Effect of electrical stimulation waveform on ... - Wiley Online Library

Effect of electrical stimulation waveform on healing of ulcers in human beings with spinal cord injury LUCINDA L. BAKER, PT, PhD°; SALAH RUBAY1, MDb; FERNANDO VILLAR, PT, PhD°; SHARON K. DEMUTH, PT, MS°

stimulation waveforms have been used to enhance wound healing, with little consideration for potential differences in their physiologic effect . The present study evaluated the effect of stimulation waveform and electrode placement on wound healing. Eighty patients with spinal cord injury and one or more pressure ulcers were treated. A total of 185 ulcers received 45 minutes of stimulation daily, Each ulcer was subjected to one of four treatment protocols: asymmetric biphasic waveform, symmetric biphasic waveform, microcurrent stim-ulatlon, or a sham control protocol . Electrodes were placed outside the wounds, over intact skin and surrounding the area of the ulcer. Data were categorized by ulcers which healed during the protocol and those which did not. Analysis of the °good response" ulcers (n = 104) showed significantly better healing rates for those receiving stimulation with the asymmetric biphasic waveform, compared with the control and microcurrent groups . Mean healing rates from the present study were similar to previously reported measures . The waveforms studied possessed minimal polar capabilities, and the electrodes were placed outside the wound. These data show that electrical stimulation clearly enhanced healing of pressure ulcers in a significant number of individuals with spinal cord injury ; the physiologic implications of these findings relative to the mechanism(s) by which electrical stimulation enhances wound healing are discussed, However, extrapolation of these results to patients with other types of wounds must await further study. (WOUND REP REG 1996 ;4 :21-8)

Electrical stimulation has been used to facilitate wound healing for more than 30 years.' Although a number of clinical studies have been published reectrical currents on tissue healing," the practice remains restricted to a few centers or as a final therapeutic effort . The low inci dence of electrical stimulation as a part of standard ent programs is largely due to ambiguity associated with the stimulation itself and the

From the Department of Biokinesiology and Physical Therapy, University of Southern California, Los Angeles, Calif, ; and the Pressure Ulcer Manage ment Service, Rancho Los Amigos Medical Center, Downey, Califb Baker, PT, PhD, Department of Biokinesiology & Physical Therapy, University of Southern California, 1540 Alcozar CHP 755, Las Angeles, CA 90033. Copyright © 1996 by The Wound Healing Society. 1067-7927196 $5.00 + 0 3611171095

types of patients most successfully managed with this adjunct treatment. Stimulation programs have ranged from the con° ' " ents over a wound for several hours a day' -' to the use of one or more short sessions of pulsed stimulation 5 days a week.' -,' Protocols have included polarity changes' -',9 when the probability of polar effects from some forms of stimulation appears to be minimal.61 ,9 A recent study has reported enhanced healing of decubital ulcers with the use of a waveform with minimal polar capabilities ." ," In short, the characteristics of stimulation which most affect the healing process are unknown. Although a variety of clinically based procedures have been developed, they lack either a physiologic rationale or experimental support. At least two aspects of electrical currents have the potential to influence healing tissue . The first of these is the ability of certain types of electrical currents to attract opposite charged particles, thus possibly en-

21

22

WOUND REPAIR AND REGENERATION JANUARY-MARCH 1996

BAKER ET AL .

Table 1 . Electrical stimulation parameters

Parameter Amplitude Phase duration (Nsec) Frequency (pulses/sec) Oil/off time (sec) Waveform

Treatment group Asymmetric Symmetric Microcurrent biphasic (A) biphasie (B) (MG) Below contraction 100

Below contraction 300

4 mA

50

50

1

7 :7

7 :7

7 :7

10

-~~-

hancing the migration of cells such as macrophages and fibroblasts ."-16 These effects of polarity are strongest when direct current stimulation is used, as reported by Woolcot et al .2 in the 1960s. Theories abound regarding the current of injury or the flow of charged particles from an uninjured to an injured area of dermis and the effect this current of injury has on the migration of cells or proliferation of tissue . 12,13 Wounds which fail to heal have been reported to display reduced levels of the current of injury ." As a wound heals and decreases in size, the current of injury decreases as well . The concept of a current of injury facilitating the normal healing process, and the enhancement of that potential by the application of external currents, is what directed early research efforts.'-' I techniques have been developed using electrical stimulation in ways relatively independent of this acknowledged physiologic phenomenon . The original investigators, using direct current stimulation, proposed a protocol which included periodically changing the polarity of stimulation,' although no such changes had been reported physiologically. Later investigators began to use short pulsed monophasic waveforms, with the apparent assumption that the polar effects from direct currents were equally likely to occur with the short pulsed stimulation.6°' 6 The actual polar effects of very short pulses (2 to 20 psec) have never been directly addressed in the literature . A second aspect of many electrical currents which could affect the healing process is the activation of cutaneous nerves, which have been reported to create a centrally mediated increase in circulation .'' Although the location and specific type of nerves excited for this response appear to be rather nondiscrete, the level of activation appears to be more of a factor. Comfortable levels of sensory stimulation appear to trigger an increased circulatory response, but intense or painful stimulation or, under some conditions, vigorous activation of muscular contraction may decrease the circulation to the skin ." Thus, low level sensory stimula-

tion, which is used to reduce ischemic pain, may provide a key to increased availability of nutrients for a healing wound and the removal of toxic waste products . Direct currents, which are applied at exceedingly low amplitudes, are unlikely to trigger the sensory nerves required to establish this more regional response ."' The short pulsed stimulation, which has largely superseded the use of DC stimulation, is generally provided at a low sensory level, with care taken to avoid muscular contractions .6, ' Although activation of peripheral nerves is greatest under the negative pole of pulsed monophasic stimulation, there is also activation at the positive pole, with only a slight decrement in effectiveness.19 Most electrical stimulators designed to excite peripheral nerves use waveforms which are said to be "balanced," thus minimizing the polar effects of the electrical currents ." Hence excitation of the peripheral nerve is not wholly dependent on the polar nature of an electrical current, and symmetric biphasic waveforms are often used to increase the area of neural excitation during an applied electrical current." Clarification of which aspects of electric currents are most important in providing enhanced healing effect would provide the clinician better guidelines for the development of facilitated healing treatments . The purpose of the present study was to explore the efficacy of the neurally regulated regional influences of electrical stimulation for enhanced wound healing, while minimizing the possible polar characteristics of the stimulation, in spinal cord injured individuals with pressure sores.

METHODS

A total of 80 patients with spinal cord injuries were treated with electrical stimulation to enhance wound healing during the study. These patients had 192 wounds, all of which were subjected to one of four treatment protocols. Before beginning the study protocol, each patient was medically cleared by a physician, provided a detailed explanation of the treatment in their own language, and provided written informed consent.

Study protocol

Patients were randomly assigned to one of three stimulation groups (Table 1) or a fourth control (C) group. Two stimulation treatment groups were treated with currents assumed to be therapeutic, on the basis of previous studies6-1,10," (A and B groups). The third ulation treatment, microcurrent (MC), was o nally thought to incorporate stimulation below effective levels and would thus serve as a "stimulated con-

WOUND REPAIR AND REGENERATION VOL. 4, NO . 1

trol" protocol . 2°,21 Preliminary data analysis appeared to show some therapeutic effect in this group, thus the treatment was continued throughout the study similar to the other stimulation protocols . Patients were maintained on their respective stimulation protocols until their ulcers healed, the physician decided to intervene, or the patient withdrew from the study. The control group received the same stimulation procedures as the MC treatment group, but special leads were used to interrupt the passage of current so the patient received no electrical stimulation. Patients were maintained on the control protocol for 4 weeks (28 days) or until their ulcer had healed, whichever occurred first. After the period of control management, patients whose ulcers remained unresolved were randomly assigned to either the A or B treatment groups for continued management . Only the therapist doing daily stimulation treatment and weekly tracings knew the group assignment of each subject. All inpatients were seen 5 days a week by a physical therapist working on the research project. Three treatment sessions of 30 minutes duration were pro vided, with a short break between sessions. After each break, the stimulator (UltraStim ; Henley Int., Houston, Tex.) was programmed to automatically restart the treatment session. The patient was instructed to remove the stimulator after three sessions. Compliant stimulation time was considered to be 1.5 hours per day, with half that amount (45 minutes) defined as semicompliant stimulation. Some patients chose to remain on stimulation for longer periods of time, and this was monitored by the therapist each day through the compliance feature of the stimulation unit. Subjects treated as outpatients were monitored regularly through clinic appointments, home visits, and frequent phone calls . Compliance to the stimula tion treatment was monitored through the compliance meter on the stimulator whenever the patient was seen by the research therapist. Follow-up was done every 2 to 4 weeks. Electrical stimulation characteristics

Electrical stimulation was provided through surface electrodes made ofcarbon-rubber. The sizes ofthe electrodes varied, depending on the size and location of the ulcer, but ranged from 2 .5 x 2 .5 to 5 x 10 cm. Electrodes were placed over intact skin less then 1 cm from the edges ofthe ulcer. Electrodes were generally placed proximal and distal to the treated ulcers, but medial and lateral placements were used in some regions (coxygeal ulcers). The electrode configuration for patients receiving the A protocol consistently placed the

BAKER ET AL .

23

Table 2. Descriptive variables for patients Gender Age Ethnicity Onset of spinal cord injury Level of injury Duration of ulcer Albumin concentration Hematocrit and hemoglobin Vital capacity and peak air flow Standard ulcer treatment Ulcer location Initial ulcer area Initial healing rate

electrode which was negative during the leading phase of the waveform proximal to the wound, with the more positive electrode placed distally . This was the only treatment protocol which used a waveform with a potential for polar effects. Stimulation amplitude was set for each subject and each wound by increasing the intensity until a minimal muscle contraction was observed . The intensity was then decreased until the contraction was no longer present. This procedure was followed for patients treated in the A and B protocols only. Stimulation amplitude was fixed at 4 mA for the MC and C protocols, the minimal intensity necessary to allow the stimulator's compliance monitor to function. Wound evaluation

The primary technique used to evaluate wound size was a tracing- of the wound edge onto a clear acetate sheet. Wound evaluation was done every week for inpatients and every 2 to 4 weeks for outpatients. The tracing was accompanied by a calibrated photograph, to assist in the later interpretation ofthe tracing. Data from each patient were not analyzed until that patient had completed the study protocol, which sometimes resulted in a delay of several months before the tracing interpretation was attempted. Because of this delay, the color photograph was invaluable . The surface area of the wound was digitized from the tracing by a technician who was not knowledgeable about the treatment received by the patient. When there was significant depth to an ulcer, several techniques were attempted for documentation. The volume of sterile saline solution required to fill the wound was attempted in some patients but was limited in many others by the inability to position the ulcer perpendicular to gravity. Depth measures were made from several wounds, but no standard formula was found in the literature to compute a valid volume

24


BAKER ET AL.

Table 3 . Demographic data for wound healing patients Treatment group Descriptive variable

A Protocol

No . of subjects Gender M F Age (range, yr)

B Protocol

MC Protocol

C Protocol

17 3 34 ± 2 (19-64)

16 5 40±2 (21-64)

17 3 36±2 (17-64)

16 3 33±4 (19-76)

10 5 5

11 2 7 1 153±18 (4-420)

8 8 4

6 4 7 1 57±13 (3-204)

20

Race White Hispanic Black Asian Duration of SCI (range, mo) Level of SCI Cervical Thoracic Lumbar Complete SCI Yes No Albumin (N = 4-5, gm/100 ml) Hemoglobin (N = 12-16, dug/100 ml) Vital capacity (N = 2250-3990 ml)

124 t 17 (1-396)

77±11 (1-252)

6 14

9 9 2

10 8 2

8 10 1

10 8 3 .2 ± 0 .11 11 .8 ± 0 .2

15 4 3 .4+0 .19 12 .0 ± 0 .3 1832+174

15 5 3 .2+0 .11 11 .1+0 .3 2900±245

12 2 3 .3 ± 0 .13 11 .6 ± 0 .3

2286 :t302

2362± 211

Values represent mean t standard error of the mean. SCI, Spinal cord injury.

measure. Thus, while acknowledging that surface area measurements represent only part of the healing process in most wounds, we note that they were the only reasonably reliable measure available . Data analysis Data from each patient and each wound were first digitized from the acetate then entered into a master database for analysis . Descriptive information collected for all patients is shown in Table 2. Compliance to the stimulation protocol was evaluated through the average current recorded by the stimulator and the duration of stimulation time accrued throughout the treatment program. Both of these measures were recorded at each evaluation session. The mean healing rate was computed for each ulcer . s -8 Sequential area measures were subtracted from the preceding measure, and the difference was char acterized as the percentage of closure. This value was then divided by the number of days between measures and multiplied by 7 . This value was identified as the weekly healing rate . The average of these sequentially computed weekly healing rates was calculated for the overall mean healing rate. Area i - Areai+ X4 = (1/n-1) I

-

Areai

Day;+l - Day;

x7

x100

In the above formula, Areai is the wound area measured on Day i, and Area;+l is the wound area measured i + 1 days later. The values at i = 1 represent the first day treatment was initiated. Comparison of mean healing rates was done with a one-way analysis of variance . An analysis of variance with repeated measures design and a covariate was used when comparing ulcers which were treated with both control and stimulation protocols. The initial ulcer area at the inception of the two treatment programs was used as the covariate . Comparison of descriptive data was also done with analysis of variance . Further analytical techniques included both multiple and stepwise regression analyses . All statistical examinations were done with the BMDP statistical package (University of California Press, Los Angeles, Calif.) . Statistical significance was determined at p < 0.05, with the use of values corrected for multiple comparisons when appropriate.

RESULTS

The 80 patients were originally distributed between the treatment groups as shown in Table 3. This table also includes the mean descriptive data from each patient. No significant differences were found between any of the treatment groups on the basis of comparison of the descriptive data .

minimum Prog, the initial areas Prog large A to wounds todifferences 4Sulfadiazine treated found location source 34 multiple change or inception and 4M? afor 2the healing areas ofstimulation REPAIR treatment therapy (days) Thus, to BDescriptive patients discharge or variability statistical time ulcer patients reasonable Ibetween semicompliant groups study protocols of12) of in Control are wounds program of 45 cream (hr/day) AND there the (range, wounds and rates found of were the minutes from for managed mean the protocol This with REGENERATION groups 11 OD, data was patients comparisons in after ulcers days) were are new parity found in patients stimulation the occlusive the and spinal healing finding aTable for shown of treatment 28 wounds total in 185 counted data were Descriptive all stimulation and in for among any days, were dressing wounds 4cord initial wounds of changed more in rates was the descriptive No frequently 192 Protocol The intervention Table after also and as injuries ±levels) treatment major ±WD-S, computed probably wounds 0among or than 42 new which initial attheir data reassigned per from discharge having 5the wet-to-dry differences treatment were No variables, day one tended readmitfrom stimulareceived inception and the the treated, groups due statismean (comulcer been used four conwith disthe ulto tosaline of Protocol stimulation *solution 5 this inability each stratification levels stratified sample termed groups in different stimulation DD, and analysis to the which represented treated or those the of of dry 154 or showing between the data these of to stimulation for subdivided "poor analysis semicompliant dressing data, received for which the identify each poor reactions and were was All and this responses," were identified extreme MOM there healing in the ±control Non-C, meeting protocol response v±appropriate failed as 12 reanalyzed those 5each, analysis 39 45 obvious by identified atime healing were to minutes noncompliant single those variability to which treatments) wounds all before and are and all groups, The groups, close statistically forms Wounds subgroups wounds group the rates selection shown comparing for fulfilling as ofstimulation mean failed that with during receiving "good percentage defining of the indicating of The Protocol were treatment stimulation time in which which healing to the the ± ET criteria responses", number 42 Table within the significant heal number treatment the AL data included were good two complihealed 28 healed of treatrates Only 6daily that disand day the deinreIn of

WOUND VOL,

BAKER

.

25

Table . Parameter

A

67 183 (2-454)

of Standard Sulfa OD WD-S DD Other Ulcer Foot Thigh Ischial Sacral Other Ulcer Surgery Pressure Infected Yes No Duration therapy Stimulation Reason Held Non-C A Withdrew Sulfa, A These have ted discharged. the tion wounds . with (range trol 192 were indicating at a pliant for charge weekly tical cer treatment the

Treatment group MC

13

C

58 231,38 (2-1095)

42 154 (5-961) 20 4 12 1 5

12

4

17 6 17 5 10

9 10 20 24 3

5 13 18 19 3

3 11 12 10 6

2 4 10 9

31 36

41 17

17 25

16 9

24 34 42

21 21 38

11*11

IN

29 2 10

47 19 3445 1.4

.1

25

9 2

2012 1t 6

7 ;

;

.

;

;

;

Because

an sample, ment

.

. .

data . . .

.

atment

wounds (compliant in ant protocol cluded wounds creased during whereas were for wounds total this differences sponse the tinctly

.

. . .

.

.

26


BAKER ET AL.

Table 5. Wound areas and healing rates-all wounds Treatment group Protocol A

Wound parameter Initial area Final area XA (cm2 ) (cm2) * (c/cA/wk) 6.6 ± 1 .4

4.4 ± 1 .8

36 .4 ± 6.2

B

2.4 ± 0.4

1.1 ± 0.3

29 .7 ± 5.1

MC

8.5 ± 1 .9

3.4 ± 1.4

23 .3 ± 4.8

C

8.6 ± 2.6

5.5 ± 1.5

32 .7 ± 7.0

*Final area denotes area when electrical stimulation was discontinued .

Comparison of all descriptive data by classification of "good" or "poor" healing responses failed to identify any statistically significant differences between these two groups . This was true for the demographic data, the blood values, and the respiratory measures . Similarly, no differences were found between the groups regarding the standard wound care treatment being received . Multiple regression analysis of key descriptive data indicated that the combined effect of 11 descriptive variables (Table 2) had no significant correlation with the mean healing rates for either the "good" or "poor" response groups . When the weekly healing rate from the initial treatment segment was added to the analysis, significant correlations were seen in both groups . However, the relationship was much stronger for the poor healing response group (Mr = 0.82) compared with the good response group (Mr = 0.59). Thus, the addition of the initial response to treatment provided a greater ability to predict the eventual outcome of the treatment, especially in the group which showed a poor response to treatment. Direct comparison of initial healing rates were not, however, significantly different between the good and poor response groups, indicating a high degree of variability in this factor .

Stepwise regression was used to further delineate the effect of the various descriptors on the mean healingrate. The initial healing rate was invariably the first variable with the most powerful effect in describing the mean healing rate . These data indicate that, although the initial response to the stimulation was not always directly predictive of the long-term effect, it did explain a significant portion of the variability of the mean healing rate, especially in the poor response group. Statistical analysis of only the good response group showed a significant difference between the A protocol and the MC and C protocols (Table 6) . No significant differences were found between the B protocol and the other treatments . Despite the persistent variability, there appeared to be a continuum of increased healing rates, at least for those wounds identified as responding to any form of treatment.

Table 6. Stratified mean healing rates Good response XA (%/wk) %*

Treatment group Protocol A

35

63 .7 ± 7.2

61

22

B

32

50 .6 t 5.6

14

MC

18

38 .5 ± 5.8'

70 56

14

C

19

29 .2 ± 8.1'

n

n

Poor response XA (%/wk) % 0.0 ± 7.2

-3 .1 ± 5.3

18 .0 ± 8.8

39 30 44

-

*Percentage of patients in that protocol who had good responses. 'Significantly less then the A Protocol good response . 'The only wounds which did not either heal or complete the 28-day protocol were 6 which were withdrawn from the study. Because of the short treatment durations, these wounds were not included in this analysis .

The A stimulation protocol was most Pffective for enhancing wound healing. Eleven subjects had wounds which were treated under both the control protocol and one of the stimulation protocols (A or B) . The mean healing rate for these patients was statistically greater during the stimulation protocol (43.3% ± 12 .5% A/week) than it had been during the control period (9 .7% ± 3.4% A/week), even when the effect of initial wound size was minimized by the use of a covariate of that measure. Most descriptive data were similar between the control period and the subsequent stimulation period ; the single exception being the size of the initial wound. Because patients had been on standard treatment for 4 weeks, the wounds were significantly smaller at the beginning of the stimulation period. However, the size of the wounds was not a significant indicator of their healing rate when the full data set was analyzed . On the basis of this observation and the comparison with the use of initial area as a covariate, the difference in initial wound size between the two management periods was not a factor in the significant increase in mean healing rate when stimulation was added. Seven of the original eleven wounds closed during the stimulation period .

DISCUSSION

The results from this study clearly show the effectiveness of an electrically balanced asymmetric biphasic waveform for enhancing wound healing in a subset of patients . These findings are in agreement with a recent study which compared the effects of asymmetric biphasic pulsed stimulation with both a direct current stimulated and control patient group." Although the computation used to generate the healing rate was unique l° and not directly comparable with the mean healing rate used in the present study, these authors reported significantly greater healing in patients with spinal cord injuries treated with the asymmetric pulsed

WOUND REPAIR AND REGENERATION VOL. Q NO . 1

stimulation when compared with either the direct current treated or control groups." The pulsed waveform reported by Vodovnik's research,lo,'1 although identified as an alternating current, was similar to the waveform used in the present study as the asymmetric biphasic waveform . The most likely explanation of the enhanced healing responses in the presence of this essentially nonpolar waveform is that the electrical stimulation triggered a neural change which resulted in an augmented healing process . The mechanism(s) associated with a neurally triggered response to stimulation may include a sympatholytic effect which could increase the blood supply to the area of a wound." Other possible mechanisms remain unspecified. Along with questions regarding the mechanism of increased healing are issues related to the size of the region affected by the stimulation . These same neural changes undoubtedly contribute to the altered wound healing reported when short monophasic pulsed stimulation is used because stimulation levels are typically set at or above sensory perception thresholds . Electrode placement outside of the wound in the present study, as well as in the reports by Vodovnik's group,", " would also indicate a neuronally triggered affect to explain the enhanced wound healing . Because electricity will universally take the pathway of least resistance, current density in the wound would logically be less with electrodes placed around the ulcer area, when compared with placing one electrode directly over the wound. However, no direct measures ofcurrent density were taken at the site of the wound in this study. This probable decreased current density in the wound would not be an issue if the mechanism of enhanced wound healing was through activation of the peripheral sensory nervous system . If the sensory nerves are the target for the stimulation, placement of both electrodes over intact skin with normal innervation may enhance the healing process to a greater degree then forcing the current directly into the wound, where sensory nerves may be lacking . The healing rates reported in the present study are similar to, or greater than, those reported in most stimulation enhanced healing studies . This result was true even when the entire study sample was reported, with mean healing rates being far greater when only the wounds which responded to treatment were assessed . This finding indicates that the stimulation provided over intact tissue was as effective for enhancing wound healing as stimulation forced through the ulcer tissue itself. The advantages ofplacing electrodes beside the wound include a reduced probability of cross-contamination between the wound and the elec-

BAKER ET AL .

27

trode and decreased disturbance of the healing tissue when electrodes are introduced and removed daily. No study to date has directly compared the effect of electrode placement on healing rates . The possibility of some polar effects from stimulation with the asymmetric biphasic waveform may be found in the continuum of enhanced healing . In the present study, the asymmetric biphasic waveform was clearly superior to the microcurrent and control treatments for enhancing healing rates but was not different from the slightly lower healing rate of the wounds receiving symmetric biphasic stimulation . If the effects of the electrical stimulation were purely through the excitation ofneural tissue, the symmetric biphasic waveform might have produced a greater response because both electrodes would be equally effective in recruiting sensory nerves . There can be no polar effects with a symmetric biphasic waveform, where the positive and negative phases are mirror images of each other. Although the asymmetric biphasic waveform was electrically balanced, because the delivery time of the primary pulse was so much shorter than the secondary balancing pulse (100 psec versus 20 msec), there may be some slight polar responses, particularly in tissues deep to the electrodes . These polar effects may be relatively small but cannot be entirely discounted. The continuum of wound healing seen in the present study tends to indicate the presence of more than one mechanism through which electrical stimulation may enhance wound healing. A combined effect ofpolar and nonpolar and local and more regional responses may function together when optimal healing is achieved through electrical stimulation . An additional major finding of the present study was the presence oftwo groups ofwounds : those which responded to any form of treatment intervention and those which did not. Attempts to characterize these two groups by their descriptive data failed to distinguish between them. The stepwise regression identified the healing rate during the first week as being significantly related to the overall healing rates ofboth good and poor healing response groups . Direct statistical comparison of this variable, however, failed to identify a significant difference between the two groups . Thus, because of a high degree of variability, the early response to treatment intervention, by itself, was not an adequate predictor of the eventual outcome of the treatment . No major difference was found in the percentage of wounds subjected to the three stimulation programs which showed a good response . Thus the addition of the best stimulation protocol to a standard treatment

28


BAKER ET AL .

intervention tended to enhance the healing rate, but those wounds which healed were already slowly healing and would ultimately have done so with or without the stimulation. In the same vein, those wounds which were recalcitrant appeared to remain recalcitrant even with the addition of stimulation to their treatment. This pattern of enhancing an ongoing process with the addition of electrical stimulation was also observed in the 11 patients who were treated during both con trol and stimulation periods . There was significant healing during the control period, as evidenced by the decreased wound size at the inception of the stimulation period . Although these eleven patients were clearly healing during the control period, their healing rates increased dramatically with the addition of stimulation. A clinical recommendation could therefore be offered that the addition of stimulation to a slowly healing wound would be expected to expedite the healing process . For the wound which has failed to respond to traditional intervention, the addition of electrical stimulation would not necessarily create a responsive, healing environment. Finally, early studies reporting effects of electrically stimulated enhanced wound healing have described an extremely heterogenous sample, including patients with diabetes, spinal cord injury, various forms of brain injury, and patients with general senility and overall debilitation .' -',9 The pathophysiology of wound creation, as well as wound healing, can vary from one diagnostic group to another. The present study is the third to report a diagnostic specific sample of adults with pressure ulcers secondary to spinal cord injury .'," All three studies have been successful in showing the efficacy of stimulation in at least some part of the study sample . Thus it appears that the addition of electrical stimulation to standard wound management programs for individuals with pressure sores secondary to spinal cord injury is warranted. Further characterization of that part of the spinal cord injury population which fails to show expected responses to wound care is needed so that alternative healing techniques (e .g ., surgical intervention) can be identified in a timely fashion. Successful application of electrical stimulation to enhance healing in the population with spinal cord injuries does not guarantee its success in patients with hard to heal wounds as a result of other pathologic conditions .

ACKNOWLEDGMENTS

This study was supported in full by a grant from the National Institute on Disability Research and Reha-

bilitation, Department of Education No . H133A80040 . We thank M. J . Scofield of the Rancho Rehabilitation Engineering program and J. Burnfield of the Physical Therapy Department of Rancho Los Angeles Medical Center for their tireless assistance during data collection . We also thank M. J. Scofield and D. R. McNeal for assistance in completing and critically evaluating the manuscript .

REFERENCES

1. Carey LC, Lepley D. Effect of continuous direct electric current on healing wounds . Surg Forum 1962 ;13:33-5. 2. Wolcott LE, Wheeler PC, Hardwicke HM, Rowley BA. Accelerated healing of skin ulcers by electrotherapy: preliminary clinical results. South Med J 1969 ;62:795-801 . 3. Page CF, Gault WR . Managing ischemic skin ulcers . Am Fam Physician 1975 ;11:108-14 . 4. Gault WR, Gatens PF . Use of low intensity direct current in management of ischemic skin ulcers . Phys Ther 1976 ;56:265-9 . 5. Carley PJ, Wainapel SF . Electrotherapy for acceleration of wound healing: low intensity direct current. Arch Phys Med Rehabil 1985;66 :443-6 . 6. Kloth LC, Feedar JA . Acceleration of wound healing with high voltage, monophasic, pulsed current. Phys Ther 1988 ;68:503-8 . 7. Feedar JA, Kloth LC, Gentzkow GD . Chronic dermal ulcer healing enhanced with monophasic pulsed electrical stimulation . Phys Ther 1991 ;71:639-49. 8. Griffin JW, Tooms RE, Mendlus RA, Clifft JK, Zwaag RV, ElZeky F. Efficacy of high voltage pulsed current for healing of pressure ulcers in patients with spinal cord injury. Phys Ther 1991 ;71:433-42 . 9. Thurman BF, Christian EL . Response of a serious circulatory lesion to electrical stimulation. Phys Ther 1971 ;51:1107-10 . 10. Karba R, Vodovnik L, Presern-Strukelj M, Klesnik M. Promoted healing of chronic wounds due to electrical stimulation . Wounds 1991 ;3 :16-23 . 11 . Stetanozaka A, Vodovnik L, Benko H, Turk R. Treatment of chronic wounds by means of electric and electromagnetic fields . II . Value of FES parameters for pressure sore treatment. Med Biol Eng Comp 1993 ;31:213-20. 12. Burr HS, Harrey SC, Taffell M. Bio-electric correlates of wound healing. Yale J Biol Med 1938 ;11:103-7 . 13. Becker RO . The bioelectric field pattern in the salamander and its stimulation by an electronic analog . IRE Trans Med Electron 1960;7 :202-7 . 14. Erickson CA, Nuccitelli R. Embryonic fibroblast motility and orientation can be influenced by physiological electric fields . J Cell Bio11984 ;98:296-307 . 15 . FukushimaK, Senda N, Iuri H, Mirua H, TamaiY, Murakami U. Studies on galvanotaxis of human neutrophilic leukocytes and methods of its measurement. Med J Osaka Univ 1953 ;4:195-208 . 16 . Orida N, Feldman JD . Directional protrusive pseudopodial activity and motility in macrophages induced by extracellular electric fields . Cell Motil 1982 ;2 :243-55. 17 . Kaada B. Vasodilation induced by transcutaneous nerve stimulation in peripheral ischemia (Raynaud's phenomenon and diabetic polyneuropathy) Fur Heart J 1982 ;3 :303-14. 18 . Baker LL, McNeal DR, Benton LA, Bowman BR, Waters RL . Neuromuscular electrical stimulation-a practical guide. 3rd ed. Downey, Calif.: Los Amigos Research & Education Institute, 1993 . 19 . McNeal DR, Baker LL . Effects of joint angle, electrodes and waveform on electrical stimulation of the quadriceps and hamstrings. Ann Biomed Eng 1988;16:299-310 .