Secondary Logo

Journal Logo


Measuring Residual Limb Skin Temperatures at the Skin-Prosthesis Interface

Huff, Elizabeth A. BSE; Ledoux, William R. PhD; Berge, Jocelyn S. MSE; Klute, Glenn K. PhD

Author Information
JPO Journal of Prosthetics and Orthotics: October 2008 - Volume 20 - Issue 4 - p 170-173
doi: 10.1097/JPO.0b013e3181875b17
  • Free

For many amputees, elevated temperatures at the skin-prosthesis interface cause an increase in skin problems and a decrease in quality of life. Survey results from a sample population of transfemoral amputees (n = 97) revealed that excessive heat and perspiration within the prosthesis was the number one quality of life complaint (72%) followed by sores, chafing, and skin irritation of the residual limb (62%).1 Exacerbating the sense of discomfort is the increased potential for residual limb skin problems. Trapped heat and perspiration within the prosthesis promotes maceration (softening) of the skin, which can create an ideal environment for a bacterial invasion of the hair follicles. Hair follicle inflammation (folliculitis) is a well-known source of irritation for many amputees.2 The accumulation of sweat on the residual limb can also serve as an irritant and a possible culture medium for other bacterial growth.3 There is even evidence to suggest that elevated skin temperature and excessive perspiration may promote the formation of friction blisters on the residual limb4,5 eventually restricting or preventing the amputee’s use of their prosthesis. To help create a more comfortable and innocuous prosthesis, the temperature measurements at the interface of the residual limb skin and prosthesis are a necessary first step.

Our group conducted a preliminary investigation that measured temperatures at the skin-prosthesis interface.6 Five transtibial amputees with various socket systems were asked to sit quietly in a chair for 15 minutes and then walk on a treadmill for 10 minutes. Skin temperatures were measured using 14 thermistors and 38-gauge wires routed through a multiplexor to a single-channel data acquisition system. The results from this study indicated that temperatures at the skin-prosthesis interface vary with activity and location on the limb. However, the data collections were hindered by multiple wire failures at the distal posterior location, and the protocol was too short for the skin temperatures to reach a steady state in either activity period (resting or walking).

This article describes the design, validation, and application of a more robust instrument to accurately and reliably measure temperature at the skin-prosthesis interface during a 2.5-hr experimental protocol that included periods of resting and activity.



To measure temperatures at the skin-prosthesis interface, 16 thermistors (type MA100GG; Thermometrics, Edison, NJ) with 2 mm sensor head diameters, were each connected through a linear voltage divider (10 kω ± 5% precision resistors) to a data acquisition system. The data acquisition system consisted of a BNC-2090 A/D board (National Instruments Corporation, Austin, TX) and Labview (National Instruments Corporation, Austin, TX) software. To reduce the incidence of wire failure, the thermistors were connected to the data acquisition system with 28-gauge wires. The 16 voltage dividers were arranged in parallel, and the circuit was powered at 0.5 V (voltage supply model PS280; Tektronix, Inc., Beaverton, OR) to minimize thermistor self-heating (<0.01°C). Data from all 16 channels were simultaneously collected at 1 Hz, and the manufacturer-supplied curve fit was used to calculate temperature from the measured output voltage.7


To determine the accuracy of the instrument, the thermistors were placed in a water bath and run throughout the range of temperatures expected during human subject testing (25°C–45°C). Bath temperature was controlled using a temperature regulator (TE-10D Thermoregulator; Techne, Burlington, NJ) which is accurate within 1°C (manufacturer’s specification). To precisely validate the thermistors, a National Institute of Standards and Technology (NIST) calibrated instrument was used as the temperature gold standard. This system consisted of a precision thermistor probe (Model No. ON-403-PP; Omega Engineering Inc., Stamford, CT) and an ultra-high accuracy thermistor thermometer (Model No. HH41; Omega Engineering Inc., Stamford, CT). The NIST calibration revealed that this device had a maximum error of 0.08°C with a maximum calibration uncertainty of 0.3°C. The data acquisition system was turned on 1 hr before the experiment for the system to reach thermal equilibrium. The bath was filled with tepid water and the thermistor probe was immersed in the water. The temperature on the regulator was gradually increased until the thermistor probe read the desired temperature ±0.25°C. The water bath was then allowed a quiescent period of at least 5 minutes to insure steady state conditions before the thermistors were immersed in the water. The thermistors were given one additional minute to reach steady state conditions before a 1-minute temperature recording was collected. This process was conducted at 25°C and repeated at 30°C, 35°C, 40°C, and 45°C.


To demonstrate the practicality of the instrument and its ability to reliably perform during a human subject experiment, one unilateral, transtibial amputee was recruited to participate in a 2.5-hr laboratory protocol. The subject was a 59 year-old male who weighed 84.5 kg and wore a 9 mm Alpha® (Ohio Willow Wood, Mt. Sterling, OH) liner with a distal locking pin and a carbon fiber socket. He was 33 years postamputation of traumatic etiology, an independent ambulator, and his residual limb was of typical dimensions and in good condition. An approved University of Washington Human Subjects Review Committee protocol was followed for recruitment and consent.

Before data collection, the 16 thermistors were taped onto the residual limb in four columns with four thermistors in each column (see Figure 1). The first column was placed on the tibial crest and the second column was placed over the tibialis anterior. Columns three and four were placed on the medial and lateral gastrocnemius, respectively. Sensors were evenly spaced from the tibial tuberosity to the distal end of the limb. After the thermistor heads were secured with tape, the prosthesis was donned with the thermistor wires exiting the proximal edge of the liner.

Figure 1.
Figure 1.:
Representative thermistor placement on the residual limb. The anterior view shows thermistor columns one and two placed over the tibia and the tibialis anterior, respectively. The posterior view shows thermistor columns three and four placed over the medial and lateral gastrocnemius, respectively. The lines on the residual limb were drawn to help guide thermistor placement.

The subject was asked to complete the following protocol: sit for 1 hr, walk on a treadmill at his self-selected speed (0.4 m/s) for 30 minutes, and finally rest for one more hour. The first 2 minutes of treadmill walking were used to self-select a comfortable walking pace that the subject could maintain for 30 minutes. Residual limb skin temperature was sampled at 1 Hz for the entire 2.5-hr protocol. For analysis purposes, three steady state periods were defined as the last minute of each period: initial rest, treadmill walking, and final rest.



The purpose of the water bath test was to assess the accuracy and linearity of the thermistors throughout their expected working range. The temperature regulated water bath revealed the average temperature error was no greater than 0.31°C over the entire range tested (see Table 1).

Table 1
Table 1:
The water bath temperature (i.e., the NIST gold standard) and the means and standard deviations from 1 minute of thermistor data (60 samples) measured at various temperatures


After donning the prosthesis, the subject’s mean residual limb temperature dropped approximately 0.4°C until it reached an initial resting steady state temperature of 29.5°C ± 0.9°C during the last minute of the initial 1-hr rest period (see Figure 2). Once the exercise period began, the mean residual limb temperature rose steeply and did not reach a plateau even after 30 minutes of treadmill walking at the subject’s self-selected speed. The residual limb temperature during the last minute of the 30-minute treadmill walk was 32.6°C ± 0.8°C. After immediately continuing up to an overall peak of 32.8°C, the temperatures began to decrease during the final rest period. The residual limb temperature during the last minute of the final rest period was 32.6°C ± 0.6°C, 3.1°C higher than the mean temperature before exercise.

Figure 2.
Figure 2.:
Average residual limb temperature from all 16 thermistors during the entire data collection period (mean: solid line, ±1 standard deviation: dashed lines). The subject sat in a chair for 1 hr (initial rest), walked on a treadmill for 30 minutes (exercise), and finally sat in a chair for an additional hour (final rest). Three steady states periods were defined as the last minute of each activity and are marked with circles.

To visually examine the subject’s steady state residual limb temperatures, three thermal contour plots of the limb were created using Matlab software (The MathWorks, Natick, MA). The three plots contain data collected during the last minute of the initial rest period, the treadmill walk, and the final rest period (see Figures 3–5). Each figure consists of an anterior and posterior view where the anterior view contains the data from sensors superficial to the tibial crest and the tibialis anterior, and the posterior view contains the data from the sensors superficial to the lateral and medial gastrocnemius.

Figure 3.
Figure 3.:
Residual limb temperature during the initial resting period. After donning the prosthesis, subject sat for 1 hr while thermistors collected residual limb temperature at 1 Hz. Temperatures varied by location with skin temperatures coolest over the most bony region (tibial crest: anterior, medial section) and warmest over the most muscular region (lateral gastrocnemius: posterior, lateral section).
Figure 4.
Figure 4.:
Residual limb temperature during the exercise period. The subject walked on a treadmill at his self-selected speed (0.4 m/s) for 30 minutes while thermistors collected residual limb temperature at 1 Hz. Temperatures increased relatively uniformly from the previous condition.
Figure 5.
Figure 5.:
Residual limb temperature during the final resting period. After the 30-minute exercise portion, the subject sat for an additional hour while thermistors collected residual limb temperature at 1 Hz. The skin temperatures still show a location variation; the bony region (tibial crest, anterior, medial section) is slightly cooler then the more muscular region (lateral gastrocnemeius, posterior, lateral section).

The initial resting contour plot (see Figure 3) shows much cooler temperatures superficial to the bony region of the residual limb (tibial crest) and much warmer temperatures superficial to the muscular region (gastrocnemius). The minimum temperature during the last minute of initial rest was 28.1°C and the maximum was 31.3°C. The treadmill walking contour plot (see Figure 4) shows that skin temperatures rose fairly uniformly about the limb with a minimum of 31.1°C and a maximum of 34.1°C during the last minute of walking. The final rest contour plot (see Figure 5) revealed that the residual limb had cooled rather evenly with a minimum temperature of 31.4°C and a maximum of 33.9°C.


Excessive heat and perspiration within the prosthetic limb are significant problems for many amputees due to the negative effect on their quality of life and the increased potential for skin problems. The first step in alleviating this problem is to understand the heat environment at the skin-prosthesis interface. This manuscript describes an instrument to measure temperatures at the skin-prosthesis interface using a network of 16 thermistors operating at low power.

The main limitation of this device is the short period of time that amputees can walk without risk of injury while wearing the thermistors. Any foreign particle at the skin/socket interface will create some irritation, and a possible skin break if left in place long enough. The subject involved in this study experienced only minor skin irritation during the data collection. Longer durations should be explored with caution.

A validation test of the instrument using a temperature-controlled water bath revealed accurate and linear results over the working range. There was little variation at each of the five temperature levels tested.

The human subject test results indicated that the prosthesis is very effective at trapping heat. During the initial rest period, the temperatures at the skin-prosthesis interface varied by location; skin temperatures superficial to bony regions were cooler than areas over large muscle masses. One hour of initial rest was more than sufficient to reach steady state conditions during initial rest.

Walking with the prosthesis caused a rapid increase in temperature, but the temperature did not reach a steady state even after 30 minutes of walking. Although a longer period may be necessary to reach a plateau temperature during walking, a study of lower limb amputees (n = 12 transtibial and n = 5 transfemoral) suggested that a 30-minute walk is a rare event.8 Most activity bouts were of 2- or 3-minute duration, and only about once per day did a lower limb amputee walk for a full 10 minutes.

At the end of the final 1-hr rest period, the skin temperatures had not returned to their initial rest period values. This result demonstrates the difficulty of releasing heat through the prosthesis. A significantly longer rest period after exercise may be necessary to return the limb to a pre-activity temperature state.


The instrument described in this article provides a mean to quantify the thermal environment at the skin-prosthesis interface; a necessary step before further research can be conducted to explore the effects of different socket and suspension system materials and geometry intended to provide a more thermally comfortable and innocuous prosthesis.


1. Hagberg K, Brånemark R. Consequences of non-vascular trans-femoral amputation: a survey of quality of life, prosthetic use and problems. Prosthet Orthot Int 2001;25:186–194.
2. Levy SW. The skin problems of the lower-extremity amputee. Artif Limbs 1956;3:20–35.
3. Barnes GH. Skin health and stump hygiene. Artif Limbs 1956;3:4–19.
4. Akers WA, Sulzberger MB. The friction blister. Mil Med 1972;137:1–7.
5. Naylor PF. The skin surface and friction. Br J Dermatol 1955;67:239–48.
6. Peery JT, Ledoux WR, Klute GK. Residual-limb skin temperature in transtibial sockets. J Rehabil Res Dev 2005;42:147–154.
7. Manufacturer’s Specifications on NTC Thermistors (Material F). Thermometrics; 2001. Available at:
8. Klute GK, Berge JS, Orendurff MS, et al. Lower limb amputee activity unaffected by shock-absorbing pylon or c-leg knee. The International Society of Biomechanics XXth Congress; Cleveland, OH; 2005.

lower limb amputee; residual limb; skin temperature; prosthesis; lower extremity

© 2008 American Academy of Orthotists & Prosthetists