The effects of transcutaneous electrical nerve stimulation on skin temperature in asymptomatic subjects.

Key Words: Skin temperature, Thermography, Transcutaneous electrical nerve

Transcutaneous electrical nerve stimulation (TENS) has been used therapeutically in the management of pain for approximately 20 years. A number of theories about the mechanism of TENS exist. The gate control theory of pain modulation has commonly been used to describe the mechanism of pain relief for high-frequency or "conventional" TENS.[1,2] This mode of TENS was designed to block the transmission of nociceptive afferent fibers in the spinal cord by stimulating large-diameter, myelinated afferent fibers.[3] The endogenous opioid theory of pain modulation[4] is associated with low-frequency or acupuncture-like" TENS. Low-frequency TENS is designed to stimulate A-delta and C afferent fibers that in turn are thought to stimulate the release of endogenous opioids in the central nervous system.[5] The blockade of peripheral afferent fibers has also been postulated to explain the mechanism of TENS, by the suppression of nociceptive transmission in the periphery.[6] The blockade of peripheral nociceptive afferent fibers may be another mechanism through which TENS exerts its effects.

Within the last 10 years, research into the role of the sympathetic nervous system (SNS) in the generation and maintenance of pain has increased.[7] One main function of the SNS in the periphery involves the the thermoregulatory control of blood flow through the skin.[8] Janig[9] has proposed that neural connections exist in the spinal cord between sensory afferent fibers and preganglionic sympathetic fibers. Transcutaneous electrical nerve stimulation may indirectly affect the sympathetic outflow from the spinal cord by stimulating peripheral afferent fibers. High-intensity TENS may also stimulate sympathetic efferents and therefore may directly affect blood flow in the periphery. in an attempt to examine blood flow, some researchers have examined skin temperature, as skin temperature has been shown to be highly correlated with superficial blood flow.[10,11]

Several studies have examined the effect of TENS on skin temperature. Patients with Raynaud's disease, diabetic polyneuropathy, and reflex sympathetic dystrophy have been studied.[12-19] Controlled clinical studies show differing patterns of temperature change in response to TENS.[12-19] Abram et al[12] reported a 2.5[degrees][+ or -]0.7[degrees]C (X [+ or -] SEM) increase in finger or toe skin temperature following TENS. No clinically important changes in skin temperature, however, were reported by either Mulder et al[13] or Ebersold et al[14] following TENS. Most of the research in the area has been performed by Kaada and colleagues.[15-19] These investigators reported increases in finger skin temperatures from 1.9[degrees]C to 13[degrees]C. An important factor, unique to the work by Kaada and co-workers, was the use of a 2-hour stabilization period before TENS. This time was used to allow the skin temperature of the extremities to reach a stable level, and to allow any painful, ischemic symptoms to appear.

Due to the disparate results that have been found within the area of TENS and skin temperature in subjects, a number of researchers have evaluated the effects of TENS on skin temperature in asymptomatic subjects.[18,20-26] The overall intent of these studies has been to understand what a "normal" physiological response would be to TENS. The knowledge of the physiological changes that are produced in asymptomatic subjects could then be used to compare the effects that TENS has on patients. Even with asymptomatic subjects, however, the results are disparate. Transcutaneous electrical nerve stimulation has been reported to increase skin temperature,[18,20-22] decrease skin temperature,[23,24] or even produce no change in skin temperature.[22,25,26] For example, in a recent study, Indergand and Morgan[27] found a decrease in blood flow and skin temperature during high-frequency TENS applied at intensities above and below motor threshold. The authors discussed, however, that the decrease continued after the stimulation had ceased and concluded that the TENS itself had no effect on blood flow or skin temperature in the area innervated by the stimulated nerves.

Many factors make comparisons among these studies very difficult. First, the studies used only one temperature measurement technique, either a contact probe or infrared thermography. The use of both techniques simultaneously allows for a comparison between measurements attained from two different sources. Frequently, the results of studies are compared when two skin temperature measurement techniques are used. The direct comparison of a measurement obtained from a very small area and a large area may be inappropriate. Second, the TENS modes used and the electrode placements were not consistent across studies. Third, some studies did not incorporate a control group.[22,24] Even with the exclusion of the uncontrolled studies, disparate findings remain. The stabilization periods in these studies, allocated to allow the temperature of the skin to reach a stable level, ranged across studies from 5 minutes to 120 minutes. This period before TENS may be important in achieving accurate results in experimental trials. Lastly, most studies do not mention the events that may occur prior to TENS. Heavy meals, strenuous exercise, and smoking alter the peripheral circulation and, therefore, affect the temperature of die skin.[11,28]

The effects of TENS on the skin temperature of asymptomatic subjects have not yet been established. No pattern has been found to account for the differences reported in the literature. In our study, we attempted to control for many factors that may affect the accurate measurement of skin temperature in response to TENS. The primary objective of our study was to examine the skin temperature on the dorsum of the hand and of the proximal phalanx of the second finger under three conditions: high-frequency TENS, low-frequency TENS, and a no-stimulation control. The secondary objective was to determine whether the location at which temperature measurements were taken resulted in different conclusions.

Method

Subjects

Twenty-four asymptomatic subjects (23 female, 1 male) who had no previous experience with TENS provided informed consent to participate in the study. They ranged in age from 19 to 28 years (X=23.0, SD=2.44). The sample was one of convenience; all of the subjects were students or staff at the University of Western Ontario, London, Ontario, Canada. Two hours before each study session, the subjects avoided eating a heavy meal, smoking, ingesting caffeine, or exercising. No analgesic medication or medication affecting circulation was taken within 24 hours before a study session. The time between sessions for each subject was 24 hours ([+ or -] 1 hour).

Design

An experimental, repeated-measures, controlled design was used. All subjects participated in each of the three study conditions (high-frequency TENS, low-frequency TENS, control). The order in which the conditions were assigned was randomized using a 3X3 Latin square design.[29] The mean skin temperature of the dorsum of the left hand and the skin temperature of the proximal phalanx of the second finger of the left hand were the primary outcome measures. Several investigators who have examined the effects of TENS on skin temperature reported either the hand temperature or the finger temperature. The comparison of studies that measure hand temperature with studies that measure finger temperature may be inappropriate. Transcutaneous electrical nerve stimulation may have differential effects on the hand and finger temperature. Therefore, the skin temperature of both the hand and the finger was measured in this study to afford the comparison of the temperatures at the two sites.

Instrumentation

One channel of a Medtronic+[R] dual-channel, constant-current TENS machine(*) was used to deliver the high-frequency TENS (100 Hz, 100-microsecond pulse width) and the low-frequency TENS (4 Hz, 250-microsecond pulse width) according to specifications supplied by the manufacturer. Two Medtronic SnapEase reusable, self-adhesive electrodes(*) were positioned over the web space between the first and second metacarpals and over the ulnar border of the left hand of each subject. The temperature of the dorsum of the left hand was measured using a Thermovision[R] 450 infrared thermography camera.(dagger) The camera has a sensitivity of 0.1[degree]C at 30[degrees]C and an accuracy of [+ or -] 2%. The finger temperature was measured using a YSI 2100 telethermometer(double dagger) and a YSI series 400 contact probe.(double dagger) The system has a sensitivity of 0.1[degree]C and an accuracy of 0.3[degree]C.

Procedure

The positioning of the subject and the equipment was standardized for all study sessions. The subject was seated in a comfortable chair, with the left arm abducted to approximately 45 degrees and the elbow flexed to approximately 75 degrees. The forearm was positioned in pronation, with the wrist and fingers in a neutral position. A pillow was used to support the forearm, hand, and fingers (Fig. 1). A piece of rigid plastic was positioned between the pillowcase and the pillow to prevent the hand from sinking into the pillow. The TENS electrodes and the skin probe were then positioned. The probe was fastened to the dorsolateral aspect of the proximal phalanx of the second digit, using two small pieces of Micropore tape.(sections) The infrared camera was positioned over the hand, perpendicular to the hand and fingers.

After positioning, the subject sat for an initial 60-minute stabilization period. Following the stabilization period, the 30-minute stimulation period began. Thirty minutes of stimulation was used because this time period was analogous to a typical treatment period. During this period, the subject received high-frequency TENS, low-frequency TENS, or no stimulation (ie, the control condition). The intensity of the high-frequency TENS was intended to produce a strong, but comfortable, sensation, with no muscle contraction. The intensity of the low-frequency TENS was intended to produce an uncomfortable, but tolerable, sensation, with rhythmic muscle contractions. When the perception of the intensity of the stimulation fell below these intended levels, the subject was instructed to advise the investigator, so that the intensity could be increased to the desired levels. The TENS machine was not turned on during the control condition.

We did not seek to assess the effects of thoughts, mood state, or mental imagery on skin temperature. No instructions, therefore, were given to the subjects with regard to thoughts or feeling during the study period.

Following the stimulation period, the subject continued to sit for an additional 30 minutes with the equipment in place. Hand and finger temperature readings were recorded at baseline (0 minutes), after 30 minutes of stabilization, after 60 minutes of stabilization (prestimulation), at 90 minutes (post-stimulation), and at 120 minutes (30 minutes after the end of the stimulation).

Data Analysis

A two-factor, repeated-measures analysis of variance (ANOVA) was used to analyze the data using the SPSS/PC+ (version 5.0) software program.(vertical bars) The hand temperature and the finger temperature data were analyzed separately. An alpha level of .05 was used as the critical value at or below which the results were accepted as statistically significant. One-way ANOVAs and post hoc tests (Student-Newman-Keuls) were performed to determine the difference among the three conditions at three times (60 minutes, 90 minutes, 120 minutes) and to compare the effects for title for each condition.[30]

Results

The data for the hand and the finger temperatures are summarized in Figures 2 and 3. The ambient temperature of the study room for all of the sessions ranged from 22.0[degrees]C to 23.0[degrees]C (X=22.54[degrees]C, SD=0.23[degrees]C).

Hand Temperature

The repeated-measures ANOVA for the hand temperature data showed an effect for time (F=86.32; df=2,138; P<.001) and a condition X time interaction F=5.14; df=4,138; P=.001). To analyze the condition X time interaction, the data were graphed (Fig. 2) and one-way ANOVAs were performed to examine the three conditions at each time. No difference was found among the three conditions at 60 minutes (F=0.51; df=2,69; P=.602) or at 120 minutes (F=1.12; df=-2,69; P=.337). At 90 minutes, however, the mean hand temperatures for the three conditions were different (F=4.64; df= 2,69; P=.013). Post hoc analyses showed that at 90 minutes, the mean hand temperature was higher after the low-frequency TENS than after the high-frequency TENS and the control condition (P<.05). No difference in hand temperature was found between the high-frequency TENS and control conditions at any time (P>.05).

Hand Temperature-Effect for

Time

The one-way ANOVAS showed differences in hand temperature over time for the low-frequency TENS condition (F=5.08; df=2,69; P=.009) and the high-frequency TENS condition (F=4.61; df=2,69; P=.013). The results of the post hoc analyses are displayed in Figure 2. The one-way ANOVA for time for the control condition came very close to showing significant results (F=3.08; df=2,69; P=.052); therefore, post hoc analyses were performed for the control condition (Fig. 2).

Finger Temperature

The two-factor, repeated-measures ANOVA for the finger temperature data showed an effect for time (F=81.02; df=2,138; P<.001; Fig. 3). One-way ANOVAs showed differences in finger temperature over time for the low-frequency TENS condition (F=4.34; df=2,69; P=.017) and the high-frequency TENS condition (F=4.20; df=2,69; P=.019). The results of the post hoc analyses are displayed in Figure 3. For the control condition, no differences were found among the three times (F=2.39; df=2,69; P=.11).

Stabilization Period

No differences in hand temperature among the three conditions were found at the beginning of the stabilization period (F=1.08; df=2,69; P=.35) or at 30 minutes (F=1.56; df=2,69; P=.22). In our study, a stabilization period of 60 minutes was used to assess the change in the skin temperature of the hand and finger. For each of the three study conditions, the mean hand temperatures for each condition were lower at baseline than at 30 minutes of stabilization (Fig. 2). The mean finger temperatures at baseline were lower than the mean finger temperatures at 30 minutes and at 60 minutes (Fig. 3). No difference was found between 30 minutes and 60 minutes for any of the conditions. Similarly, no differences in finger temperature among the three conditions were found at the start of the stabilization period (F=.92; df=2,69; P=.40) or at 30 minutes (F= 1.01; df=2,69; P=.37). No differences were found between 30 and 60 minutes for the control condition. There was, however, a monotonic or steady decrease in the temperature of the hand and of the finger following 60 minutes. A drop of 1.68[degrees]C and 1.94[degrees]C, for the hand and the finger, respectively, occurred between 60 minutes and the end of the study session. This decrease in temperature may represent the normal physiological cooling that occurred at rest after the 60-minute stabilization period.[25] As a result, the effects of the two TENS conditions on skin temperature could be assessed in comparison with the normal physiological effect.

Discussion

The results of our study showed that low-frequency TENS, applied at an uncomfortable level, produced an effect on hand temperature that was different from that of the high-frequency TENS and control conditions. The mean hand temperature following low-frequency TENS was wanner than the hand temperature after the high-frequency TENS or the control period. This elevation of temperature immediately began to reverse once die stimulation stopped. The hand temperature of the control group displayed a monotonic cooling. A similar pattern of cooling of the hand was found during the high-frequency TENS condition. The finger temperature was not affected by either mode of TENS when compared with the control condition. During all three conditions, the finger temperature decreased in a similar manner during both the stimulation period and the poststimulation period.

The monotonic decrease in skin temperature, or physiological cooling over time, may have occurred for two reasons. The room temperature was approximately 22.5[degrees]C. Both finger and hand temperatures were considerably higher than the room temperature throughout the study period. Therefore, it could reasonably be expected that skin temperature would fall due to temperature differences between the exposed skin and the surrounding environment. As well, during the study session, the arm was at rest, producing no voluntary muscle action. Consequently, the blood flow to the hand and fingers could decrease, which might also lower the skin temperature.

Several explanations may account for the effect of low-frequency TENS on skin temperature of the hand. Low-frequency TENS may have produced an increase in blood flow to the stimulated muscles. The increased temperature of the skin may reflect the increase in blood flow. Some studies[18,22,23] have assessed the circulatory effects of low-frequency, or burst, TENS, which produce local muscle contractions. Muscle blood flow was not directly measured in these studies, and skin temperature measurements were obtained from areas distant from the area of the stimulated muscles. Therefore, the relationship between muscle blood flow and skin blood flow could not be established.

Investigators[31,32] have evaluated the circulatory effects of electrical current designed for muscle stimulation, but not current designed specifically for pain modulation. Currier et al[31] reported an increase in blood flow through the popliteal artery following stimulation of the soleus muscle. They did not, however, find an increase in skin temperature. Tracy et al[32] reported that quadriceps femoris muscle stimulation at frequencies of 10, 20, and 50 Hz, but not at 1 Hz, produced an increase in blood flow through the femoral artery. They did not, however, determine the skin temperature of the extremity. Little scientific evidence, therefore, is available to conclude that the electrically stimulated muscles produce warming of the skin overlying the stimulated muscle.

Another possible explanation for the higher hand temperature immediately following the low-frequency TENS condition is the higher current intensity used with this condition. The low-frequency TENS was delivered at an uncomfortable level, which stimulates A-delta and C afferent fibers.[5] Local vasodilatation has been shown to occur following both prolonged intraneural and transcutaneous stimulation applied at intensities sufficient to stimulate an axon response.[21,33] The low-frequency TENS may have produced a local axon reflex[34] in the area around the electrodes by antidromically stimulating C afferent fibers. This local vasodilatation during the low-frequency TENS produced a warming effect that may have counterbalanced the normal physiological cooling in the area between the two electrodes. This axon response may not have occurred in the finger in our experiment, because the finger temperature readings were taken at a site farther from the electrodes.

A third possible explanation is that the low-frequency, high-intensity TENS produced a segmental inhibition of the SNS. Research has recently been performed to determine the role of the SNS and the peripheral afferent system in the establishment and maintenance of chronic pain.[7] Cutaneous vasoconstriction is caused by an increase in the activity of postganglionic sympathetic nerve fibers and a subsequent release of norepinephrine from their terminals.[9,35] Norepinephrine has also been implicated in the maintenance of peripheral pain.[9] The low-frequency TENS may have produced a vasodilatation following the prolonged stimulation of nociceptive afferents in the hand and the inhibition of the cutaneous sympathetic efferent fibers. Transcutaneous electrical nerve stimulation may play a role in the reduction of sympathetically related pain by its possible effects on the SNS.[15]

Several investigators[36-38] have reported a normal physiological cooling of the skin of the extremities, that ranged from 0.80[degrees]C to 5.00[degrees]C, during a period of rest. The data from the control group in our study further support this idea. An important factor in the divergent results among the various studies may be the skin temperature prior to electrical stimulation.[39] Kaada et al[18] reported a 1.9[degrees]C increase in the temperature of the second digit of asymptomatic subjects following burst-mode TENS. This finding disagrees with our results. We found no change m the finger temperature following low-frequency TENS. The stabilization period in the study by Kaada et al was 120 minutes, at which point the prestimulation temperature of the fingers was approximately 27.5[degrees]C. In comparison, the prestimulation finger temperature in our study was 29.63[degrees]C. Kaada[15] and Kaada and Helle[19] also examined the effects of low-frequency TENS on skin temperature of the second finger in patients with Raynaud's disease and diabetic neuropathy. These two studies demonstrated a 90[degrees]C to 10[degrees]C increase and a mean increase of 8.5[degrees]C in finger temperature, respectively. The prestimulation finger temperatures of the subjects in Kaada's study[15] ranged from 22[degrees]C to 24[degrees]C. The larger change in skin temperature found in Kaada's study may have been due to both the extremely long stabilization period, which resulted in very cool starting skin temperatures, and the pathology present.

Wong and Jette[23] performed a well-controlled study to evaluate the effects of three different modes of TENS on skin temperature. They reported a decrease in finger temperature following three different modes of TENS compared with a placebo condition. A number of important differences exist between the results of Wong and Jette's study and the results of our study. Following the stimulation period or all three conditions in our study, the temperature of the finger continued to decrease. In contrast, the finger temperature for the three experimental groups in Wong and Jette's study showed a rebound warming after the stimulation ended. One difference between our study and that of Wong and Jette may be that the pre-stimulation skin temperatures were different between the two studies. Wong and Jette, however, did not report prestimulation skin temperatures. Very high or very low baseline skin temperatures might influence the rate and the extent of temperature change and thus have an effect on the results of TENS. Other differences between our study and that of Wong and Jette are the intensity of the TENS used and the placement of the electrodes. The intensity of the stimulation may be an important factor m the physiological processes that are activated, and consequently the effects that different modes of TENS have on skin temperature. More thorough and controlled investigations in these areas are required to better understand the effects that TENS has on skin temperature. This may be very important, clinically, in terms of which mode of TENS is used to treat various clinical symptoms such as inflammation and ischemic pain.

The data from the stabilization period of our study showed an initial rise in the temperature of the hand and the finger. This increase peaked at approximately 30 minutes and then slowly decreased throughout the remaining 90 minutes of the study. One possible explanation for the critical rise in skin temperature may be related to the ambient temperature outside of the study room. The ambient temperature from which most of the subjects came before the study session was cooler than the air temperature of the study room. Consequently, upon entering the warmer study room, vasodilatation in the skin occurred in reaction to the warmer room temperature.

Clinical implications

Transcutaneous electrical nerve stimulation has been used by physical therapists for many years for the management of acute and chronic pain. The main mechanisms by which TENS is thought to be working to produce pain relief are twofold: the gate control theory or some related mechanism, and the endogenous opioid theory. When compared with the control condition, the low-frequency TENS may have increased the skin temperature through local or segmental spinal reflex actions. Preganglionic sympathetic fibers in the spinal cord may have been inhibited by stimulated nociceptive afferents. A local axon reflex response to the high, uncomfortable intensity of the low-frequency TENS could also explain the warmer skin temperature of the hand compared with the high-frequency TENS and the control condition.

Based on the results of our study, low-frequency TENS applied at an uncomfortable intensity may prevent a local decrease in blood flow, as measured by skin temperature. We believe the magnitude of the difference between the low-frequency TENS condition and the control condition (1.60[degrees]C) is clinically important.[40-42] In asymptomatic subjects, Uematsu et al[40] showed that a temperature difference greater than 0.31[degrees]C ([+ or -] 0.25[degrees]C) in the hand, when compared with the opposite side, could be considered abnormal. Meeker and Gahlinger[41] concluded that a difference of at least 1[degree]C covering at least 25% of the area of the dermatome under consideration (ie, area of the skin that is supplied by cutaneous branches from a single spinal nerve) should be considered an abnormal finding suggestive of underlying pathology. Therefore, the effects produced by low-frequency TENS in our study may be of sufficient magnitude to have an effect on pathological conditions. The cooling that occurs with certain conditions such as reflex sympathetic dystrophy may be amenable to change through low-frequency TENS.

There are a number of limitations to our study. Our sample consisted mainly of young women. Consequently, the results should only be generalized to people who possess similar characteristics to the sample. Also, due to the rigid experimental control imposed on this study, the generation of the results should be restricted to similar situations, and not directly to clinical settings where differences exist. The major limitation to the study, however, lies in the use of an asymptomatic sample as opposed to subjects with relevant pathologies. Rothstein[43] commented that the role of researchers is not only to perform clinical research, but also to be a source of basic knowledge in physical therapy. This basic knowledge should include an understanding of the physiological mechanisms by which physical therapy modalities produce their effects. By testing models in asymptomatic subjects, where many of the experimental variables can be manipulated, some basic understandings can be obtained without risk to patients. A thorough knowledge of the interaction between frequencies, intensities, and durations of stimuli in asymptomatic subjects should afford help in understanding the effects of various variables.

Future studies should explore the effects of low-frequency and high-frequency TENS on the temperature of the skin at different starting skin temperatures. The relationship between skin temperature change and change in pain perception in asymptomatic subjects should be determined. The relationship between changes in skin temperature and perceived levels of pain in subjects who exhibit cold, painful limbs should also be examined. Finally, the potential effect of TENS in preventing ischemic pain in relevant patients should be investigated.

Conclusions

In young asymptomatic women, the mean temperature of the stimulated hand was warmer following 30 minutes of low-frequency TENS, applied at an uncomfortable intensity level, than the mean temperature of the hand following a similar period of high-frequency TENS or no stimulation. The low-frequency TENS produced no effect on the temperature of the finger of the stimulated hand. High-frequency TENS, however, produced no effect on either the hand temperature or the finger temperature of the stimulated hand.

(*) Medtronic Nortech Division, San Diego, CA 92121. (dagger) Agema Infrared Systems AB, Danderyd, Sweden. (double dagger) Yellow Springs Instrument Co, Yellow Springs, OH 45387. (sections) 3M Canada Inc, London, Ontario, Canada N6A 4T1. (vertical lines) SPSS Inc, Chicago, IL 60611.

Acknowledgment

We thank Medtronic of Canada for supplying the TENS machine and the TENS electrodes that were used in this study.

RJ Scudds, PT, was a student in the Department of Physical Therapy, University of Western Ontario, when this study was conducted in partial fulfillment of the requirements for the degree of Master of Science. She is currently a PhD candidate in the Department of Epidemiology and Biostatistics, University of Western Ontario. Address all correspondence to Ms Scudds at Department of Epidemiology and Biostatistics, Kresge Building, University of Western Ontario, London, Ontario, Canada N6A 5C1 ([e-mail]).

A Helewa, PT, is Professor and Chair, Department of Physical Therapy, Elborn College, University of Western Ontario, London, Ontario, Canada N6G 1H1.

RA Scudds, PhD, PT, is Assistant Professor, Department of Physical Therapy, Elborn College, University of Western Ontario.

This study was approved by the University of Western Ontario's Review Board for Health Science Research Involving Human Subjects.

The results of this research project were presented at the Fifth World Congress on Pain; August 18-25, 1993; Paris, France.

This article was submitted March 29, 1994, and was accepted February 24, 1995.

References

[1] Melzack R, Wall PD. Pain mechanisms: a new theory. Science. 1965;10:971-979. [2] Wall PD. The gate control theory of pain mechanisms: a re-examination and restatement. Brain. 1978;101:1-18. [3] Wall PD. The discovery of transcutaneous electrical nerve stimulation. Physiotherapy Canada. 1985;71:348-350. [4] Basbaum AI, Fields HL. Endogenous pain control: review and hypothesis. Ann Neurol. 1978;4:451-462. [5] Sjolund BH, Eriksson MBE. Endorphins and analgesia produced by peripheral conditioning stimulation. In: Bonica JJ, ed. Advances in Pain Research and Therapy, Volume 3. New York, NY: Raven Press; 1979:587-592. [6] Campbell JN, Taub A. Local analgesia from percutaneous electrical stimulation: a peripheral mechanism. Electrical Analgesia. 1973;28: 347-350. [7] Janig W, Blumberg H, Boas RA, Campbell JN. The reflex sympathetic dystrophy syndrome: consensus statement and general recommendations for diagnosis and clinical research. In: Bond MR, Charlton JE, Woolf CJ, eds. Proceedings of the VIth World Congress on Pain, 1990, Adelaide, South Australia. Amsterdam, the Netherlands: Elsevier Science Publishers BV; 1991:373-376. [8] Brooks GA, Fahey TD. The circulation and its control. In: Brooks GA, Fahey TD, eds. Exercise Physiology v. Human Bioenergetics and Its Applications. New York, NY: Macmillan Publishing Co; 1985:295-311. [9] Janig W. The sympathetic nervous system in pain: physiology and pathophysiology. In: Stanton-Hicks M, ed. Pain and the Sympathetic Nervous System. Boston, Mass: Kluwer Academic Publishers; 1990:17-89. [10] Stoner HB, Barker P, Riding GSG, et al. Relationships between skin temperature and perfusion in the arm and leg. Clin Physiol. 1991;11:27-40. [11] Bornmyr S, Svensson H. Thermography and laser-doppler flowmetry for monitoring changes in finger skin blood flow upon cigarette smoking. Clin Physiol. 1991;11:135-141. [12] Abram SE, Asiddao CB, Reynolds AC. Increased skin temperature during transcutaneous electrical stimulation. Anesth Analg. 1980; 59:2-25. [13] Mulder P, Dompeling EC, van Slochterenvan der Boor JC, Kuipers WD, Smit AJ. Transcutaneous electrical nerve stimulation (TENS) in Raynaud's phenomenon. Angiology. 1991; 42:414-417. [14] Ebersold MJ, Laws ER, Albers JW. Measurements of autonomic function before, during, and after transcutaneous stimulation in patients with chronic pain and in control subjects. Mayo Clin Proc. 1977;52:228-232. [15] Kaada B. Vasodilation induced by transcutaneous nerve stimulation in peripheral ischemia (Raynaud's phenomenon and diabetic polyneuropathy). Eur Heart J. 1982;3:303-314. [16] Kaada B, Eielsen O. In search of mediators of skin vasodilation induced by transcutaneous nerve stimulation, I: failure to block the response by antagonists of endogenous vasodilators. Gen Pharmacol. 1983;14:623-633. [17] Kaada B, Eielsen O. In search of mediators of skin vasodilation induced by transcutaneous nerve stimulation, II: serotonin implicated. Gen Pharmacol. 1983;14:635-641. [18] Kaada B, Olsen E, Eielsen O. In search of mediators of skin vasodilation induced by transcutaneous nerve stimulation, III: increase in plasma VIP in normal subjects and in Raynaud's disease. Gen Pharmacol. 1984;15: 107-113. [19] Kaada B, Helle KB. In search of mediators of skin vasodilation induced by transcutaneous nerve stimulation, IV: in vitro bioassay of the vasoinhibitory activity of sera from patients suffering from peripheral ischemia. Gen Pharmacol. 1984;15:115-122. [20] Owens S, Atkinson ER, Lees DE. Thermographic evidence of reduced sympathetic tone with transcutaneous nerve stimulation. Anesthesiology. 9;50:62-65. [21] Leandri M, Brunetti O, Parodi CI. Telethermographic findings after transcutaneous electrical nerve stimulation. Phys Ther. 1986;66: 210-213. [22] Nolan MF, Hartsfield JK, Witters DM, Wason PJ. Failure of transcutaneous electrical nerve stimulation in the conventional and burst modes to alter digital skin temperature. Arch Phys Med Rehabil. 1993;74:182-187. [23] Wong RA, Jette DU. Changes in sympathetic tone associated with different forms of transcutaneous electrical nerve stimulation in healthy subjects. Phys Ther. 1984;64:478-482. [24] Casale R, Gibellini R, Bozzi M, Bonelli, S. Changes in sympathetic activity during high-frequency TENS. Acupunct Electrother Res. 1985;10:169-175. [25] Ernst M, Gracely RH, Lee MHM. Influence of TENS on autonomic function. Acupunct Electrother Res. 1987; 12:257. Abstract. [26] Johnston MI, Hajela VK, Ashton CH, Thompson JW. The effects of auricular transcutaneous nerve stimulation (TENS) on experimental pain threshold and autonomic function in healthy subjects. Pain. 1991;46:337-342. [27] Indergand HJ, Morgan BJ. Effects of high-frequency transcutaneous electrical nerve stimulation on limb blood flow in healthy humans. Phys Ther. 1994;74:361-367. [28] Meis PJ, Hulcher FH, Rose JC, et al. Meal timing alters circadian rhythm of skin surface temperature in non-human primates. International Journal of Chronobiology. 1980;7:65-76. [29] Winer BJ. Statistical Principles in Experimental Design. 2nd ed. New York, NY: McGraw-Hill Book Co: 1971. [30] Witte RS. Statistics. 4th ed. Fort Worth, Tex: Harcourt Brace Jovanovich College Publishers; 1993. [31] Currier DP, Petrilli CR, Threlkeld AJ. Effect of graded electrical stimulation on blood flow to healthy muscle. Phys Ther. 1986;66:937-943. [32] Tracy JE, Currier DP, Threlkeld AJ. Comparison of selected pulse frequencies from two different electrical stimulators on blood flow in healthy subjects. Phys Ther. 1988;68: 1526-1532. [33] Blumberg H, Wallin BG. Direct evidence of neurally mediated vasodilation in hairy skin of the human foot. J Physiol (Lond). 1987; 382: 105-121. [34] Cuello AC, Matthews MR. Peptides in peripheral sensory nerve fibres. In: Wall PD, Melzack R, eds. Textbook of Pain. New York, NY: Churchill Livingstone Inc; 1984:65-79. [35] Shepherd JT, Vanhoutte PM. Local modulation of adrenergic neurotransmission. Circulation. 1981;64:655-666. [36] Barcroft H, Edholm OG. Temperature and blood flow in the human forearm. J Physiol. 1946;104:366-376. [37] Guadagni DN, Kreith F, Smyth CJ, Bartholomew BA. Contact probe for skin temperature measurements. Journal of Physics E: Scientific Instruments. 1972;5:869-876. [38] Arsenault AB, Belanger AY, Durand M-J. Effects of TENS and topical skin anesthesia on soleus H-reflex and the concomitant influence of skin/muscle temperature. Arch Phys Med Rehabil. 1993;74:48-53. [39] Skoglund CR. Vasodilatation in human skin induced by low-amplitude high-frequency vibration. Clin Physiol. 1989;9:361-372. [40] Uematsu S, Edwin DH, Janakel WR, et al. Quantification of thermal asymmetry, part 1: normal values and reproducibility. J Neurosurg. 1988;69:552-555. [41] Meeker WC, Gahlinger PM. Neuromuscular thermography: a valuable diagnostic tool? J Manipulative Physiol Ther. 1986;9:257-266. [42] Ring EFJ. Quantitative thermal imaging. Clinics in Physics and Physiological Measurements. 1990;11(suppl A):87-95. [43] Rothstein JM. Editor's note. Physical therapy research: academic inadequacies. Phys Ther. 1992;72:92-94.