Abstract
Background Transcutaneous electrical nerve stimulation (TENS) is commonly used for the management of pain; however, its effects on several pain and function measures are unclear.
Objective The purpose of this study was to determine the effects of high-frequency TENS (HF-TENS) and low-frequency TENS (LF-TENS) on several outcome measures (pain at rest, movement-evoked pain, and pain sensitivity) in people with knee osteoarthritis.
Design The study was a double-blind, randomized clinical trial.
Setting The setting was a tertiary care center.
Participants Seventy-five participants with knee osteoarthritis (29 men and 46 women; 31–94 years of age) were assessed.
Intervention Participants were randomly assigned to receive HF-TENS (100 Hz) (n=25), LF-TENS (4 Hz) (n=25), or placebo TENS (n=25) (pulse duration=100 microseconds; intensity=10% below motor threshold).
Measurements The following measures were assessed before and after a single TENS treatment: cutaneous mechanical pain threshold, pressure pain threshold (PPT), heat pain threshold, heat temporal summation, Timed “Up & Go” Test (TUG), and pain intensity at rest and during the TUG. A linear mixed-model analysis of variance was used to compare differences before and after TENS and among groups (HF-TENS, LF-TENS, and placebo TENS).
Results Compared with placebo TENS, HF-TENS and LF-TENS increased PPT at the knee; HF-TENS also increased PPT over the tibialis anterior muscle. There was no effect on the cutaneous mechanical pain threshold, heat pain threshold, or heat temporal summation. Pain at rest and during the TUG was significantly reduced by HF-TENS, LF-TENS, and placebo TENS.
Limitations This study tested only a single TENS treatment.
Conclusions Both HF-TENS and LF-TENS increased PPT in people with knee osteoarthritis; placebo TENS had no significant effect on PPT. Cutaneous pain measures were unaffected by TENS. Subjective pain ratings at rest and during movement were similarly reduced by active TENS and placebo TENS, suggesting a strong placebo component of the effect of TENS.
Transcutaneous electrical nerve stimulation (TENS) is an inexpensive, noninvasive intervention used to manage a wide variety of painful conditions. Previous studies showed that TENS increases pressure and heat pain thresholds in people who are healthy1–6 and reduces mechanical and heat hyperalgesia in arthritic animals.7,8 However, a recent systematic review showed that TENS was not effective for knee osteoarthritis (OA) pain,9 in direct contrast to an earlier systematic review that concluded that TENS was effective for knee OA pain10 and a meta-analysis that showed a significant reduction in knee OA pain with TENS.11 Several limitations in the included trials may explain the lack of effect of TENS; these include small sample size, poor methodological quality, and inadequate randomization and blinding.
A key factor that may explain the lack of effect of TENS is that pain at rest was the main outcome routinely examined in earlier studies.9 However, TENS has a greater effect on movement-evoked pain and subsequently results in improved function.12 Evoked stimuli are used to measure painlike behaviors in animal studies and commonly used to examine responses at the site of injury (ie, primary hyperalgesia) and outside the site of injury (ie, secondary hyperalgesia). Primary hyperalgesia and secondary hyperalgesia are surrogate measures for sensitization of the peripheral nervous system (at site of injury) and central nervous system (outside site of injury), respectively. It is possible that the effectiveness of TENS varies by outcome.
Dosing also is critical to the effectiveness of TENS.12–15 Animal studies show that both high-frequency TENS (HF-TENS, >50 Hz) and low-frequency TENS (LF-TENS, <10 Hz), delivered at an intensity of 90% of the motor threshold (strong sensory intensity), reduced pain sensitivity in arthritic animals.16–20 Several studies of HF-TENS in people who are healthy show that higher intensities, generally described as strong but comfortable, resulted in greater pain reduction.2,6,13,14,21,22 Similarly, studies of postoperative pain show that adequate intensities are necessary to produce analgesia and that low intensities are ineffective.12,15 It is unclear whether the frequencies and intensities used in animal studies will have similar effects in people.
The purposes of this study were: (1) to determine the efficacy of HF-TENS and LF-TENS for knee OA pain and (2) to determine which outcome measures (pain at rest, movement-evoked pain, pain sensitivity, and function) are most likely to be affected by HF-TENS and LF-TENS in people with pain to inform the design of future studies. We compared the effect of TENS, applied with parameters used in our earlier animal studies,13,14,16,22–28 with the effect of a new placebo TENS on pain at rest, movement-evoked pain, pain sensitivity measures, and function in people with OA. We hypothesized that both HF-TENS and LF-TENS would reduce pain during movement but not pain at rest, decrease pain sensitivity, and increase function.
Method
Design Overview
This study was a double-blind, randomized clinical trial that included 75 people who had knee OA and were randomly allocated to 1 of 3 groups (HF-TENS, LF-TENS, and placebo TENS). Outcome measurements were obtained before and during a single TENS treatment.
Setting and Participants
Study participants were recruited through flyers and active screening by experimenters in the orthopedic and sports medicine department of a large midwestern tertiary care center. Inclusion criteria were as follows: diagnosis of medial-compartment knee OA (radiographically and symptomatically diagnosed by an orthopedic surgeon), 18 to 95 years of age, able to ambulate to mailbox and back, stable medication schedule for 3 weeks before testing, and pain rating of greater than 3 of 10 during weight bearing (on a verbal rating scale). Lateral-compartment knee OA was excluded to standardize the test sites for pain sensitivity measures. A pain rating of less than 3 of 10 was needed to derive a clinically meaningful change attributable to the intervention.
Initial screening was performed by the recruiter, or telephone screening was conducted by the project coordinator to determine inclusion. Exclusion criteria were as follows: uncontrolled diabetes mellitus or hypertension, dementia or cognitive impairment, neurological disorder, permanent lower-extremity sensory loss, earlier TENS use, knee surgery in last 6 months, or knee injection in last 4 weeks. To avoid interactions with medications, participants were given instructions not to take any analgesic medication 4 hours before the testing session. At the first visit, we obtained a complete list of medications currently being taken and assessed bilateral recognition of sharp versus dull pressure at the L3–S2 dermatomes and proprioception of the great toe to rule out loss of sensation (an exclusion criterion). All testing was conducted in 1 of 2 dedicated research spaces at The University of Iowa.
Figure 1 shows the CONSORT diagram for this randomized controlled trial. As illustrated, 311 people were assessed for eligibility, and 87 declined to participate. We were unable to make contact with 31 people, and 116 were excluded. Seventy-five of the remaining 77 people were allocated to treatment groups and completed the testing. Two people were excluded in the secondary screening process because of reduced sensation and lateral joint pain.
CONSORT diagram. Most people were excluded from the initial screening because of earlier transcutaneous electrical nerve stimulation (TENS) use or minimal pain. However, all participants allocated to a group completed the study. CNS=central nervous system, CVA=cerebrovascular accident, OA=osteoarthritis
Randomization and Intervention
An allocation concealment protocol with sequentially numbered, opaque, sealed envelopes and permuted blocks of 3 and 6 was used to randomize participants to groups.29 Allocation envelopes were kept in a location separate from the testing location and were not available to the data collection examiner. The foil-lined envelopes were signed, dated, and opened by the allocation examiner immediately before TENS application and after the data collection examiner left the room.
A commercially available TENS unit (Rehabilicare Maxima, DJO Inc, Vista, California) was used to deliver TENS. The unit uses an asymmetrical, biphasic waveform, the pulse duration was set at 100 microseconds, and the intensity was 10% below the motor threshold. We chose to modulate frequency and keep all other parameters the same to test whether there is a frequency-dependent effect on outcomes. The same parameters were previously used in preclinical animal and human studies of TENS in our laboratory.13,14,16,23–28 The placebo TENS unit (DJO Inc) was identical in appearance to the TENS unit, and placebo TENS was applied in the same manner as active TENS (100 Hz, 100-microsecond pulse duration, and intensity 10% below motor threshold). The new transient placebo TENS unit delivered a current for the first 30 seconds and then ramped down to zero over 15 seconds. The transient placebo is valid and has no effect on pain measures in healthy controls.13 Neither the TENS allocation examiner nor the data collection examiner could differentiate between active TENS and placebo TENS and, importantly, all participants received the same set of instructions. Blinding was assessed at the end of TENS treatment and before participants left the clinic. Participants were asked if they thought they received active TENS or placebo TENS.
Transcutaneous electrical nerve stimulation was applied with 4 self-adhesive electrodes (∼5 × 5 cm [2 × 2 in]; DJO Inc) to bracket the knee with OA and to apply paresthesia encompassing the painful knee. Two electrodes were placed above the knee, and 2 were placed below. The current was delivered across the joint through 2 channels. One channel was connected to an electrode above the knee medially and below the knee laterally; the second channel was connected to an electrode above the knee laterally and below the knee medially. The specific electrode sites were determined by the allocation examiner using points of least impedance.30 To locate the points, the participant and the examiner each held a gelled electrode in his or her hand. The examiner completed the circuit by placing a water-dipped finger on the participant's skin and increasing the intensity to the experimenter's sensory threshold. The examiner then glided her finger over the area to locate the points of least impedance (ie, points at which increased sensation occurred for the examiner). The points were always located within the frontal plane and within 25 to 75 mm of the midpatella. The points were cleaned with water, and the electrodes were applied.
The TENS units were turned on for 20 minutes before testing to reach the peak effect and were turned off when testing was completed, for a total application time of 40 to 50 minutes. The greatest effects of TENS occur when the unit is on. All of our measurements were obtained in the same order so that quantitative sensory testing was done immediately after 20 minutes and function testing (Timed “Up & Go” Test [TUG]) was performed last. Therefore, we were comparing the effectiveness of the same length of TENS treatment across participants.
Outcome Measures and Follow-up
Subjective pain intensity.
Participants were asked to rate their pain on a horizontal 100-mm visual analog scale (VAS). The anchors were “no pain” and “worst imaginable pain.” The VAS is valid and reliable compared with other pain rating scales (r=.71–.78, intraclass correlation coefficient [ICC]=.71–.99).31,32 Pain was assessed at rest, during the TUG, and during heat temporal summation (HTS).
Pain sensitivity.
Pain sensitivity was measured with the quantitative sensory tests outlined below. Three sites were marked 1 cm apart at the medial joint line bilaterally. Another 3 sites were marked 2.5 cm apart on the tibialis anterior muscle bilaterally, with the top site being marked 7.5 cm below the inferior border of the patella. All pain sensitivity measurements were obtained at all 6 sites, except for the heat pain threshold (HPT) and HTS (see below), which were obtained over the middle point of the knee and tibialis anterior muscle because of the larger size of the thermode stimulator probe. All sites were assessed bilaterally.
Cutaneous mechanical pain threshold (CMPT).
The CMPT was assessed with a set of 20 von Frey filaments (North Coast Medical, Gilroy, California) applied to the test sites in ascending order (0.008–300 g). The tip of the filament was applied perpendicular to the site and pressed until bending occurred. One trial per filament was done. The CMPT has excellent test-retest reliability (r=.97).32
Pressure pain threshold (PPT).
The PPT was assessed with a handheld pressure algometer (Somedic AB, Farsta, Sweden) applied at 40 kPa/s (1-cm2 circular tip). Participants were instructed to press the handheld response switch when the sensation first became painful. Participants were familiarized with the procedure by performing a practice test on the forearm. The PPT has excellent test-retest reliability (r=.70–.94) and is a valid measure for deep-tissue hyperalgesia, as indicated by earlier studies showing decreases in chronic pain compared with the responses in healthy controls.33
HPT.
The HPT was assessed by use of a TSA II NeuroSensory Analyzer (Medoc Ltd, Ramat Yishai, Israel) with a 5-cm2 probe. The probe was placed at the middle of the 3 marks at each site. The temperature was initially set at 37°C and increased 1°C/s to a maximum of 52°C. Participants indicated when they first felt pain (1/10) by pushing a button that terminated the stimulus.
HTS.
The HTS was measured with a TSA II NeuroSensory Analyzer. A tonic heat stimulus of 45.5°C was applied for 20 seconds. After the first 5 seconds of the heat stimulus application, participants rated pain every 5 seconds for 15 seconds. A difference between the pain rating (first rating) at 5 seconds and the pain rating at 15 seconds was used for analysis. Thermal measures have good test-retest reliability (ICC=.77) in people with knee OA.32
TUG.
The TUG is a standardized test in which people arise from a chair with no arm rest, ambulate approximately 3 m (9.8 ft) as quickly as possible, turn, ambulate back, turn, and return to sitting in the chair.34 Participants were timed in a standardized fashion from the moment the upper back left the chair until return to the full sitting position with the back in contact with the chair. We previously used the TUG in people with OA and found a reduction in TUG walking time, demonstrating the sensitivity of the TUG to a single joint mobilization treatment.35 The TUG has good reliability (ICC=.92–.99) in elderly populations34,36,37 and has good construct validity and significant correlations with gait speed (r=.61–.75), the Berg Balance Scale (r=.81), and step length (r=.77).36,37
Test protocol.
A time line of the test protocol is shown in Figure 2. All testing was performed with the same equipment and by the same examiner. At the testing session, informed consent was obtained and sensory screening was done. If the sensory examination results were normal, testing began. First, participants completed a demographic questionnaire, and height and weight were recorded. Pain at rest was measured before pain sensitivity. The order of the following measurements remained consistent for all participants: CMPT, PPT, HPT, and HTS. The testing order for each of the 4 areas and 3 test sites was randomized to prevent an ordering effect of testing. Participants then completed the TUG and rated their maximum pain during this test.
Time line for the 3-hour testing session. CMPT=cutaneous mechanical pain threshold, HF=high frequency, HPT=heat pain threshold, HTS=heat temporal summation, LF=low frequency, PPT=pressure pain threshold, TENS=transcutaneous electrical nerve stimulation, TUG=Timed “Up & Go” Test, VAS=visual analog scale
Once testing was completed, the data collection examiner left the room and the allocation examiner allocated TENS. The allocation examiner stayed with the participant for 20 minutes, at which time the data collection examiner reentered and repeated the testing as described above. The same blinded examiner obtained all outcome measurements before and after TENS. Once the second testing session was completed, the TENS allocation examiner returned to remove the TENS unit and assess participant blinding.
Data Analysis
Sample size calculations were made with preliminary PPT data to compare HF-TENS and LF-TENS against placebo TENS. Pain thresholds are commonly used in studies of animals and people who are healthy.1,7,13,16,25 The study sample size of 25 participants per group was determined on the basis of an expected PPT difference of 100 kPa, a standard deviation of 110, a significance level of .05, and a power of .80. For the other outcome measures, the detectable differences were at least 10 mm (out of 100 mm) on the VAS (SD=0.9) and 1.4 seconds on the TUG (SD=1.5). Therefore, the sample size was powered to detect clinically significant differences in all variables, given that pain ratings and impairments are minimal in early OA.38
Mean and standard error of the mean were calculated for all variables before and after TENS and for the difference score before and after TENS within individual groups. In addition, 95% confidence intervals were calculated for primary outcome measures. A linear mixed-model analysis for repeated measures was used to compare mean changes among treatment groups (HF-TENS, LF-TENS, placebo TENS) in the outcome measures for the site (knee, tibialis anterior muscle), side (affected, contralateral), and time (before, after) (within-subjects effects). The model included baseline values and body mass indexes as covariates because body mass indexes differed between groups (Tab. 1). For specific comparisons of interest, a test of mean contrasts based on the fitted method was performed. To account for the multiple tests performed in relation to a specific hypothesis, P values were adjusted with the Bonferroni method. The Kruskal-Wallis test was used to detect differences in nonparametric data.
Demographic Characteristics of Study Participants and Outcome Measures Before Transcutaneous Electrical Nerve Stimulation (TENS)a
Role of the Funding Source
This study was supported by the National Institutes of Health (R03-NR010405), a Marsha and Ralph Congdon Faculty Development Fellowship in Acute Care for the Chronically Ill, and The University of Iowa Carver College of Medicine.
Results
Table 1 shows the demographic information for all treatment groups as well as the baseline values for each measure. There were no significant differences between groups in demographic characteristics, with the exception of body mass indexes (P=.027). Therefore, body mass indexes were controlled for in the analyses. All baseline measures were similar between groups, with the exception of HPT (P=.02). Baseline measures were also controlled for in the analyses. Eleven participants (15%) were taking opioid medications for pain control, and 54 (72%) were taking nonopioid medications. There were no significant differences between groups in the numbers of participants taking analgesic medications (Tab. 1).
TENS Amplitude and Blinding
The pulse amplitudes required to achieve the desired TENS treatment intensities were similar between groups; the mean amplitudes were 27.4 mA (SD=1.70) for HF-TENS, 24.1 mA (SD=1.8) for LF-TENS, and 24.5 mA (SD=1.6) for placebo TENS. Of the participants receiving placebo TENS, 57% correctly identified the treatment as a placebo, whereas 43% believed that they received the active treatment. Participants receiving active TENS correctly identified the treatment as active 92% of the time.
Pain Sensitivity
High-frequency TENS increased PPT at the affected knee (P=.002) and over the anterior tibialis muscle of the affected leg (P=.0001) compared with pre-TENS values (Tab. 2, Fig. 3). Low-frequency TENS significantly increased PPT at the ipsilateral knee only (P=.05). Placebo TENS did not significantly change PPT (Fig. 3). Pairwise comparisons of the treatment groups revealed a significant difference in the mean changes in PPT between HF-TENS and placebo TENS (P=.026); there was no significant difference between LF-TENS and placebo TENS or between LF-TENS and HF-TENS.
Primary and Secondary Measures Expressed as Difference Scoresa
Differences between pressure pain threshold (PPT) before transcutaneous electrical nerve stimulation (TENS) and PPT after TENS for both the knee and the tibialis anterior muscle both ipsilaterally and contralaterally. Data are expressed as the mean and standard error of the mean. *=significantly different from baseline, +=significantly different from placebo. HF=high-frequency TENS, LF=low-frequency TENS, P=placebo TENS
Baseline CMPT, HPT, and HTS values for each group before TENS are shown in Table 1. The HPT, CMPT, and HTS values were unchanged after all treatments, and there were no significant differences between groups (Tab. 2).
Pain at Rest
Pain at rest decreased significantly in all 3 groups (P=.001, P=.01, and P=.0001 for HF-TENS, LF-TENS, and placebo TENS, respectively). However, there were no significant differences between groups (Tab. 2, Fig. 4).
Difference scores for pain at rest in ipsilateral and contralateral knees during transcutaneous electrical nerve stimulation (TENS). Significant decreases were observed ipsilaterally for all 3 groups (placebo TENS [P], low-frequency TENS [LF], and high-frequency TENS [HF]). Data are expressed as the mean and standard error of the mean. *=significantly different from baseline
Function (TUG) and Pain During Function
Pain during the TUG decreased significantly in all 3 groups (P=.001, P=.03, and P=.001 for HF-TENS, LF-TENS, and placebo TENS, respectively). However, there were no significant differences between groups (Tab. 2, Fig. 5). The time to perform the TUG did not change significantly in any of the groups, and there were no significant differences between groups (Tab. 3).
Difference scores for movement-evoked pain during the Timed “Up & Go” Test (TUG) in ipsilateral and contralateral knees during transcutaneous electrical nerve stimulation (TENS). Significant decreases were observed ipsilaterally for all 3 groups (placebo TENS [P], low-frequency TENS [LF], and high-frequency TENS [HF]). Data are expressed as the mean and standard error of the mean. *=significantly different from baseline
Timed “Up & Go” Test (TUG) Walking Times Before and During Transcutaneous Electrical Nerve Stimulation (TENS)a
Discussion
The present study showed an increase in PPT at the knee joint with HF-TENS and LF-TENS and over the tibialis anterior muscle with HF-TENS relative to the results obtained with placebo TENS; placebo TENS had no significant effect on PPT. However, subjective pain at rest and during the TUG decreased equally with all 3 treatments. No differences were observed for CMPT, HPT, HTS, or function. These data show that TENS is effective for deep-tissue pain sensitivity induced by OA. They further show that a single treatment of TENS has minimal effects on pain and function relative to the effects of placebo TENS.
HF-TENS and LF-TENS Reduce Deep-Tissue Pain Sensitivity
The present study showed that HF-TENS and LF-TENS increase PPT at the site of injury; we interpret this finding as a reduction in primary hyperalgesia. Changes in primary hyperalgesia measures suggest changes in nociceptor sensitivity39 and parallel changes that we previously observed in animal studies.25 The present study also showed that HF-TENS increases PPT over the tibialis anterior muscle, an area outside the site of injury; we interpret this finding as a reduction in secondary hyperalgesia. Changes in secondary hyperalgesia measures suggest changes in central neuron excitability39 and parallel changes that we previously observed in animal studies.19 The increases in PPT are similar to increases in PPT during active TENS in healthy controls1–6,13,14 and in arthritic animal models examining hyperalgesia immediately after TENS with the same parameters.16,20,24 Transcutaneous electrical nerve stimulation reduces the excitability of nociceptive neurons in the central nervous system in arthritic animals,40 and higher intensities produce greater reductions in excitability.41 Therefore, changes in PPT with TENS are likely mediated by reduced central neuron excitability. The present study is the first to show the effects of TENS on hyperalgesia in people with OA; these effects may serve as a useful measure of neuron excitability. The PPT correlates with movement-evoked pain (Sluka KA, Rakel BA, et al, unpublished observations); both are evoked pain stimuli. Furthermore, palpation tenderness is an essential part of the physical examination of patients; PPT measures may offer clinicians an improved objective measure of tenderness.
The reason for the lack of changes in CMPT, HPT, or HTS may have been the fact that these measures are cutaneous stimuli and are not sensitized by knee OA. A previous study of OA showed greater enhanced temporal summation to pressure in people with more pain.42 However, a recent study showed enhanced HTS in people with OA.43 Alternatively, TENS may have minimal effects on cutaneous heat pain but be more effective for deep-tissue pain. In support of this notion, we found no change in HPT with TENS in people who were healthy44 but significant reductions in PPT and temporal summation to mechanical stimulation of muscle.13 Together, these data support the idea that TENS is more effective in reducing deep-tissue hyperalgesia.
Effects of TENS on Pain and Function
We previously showed that active TENS reduced movement-evoked postoperative pain relative to the results obtained with placebo TENS.12 Surprisingly, the present study showed equivalent reductions in pain during the TUG with both active TENS and placebo TENS but no changes in function. The reason for the lack of difference between active TENS and placebo TENS may be the fact that the TUG minimally increased pain above that at rest (<10 mm/100 mm), suggesting that the TUG did not produce enough pain in our participants to evaluate movement-evoked pain. In contrast to the present study, which involved a single visit, 2 weeks of active TENS reduced TUG walking times in people with symptomatic knee OA relative to the results obtained with placebo TENS.45 We previously showed decreases in TUG walking times in people with OA after a single joint mobilization,35 demonstrating the capacity of a single treatment to modify TUG walking times. It is possible that the difference between studies with improvements in the TUG and the present study is related to the severity of functional limitations. For instance, in earlier OA studies showing a positive effect on the TUG, the times were 20 to 24 seconds,35,45,46 whereas the times in the present study were 12 to 14 seconds, close to normal (≤10 seconds).34 Therefore, for less severe symptomatic OA, the TUG may not be an appropriate measure for examining movement-evoked pain and function. Future studies should use a function test that produces greater pain, such as stair climbing.
We previously showed no significant effect of TENS on pain at rest (VAS) relative to the effect of placebo TENS after abdominal surgery.47 In the present study, both placebo TENS and active TENS reduced ratings of pain at rest by 10 to 18 mm out of 100 mm. These changes are minimal and not clinically important when people have pain ratings below 50 of 100 mm on the VAS38; pain ratings in our participants averaged between 24 and 28 mm. Therefore, the effects on pain at rest may depend on the pain intensity.
Earlier work showed that high doses of caffeine can reduce analgesia produced by TENS.48 In the present study, we did not control for caffeine; therefore, caffeine may have influenced our results. However, our earlier studies of people who were healthy did not control for caffeine and showed positive effects of active TENS relative to placebo TENS,13,14,22,49 and we showed effects on PPT in the present study. Opioid intake also may influence results; people and animals who are tolerant of opioids showed reduced effects of LF-TENS but not HF-TENS.23,50 However, the majority of our participants (85%) were not taking opioid analgesics, and the numbers of participants taking opioid analgesics were similar between groups. Additionally, participants did not take analgesic medications for 4 hours before testing to eliminate the effect of analgesia on the test results.
Intensity Is Critical for TENS Effectiveness
Earlier studies of people with OA included in the 2009 Cochrane review9 supported the need for high-intensity TENS. Intensities in the included studies varied widely.9 High-frequency TENS was given at a strong but comfortable intensity in 7 trials,45,46,51–55 at sensory threshold or below in 5 trials,56–60 at a noxious level in 1 trial,61 and at an unreported intensity in 2 trials.62,63 Low-frequency TENS was applied at a motor intensity in 3 studies64–66 and a strong sensory intensity in 1 study.45 Trials that reported effective TENS generally used higher intensities than those that reported no effect. In the present study, TENS was applied at 90% of the motor threshold, a strong sensory intensity. In recent studies, higher intensities of TENS were shown to be more effective in reducing pain.12–15,22,67 The intensity of stimulation was positively correlated with the change in PPT produced by TENS.13,14 Another variable that may have influenced our results is the fact that we did not continuously adjust the stimulation over time, as is common in clinical practice. Titrating TENS intensity upward during treatment increased hypoalgesia in healthy controls.67
Finally, LF-TENS may not have been as effective because we used a lower pulse duration and a lower intensity than are commonly used clinically (ie, motor intensity).45,64–66 The total current applied with HF-TENS is greater than that applied with LF-TENS when the same intensity and pulse duration are used. Although LF-TENS is traditionally delivered at motor intensities, our earlier work showed that LF-TENS at a strong sensory intensity produced effects equivalent to those obtained with HF-TENS at a strong sensory intensity.16,25 We also showed opioid-mediated analgesia at sensory intensities with both LF-TENS and HF-TENS.17,20 Therefore, the present study focused on the effect of frequency delivered at a strong sensory intensity for both HF-TENS and LF-TENS.
Blinding of Active and Placebo TENS
The present study is the first to validate and test the new transient placebo TENS unit in a group of people with pain. We were able to adequately blind the participants receiving placebo TENS, with 57% correctly identifying the placebo; this value was not significantly different from chance (50:50). If blinding had been ineffective, we would have expected the percentage correctly identifying the placebo to be closer to 100%, like the value obtained for the active TENS unit. The new placebo TENS unit can completely blind the experimenter applying TENS, and we previously showed the complete inability of the examiner to correctly identify active TENS or placebo TENS.13,14,22 Adequate blinding is important because earlier TENS studies used separate instructions for active TENS and placebo TENS, and doing so can influence the expectations of patients and have a profound effect on treatment outcomes.68–72 For example, seminal work by Levine and Gordon68 revealed significant analgesia with a placebo drug given in a completely blinded manner; the placebo produced analgesia equivalent to that of low-dose morphine. On the other hand, negative or neutral instructions resulted in less reduction in pain after spinal manipulation than positive instructions.72 By giving the same instructions, we were able to show that active TENS in people with OA produced pain reduction similar to that produced by placebo TENS, suggesting that TENS has a strong placebo effect on OA pain.
It is possible that the initial placebo effect that occurs with a single treatment is reduced with repetitive TENS. In support of this notion, 10 days of electroacupuncture or 2 weeks of TENS in people with knee OA resulted in significant improvements in pain and function.45,73 Furthermore, Marchand et al74 showed a cumulative effect of active TENS but not placebo TENS given twice per week in people with chronic low back pain. Therefore, future studies should examine the effects of repetitive active TENS and transient placebo TENS when the experimenter applying TENS also is blinded with regard to group.
In the present study, participants were able to correctly identify active TENS 92% of the time. We previously reported similar responses to active TENS in healthy controls.13 Despite participants knowing that they received active TENS, there was no difference between active TENS and placebo TENS in subjective pain rating. Blinding of an electrical modality such as TENS has always been difficult, and few studies have reported blinding of active TENS.
Clinical Implications
In summary, the present randomized clinical trial examined the effects of single treatments of HF-TENS and LF-TENS on knee OA pain and function. The use of various outcome measures, different frequencies, and an improved placebo provided insight for the management of knee OA pain with TENS. We pilot tested a series of outcome measures designed to parallel and validate animal models of TENS and to test the effects of TENS in a true double-blind manner. Using PPT as an objective measure of pain sensitivity, we showed that both HF-TENS and LF-TENS reduced primary hyperalgesia and that only HF-TENS reduced secondary hyperalgesia in people with OA. Quantitative sensory testing with cutaneous mechanical and heat pain measures was not affected by HF-TENS, LF-TENS, or placebo TENS, suggesting that TENS has no effect on cutaneous hyperalgesia. Alternatively, it is possible that the participants with OA did not have cutaneous mechanical and heat hyperalgesia. All treatments had similar but minimal effects on subjective pain measures, suggesting a placebo component of the effect of TENS. None of the treatments had an effect on TUG walking times. The TUG may not be an appropriate functional outcome measure in people with early symptomatic OA of the knee because increases in pain and decreases in function are minimal. Future studies should expand outcome measures used in TENS studies to include not only pain at rest, as commonly assessed, but also pain during physical function tasks and deep-tissue hyperalgesia measures. The effects of repetitive treatments and of higher intensities should be tested in people with painful conditions to further elucidate the most effective use for TENS.
The Bottom Line
What do we already know about this topic?
A substantial body of literature shows that TENS produces analgesia in both animal and human subjects with a variety of pain conditions by using endogenous opioid pathways. Systematic reviews of TENS for knee OA offer conflicting positions. Although prior studies routinely examine resting pain as their main outcome, TENS may work more effectively for pain at rest, movement-evoked pain, and pain sensitivity.
What new information does this study offer?
This study evaluated the effect of TENS, applied with parameters used in prior animal studies, with a new placebo TENS on a variety of outcomes including resting and movement pain, pain sensitivity, and function in subjects with OA. There was a significant reduction in evoked pain measures but not in resting pain with TENS.
Footnotes
-
Dr Rakel, Dr Walsh, and Dr Sluka provided concept/idea/research design. Ms Vance, Dr Rakel, Ms Blodgett, Dr Walsh, and Dr Sluka provided writing. Ms Vance, Dr Rakel, Ms Blodgett, Dr DeSantana, and Dr Sluka provided data collection. Ms Vance, Dr Rakel, Dr Amendola, and Dr Sluka provided data analysis. Ms Vance, Dr Rakel, and Ms Blodgett provided project management. Dr Rakel and Dr Sluka provided fund procurement and facilities/equipment. Dr Amendola provided study participants. Ms Vance, Dr Rakel, and Ms Blodgett provided clerical support. Ms Vance, Dr Rakel, Ms Blodgett, Dr DeSantana, Dr Amendola, Dr Walsh and Dr Sluka provided consultation (including review of manuscript before submission).
-
The authors thank Shannon Lehman, who served as the TENS effectiveness coordinator, for assistance with the study. The TENS units and supplies used in the study were donated by DJO Inc.
-
This study was approved by the Human Subjects Institutional Review Board at The University of Iowa.
-
This study was supported by the National Institutes of Health (R03-NR010405), a Marsha and Ralph Congdon Faculty Development Fellowship in Acute Care for the Chronically Ill, and The University of Iowa Carver College of Medicine.
-
This trial is registered at ClinTrials.gov: NCT01354054.
- Received June 1, 2011.
- Accepted March 22, 2012.
- © 2012 American Physical Therapy Association
References
Issue highlights
