Secondary Logo

Journal Logo


Kinematic Evaluation of Terminal Devices for Kayaking With Upper Extremity Amputation

Highsmith, M Jason DPT, CP, FAAOP; Carey, Stephanie L. MS; Koelsch, Kip W. MA; Lusk, Craig P. PhD; Maitland, Murray E. PT, PhD

Author Information
JPO Journal of Prosthetics and Orthotics: July 2007 - Volume 19 - Issue 3 - p 84-90
doi: 10.1097/JPO.0b013e31806ada2f
  • Free

Kayaking and kayak fishing are very popular recreational activities in the United States.1,2 There are an estimated 350,000 kayaks sold every year in this country.3

Previously, kayaking has not been readily accessible to individuals with upper extremity amputations. Depending on the individual, the opportunity to learn kayaking could be a lifestyle modifier. Satisfying recreational activities are known to improve psychological well-being and reduce stress.1,2,4 Several studies indicate that persons with upper extremity amputation may reject their prosthesis because of insufficient function in one or more arenas, including recreational activities.5–9

In a general sense, cylindrical grasp can be used to hold a variety of objects, such as paddles, brooms, fishing rods, oars, weight-lifting equipment, and tools. There are few terminal devices that can perform a stable cylindrical grasp. Kayaking requires a cylindrical grasp. Currently there are two commercially available terminal devices (TDs) offering the individual with upper extremity amputation the ability to hold and manipulate a paddle for kayaking and canoeing.10

The first is a multipurpose type device not specifically designed for any one activity. The Grip 2S and Grip 3 Prehensors from TRS Inc. (Boulder, CO) use an active, body-powered, voluntary closing mechanism to grasp the kayak paddle.10 Voluntary closing terminal devices require constant scapular force to maintain grasp. The Grip 2S and Grip 3 units can be modified to be mechanically fixed in a closed position to eliminate the need to maintain force; however, the grasp area is not specifically designed for kayaking. Although this type of prehensor has its advantages (e.g., increased and graded prehensile force, proprioception), most body-powered terminal devices in service are of the voluntary opening type.11 Users tend to prefer voluntary opening devices for activities of daily living (ADLs) because grasp is maintained passively, usually through tension developed by rubber bands.11–15

The second type of terminal device available for kayaking is a passive-functioning, recreational terminal device. The Hammerhead Kayak Hand (also from TRS Inc.10) was made available on the commercial market in March 2006. Passive functioning terminal devices usually are specific to only one or two activities. Examples include bowling, billiards, weight lifting, photography, and in this case, kayaking. Passive prehension is preferred for these activities because once the recreational hand is attached to the activity object (e.g., paddle, pool cue, bowling ball, camera), no mental or physical effort is required to maintain grasp of the object. This permits mental effort to be redirected from the specific task of grasping to enjoying the activity.12,14 This new passive-functioning kayak terminal device is designed with a high-friction grip surface and a diameter larger than a standard paddle, affording rotation and movements similar to action afforded by an anatomic wrist and fingers. This unit is otherwise positioned in neutral flexion/extension and radial/ulnar deviation. It offers a user-friendly T-handle and button-closure strap to secure or quickly release a paddle with unilateral arm involvement. Concerns regarding this hand are with the initial position and its ability to afford paddle rotation while preventing mediolateral shifting.

The effort used to maintain grasp of a paddle with a mechanical, body-powered terminal device, even over short durations, can be fatiguing and awkward. Passive hands tend to be preferred when possible, particularly for recreational activity to afford maximal mental unloading. The passive-function, TRS Hammerhead kayak hand currently offers an immediate solution for a return to kayaking among interested persons with upper-extremity amputation. However, evaluation is indicated to address the concerns identified.

For the reasons highlighted, it seems logical that through systematic motion analysis of kayaking, a prehensor offering stable cylindrical grasp that prevents mediolateral paddle sliding while still allowing paddle rotation could be created. If the grasp function of the new device is successful, then perhaps this feature could be integrated with the numerous beneficial features of the TRS Hammerhead Kayak hand. Such a device, that combines all these features, might encourage maximal activity participation and ease of movement.

One purpose of this project was to develop a simply designed prehensor offering a stable cylindrical grasp specifically to facilitate kayaking with individuals having an upper extremity amputation. A second purpose was to compare the kinematic performance of the newly designed prehensor with the commercially available TRS unit and then compare both units with expert kayak kinematics.


Investigators performed a preliminary kinematic motion analysis of an expert kayak paddler and found that the joints of the nondominant upper limb move through relatively limited ranges of motion during the stroke cycle. It was additionally noted that paddle rotation is driven by the dominant hand and that the nondominant hand behaved like a cylinder rotating within a cylinder.

Using the following parameters, a concept hand was sketched, designed, and fabricated:

A kayak hand must:

  1. Maintain grasp of the paddle
  2. Allow the paddle to rotate to position the blade in the water appropriately from side to side
  3. Minimize mediolateral paddle shifting during all phases of the kayak stroke cycle
  4. Allow movement comparable to forearm rotation (pronation/ supination)
  5. Mimic anatomic wrist positioning (flexion/extension and radial/ulnar deviation)

The USF Kayak Hand was fabricated with these design parameters as guidelines (Figure 1). A pseudo-prosthesis (similar to that of Lake16) was then fabricated so the USF Kayak Hand could be tested by nonamputee expert kayakers (Figures 2 and 3). The University of South Florida’s Institutional Review Board granted permission to recruit and test the device with human subjects. Recruited subjects gave informed consent for their participation.

Figure 1.:
The USF Kayak Hand. The attachment bolt is welded to the frame. The frame has large windows cut into it to reduce mass and friction with the inner sleeve.
Figure 2.:
Expert kayaker wearing pseudo-prosthesis and TRS Hammerhead Kayak Hand during preliminary testing. Reflective markers are located on the upper limbs, prosthesis, torso, and paddle.
Figure 3.:
Expert kayaker sitting on a mat table simulating paddling in the laboratory with pseudo-prosthesis and USF Kayak Hand. Marker placement was duplicated with all subjects.


Four volunteers participated in this study (n = 4). Two participants without amputation were the control group and two subjects with upper-extremity amputations were the experimental group. All subjects were right hand dominant (prior to amputation in one case).


Control subjects (n = 2) were expert kayakers with intact upper extremities. Control subjects were required to be expert kayakers, defined as having prior kayak coaching experience and regularly participating in training and racing events during the last 4 years. Control subjects wore the pseudo-prothesis (Figure 2) on the left (nondominant) hand.


There were two subjects with unilateral amputation of the upper extremity. One subject had a right-sided transhumeral amputation (THA), and the other subject had a left-sided transradial amputation (TRA). Each subject used a body-powered prosthesis that had an interchangeable TD. Subjects expressed a desire to kayak. Subjects were not candidates if they owned only a myoelectric or externally powered prosthesis. This was because such a prosthesis should not get wet because moisture can damage the electronic circuitry, whereas body-powered prostheses may have an interchangeable wrist/hand, be very durable, and be less vulnerable to water damage.

Experimental group subjects were taken out on the water for training with expert supervision to learn to use both kayak terminal devices under real conditions. Experimental group subjects were given two training sessions, at least 1 week apart, that lasted 1.5 to 2 hours per session prior to final kinematic analysis. Training under real conditions was deemed necessary to assure a reasonable level of performance with each TD in the laboratory setting.


The motions of subjects simulating kayak paddling in a laboratory setting during various test conditions were recorded using a six-camera, infrared Vicon motion analysis system.

Test conditions are listed below.

Repeated (Within Group) Measures (Control Group):

  1. Expert Paddling—no prosthesis
  2. Expert Paddling—TRS Hammerhead Kayak TD with pseudo-prosthesis
  3. Expert Paddling—USF Kayak Hand with pseudo- prosthesis

Between-Group (Experimental) Measures:

  1. TRA—TRS Hammerhead Kayak TD
  2. TRA—USF Kayak Hand
  3. THA—TRS Hammerhead Kayak TD
  4. THA—USF Kayak Hand

Twenty-two reflective markers were placed on the subjects’ upper extremities, torso, and kayak paddle as shown in Figures 2 and 3. Prior to each data collection, the cameras were calibrated. When simulating paddling in the laboratory, subjects sat on top of a mat table (Figure 3). For each subject and experimental setup, six full strokes (left paddle, right paddle) were collected. Blinding and randomization were not necessary because this was a small sample, descriptive study.


From the six strokes collected from each trial, the middle four strokes were averaged and used for data analysis. Data presented here is the average of these four strokes. Anthropometric measurements from the subjects were recorded and used to aid in computing joint centers. From the markers and joint center calculations, segments of the torso, upper arm, forearm, hand, and kayak paddle were created. Shoulder joint angles, elbow flexion, torso rotation angles, and mediolateral hand motion along the paddle were calculated. For the shoulder joint, only coronal and sagittal plane movements were considered. The mediolateral shifting of the hand/terminal device on the paddle was determined by calculating the distance of movement of the hand center along the mediolateral axis of the kayak paddle coordinate system. The hand’s center was estimated using the hand marker and the thickness of the hand (measured for each subject). Although data were analyzed for both sides, only data from the prosthetic side are presented. All subjects, with the exception of the subject with THA, wore the pseudo-prothesis or prosthesis on the left side.


Results are based on the four averaged strokes for each subject. The values for control subject 1 and control subject 2 were averaged together and represent the control group throughout the data. Data are presented in the following sequence: trunk rotation, shoulder motion, elbow flexion, and mediolateral sliding. Forearm rotation was effortlessly simulated by leaving the friction setting loose in the standard wrist units of all three prostheses: pseudo-prosthesis, TRA, and THA (all comparable to Hosmer Friction Wrist Unit 52147, Hosmer, Campbell, CA). There were no obvious safety issues, drawbacks, or complaints associated with allowing the TD to rotate within the wrist, and thus kinematic data are not reported for this aspect of the articulation.

Maximum trunk rotation is illustrated in Figure 4. In all conditions except one, right peak trunk rotation exceeded left. With THA using the TRS hand, maximum left trunk rotation exceeded that of the right. The TRS hand seemed to afford a consistent amount of peak trunk rotation across all users, whereas the USF hand did not.

Figure 4.:
Maximum trunk rotation.

Despite the TD, control subjects maintained nearly normal shoulder kinematics in both the sagittal and coronal planes, as shown in Figures 5 and 6 (depicting an example kayak stroke from control subject 1). In the control group, regardless of the TD, the shoulders never went into extension. The subject with a TRA demonstrated less shoulder flexion with both TDs, but the TRS hand actually yielded shoulder movement into extension. Regarding the TRA subject’s coronal plane motion, both TDs yielded movement at the upper limits of the normal range compared with control subjects. The subject with THA demonstrated the largest range of shoulder abduction motion regardless of TD, as shown in Figure 7. However, in this instance the USF hand actually yielded shoulder extension.

Figure 5.:
A comparison of shoulder flexion during one stroke for control subject 1 under three conditions: no prosthesis, pseudo-prosthesis with TRS hand, and pseudo-prosthesis with USF hand.
Figure 6.:
A comparison of shoulder abduction during one stroke for control subject 1 under three conditions: no prosthesis, pseudo-prosthesis with TRS hand, and pseudo-prosthesis with USF hand.
Figure 7.:
Range of shoulder motion in the coronal plane for all subjects, all conditions. Positive value = shoulder abduction. Negative value = shoulder adduction.

Sagittal plane elbow motion is depicted in Figure 8. The greatest range of elbow motion was seen in control subjects when no prosthesis was used (15° to 58°). The USF hand yielded less elbow movement range, but the movement stayed within the control group’s range (no-prosthesis condition) with control and TRA subjects. The TRS hand yielded an elbow range of motion similar to that of the USF hand but resulted in elbow flexion beyond the range defined by the control group (no-prosthesis condition) with control and TRA subjects. Overall elbow range in THA was negligible because the elbow was locked at 40°. Therefore, any motion observed in this group is likely the combined result of motion between the interface and body and unwanted movement in the prosthetic elbow joint.

Figure 8.:
Range of elbow flexion for all subjects, all conditions.

Figure 9 shows the mediolateral shift of the hand along the kayak paddle. These distance values may be elevated because of the possibility of the hand pivoting on the paddle at points not located at the hand’s center. However, the trends showing the TRS hand having a greater mediolateral shift than the USF hand should be noted. The greater mediolateral shift in the control group without a prosthesis may also be attributable to loosening of the grip of the nondominant hand to allow greater paddle control to the dominant hand.

Figure 9.:
Mediolateral paddle slide for all subjects, all conditions. Distance displayed in millimeters that the hand or TD slides along the paddle.


The USF Kayak hand offered a stable cylindrical grasp as intended. It displayed superior mediolateral capture under all conditions. Mediolateral grasp was maintained while still allowing rotation of the paddle under wet and dry conditions. However, the USF hand, prepositioned in 15° of radial deviation, did not afford any motion consistent with radial and ulnar deviation, which led to complaints of being “stiff.” The tradeoff benefit to being “stiff” was reportedly an ability “to feel the water.” The USF hand was also reportedly more difficult to apply to the paddle than was the TRS hand. It has two hook-and-loop closure straps and an inner sleeve that can be dropped during paddle application.

Another dilemma with the USF hand is that it is a custom-fabricated item that involves welding. This requires technical skills not readily available in every prosthetic facility. The final problem is that prototypes for evaluation in this study were fabricated from mild steel pipe, so corrosion could be a problem in the long term. Thus, revisions to the USF hand would include the addition of a compliant wrist joint that offers radial and ulnar deviation and fabricating the frame out of aluminum or thermoplastic to prevent corrosion and further reduce mass (although mass was not specifically a complaint).

Regarding the TRS hand, this product is commercially available, presently making it the go-to TD option for kayaking. It was reportedly easier to apply to the paddle than the USF hand because of its single, button-closure strap and having no additional parts that could “fall in the water” or otherwise be dropped during paddle application. The dilemma with this TD is that its easy closure strap presented difficulty in terms of maintaining mediolateral hand position along the paddle. The TRS TD was found to allow the paddle to slip when dry in the laboratory, so when the paddle was wet from kayaking, the problem seemed to magnify.

Amputee participants reported that the TRS TD was more forgiving of error in form than the USF hand, but the tradeoff was that the sound hand was perceived to work a bit harder to maintain mediolateral paddle position. Despite individual drawbacks, both TDs allowed independent donning and paddle application. Both devices also resulted ultimately in successful activity participation at both the transhumeral and transradial levels of amputation.

Although observational, kinematic data were collected from only two subjects with different levels of upper limb amputation, these data may be widely generalizable to other persons with comparable levels of upper limb amputations. The issue of sample size is an obvious limitation of this and several other studies involving persons with upper limb amputation. Other limitations concern marker placement that does not permit thorough explanation of mediolateral paddle sliding in the nonprosthetic, control situation that yielded high mediolateral slide values. Our intention was to focus on gross movements of the entire trunk and upper extremities during kayaking, whereas a more in-depth study of hand/wrist movement with more markers on the hand/wrist complex is required to fully explain this phenomenon.


From a gross perspective, objective kinematic differences regarding joint motion were subtle between the two TDs evaluated across all three test groups. As predicted, mediolateral paddle slide was the most obvious difference of the parameters measured. Regarding this sliding, the TRS hand yielded the most slide compared with the anatomic hand and the USF TD. To prevent excess sliding when the TRS hand gets wet during kayaking, the uninvolved hand will likely be required to maintain mediolateral paddle position. Such effort could potentially add stresses to the sound hand in extended bouts of kayaking. In terms of nonmeasured parameters, the TRS hand was reportedly easier to apply to the paddle and more forgiving of technical errors in form. From this analysis, it is not possible to determine if such form errors and sound hand grasp maintenance (to prevent mediolateral sliding) could result in overuse injuries in a more seasoned athlete after an extended bout of kayaking. Under these brief evaluative conditions, all subjects were successful with both TDs during on-water kayak training and during simulated paddling in the laboratory. Additional study is not required to recommend either TD for use with the novice, recreational athlete. However, additional study on this topic should consider the effect of a true wrist articulation, rather than simulating wrist action in the grip area, a detailed kinematic analysis and breakdown of the stroke cycle in the nonamputee population, and the evaluation of form errors and overuse injuries associated with use in the amputee populations with select terminal devices.


The authors acknowledge Bob Radocy, of TRS, for consultation and for providing a terminal device for comparison/evaluation; Westcoast Brace & Limb (Tampa, FL) for technical fabrication assistance and subject recruitment; Sam Phillips, dean of St. Petersburg College of Orthotics and Prosthetics, for methodology insight; and the University of South Florida’s Outdoor Recreation Staff and School of Physical Education and Sport Studies’ Motion Analysis Laboratory for facilities to train and evaluate participants for the project.


1. Coastal Kayak Fishing Inc. Available at: Accessed November 10, 2006.
2. Florida Kayak Fishing/Flyfishing. Available at: Accessed November 10, 2006.
3. United States Coast Guard. Available at: Accessed November 10, 2006.
4. Lee IM, Sesso HD, Paffenbarger RS Physical activity and coronary heart disease risk in men: does the duration of exercise episodes predict risk? Circulation 2000;102 (9):981–986.
5. Stein RB, Walley M. Functional comparison of upper extremity amputees using myoelectric and conventional prostheses. Arch Phys Med Rehabil 1983;64 (6):243–248.
6. Datta D, Selvarajah K, Davey N. Functional outcome of patients with proximal upper limb deficiency: acquired and congenital. Clin Rehabil 2004;18 (2):172–177.
7. Michael JW, Gailey RS, Bowker JH. New developments in recreational prostheses and adaptive devices for the amputee. Clin Orthop Relat Res 1990;256:64–75.
8. Durance JP O’Shea BJ. Upper limb amputees: a clinic profile. Int Disabil Stud 1988;10 (2):68–72.
9. Rubin G, Fleiss D. Devices to enable persons with amputation to participate in sports. Arch Phys Med Rehabil 1983;64 (1):37–40.
10. TRS, Inc. Available at: Accessed November 10, 2006.
11. Frey DD, Carlson LE, Ramaswamy V. Voluntary-opening prehensors with adjustable grip force. J Prosthet Orthot 1995;7 (4):124–131.
12. Northwestern University Prosthetic Orthotic Center. Northwestern University Upper Limb Prosthetics for Prosthetists Course Manual. Chicago, IL: Northwestern University: Spring 2004.
13. Wilson AB. A Primer on Limb Prosthetics. Springfield, IL: Charles C. Thomas Publisher LTD; 1998.
14. Radocy R. Prosthetic adaptations in competitive sports and recreation. In: Smith DG, Bowker J, Michael J, eds. Atlas of Amputations and Limb Deficiencies: Surgical, Prosthetic, and Rehabilitation Principles. 3rd ed. Rosemont, IL: American Academy of Orthopaedic Surgeons; 2002:327–337.
15. Procter S, LeBlanc M. Clinical evaluation of a new design prosthetic prehensor. J Prosthet Orthot 1991;3 (2):79–83.
16. Lake C. Effects of prosthetic training on upper-extremity prosthesis use. J Prosthet Orthot 1997;9 (1):3–9.

cylindrical grasp; kayak; passive function; passive terminal device; prosthesis; recreational terminal device; terminal device; upper extremity; upper limb

© 2007 American Academy of Orthotists & Prosthetists