on the simulated output from FE analysis prior to actual construction. Such a strategy using FE analysis prior to construction of a new product is the standard used in general industry. In our cadaveric experiments, the suture tensions changed depending on the soft tissue material. Maximum suture tension in the skin was 4.4 N. The skin did not interfere with joint motion, and it was loose and wrinkled when the knee joint extended. Therefore, the tension of the skin was not overly large, even when the knee joint was flexed. Compared with the skin, the capsule showed a larger tension in the proximal location of the sensor than when the sensor was located distally. The proximal sensor had 10 times larger tension when compared with the distal sensor, while the distal sensor showed a smaller tension similar to that of the skin. The joint capsule is also an envelope surrounding the joint, but the outer layer is a fibrous membrane. The membrane has a thin structure and high tensile strength because the capsule is supported by ligaments surrounding the knee joint.15 In particular, the medial patellofemoral ligament reinforces the structure and the stiffness of the capsule.16 Hence, the
high tension (44 N) in the proximal capsule arises from the high stiffness of the medioproximal capsule because it was assumed that the area was stretched and pulled the suture tension with the knee flexion. Suture tension changed nonlinearly during knee flexion in all conditions in this study. Moreover, when comparing MIS-skin and CS-skin, the suture tension of the CS-skin had a sharp change. In CS-skin, since even distal sensor shows a sharp change, the sharpness could come from the presence of incision and dependence on the incision length. Assuming that the tissues had homogeneous material properties, the sharp change in the suture tension means that the proximal skin was stretched more rapidly in the late flexion phase. In the proximal capsule, the tension increased steeply, similar to the CS-skin. This is because the superior area of the medial patellar surface pushes and then stretches the capsule directly in the late flexion (near 150°). Hence, the proximal sensor showed large and rapid changes in tension. Actually, the largest tensions in the CS-capsule were measured with the sensor placed on the patella. In this study, changes in tension can be described as a hysteresis loop. The skin and the capsule are viscoelastic materials, and hysteresis might arise from structure changes of the soft tissues because the skin and capsule of the knee might be relaxed and wrinkled around the joint. Funk et al.17 showed that the capsular volume decreased with knee flexion, which reflects the degree of capsular or skin structural looseness. These structural alternations likely contribute to the hysteresis loop. In practical procedure, the surgeon made a knot on the sensor smoothly, although the surgical needle had to be removed. Larger hole size would allow surgeon to make a knot more easily. Also, the sensor needed 6 mm of free unknotted span, which is a distance between both of the hole centers. Hence, the requirement of 6 mm did not necessarily simulate the actual surgical condition, and the suture tension might be different from true tension. In the orthopedic field, MIS is getting a standard technique, and appropriate suture tension and surgical techniques are needed to ensure wound healing for early recovery. The size of the sensor is 410 mm, which is clinically acceptable to minimize damage to the soft tissues. Indeed, this sensor is suitable for use even within a small incision or the small working space of MIS. This study has several limitations. One limitation was the method used to detach the sensor from the suture. Specifically, the surgeon must cut the thread and remove the knot to detach the sensor because the sensor had to be set under a suture knot. This cutting process restricts to use the sensor in situ or during healing condition potentially. Second, only one specimen was used in this study. To reproduce the natural condition of the soft tissues, we used unfrozen fresh cadavers. Moreover, to keep a structural continuity of the skin, we prepared the whole lower extremity with the
Downloaded from pih.sagepub.com at UNIV OF PENNSYLVANIA on June 14, 2015
Kiriyama et al.
147
intact hip and ankle joints. The purpose of this study, however, was to develop a new miniature tension sensor and to evaluate the reliability biomechanically. Thus, the results from this specimen were likely appropriate to yield reliable data, and the low number of specimens does not obviate the demonstrated utility of the sensor. Third, our specimen was elderly, and thus, the material properties of this specimen might be different from normal adult tissues. Regardless, TKA is often performed in individuals of advanced age, and thus, we believe that the age of the specimen was relevant to an important clinical population. Regardless of the limitations, the usefulness of the suture tension sensor could indicate plans to investigate the usage of this sensor in other application in a future work. In the knee joint capsule, the structure and intraarticular pressure have been studied from the anatomical16 and pathological point of view.17–20 The anatomical structure is complicated;16 thus, biomechanical approaches to characterize mechanical properties have been limited to the investigation of the relationship between intra-articular pressure and the capsular volume.17–20 Our miniature sensor can evaluate a part of load transmitted in soft tissues via suture tension. For example, when sensors are placed on an intact tissue and sutured there, the sensor measures a part of load sharing between the tissue and the device because the deformed soft tissue changed suture tension. This helps us to understand the distribution of the tensile load of soft tissues. Hence, this miniature tension sensor represents a novel approach to measure forces applied to soft tissues. Orthopedic surgeries on tendons are often necessary, and in situ measurement of the surgical suture is required to obtain optimal initial strength. In many cases, the surgeon determines the suture technique and the number and type of suture thread and knots; thus, suture tension and the material affect initial strength of the suture and those rely on the surgeons themselves. Since the sensor needs to be removed when our suture tension sensor is used, surgeon does not obtain actual final tension. However, surgeons are possible to reproduce suture tension by their hand feeling; they can quantify the suturing tension using our sensors. Although this is indirect quantification, the present miniature sensor could provide basic biomechanical information to help surgeons to select suturing methods for tendon surgeries, such as tendon suture and transfer, primary repair of the Achilles tendon,21 and rotator cuff repairs.22 Hence, regardless of some limitations, this miniature tension sensor could provide basic biomechanical information and has a potential to elucidate a biomechanical function of the soft tissues resulting in improvement of surgical outcomes.
Conclusion We developed a miniature sensor suitable for use during space-constrained surgery that was able to continuously
measure the tension of deformable soft musculoskeletal tissues. This miniature tension sensor allows quantitative measurement of tension within soft tissues and therefore should provide basic biomechanical information to improve patient outcomes. Acknowledgements The cadavers used in our study were donated to the Clinical Anatomy Laboratory, Keio University School of Medicine with the consent of the families. The authors are grateful to Sadakazu Aiso, MD, PhD, Nobuaki Imanishi, MD, PhD, and Naoto Hirakata, MD, PhD, for approving the use of the laboratory. Declaration of conflicting interests The funding source did not influence the study design, collection, analysis, or interpretation of data. Ethical approval This study was approved by a Review Board of Keio University, School of Medicine. Funding A portion of this study was sponsored by ETHICON in Japan. References 1. Ogawa R, Akaishi S, Huang C, et al. Clinical applications of basic research that shows reducing skin tension could prevent and treat abnormal scarring: the importance of fascial/subcutaneous tensile reduction sutures and flap surgery for keloid and hypertrophic scar reconstruction. J Nippon Med Sch 2011; 78: 68–76. 2. Bisson LJ, Sobel AD and Godfrey D.Effects of using a surgical clamp to hold tension while tying knots with commonly used orthopedic sutures. Knee Surg Sports Traumatol Arthrosc. 2012; 20: 1673–1680. 3. Karahan M, Akgun U, Turkoglu A, et al. Pretzel knot compared with standard suture knots. Knee Surg Sports Traumatol Arthrosc. 2012; 20: 2302–2306. 4. Markolf KL, Willems MJ, Jackson SR, et al. In situ calibration of miniature sensors implanted into the anterior cruciate ligament, part II: force probe measurements. J Orthop Res 1998; 16: 464–471. 5. Meyer DC, Jacob HAC, Nyffeler RW, et al. In vivo tendon force measurement of 2-week duration in sheep. J Biomech 2004; 37: 135–140. 6. Kocabey Y, Taser O, Nyland J, et al. Horizontal suture placement influences meniscal repair fixation strength. Knee Surg Sports Traumatol Arthrosc. 2013; 21: 615–619. 7. Staerke C, Brettschneider O, Grobel KH, et al. Tensile forces on sutures in the human lateral knee meniscus. Knee Surg Sports Traumatol Arthrosc 2009; 17: 1354–1359. 8. Bassini R, Cantone MC, Gambarini G, et al. Method for measuring suture tension in surgery. Med Biol Eng Comput 1988; 26: 451–454. 9. Schubert HM, Moser TM, Buchegger JW, et al. Tyrolean tensiometer: a new instrument for easy
Downloaded from pih.sagepub.com at UNIV OF PENNSYLVANIA on June 14, 2015
148
10.
11. 12.
13.
14.
15.
16.
Proc IMechE Part H: J Engineering in Medicine 228(2) intraoperative tension measurement before nerve coaptation. J Trauma 2006; 61: 760–763. Schubert HM, Hohlrieder M, Buchegger JW, et al. Tyrolean tensiometer: a new instrument for easy intraoperative tension measurement before vascular anastomosis. J Plast Reconstr Aesthet Surg 2007; 60: 311–315. Cummings JF.A miniature suture tensiometer for laparoscopic applications. J Invest Surg 2000; 13: 253–258. Kiriyama Y, Matsumoto H, Imanishi N, et al. Development of a miniature tension sensor to monitor suture tension during the joint motion. J Biomech 2008; 41(Suppl. 1): S503. Keegan KG, Baker GJ, Boero MJ, et al. Evaluation of support bandaging during measurement of proximal sesamoidean ligament strain in horses by use of a mercury strain gauge. Am J Vet Res 1992; 53(7): 1203–1208. Roberts CS, Cummings JF, Grood ES, et al. In-vivo measurement of human anterior cruciate ligament forces during knee extension exercises. Trans Orthop Res Soc 1994; 19: 84. Merican AM, Sanghavi S, Iranpour F, et al. The structural properties of the lateral retinaculum and capsular complex of the knee. J Biomech 2009; 42: 2323–2329. Mochizuki T, Nimura A, Tateishi T, et al. Anatomic study of the attachment of the medial patellofemoral
17.
18.
19.
20.
21.
22.
ligament and its characteristic relationships to the vastus intermedius. Knee Surg Sports Traumatol Arthrosc. 2013; 21: 305–310. Funk DA, Noyes FR, Grood ES, et al. Effect of flexion angle on the pressure-volume of the human knee. Arthroscopy 1991; 71(1): 86–90. Geborek P, Moritz U and Wollheim FA.Joint capsular stiffness in knee arthritis. Relationship to intraarticular volume, hydrostatic pressures, and extensor muscle function. J Rheumatol 1989; 16: 1351–1358. Jawed S, Gaffney K and Blake DR.Intra-articular pressure profile of the knee joint in a spectrum of inflammatory arthropathies. Ann Rheum Dis 1997; 56: 686–689. Jayson MIV and Dixon ASJ. Intra-articular pressure in rheumatoid arthritis of the knee. I. Pressure changes during passive joint distension. Ann Rheum Dis 1970; 29: 401–408. Anderl W, Heuberer PR, Laky B, et al. Superiority of bridging techniques with medial fixation on initial strength. Knee Surg Sports Traumatol Arthrosc. 2012; 20: 2559–2566. Garofalo R, Castagna A, Borroni M, et al. Arthroscopic transosseous (anchorless) rotator cuff repair. Knee Surg Sports Traumatol Arthrosc. 2012; 20: 1031–1036.
Downloaded from pih.sagepub.com at UNIV OF PENNSYLVANIA on June 14, 2015
