Does Activity Affect Residual Limb Skin Temperatures? : Clinical Orthopaedics and Related Research®

Secondary Logo

Journal Logo

Symposium: Recent Advances in Amputation Surgery and Rehabilitation

Does Activity Affect Residual Limb Skin Temperatures?

Klute, Glenn K. PhD1, 2, a; Huff, Elizabeth MS1; Ledoux, William R. PhD1, 2, 3

Author Information
Clinical Orthopaedics and Related Research 472(10):p 3062-3067, October 2014. | DOI: 10.1007/s11999-014-3741-4
  • Free



Many lower extremity amputees report elevated residual limb skin temperatures and accompanying perspiration. Sixty-five of 90 respondents (72%) to a survey claimed the most common cause for reporting a moderate or worse reduction in their quality of life during the summer months was heat and perspiration within the prosthetic socket [4]. Simply donning a prosthesis will cause an increase in residual limb skin temperatures and a short subsequent bout of activity will cause a further increase [8]. The material properties of prosthetic limbs certainly contribute to observed increases in residual limb skin temperatures. Although the elastomeric liners worn by amputees between their residual limb and prosthetic socket may help maintain secure adherence during ambulation, these materials are nearly impermeable to moisture [3] preventing thermal relief through evaporative cooling while their low thermal conductivity [7] prevents conductive heat transfer.

Nonamputees can make behavioral choices to aid in maintaining thermal comfort, donning and doffing garments as desired. Lower limb amputees have fewer choices because they cannot doff their prosthesis and remain ambulatory. Whereas skin temperatures increase after donning a prosthesis [8], it is unknown for how long skin temperatures will continue to increase before reaching a steady-state condition. With activity comes an increase in residual limb skin temperatures, even with bouts as short as 10 minutes [8], but will temperatures plateau with longer bouts of activity or will they continue to rise? After activity, circumstances may be such that an amputee cannot doff their limb to seek thermal relief. Is rest alone after activity sufficient to return skin temperatures to their baseline skin temperatures or will an intervention be necessary? If there is a need for a prosthetic intervention, the design may be challenging because skin temperatures may not be evenly distributed. Regions with greater muscle mass may give rise to higher temperatures and require more heat transfer than regions with little muscle mass.

The purpose of this study was to characterize residual limb skin temperatures inside the socket and suspension system of lower extremity amputees during activity and after activity ceases. We made observations on a sample population of unilateral transtibial amputees to answer the following questions: (1) Does residual limb skin temperature change as a function of activity and its cessation? (2) If changes occur, are there regional differences (circumferential or proximal-distal) in temperature?

Patients and Methods

Fourteen male unilateral transtibial amputees provided informed consent to participate in this institutional review board-approved observational study. All were moderately active community ambulators by self-report. Exclusion criteria included: (1) poor socket fit that limits mobility or might contribute to a skin or soft tissue injury from study instrumentation; (2) physical inability to walk for 30 minutes; and (3) a diagnosis of diabetes. After enrollment, a screening test was performed to ensure participants could sense if the temperature measurement system was irritating their skin. A Semmes-Weinstein 5.07 monofilament (CHS Services, Inc, East Setauket, NY, USA) was used to make skin contact, bend, and depart the skin at each intended sensor site. Subjects were excluded if they could not sense the monofilament at all of the sites. Four participants were excluded from participation or data analysis: one participant had a poor socket fit requiring clinical attention, one participant was unable to walk for 30 minutes, and two participants acknowledged a diagnosis of diabetes after screening. A fifth participant’s data were lost as a result of problems with the data collection system.

The nine participants who completed the protocol were 49 ± 15 years of age (mean ± SD), 16 ± 13 years postamputation with a height of 1.81 ± 0.04 m and a body mass of 93 ± 14 kg. Their prosthetic prescriptions (Table 1) all included: a modified patellar tendon-bearing, carbon fiber-laminated socket, an elastomeric liner (except for one who wore a foam liner) of uniform thickness, and socks of various plies to accommodate changes in socket fit (except for one who wore no sock).

Table 1:
Participant demographics and prosthetic prescriptions*

On arriving in the laboratory, the participants were asked to sit in a chair and remove their prosthesis. Sixteen thermistors (Type MA100GG; Thermometrics, Edison, NJ, USA), validated before use with a water-bath procedure and a National Institute of Standards and Technology-calibrated instrument [6], were secured with medical-grade tape onto the participant’s residual limb in four anatomically defined circumferential locations consisting of four columns of four sensors each (Fig. 1). The first column of four sensors was over the tibial crest, the second over the tibialis anterior, the third over the medial gastrocnemius, and the fourth over the lateral gastrocnemius. The four sensors were evenly spaced by observation proximal-distal in each column from the transverse plane transecting the tibial tuberosity (Sensor 1) to the distal end of the limb (Sensor 4). The prosthesis was then donned over the sensors with the sensor wires exiting along the proximal edge of the liner. This process took approximately 15 minutes during which time the participants became acclimated to the laboratory environment (23 ± 3° C and 50% ± 8% relative humidity).

Fig. 1A-B:
Representative images show the thermistor placement on a left residual limb. The anterior view (A) shows thermistors placed over the tibial crest and the tibialis anterior. The posterior view (B) shows thermistors placed over the medial and lateral gastrocnemius. The lines on the residual limb were drawn to help with thermistor placement.

The experimental protocol consisted of three activities performed in order: an initial 60-minute seated rest, a 30-minute treadmill walk at self-selected speed, and a final 60-minute seated rest. At the beginning of the treadmill walk, the speed of the treadmill was adjusted until the participant felt confident they could walk at that speed for 30 minutes. This adjustment was accomplished within 3 minutes. Temperature data were collected (SmartReader Plus 8; ACR Systems Inc, Surrey, British Columbia, Canada) at 0.125 Hz throughout the experiment.

Linear mixed effects regression was used to determine if changes in skin temperatures were statistically significant (p < 0.05) by anatomical circumferential location (ie, tibial crest, tibialis anterior, medial gastrocnemius, and lateral gastrocnemius), proximal-distal sensor position (ie, Sensor 1 through Sensor 4), and activity (ie, initial rest, treadmill walk, final rest) with temperature as the dependent variable and location and sensor position and activity as independent fixed effects. The hypothesis test for the association between temperature and each independent effect was adjusted for the other two effects. Random effects were estimated for differences in average temperature across subjects. Additional random effects were estimated for variation in the relationship between temperature and each effect across subjects because each subject experienced all levels of each of the three effects. Normality of the residuals was assessed using normal plots. Effect sizes are presented as mean differences in temperature across relevant categories and 95% confidence intervals (CIs). If the association between temperature and either location or activity was significant, pairwise comparisons were carried out using simultaneous inference [5]. The association between temperature and sensor position was also modeled as a polynomial trend from proximal to distal. The linear mixed effects regressions were estimated using R 3.0.1 (; R Core Team, Vienna, Austria) and the lme4 and nlme packages. Pairwise comparisons were performed with the multcomp package. Data from the last minute of each of the three activity periods were used for analysis. To provide a visual representation of the results, a thermal contour plot was created using a cubic interpolation of the 16 thermistors to map the temperatures using the last minute of data from each of the three activity periods (Matlab; The MathWorks, Natick, MA, USA).


Activity rapidly increased the residual limb skin temperatures and cessation of activity caused a moderate decrease (linear mixed effects model adjusting for proximal-distal sensor position and anatomical location; p < 0.001) inside the prosthesis (Fig. 2). After beginning the initial rest period, the skin temperature reached steady state after approximately 25 minutes. At the end of the initial rest period, the skin temperature was 31.0° ± 1.5° C. Once treadmill walking began, the skin temperature rapidly increased. After 30 minutes of treadmill walking at a self-selected speed (0.7 ± 0.3 m/s), the mean skin temperature had increased to 34.1° ± 1.3° C, an increase of 3.1° C (95% CI, 2.4-3.8). After the final 60 minutes of rest, the skin temperature was 33.2° ± 1.2° C, 0.9° C lower (95% CI, 0.5-1.2) than at the end of treadmill walking but 2.2° C higher (95% CI, 1.4-2.9) than the temperature observed at the end of the initial rest period. All pairwise comparisons among activity periods (ie, initial rest, treadmill walking, and final rest) were significant (p < 0.001).

Fig. 2:
Mean (solid line) residual limb skin temperature (± 1 SD, dotted lines) from 16 sites inside the prosthesis was collected during a 150-minute protocol, which began with an initial 60-minute seated rest followed by a 30-minute treadmill walk at self-selected speed and then a final 60-minute seated rest. Three steady-state periods were defined as the last minute of each activity period and are marked with circles.

Temperatures were higher in the tibialis anterior region than other circumferential regions (linear mixed effects model adjusting for activity and proximal-distal sensor position; p < 0.006). The tibial crest temperature was 32.6° ± 2.2° C; the tibialis anterior temperature was 33.3° ± 1.6° C; the medial gastrocnemius temperature was 32.5° ± 1.9° C; and the lateral gastrocnemius temperature was 32.5° ± 1.7° C. The tibialis anterior temperature was higher (p < 0.001) by 0.8° C than the tibial crest (95% CI, 0.4-1.3), medial gastrocnemius (p = 0.002), and lateral gastrocnemius (95% CI, 0.4-1.2) and by 0.9° C than the medial gastrocnemius (95% CI, 0.4-1.3); there were no other pairwise differences observed. This higher temperature can be seen in all three thermal contour plots (initial rest, Fig. 3; treadmill walk, Fig. 4; final rest, Fig. 5).

Fig. 3:
Thermal contour plot shows the mean (n = 9) residual limb temperature (°C, legend on right) while wearing their prosthesis at the end of the initial 60-minute seated rest period.
Fig. 4:
Thermal contour plot shows the mean (n = 9) residual limb temperature (°C, legend on right) while wearing their prosthesis at the end of the 30-minute treadmill walk at self-selected speed.
Fig. 5:
Thermal contour plot shows the mean (n = 9) residual limb temperature (°C, legend on right) while wearing their prosthesis at the end of the final 60-minute seated rest period.

Temperatures were higher in the most proximal regions than the most distal locations (linear mixed effects model adjusting for activity and anatomical location; p = 0.001). Sensor 1, the most proximal position, was 33.1° ± 1.6° C. Progressing distally, Sensor 2 was 33.0° ± 1.7° C, Sensor 3 was 32.7° ± 1.9° C, and Sensor 4, the most distal, was 32.2° ± 2.1° C. Overall, the temperature difference from proximal (Sensor 1) to distal (Sensor 4) was 0.9° C (95% CI, 0.5-1.4) and reflected a linear trend (p < 0.001). The temperature gradient can be seen in all three thermal contour plots (Figs. 3-5), particularly in the lateral gastrocnemius and medial gastrocnemius regions.


Behavioral choices such as what clothes to wear can have a powerful effect on thermal comfort. Unfortunately for lower limb amputees who want to remain ambulatory, simply choosing to wear a prosthesis means accepting some thermal discomfort. A better understanding of the thermal environment at the skin-prosthesis interface will help in designing interventions intended to address this quality-of-life issue. The purpose of this study was to discover the effect of activity and its cessation on residual limb skin temperature and map its variance across the skin-prosthesis interface.

A limitation of this study is the small sample size (n = 9) of transtibial amputees who were moderate community ambulators. Applying these results to the broader population of amputees who have higher or lower mobility, different etiologies, different amputation levels (eg, transfemoral), or comorbidities should be done with caution. Our experimental protocol did not control for medications, antiperspirants, powders, showering before experiments, or diet. Other factors that may also govern the physiological response to heat stress include heat acclimation and exercise training [9] and time of day [1], which were not specifically studied. Additionally, all of this study’s participants were men, and although there is every reason to believe that temperatures would rise with activity in women with transtibial amputations in this setting, the degree to which they may change and the speed with which they might return to baseline values would need to be specifically determined. Finally, this study did not measure the participant’s perception of temperature or thermal comfort. An experiment controlling residual limb skin temperature and reporting outcomes (eg, perception or comfort) would aid in understanding this relationship.

Our observations during the initial rest period confirm that simply donning a prosthesis causes an increase in skin temperature. Earlier work, with a shorter, 15-minute initial rest period (n = 5 transtibial amputees), reported a 0.8° C temperature increase but did not reach steady state [8]. The results of the current study, using an initial rest period four times as long, revealed approximately 25 minutes was needed to reach steady-state conditions but the temperature increase was only approximately 0.6° C. These results suggest that simply donning a prosthesis will cause a slow increase in skin temperature because the prosthesis blocks heat transfer from convective or evaporative processes and conductive heat transfer from blood flow is insufficient to prevent the increase.

Activity causes a rapid increase in skin temperature. A 10-minute treadmill walk resulted in a 0.8° C increase [8], whereas a 30-minute bout caused an increase of 3.1° C. The consequence of longer activity periods is greater thermal discomfort. Amputees capable of longer activity bouts may have greater need for a thermal intervention. Cessation of activity during the final rest period caused a decrease in skin temperature, but although the final rest period was twice the duration of the treadmill walk, the temperature did not return to the initial resting period temperature. These results suggest that after activity, substantially long periods of inactivity would be needed to return the residual limb to a resting state temperature. Such long rest periods might not be practical, further supporting the need for a thermal intervention.

The physiological response to local skin temperature increases includes local cutaneous vasodilation [2] with a graded response between 29° and 40° C [10]. Vasodilation and perspiration are functionally related and often temporally coincident, but although short-term exercise can increase the threshold temperature for vasodilation, it does not affect the threshold for perspiration [2]. For some individuals, the skin temperatures reached in a 30-minute walk during their daily activities might cross these thresholds and result in even greater discomfort caused by the accumulation of perspiration inside an impermeable prosthesis. For others, if their rest periods between activities were of insufficient length, the cumulative effect across multiple activity and rest periods would increase the probability of surpassing the perspiration threshold. A thermal intervention that prevents or reduces increases in skin temperatures might also prevent a perspiration response.

The observations also reveal circumferential and proximal-distal differences. The residual limb temperature was higher at the tibialis anterior location than at other locations circumferentially around the limb. This result was unexpected as a result of the smaller muscle mass at the tibialis anterior location than at the medial gastrocnemius or lateral gastrocnemius locations. Additionally, residual limb skin temperatures decreased 0.9° C from the most proximal to the most distal locations on the residual limb. These regional differences suggest a varying need for heat transfer that depends on location. An intervention intended to address thermal comfort could exploit these differences to reduce power consumption or weight by providing greater capacity at the only sites that require it.

A common complaint of lower extremity amputees is that excessive residual limb skin temperatures reduce their quality of life [4]. The results presented here indicate that walking with a prosthesis causes a dramatic increase in temperature, and subsequent rest fails to return skin temperatures to their preactivity levels. In comparison with previous work [8], longer bouts of activity appear to exacerbate the problem. The development of a cooling intervention might improve thermal comfort and improve prosthesis adherence by preventing or reducing local perspiration, particularly if it can provide appropriate cooling capacity for regions of the residual limb that need it. Knowledge of amputee perception of thermal comfort in response to changes in skin temperature would aid in the development of such a device as would field trials that include reporting of outcome measures (eg, thermal comfort, use, activity levels, environment temperature and humidity).


We thank the Department of Veterans Affairs, Rehabilitation Research and Development Service, for support through Center of Excellence grants A9243C and A4843C (PIs: Bruce J. Sangeorzan MD, and Joseph M. Czerniecki MD). We also thank Jane Shofer MS, for statistical analysis.


1. Aoki, K., Stephens, DP. and Johnson, JM. Diurnal variation in cutaneous vasodilator and vasoconstrictor systems during heat stress. Am J Physiol Regul Integr Comp Physiol. 2001; 281: R591-R595.
2. Charkoudian, N. Skin blood flow in adult human thermoregulation: how it works, when it does not, and why. Mayo Clin Proc. 2003; 78: 603-612. 10.4065/78.5.603
3. Hachisuka, K., Matsushima, Y., Ohmine, S., Shitama, H. and Shinkoda, K. Moisture permeability of the total surface bearing prosthetic socket with a silicone liner: is it superior to the patella-tendon bearing prosthetic socket? J UOEH. 2001; 23: 225-232.
4. Hagberg, K. and Branemark, 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. 10.1080/03093640108726601
5. Hothorn, T., Bretz, F. and Westfall, P. Simultaneous inference in general parametric models. Biom J. 2008; 50: 346-363. 10.1002/bimj.200810425
6. Huff, EA., Ledoux, WR., Berge, JS. and Klute, GK. Measuring residual limb skin temperatures at the skin-prosthesis interface. Journal of Prosthetics and Orthotics. 2008; 20: 170-173. 10.1097/JPO.0b013e3181875b17
7. Klute, GK., Rowe, GI., Mamishev, AV. and Ledoux, WR. The thermal conductivity of prosthetic sockets and liners. Prosthet Orthot Int. 2007; 31: 292-299. 10.1080/03093640601042554
8. Peery, JT., Ledoux, WR. and Klute, GK. Residual-limb skin temperature in transtibial sockets. J Rehabil Res Dev. 2005; 42: 147-154. 10.1682/JRRD.2004.01.0013
9. Roberts, MF., Wenger, CB., Stolwijk, JA. and Nadel, ER. Skin blood flow and sweating changes following exercise training and heat acclimation. J Appl Physiol Respir Environ Exerc Physiol. 1977; 43: 133-137.
10. Stephens, DP., Charkoudian, N., Benevento, JM., Johnson, JM. and Saumet, JL. The influence of topical capsaicin on the local thermal control of skin blood flow in humans. Am J Physiol Regul Integr Comp Physiol. 2001; 281: R894-R901.
© 2014 Lippincott Williams & Wilkins, Inc.