Dimethyl Sulfoxide (DMSO) Blocks Conduction in Peripheral Nerve C Fibers: a Possible Mechanism of Analgesia
M. Steven Evans,a, Kenneth H. Reid,b, and James B. Sharp Jr.c
Received June 29, 1992; Revised version received November 3, 1992; Accepted November 5, 1992.


Dimethyl sulfoxide (DMSO) is readily absorbed through skin, and relieves musculoskeletal pain when applied topically to painful areas. We studied the effects of DMSO on C-type nerve fibers, which mediate pain sensation. DMSO was applied directly to exposed cat sural nerves. C fiber conduction velocity was slowed by DMSO, even in low concentrations (5-7% v/v). Higher concentrations completely blocked C fiber conduction, with a minimum blocking concentration of 9%. Onset of nerve block was almost immediate with 15% DMSO or higher concentrations. C fiber blockade may account for analgesia with DMSO.

Dimethyl Sulfoxide, Analgesia, Pain, C fiber, Nerve block


Dimethyl sulfoxide (DMSO) is a solvent with local analgesic properties. It penetrates skin quickly,11 so for analgesia it is applied directly to skin over the painful area. Initial reports on DMSO emphasized its usefulness as a local analgesic,10 and it became a popular remedy for relief from the musculoskeletal pain of arthritis, sprains and strains,13 even though never approved by the Food and Drug Administration for this use. 7 Clinical studies of DMSO found it to be analgesic,241214 although some studies found no benefit. 16 Somatic pain is mediated by unmyelinated C type nerve fibers terminating in the skin, muscle and joint capsules. 5 We have studied a possible mechanism of DMSO-induced analgesia by observing the effects of DMSO on conduction in C fibers of cat cutaneous nerves. (Some of these results were presented previously in abstract form.)

Ninety percent DMSO in water (Eqigisic, Burlington Bio-Medical Co.) was diluted to the desired concentration (expressed as v/v) with deionized water and Hank's Balanced Salt Solution concentrate (Grand Island Biological Company).

The final concentrations were (in mM):

NaCl 136.9
glucose 11.0
NaHCO3 15.1
KCl 5.4, CaCl2 1.8
MgSO4·7H2O 0.8
KH2PO4 0.7
NaH2PO4·2H2O 0.6

The solute concentration was 280 mM, but total osmolarity was greater due to the DMSO (1588 mosmol for a 10% DMSO solution). If necessary, the solution was titrated to pH 7.4 with small amounts of NaH2CO3.

Sural nerves of twenty-five random source 2-4 kg cats were studied. Cats were anesthetized with 40 mg/kg sodium pentobarbital intravenously, supplemented as needed to maintain a level of anesthesia sufficient to eliminate withdrawal reflexes. The left sural nerve and its epineurial sheath were carefully microdissected from the underlying muscle and fascia, but left in situ. Proximally, the sural nerve remained connected to the sciatic nerve, and distally remained covered by fascia overlying the gastrocnemius muscle. The nerve's blood supply was left intact. The nerve was placed in a small bath, a slotted glass cylinder with an internal length of 2.2 cm. Bathing medium was injected into the nerve bath manually with a syringe, and removed at a rate of 0.08 ml/min with an infusion pump. The volume of the nerve bath was 0.4 ml, so that a complete change of solution took 5 minutes. A bipolar stimulating electrode was placed on the nerve proximal to the bath, and a bipolar recording electrode distal to it (antidromic stimulation). The cat limb, electrodes and bath were then enclosed in a plastid tent kept at 100% humidity, 36°C.

Stimulating and recording electrodes were fabricated from chlorided silver wires. Constant-voltage capacitance-isolated square wave stimuli, 0.5 ms duration, were given at 0.3-3 Hz. Stimulus intensity was set to twice the voltage at C fiber threshold, which ranged from 5 to 10 V (mean of 7.5 V ± 0.37 S.E.M.). C fiber responses were identified by their slower conduction velocity, smaller amplitude and higher stimulus threshold compared to A fibers. Conventional electronic devices were used for recording and displaying responses. A general-purpose microcomputer averaged 16 responses, which were then plotted with a chart recorder. The gain and timebase needed to display C responses were incompatible with simultaneously recording A fiber responses.

Figure 1

Sural Nerve Responses

sural nerve responses

A: sural nerve responses are stable. Top: C fiber waveform at beginning of experiment. A fiber responses at the beginning of the record are distorted because of the high gain (which saturates the signal averager) and slow timebase needed to demonstrate C responses. Positive is upward in all traces. Middle: response 9 h later, with no DMSO treatment. Bottom: same nerve, C fiber response blocked 10 min after application of 10% DMSO.

B: DMSO transiently increases spontaneous firing. Top: control sural nerve C fiber response. Middle: 1 min after 5% DMSO, background noise is dramatically increased by spontaneous action potential firing. Bottom: after 20 min in 5% DMSO, spontaneous firing is not apparent.

C: DMSO blockade is reversible. A nerve was exposed to DMSO during the times indicated by dotted lines. Control responses (0-20 min) varied in amplitude. 30% and 40% DMSO quickly abolished C fiber responses, with the blockades reversed by washing. In this example, 50% DMSO caused irreversible blockade.

C fiber compound action potential amplitude (most positive point minus most negative point) and latency (to most positive point) were measured. In agreement with previous investigators,56 we found that conduction velocity was stable, but C fiber response amplitudes and shapes varied from minute to minute. Because of this, we used complete block of C fiber responses as an endpoint (Fig. 1A), since this never occurred normally. Complete block was defined as loss of any organized trace of the original waveform. The sural nerve preparation was stable in long-term experiments. Two nerves were followed without drug treatment for nine hours, with no systematic changes in response amplitude or latency (Fig. 1A). During experiments, the health of the nerve was assessed by stroking hairs on the foot, causing action potentials to fire. In no case was responsiveness to this natural stimulus diminished, even after complete blockade of nerve conduction through the bath.

Experiments were performed by increasing the DMSO concentration until C fibers were blocked, then lowering the concentration until the block reversed. Immediately after DMSO application, a 1-5 min transient period of increased noise was always seen with concentrations of 5% or more (Fig. 1B). This was due to asynchronous spontaneous firing of axons, transiently reducing the amplitude of evoked C fiber responses.

DMSO effects on C fiber response amplitudes were concentration-dependent. Five percent and 7% DMSO tended to reduce C fiber response amplitudes, but the minimum blocking concentration was 9% (Figs. 1A, 2B, and C). In 14 experiments using 9% or 10% DMSO, C fiber responses were invariably reduced, and were completely blocked in 9 experiments (Figs. 1A and 2A). In these 9 experiments, blockade occurred 22.0 min ± 2.0 after DMSO application. DMSO concentrations of 15% or more reliably blocked C fibers. All nerves exposed to 15% or more DMSO were quickly blocked (mean 6.3 minutes ± 0.3, n = 15 experiments) (Fig. 2B).

Five percent DMSO relieved block caused by higher concentrations (Fig. 2A), but reversal of established blockade was never seen with 9% DMSO or higher concentrations. In 11 experiments (8 nerves), C fiber blockade was produced by 9% to 26% DMSO, followed by application of 5%; recovery was good in 10 of the 11 experiments. Because blockade was usually reversible, two or three experiments were performed on some nerves that recovered to 75% of their maximal pre-DMSO response amplitude.

Six nerves were exposed to low concentrations of DMSO for several hours (not illustrated). With such prolonged exposure, 1% to 7% DMSO blocked four nerves 6.25-8 h after application of drug, suggesting that long exposure to low concentrations of DMSO can reduce C fiber transmission. In three of the four nerves, blockade reversed after extended washing.

Figure 2

DMSO reduces C fiber response amplitude and conduction velocity.

c fiber response A: response in 5% DMSO. Middle: 10% DMSO blocked C fiber responses. Bottom: two minutes after reapplication of 5% DMSO, response returns.
c fiber response B: time-response curves for DMSO are shown. Changes are expressed as percentage of the mean amplitude of the 3-5 measurments immediately preceding DMSO application. Seven percent and lower concentrations did not block over 3-4 h. The minimum blocking concentration of DMSO was approximately 9%. Most nerves were blocked by a 55 min application of 9% DMSO, and drug application was not continued past that time. Concentrations of 15% and more blocked C responses quickly. Bars indicate standard error of the mean. For clarity, bars are not shown for 0% and 9% DMSO at times less than 35 min. For other observations, if bars are not shown, standard error or the mean is less than symbol size. Small numerals indicate number of experiments.
c fiber response C: Concentration-response curve for DMSO. Responses measured 35 min after drug application. 5% and 7% DMSO slightly reduced response amplitudes, but the difference between these and the control values was not statistically significant (two-tailed t-test). Nine percent DMSO reduced C responses by 50% after 35 min, and eventually blocked most nerves. The difference between 9% DMSO and control is significant (P < 0.05).
c fiber response D: 7% DMSO increased C fiber response latency by 43% over control in this nerve (control latency 78 ms, 7% DMSO 112 ms).
c fiber response E: response latency did not change with time.
c fiber response F: increase in latency is related to DMSO concentration, but not to duration of DMSO exposure. Five percent to 10% DMSO all caused significant increases in response latency over control (P < 0.01). There was no difference in the latencies at 5 min compared to 35 min of DMSO exposure.

DMSO reduced C fiber conduction velocity (Fig. 2D- F). Concentrations as low as 2% DMSO produced definite increases in response lattency. Unlike the effect on response amplitude, which took tens of minutes to develop with 9% DMSO, the effect on response latency was almost instantaneous. This effect was stable during the time of drug exposure, and reversed after drug removal. Response latency increased by almost 200% with 10% DMSO, implying that overall conduction velocity in the nerve segment between stimulating and recording electrodes was reduced by 50%. However, this represents a conservative estimate of the effects of DMSO, since the nerve segment studied, usually 5 cm long, was exposed to DMSO only within the 2.2 cm long nerve bath.

The mechanism of the conduction block induced by DMSO may involve potassium channel blockade. Sawada and Sato18 found that DMSO caused a rapid membrane depolarization and decrease in membrane conductance of Aplysia neurons, consistent with blockade of leak potassium channels. DMSO slowed action potential repolarization, and voltage clamp study showed that DMSO inhibited the delayed rectifier potassium channel. Their study implies that DMSO can produce a depolarizing nerve block through inhibition of both voltage-sensitive and voltage-insensitive potassium channels. Depolarizing block is consistent with results of the present experiments, in which the transient increase in electrical noise seen immediately after application of DMSO suggests that fibers have depolarized sufficiently to allow spontaneous action potential firing. How DMSO might specifically affect potassium channels is not clear, but Mayer and Avi-Dor15 have suggested that DMSO can change the state of hydration of the potassium ion, which might alter its ability to penetrate ion channels.

The mechanism of DMSO's ability to profoundly reduce C fiber conduction velocity is unclear. This effect is fast, stable, and reversible. It is detectable with low concentrations of the drug. Conduction velocity reduction may also be caused by effects on potassium channels. One possibility is that persistent membrane depolarization could facilitate sodium channel inactivation, with consequent slowing of action potential generation. A second possibility is that an increase in membrane resistance due to blockade of potassium channels might affect conduction velocity, but in this case the magnitude and direction of the effect are not easily predicted. 9 Conduction velocity could also be reduced by increasing extracellular electrical resistance,8 but under normal conditions, conduction velocity changes appreciably only with manyfold increases in extracellular resistance.

The ability of 9% DMSO to block C fibers contrasts with previous studies on myelinated A type nerve fibers, in which much larger concentrations were needed. Davis et al. 3 found that 50% DMSO was necessary to block muscle twitch in a frog sciatic nerve-gastrocnemius muscle preparation. Becker1 found that 75% DMSO blocked both A and C fibers. Shealey19 studied a prolonged small fiber afterdischarge in spinal cord, medulla and tegmentum following stimulation of the superficial radial and sural nerves, and found it to be blocked by exposing the peripheral nerve to 5-10% DMSO, a result that could be due to blockade of C fibers.

The present experiments suggest that DMSO can be an effective topical analgesic. If analgesia with DMSO depends on blockade of C fibers, one requirement is that the drug concentration around C fibers leading from the painful area be brought to 10% or more for several minutes. It is likely that this situation occurs clinically, since DMSO penetrates skin easily, and topical application of 70-90% is commonly used. Insofar as DMSO selectively affects somatic C fibers, a selective analgesia and loss of temperature sensation, rather than complete anesthesia, may be expected. Lightly-myelinated A-delta fibers that mediate sharp (first) pain were not investigated in this study, but if unaffected by DMSO, would allow sharp pain to persist, and deep, aching pain to be abolished. Deep, aching pain is present in arthritis and musculoskeletal injuries, the conditions for which DMSO has often been used.

This work is part of a thesis submitted by M.S.E. to the Graduate School of the University of Louisville, August, 1984. The authors thank Jennifer Evans and Dean Naritoku for critical reading of the manuscript.

  1. Becker, D.P., Young, H.F., Nuisen, F.E., and Jane, J.A., Physiological effects of dimethyl sulroxide on peripheral nerves: possible role in pain relief. Exp. Neurol. 24 (1969) 272-276.
  2. Brown, J.H., DMSO--its efficacy in acute musculoskeletal problems as evaluated by a double blind study. Industrial Medicine and Surgery. 35 (1967) 777-781.
  3. Davis, H.L.. Davis, N.L., and Clemons, A.L.. Procoagulant and nerve-blocking effects of DMSO. Ann. N.Y Acad. Sci. 141 (1967) 310-325.
  4. Demos, C.H., Beckkloff, G.L., Donin, M.N., and Oliver, P.M., Dimethyl sulfoxide in musculoskeletal disorders. Ann. N.Y. Acad. Sci. 141 (1967) 517-523.
  5. Douglas, W.W. and Ritchie, J.M., Mammalian non-myelinated nerve fibers. Physiol. Rev. 42 (1962) 297-334.
  6. Gasser, H.S., Unmedullated fibers originating in dorsal root ganglia. J. Gen. Physiol. 33 (1950) 650-690.
  7. Harter, J.G., The status of dimethyl sulfoxide from the perspective of the Food and Drug Administration. Ann. N.Y. Acad. Sci. 411 (1983) 1-5.
  8. Hodgkin, A.L., The relation between conduction velocity and the electrical resistance outside a nerve fiber. J. Physiol., 94 (1939) 560-570.
  9. Hunter, P.J., McNaughton, P.A., and Noble, D., Analytical models of propagation in excitable cells. Prog. Biophys. 30 (1975) 99-144.
  10. Jacob, S.W., Bischel, M. and Herschler, R.J., Dimethyl sulfoxide (DMSO): a new concept in pharmacotherapy. Current Ther. Res. 6 (1964) 134-135.
  11. Kolb, K.H., Jaenicke, G., Kramer, M. and Schulze, P.E., Absorption, distribution, and elimination of labeled dimethyl sulfoxide in man and animals. Ann. N.Y. Acad. Sci. 141 (1967) 85-95.
  12. Lockie, L.M. and Norcross, B.M., A clinical study on the effects of dimethyl sulfoxide in 103 patients with acute and chronic musculoskeletal injuries and inflammations. Ann. N.Y. Acad. Sci. 141 (1967) 599-602.
  13. MacGrady, P, The Persecuted Drug: The Story of DMSO, Doubleday and Company, Garden City, New York, 1973, 333 pp.
  14. Matsumoto, J., Clinical trials of dimethyl sulfoxide in rheumatoid arthritis patients in Japan. Ann. N.Y. Acad. Sci., 141 (1967) 560-568.
  15. Mayer, M. and Avi-Dor, Y., Interaction of solvents with membranal and soluble potassium ion-dependent enzymes. Biochem. J. 116 (1970) 49-54.
  16. Percy, E.C. and Carson, J.D., The use of DMSO in tennis elbow and rotator cuff tendonitis: a double-blind study. Med. Sci. Sports Exercise. 13 (1981) 215-219.
  17. Reid, K.H., Evans, M.S., and Sharp Jr., J.H., Analgesic-like effects of dimethyl sulfoxide on C fiber activity in the cat: evidence for more than one mechanism of action. Soc. Neurosci. Abstr. 8 (1982) 854.
  18. Sawada, M. and Sato, M., The effect of dimethyl sulfoxide on the neuronal excitability and cholinergic transmission in Aplysia ganglion cells. Ann. N.Y. Acad. Sci., 243 (1975) 337-357.
  19. Shealey, C.N., The physiological substrate or pain. Headache. 6 (1966) 101-108.
Author Information

aM. Steven Evans, Division of Neurology, Department of Internal Medicine, Southern Illinois University School of Medicine, Springfield. IL (USA).

bKenneth H. Reid, Department of Anatomy and Neurobiology, University of Louisville School of Medicine, Louisville, KY (USA).

cJames B. Sharp Jr., Department of Veterinary Medicine, University of Louisville School of Medicine, Louisville, KY (USA).


Neuroscience Letters, Volume 150, 1993, pp. 145-148., Elsevier Scientific Publishers, Ireland Ltd.

Direct all correspondance to: M.S. Evans, Box 19230, Division of Neurology, Southern Illinois University School of Medicine, Springfield, IL 62794, USA. Fax: (1) (217) 788-5567.

DMSO Organization wishes to thank Elsevier Scientific Publishers for allowing this article to be placed on our World Wide Web site. The publisher retains all copyright. To copy any portion of this article, please obtain permission from the publisher.

© 2001-2014 All rights reserved