Chapter 6.5

Methods of stimulation and measurement

	CRANIAL ELECTROSTIMULATION (CES)
	THE MEANS WHEREBY

MA or EA?

Comparisons of EA with retained needling (RN)
Studies where EA is more effective than MA
Studies where EA and MA are equally effective
Studies where MA is more effective than EA
Studies where EA is more effective than moxibustion

Focus on the needle

To TENS or not to TENS?

Studies where EA is more effective than TENS

Studies where EA and TENS are equally effective

Studies where TENS is more effective than EA

Studies where MA is more effective than TENS

Studies where TENS is more effective than MA

Other comparisons

Other forms of electrical point or meridian stimulation

OTHER METHODS OF STIMULATING ACUPOINTS AND MERIDIANS

Heat and cold
Microwaves
Millimetre waves (Extremely high frequency, EHF)
Light

Low-intensity laser and polarised light

Comparative studies of ‘laser acupuncture’

Low-intensity pulsed electromagnetic fields

Permanent magnets
Vibration
Sound
Ultrasound

POSSIBLE ADVERSE EFFECTS OF STIMULATION

Electrical stimulation

LILT

The vexed question of epilepsy

Factors having possible adverse effects on stimulation

The state of the patient
Drug interactions
Other treatments and environmental factors

EXPERIMENTAL MEASUREMENTS

Electrodermal measurements at AcPs and TrPs in response to treatment: in search of the objective

Preamble
Back to the point
Electroacupuncture according to Voll (EAV)

Dutch EAV studies
A British EAV study
An American EAV study

Other electrodermal measurement systems

Measuring temperature
Electrical imaging methods

Section summary

Measurements of magnetic fields
Box 6.5.1 A note on muscle and magnets

In Conclusion

Other icons used in this chapter:

Cranial electrostimulation (CES)

Auricular EA, so often used in the treatment of drug withdrawal, is not without precedent. As outlined in Ch. 2, the search for an effective electrical method for inducing analgesia, anaesthesia or generalised relaxation (‘electrosleep’) has resulted in a long tradition of applying stimulation to the head, generally via surface electrodes, although the relationships between the various methods used are not always clear.¹

In the mid nineteenth century, Julius Althaus (see Ch. 4, ‘late-nineteenth-century clinicians’) applied DC between the forehead and occiput. Stéphane Leduc around 1900 first used interrupted DC applied on the midline between forehead and lumbar region, and then AC.² Others, like a number of late nineteenth century clinicians,³ have explored linking forehead and leg⁴ (see Ch. 4, ‘late-nineteenth-century clinicians’) or forehead and wrist,⁵ and even mouth and anus⁶^,⁷ (not recommended!). By the early 1950s, at least in the former Soviet Union, electrode positions had become standardised: two electrodes (generally cathodes) over the forehead or closed eyes, and two (the anodes) at the occiput.

However, as one influential American researcher, Robert Smith, pointed out,⁸ in dogs it was a lot easier to get the electrodes to stay on if they were positioned bitemporally. With the development of auricular acupuncture, a transverse location became more popular, placed just in front of the top of the ears (Saul Liss), on the occiput just behind the ears (Margaret Patterson, Bob Beck), or of course between AcPs in the ears themselves (Wen Hsiang-lai, Lorenz Ng). In the former USSR the forehead/occiput arrangement remains common. One commercially available Swiss device even uses an intraoral electrode! Cranial and body AcP locations have been compared; Liss determined that changes in circulating neurochemicals were greater using the transcranial locations than LI-4 bilaterally, for instance.⁹

All in all, apart from the obvious division into two ‘families’ of CES, with electrodes positioned either ‘fore and aft’ or transversely, it would seem that the actual locations used for stimulation (whether AcPs or not) might be less significant than the electrical parameters of the stimulation,¹⁰ provided the factor of skin impedance is overcome.¹¹

Perhaps because Leduc began by using 100 Hz, most Russian electrosleep (‘electronarcosis’ or ‘central electroanalgesia’) research also used this frequency, although probably only because, with the large currents involved at that time, it was high enough to avoid muscle twitching.¹² Wen too used frequencies around 120–125 Hz (see SubCh. 6.3, ‘no effect on plasma or CSF βEP in addicts’), again only because of historical accident: the output of the first Chinese EA device that came to hand went no higher than 111 Hz.¹³ Even so, results of a number of early controlled studies seemed to favour 100 Hz¹⁴ over other frequencies, as has research on sinewave cranial TENS that supports the use of 100 Hz (as opposed to 5 Hz or 2 kHz), in terms of its effect both on blood pressure (BP) and heart rate (HR)¹⁵ and on experimental trigeminal pain.¹⁶

However, some researchers have felt that a slower, rhythmically repetitive stimulation would be more effective¹⁷^,¹⁸ (see SubCh. 6.3, ‘animal hypnotism’), with frequencies around 100 Hz considered to be more ‘activating’.¹⁹ There are different interpretations of ‘low’ frequency in this context; in Russia, for instance, frequencies from 1–2 Hz up to 100–130 Hz are considered by some authorities as suited to inhibition of CNS activity,²⁰ whereas in one small informal American study, 100 Hz (0.5–1.5 mA, 1 ms) was unexpectedly found to enhance alertness and even induce a sense of euphoria in some subjects.²¹

Applying strong currents transcranially can result in adverse effects such as spasm and even convulsion (below, ‘may well induce convulsions’). Attempts were made to reduce these by ramping current intensity,²² combining biphasic and monophasic currents (Anan’ev,²³ Lebedev²⁴^,²⁵ and others²⁶), or using high (Knutson 700–1500 Hz) or very high frequencies (Liss 15 kHz, Kuster 20–100 MHz,²⁷ Limoge 100 or 166 kHz), generally modulated at lower frequencies. For this purpose, Saul Liss has used 15 Hz (‘the bioactive frequency’²⁸) and Aimé Limoge 100 Hz²⁹ or 77 Hz³⁰ (77 Hz has also been used in a Russian device³¹). Magnes’s Israeli group, for instance, found a 3.5–4.5 kHz sinewave modulated at 100 Hz gave analgesia together with signs of anaesthesia (stimulation through electrodes placed at various locations actually in the brain stem resulted in analgesia only).³²

Liss, in experiments on GABA synthesis, defined the optimum parameters for his ‘Liss cranial stimulator’ as 15 kHz modulated at 15 Hz, adding in 500 Hz to reduce the amount of charge transferred by 50%³³ (even though effects were not quite as good, changes in circulating neurochemical levels were better than those obtained using 80 Hz TENS³⁴). Lebedev’s Moscow group, in contrast, found simple 70 Hz (3–3.5 ms) rectangular pulses combined with DC (with a ratio of DC to pulse current of 2:1) to be optimal for analgesia in rats as well as other animals.³⁵ Some American authors considered 100 Hz sinewaves produce greater analgesia, with 700 Hz – 10 kHz resulting in deeper relaxation.³⁶

Over the years, more and more CES devices have become available that have effects even when the currents applied are below the threshold for sensory stimulation (Ifor Capel, Daniel Kirsch), at 1 mA or below.³⁷ Although no longer producing sufficient analgesia for surgical procedures (as with EAA, this requires strong stimulation), they have useful applications and are safe to use in everyday clinical practice as well as by patients at home. Even where electroanaesthesia is concerned, as one prominent early Russian researcher stated, ‘it is not the amount of electric current flowing through the brain that matters, but the excitatory, stimulating effect of a steep increase in current’.³⁸

Thus waveform may in some circumstances be more important than intensity. Most CES systems use squarewaves in some form.³⁹ However, one American device using 100 Hz sinewave stimulation has been well researched,⁴⁰^,⁴¹^,⁴² a few have used random noise rather than repetitive signals⁴³^,⁴⁴ and one a beat pattern between two different biphasic spike frequencies (0.4 and 0.5 Hz) that only repeats every 20 seconds or so.⁴⁵ EEG-modulated trains have also been employed, giving better effects in cats (in terms of relaxation) than unmodulated trains.⁴⁶

Treatment duration may also be an important factor: 20 minutes of Russian style CES inhibits high-threshold nociceptors, short (5-minute) or very short (1-minute) treatments inhibit only medium- or low-threshold mechanoreceptors respectively.⁴⁷ Furthermore, in transverse CES systems using DC or asymmetric biphasic stimulation (even if charge balanced), which side of the head is positive (generally the right) can be an important factor.⁴⁸^,⁴⁹ Indeed, it is noteworthy that unilateral stimulation may be quite ineffective using some forms of CES.⁵⁰ Hence, whereas forehead to limb electrode arrangements may produce results (above, ‘forehead and wrist’), single ear to foot or tail may not.⁵¹

Ifor Capel, with colleagues in both Britain and the USA, has carried out a meticulous series of studies over more than 20 years using earlobe CES. He calls this SPES (for ‘subperception electrical stimulation’). Unlike many forms of CES, stimulation is applied via needle electrodes, but these only penetrate the epidermis (not the dermis⁵²) sufficiently to overcome skin impedance differences. The stimulation itself is very precisely defined, and is one that Capel’s team has found to be very effective for pain. It consists of a biphasic rectangular charge-balanced waveform with a first (positive) phase duration of 2–3 ms, amplitude 10 µA and repetition rate 10 Hz. These parameters were arrived at after testing a wide range of options: frequencies of 7.5 or 50 Hz were virtually ineffective, for instance, and other currents in the 5–20 µA range also less effective,⁵³^,⁵⁴ as were sine or spike waveforms.⁵⁵^,⁵⁶

To test for pain, Capel’s coworkers have used the rat tail flick latency (TFL) model, among others (subthreshold vagal stimulation is well known to decrease TFL⁵⁷). In contrast to Ng’s auricular EA (see SubCh. 6.3, ‘animal hypnotism’), the effects of this form of CES on opiate withdrawal treatment appear quite robust and immune to environmental disturbances. However, excessive noise or disturbance can affect results when working with pain, perhaps because of the sensitivity of the TFL model to stress (see SubCh. 6.3, ‘Stress also may invalidate HP or TFL testing’)⁵⁸^,⁵⁹ (this is true of other pain measures as well, such as the radiant heat snout test⁶⁰).

SPES may indeed be more potent in its antinociceptive effects than focal electrical stimulation of the NRD or lH within the brain,⁶¹^,⁶² with particular effects on nociceptive activity in the habenula,⁶³ and NPf.⁶⁴ SPES analgesia, like LF EA (see SubCh. 6.4, ‘a delayed onset’), is slow in onset (although this may not always be the case⁶⁵) and with an after-effect that may last several hours, possibly with cycles of greater and less effect until dying away altogether.⁶⁶

With SPES administered before anaesthetic agents such as hexobarbital, the 10 Hz signal also decreased the sleeping time of the anaesthetised rats, as did 500 Hz (although curiously only the latter was effective if applied peripherally via the paws).⁶⁷ In contrast, pentobarbital may have a reverse effect, negating SPES antinociception if given before SPES is started.⁶⁸^,⁶⁹ SPES may also affect blood sugar levels and appetite, in both rats and humans.⁷⁰ The 10 Hz frequency (but with a stimulus very different from SPES in other respects) has also been found to enhance immune function in stressed rats, although 1 kHz proved more effective prophylactically.⁷¹ Other forms of CES may also have immune effects,⁷² with strong electronarcosis temporarily affecting lymphocyte: Neutrophil ratios, for example.⁷³

A series of Japanese–American studies has explored the role of scalp sensory nerves in CES.⁷⁴^,⁷⁵^,⁷⁶ Although the simplest pathway between cranial electrodes is over the skin surface, a certain amount of current (maybe a fifth,⁷⁷ or even more⁷⁸^,⁷⁹) does in fact penetrate the skull via more complex pathways (through the orbital and occipital foramina, via the auditory meatus, or by stimulating superficial branches of the cranial nerves – V, VII, IX or X in the case of auricular stimulation⁸⁰). Thus even when electrodes are applied only on the forehead (in an attempt at ‘sham’ CES), current may penetrate the skull particularly if points of low skin resistance (SR) over cranial nerve branches are used.⁸¹ Skull thickness is not important.⁸² Also, if the cranial nerves are stimulated their cell bodies may be directly activated, with a cascade of effects in the brainstem where they are located; that some forms of unilateral CES (but not all, see above) are in fact effective would seem to support this possibility.⁸³

The SPES signal in particular may penetrate the brain and directly affect the CNS,⁸⁴ as preliminary peripheral nerve block studies did not seem to reduce its efficacy.⁸⁵ This seems to be the case for other forms of CES as well. In rabbits, for instance (using 4 and 4.2 kHz with a resulting 200 Hz interference current), endogenous electrical activity was found to be affected primarily in the hypothalamus, thalamic centromedian nucleus (NCm), periaqueductal grey (PAG) matter and reticular formation – areas also implicated in EAA. However, in dead rabbits current only flowed over the skull surface.⁸⁶ In rats, sinewave CES has been shown to affect metabolic activity in the PAG, solitary tract nucleus (NST), reticular formation, trigeminal nucleus and elsewhere.⁸⁷ Theoretically, CES should affect the thalamus;⁸⁸ indeed in monkeys some 42% of the applied current from strong (100 mA, 20–2000 Hz) TCET enters the brain as a whole, with slightly higher current concentrations measured in the thalamus compared with those in the cortex (although this difference may not be significant).⁸⁹ Currents have also been detected during such treatment in the frontal lobes and brainstem.⁹⁰

Methods have thus been developed for positioning electrodes in order to maximise current density,⁹¹^,⁹² particularly in the thalamus.⁹³ One study noted that synchronised 5–7 Hz discharges persisted in non-specific thalamic regions after 5–10 Hz CES, later spreading to other regions of the brain as well,⁹⁴ and changes in thalamocortical EP with strong CES have also been investigated.⁹⁵ Current changes have been measured too in the hippocampus and associated areas.⁹⁶ With strong cranial electroanaesthesia,⁹⁷ autoradiography has showed greatly increased glucose metabolism in the ventral PAG (vPAG), red nucleus and cerebellar cortex (rather different regions from those involved in EA).

These changes do not appear to damage the brain physically at all.⁹⁸^,⁹⁹ Early reports of adverse effects when electrodes were used over the eyes¹⁰⁰ turned out to be merely due to mechanical pressure.¹⁰¹

Distally applied EA influences cerebral circulation and the EEG. It is thus hardly surprising that rhythmic treatment applied directly to the head has effects on cranial circulation,¹⁰²^,¹⁰³^,¹⁰⁴ and the EEG, as reported in many papers presented at the first international electrotherapeutic sleep and electroanaesthesia symposium in Graz in 1966,¹⁰⁵ and elsewhere.¹⁰⁶^,¹⁰⁷^,¹⁰⁸^,¹⁰⁹^,¹¹⁰^,¹¹¹^,¹¹²^,¹¹³^,¹¹⁴^,¹¹⁵^,¹¹⁶^,¹¹⁷^,¹¹⁸^,¹¹⁹^,^120,¹²¹ However, EEG changes have not always been detected with 100 Hz CES, despite treatment effectiveness.¹²²

Notwithstanding this plethora of studies, the EEG changes found with CES are not always easy to interpret,¹²³ although in general they do appear to mirror the state of the subject (excitable, drowsy, etc., but not necessarily completely asleep) and would seem to indicate that CES effects are not just the result of suggestion¹²⁴ (but see the critique by Douglas Taylor¹²⁵). The evoked potential (EP) changes that occur with CES are also complex, although it is claimed that CES can affect the amplitude of the P300 EP, which is supposedly an important marker of drug abuse and risk for drug abuse.¹²⁶ In one SPES study of four heroin abusers, for instance, increased P300 amplitude was interpreted as indicating enhanced ‘cognitive awareness’.¹²⁷

With subthreshold stimulation, these findings become more interesting. Thus, using the variable biphasic spike waveform of the Alpha-Stim 100 set at an ‘average’ 0.5Hz frequency, Michael Heffernan found evidence of increased correlation dimension (see SubCh. 5.1, ‘variability expressed in mathematical terms’) and beneficial ‘spectral smoothing’ (smaller standard deviations in the EEG FFT spectrum) in a double-blind placebo-controlled protocol. He was surprised to note better spectral smoothing when the signal was applied through bilateral trapezius AcPs (low SR points) rather than via the earlobes, although correlation dimension was comparable at the two sets of points.¹²⁸ (This is somewhat in contrast to Liss’s comparison of cranial and LI-4 effects, mentioned above.)

Heffernan reports increased power in the alpha range with both 0.5 and 1.5 Hz frequencies, but possibly even some sympathetic arousal with 100 Hz.¹²⁹ (Daniel Kirsch, creator of the Alpha-Stim device, suggests that frequencies above the conventional EEG range should not be used.¹³⁰) Even though at levels too low to evoke electrophysiological activity directly,¹³¹ SPES too may affect the EEG, especially increasing posterior theta and frontal delta, in line with increased drowsiness in normal subjects.¹³²

Skin conductivity, like the EEG, changes in keeping with greater relaxation when true (but not sham) CES is applied.¹³³^,¹³⁴^,¹³⁵ Heart rate, EMG and finger temperature have also been shown (in a pilot single-blind study) to change significantly with low-intensity CES,¹³⁶ and in a double-blind study this improved concurrent thermal biofeedback training (in patients with migraine).¹³⁷ Other studies too have reported EMG changes indicative of relaxation,¹³⁸^,¹³⁹^,¹⁴⁰ or a general anxiolytic response,¹⁴¹^,¹⁴² with significant reductions in HR, anxiety and other measures in one single-blind placebo-controlled study of 100 Hz sinewave CES, for example.¹⁴³ CES may also reduce stomach acidity in monkeys,¹⁴⁴^,¹⁴⁵^,¹⁴⁶^,¹⁴⁷ as well as in humans¹⁴⁸ (with currents of 0.9 mA or more applied transcranially, but not if just frontally). Also in humans, CES may regulate gastrointestinal motility,¹⁴⁹ and reduce muscle spasticity¹⁵⁰^,¹⁵¹^,¹⁵² (perhaps because of its effects on GABA¹⁵³), as well as tremor or other involuntary movements¹⁵⁴ (as part of a general ‘quieting response’¹⁵⁵). In contrast, those withdrawing alcoholics who tremor little initially may tremor more with CES.¹⁵⁶

Inhibition of blood pressure increase in rabbits, rats and cats subjected to nociceptive stress has also been noted,¹⁵⁷^,¹⁵⁸ as well as beneficial changes in BP¹⁵⁹^,¹⁶⁰ and peripheral circulation in humans.¹⁶¹^,¹⁶²^,¹⁶³ On the one hand, decreased blood pressure occurs after about 20 minutes of SPES treatment, for instance, concurrent with a reduction in salivary cortisol. On the other, with the strong currents used in electroanaesthesia, BP predictably rises (but less so with sine than square or triangular waves, and less in any case than with neuroleptanalgesia).¹⁶⁴^,¹⁶⁵

A number of other intraoperative stress measures may also increase with strong CES compared with chemical anaesthesia. In one study of strong very-high-frequency CES for anaesthesia, for example, stress markers still increased despite the concurrent use of neuroleptic, benzodiazepine, curare and nitrous oxide/oxygen medication,¹⁶⁶^,¹⁶⁷ and one early researcher even felt the cardiovascular complications of strong CES were too serious to warrant continued investigation in humans.¹⁶⁸

As so often with EA, CES appears to have some effect only when homeostasis is disturbed. For instance, Limoge’s group found that morphine analgesia in rats was potentiated by CES, but that this had no effect on pain threshold (PT) in drug-free animals,¹⁶⁹^,¹⁷⁰ and Eric Braverman noted that changes in the direction of EEG normalisation were more marked in subjects where this was initially abnormal.¹⁷¹ Other EEG changes have been reported to occur only during a reaction time test, with no difference between real and sham CES evident while resting quietly.¹⁷² Counterintuitively, in another study using a different form of CES (100 Hz sinewave), whereas anxiolytic effects were ‘interesting’ in subjects at rest, they were not impressive during a brief stressful procedure.¹⁷³

Within the brain, SPES suppresses noxious but not spontaneous habenular activity,¹⁷⁴ and appears to have no effect on neurotransmitter levels in rats where these are normal.¹⁷⁵ (However, although overall concentrations may not change, the presence of increased breakdown products indicates that turnover of some neurotransmitters is still accelerated in normals.¹⁷⁶)

The regulatory effect of CES, as sometimes with EA, may depend on the extent to which homeostasis has been disturbed. For instance, CES may have a greater effect on the spasticity of cerebral palsy when this is more severe.¹⁷⁷

Furthermore, as with EA, there seems to be considerable individual variation in effectiveness with CES,¹⁷⁸^,¹⁷⁹ as well as variation between sessions in the same individual.¹⁸⁰

Again as with EA, results may be delayed, cumulative¹⁸¹ and last for hours¹⁸² or even days after treatment.¹⁸³^,¹⁸⁴

Some CES may affect levels of β-endorphin (βEP)¹⁸⁵ and met-enkephalin (ME)¹⁸⁶ in the cerebrospinal fluid (CSF). Opioid mechanisms are probably implicated in CES analgesic effects that involve the PAG¹⁸⁷^,¹⁸⁸ as well as peripherally.¹⁸⁹ Opioids may also be involved in CES effects on withdrawal.¹⁹⁰ Limoge’s CES analgesia appears to be reversed by naloxone,¹⁹¹ for instance, although not all CES appears to involve endorphin mechanisms.¹⁹²^,¹⁹³

For example, whereas 10 Hz SPES may have some naloxone-reversible effects¹⁹⁴^,¹⁹⁵^,¹⁹⁶ (more clearly so with high doses of naloxone¹⁹⁷), it shows no cross-tolerance with morphine,¹⁹⁸ and 500 Hz SPES may in fact be enhanced by naloxone.¹⁹⁹ Furthermore, plasma βEP levels seem unrelated to SPES analgesia.

As already described, EA shows a significant decrease in effect (tolerance) with continued use. This is the case for stress-induced analgesia (SIA) as well, but not for SPES,²⁰⁰^,²⁰¹^,²⁰² which may itself even inhibit the early effects of SIA.²⁰³ The optimal treatment duration for SPES is around 30 minutes, and SPES analgesia lasts for a further 200 minutes beyond that.²⁰⁴ Thus, if treatments are repeated more frequently than 3-hourly, different neurochemical pathways may become involved. In these circumstances SPES may have more obvious endorphinergic effects.²⁰⁵ In line with its dose-dependent interaction with naloxone, it has been suggested that SPES may possibly involve dynorphin rather than βEP receptors.²⁰⁶

With Saul Liss’s variant of CES (15 kHz/500 Hz/15 Hz), βEP increases in both CSF and blood plasma.²⁰⁷^,²⁰⁸ This also occurs with VHF high-intensity currents (167 kHz/77 Hz, 250–300 mA).²⁰⁹ Other forms of CES may also increase plasma βEP.²¹⁰ This was the case in one Italian study, for instance, where 4 Hz CES (sufficient to produce intense tingling – at a mean of 4.9 mA) was utilised.²¹¹ However, this does not always happen.²¹²

Russian researchers have found, as has Ifor Capel, that plasma opioid peptide (OP) changes may be inconsistent, although plasma ACTH may remain fairly steady during surgery under electroanaesthesia (almost as though the procedure were not experienced as stressful).²¹³ With SPES, over 2 hours plasma ACTH may initially rise and then decline²¹⁴ (the level correlating with its antinociceptive action²¹⁵). With the Liss device, plasma ACTH increases.²¹⁶

Some Russian forms of CES may reduce hypothalamic secretion of corticotropin-releasing factor (CRF) and vasopressin (Vas), both of which are associated with stress.²¹⁷ Nevertheless, intra- or postoperative use of electroanaesthesia,²¹⁸ even if not hugely strong (15 Hz, 30 mA, applied paraspinally),²¹⁹ may still result in raised levels of urinary stress hormones such as cortisol, noradrenaline (NA, norepinephrine) and adrenaline (Adr, epinephrine), and possibly even growth hormone (GH). However, this does not conclusively indicate that the procedure is experienced as stressful: GH may also be inhibited by stress, and the use of urinary NA and Adr as stress markers has been disputed.²²⁰ Changes in levels of CRF, Vas and GH may themselves be mediated by OPs.²²¹

Effects may be very different with an intervention like SPES, at the other end of the intensity spectrum. Yet, despite the fact that SPES provides subthreshold stimulation, cortisol and ACTH are clearly implicated²²² (as with auricular EA in the treatment of addiction (see SubCh. 6.3, ‘significantly reduced following auricular EA in addicts’)) although their levels do not necessarily change in parallel.²²³ Whereas both the pituitary and adrenal cortex are involved (but not the adrenal medulla²²⁴), freely circulating cortisol generally increases, yet without enhancing hypothalamic or hippocampal receptor availability.²²⁵ In other words, the cortisol increase is not associated with increased stress,²²⁶ and any cortisol-like agents that bind to type I or type II corticosteroid receptors in these brain areas may interfere with SPES.²²⁷ Salivary rather than plasma cortisol level may in fact be a more sensitive indicator of the anxiolytic efficacy of SPES.²²⁸^,²²⁹ However, plasma cortisol levels do appear to decrease with use of the Liss device.²³⁰^,²³¹

Serotonin (5HT) may also be involved in some CES analgesic²³²^,²³³^,²³⁴^,²³⁵ and antidepressant effects.²³⁶ SPES analgesia, for instance, is partially reversed by the 5HT antagonist pCPA,²³⁷ and may increase synthesis or release of midbrain 5HT, dopamine (DA) and NA, as well as hypothalamic 5HT and DA.²³⁸^,²³⁹ However, these 5HT changes may be inconsistent and only transient.²⁴⁰

Whereas NA concentration may not change noticeably overall (SPES may inhibit anxiety-associated NA, but then increases NA turnover compared with sham treatment), its principal breakdown product MHPG (3-methoxy-4-hydroxyphenylglycol) does increase significantly in blood, urine and various brain regions.²⁴¹^,²⁴²^,²⁴³^,²⁴⁴ However, these increases are not necessarily in line with PT changes, for instance,²⁴⁵ so are not necessarily diagnostic of efficacy.²⁴⁶ Nevertheless, low baseline monoamine levels do decrease SPES efficacy.²⁴⁷ (It is perhaps an intriguing irrelevance that the locus coeruleus, the source of much noradrenergic activity, is essential to animal hypnosis.²⁴⁸)

Histamine too appears to play a key role in SPES, both centrally and peripherally²⁴⁹ (and possibly in other forms of CES²⁵⁰).

Catecholamines have been implicated in the circulatory effects of older, stronger forms of electronarcosis, too,²⁵¹^,²⁵² although changes in blood levels of 5HT and DA, for instance, are not always found.²⁵³

Serum cholinesterase may decrease significantly when the Liss pain suppressor is used to treat depressed patients.²⁵⁴

As with EA (see SubCh. 6.2, ‘also protect against degradation of SP and CCK’), some enkephalinase inhibitors may potentiate Russian-style CES,²⁵⁵ Limoge’s CES²⁵⁶ and SPES analgesia in rats.²⁵⁷^,²⁵⁸ SPES may also be enhanced by L-tryptophan,²⁵⁹ although this has been questioned.²⁶⁰ Using the Liss device, plasma levels of L-tryptophan may decrease (with increases in 5HT in both CSF and plasma), while plasma DHEA and GABA increase. The Liss device is claimed even to increase GABA levels in vitro.²⁶¹^,²⁶² Blood GABA levels may also increase with Russian-style CES.²⁶³

Other forms of CES have shown short-term plasma decreases and longer-term (1-week) increases in L-tryptophan.²⁶⁴ 5HTP and (surprisingly, see SubCh. 6.2, ‘depressing the nociceptive response’) allopurinol may enhance some CES analgesia.²⁶⁵

Strong CES may temporarily reduce blood calcium and phosphorus levels,²⁶⁶ and calcium channel blockers may totally inhibit SPES efficacy whereas caffeine may increase it.²⁶⁷^,²⁶⁸ And there may as indicated above be interactions between barbiturates and SPES. As with EAA, the combination of strong CES with premedication for electroanaesthesia²⁶⁹ or with concurrent chemical anaesthesia has been explored.²⁷⁰

Importantly, SPES at least appears to be unaffected by lesion of the usual descending pain modulation pathway, the dorsolateral funiculus,²⁷¹ which indicates that others must be involved.

Some CES studies have been carried out on experimental pain,²⁷² as well as on experimentally induced motion sickness (nausea and parasympathetic cardiovascular reactions),²⁷³^,²⁷⁴ just as they have been with P-6 EA.

Unusual (non-cutaneous) sensations in the head with one form of sinewave CES have been found to correlate with peripheral sensory nerve impairment (in dialysis patients).²⁷⁵ In HIV patients without overt neurological symptoms, sensory lack of persistence to this type of stimulation (an inability to maintain perception of the stimulus over repeated trials) may reflect subclinical impairment, increasing significantly with the severity of HIV infection.²⁷⁶ Such sensations varied specifically with the sinewave frequency used and were not obtained with rectangular waveforms, but did occur with squarewaves although not in a frequency-dependent manner.²⁷⁷

Because some forms of CES include a DC component, ions can be introduced through the skin if they are present in the electrolyte beneath the electrodes. This process, iontophoresis, has been used experimentally with lithium chloride in dogs (lithium was found to be present in many parts of the brain afterwards), so avoiding the systemic toxicity associated with its oral administration.²⁷⁸

The similarities and differences between EA (particularly auricular EA) and CES are tantalising. In some ways they have similar effects (described by one researcher with intimate knowledge of both as ‘modest’²⁷⁹) and probably share some neurochemical pathways, but not all (it is highly significant in this regard that SPES continues to be effective even after dorsolateral fascicle lesion). And clearly there are different ‘families’ of CES, from strong and potentially dangerous, rather than therapeutic, to subthreshold and almost unmeasurable. The divisions between strong and weak, HF and LF will already be familiar from the discussion of EA parameters above.

The means whereby

MA or EA?

	‘While EA has the advantage of precise characterization of stimulation parameters, it could hardly replace manual needling at least in the present stage.’
	Han Jisheng²⁸⁰

An important question is how EA and MA (manual acupuncture) differ, both in mechanism and in effect. Although it is tempting to compare EA studies with those on MA, the very different procedures adopted by different researchers could lead to conclusions that are not particularly useful. Here the focus is on those studies where both EA and MA were used in similar protocols. Even so, given the variety of EA approaches covered, the results of the following survey should be taken only as pointers to appropriate clinical practice, and not necessarily as hard and fast evidence one way or the other.

Comparisons of EA with retained needling (RN)

EA may be more effective than simply inserting a needle at the same point for the same period (retained needling, RN). There may be different local reactions²⁸¹ and also central ones. For example, EA produces frequency-dependent effects on opioid messenger RNA (mRNA) levels, whereas RN does not,²⁸² spinal cord levels of met-enkephalin (ME) and dynorphins (Dyn) are enhanced by EA but not RN,²⁸³ and 100 Hz ST-36/SP-6 EA reduces spinal cord C-fibre evoked potentials (EPs) more than RN.²⁸⁴

EA is more effective than RN in the radiant heat avoidance test in rabbits.²⁸⁵ EA (unilateral) also more effectively reduces arthritic pain in rats,²⁸⁶ or inhibits a vibration-induced finger reflex (with needles insulated except at the tip giving the best results).²⁸⁷ EA also has a greater effect on regional cerebral blood flow.²⁸⁸ And EA (2/15 Hz, 1–2 mA, 300 µA; GB-24, LIV-14) has more effect on the motility of Oddi’s sphincter.²⁸⁹

Studies where EA is more effective than MA

Auricular EA (7 Hz) more effectively affects CNS endorphin levels than MA.²⁹⁰ In particular, EA may affect cerebrospinal fluid (CSF) βEP (see SubCh. 6.2, ‘used as measures of PT’) or Dyn levels more than MA (see SubCh. 6.2, ‘found in response to MA given antenatally’). Whereas EA may increase plasma βEP, β-lipotropin and ACTH, MA has only a transient effect on the first two²⁹¹ and, although plasma βEP levels increased significantly with both LF EA and MA in another study, there was a correlation with pain relief only with the EA.²⁹² GH changes with LF EA were not found with RN in one Australian report.²⁹³ Repeated EA treatment may significantly increase hippocampal levels of substance P (SP), neurokinin A (NKA) and neuropeptide Y (NPY) as well as elsewhere in the brain, but such changes do not occur with repeated MA (or exercise).²⁹⁴ EA significantly increased CSF calcitonin gene-related peptide (CGRP) in normal adult rats, as well as non-significantly in serum, urine and some parts of the brain, but MA had no such effects.²⁹⁵

The effects of HF EA on the EEG (decreasing slow activity) may be quite different from those of MA (affecting only EEG frequencies above 90 Hz).²⁹⁶ Cortical EPs changed more in response to LF EA at contralateral ST-36/ST-40 than with strong bilateral ST-36 MA in rats (a rather uneven comparison, however).²⁹⁷

In rats, EA prolongs TFL more than MA and RN,²⁹⁸^,²⁹⁹ with little difference between the last two³⁰⁰ (although vocalisation threshold, which involves supraspinal pathways, may be the same for EA and MA³⁰¹). In humans, EA (100 Hz, 4 ms, triangular biphasic waveform, sufficiently intense to cause cause strong but not painful tingling at needles inserted 5 mm apart at LI-4, ST-2 and ST-44) is more effective than MA in increasing dental pain threshold.³⁰² EA is also more effective than MA at reducing subjective intensity of experimental wrist pain.³⁰³ Whereas MA alone may decrease experimental pain, the addition of electrical stimulation (in the form of TEAS) can make such decreases statistically significant, and as marked as those experienced with intramuscular morphine.³⁰⁴^,³⁰⁵ LI-4 EA has also been reported as decreasing experimental lip pain more than MA (with corresponding fMRI differences).³⁰⁶ Bilateral LF ST-36/ST-37 EA reduces EP responses to right median nerve stimulation more than MA at the same points.³⁰⁷ In rabbits, EA (10 kHz, 200 µA) resulted in more widespread analgesia than MA³⁰⁸ and was also more practical (in this particular protocol, time to onset of adequate analgesia was 45 minutes using MA, and the effect lasted only a few minutes once manipulation was stopped).³⁰⁹

MA in the form of RN may have no effect on PT.³¹⁰^,³¹¹ Thus in one French study, RN did not affect human flexion reflex to pain, whereas 500 Hz LI-4 EA did.³¹² In a Japanese report, LF EA increased PT, whereas RN did not.³¹³ However, MA may elevate PT (in monkeys),³¹⁴ although not significantly (in humans),³¹⁵ and very strong EA 3 Hz, 0.2 ms, ‘from uncomfortable to frankly painful’, was also found in one very small study not to affect PT.³¹⁶

EA may more effectively increase microcirculatory blood flow in the cerebral pia mater than MA.³¹⁷ On the other hand, in healthy humans, although MA was not found to affect skin blood flow in one comparative study, EA (2 Hz, motor level, at left LI-4 and LI-10) temporarily reduced it before it increased again (significantly) following treatment.³¹⁸ (Perhaps this reduction occurred because of the more continuous nature of EA, or because it was experienced as ‘uncomfortable’.)

Pretreatment with EA (5 Hz, 5 mA) has more effect than MA on experimental paw oedema in rats, possibly because of the greater intensity of stimulation possible.³¹⁹

In rats, auricular EA decreased the duration of cardiac arrhythmia in acute ischaemia and subsequent reperfusion, whereas MA did not although decreasing its severity.³²⁰

Non-noxious 100 Hz EA may enhance leukocyte count more than MA (even at non-AcPs).³²¹

High-intensity EA (2 or 80 Hz, 0.2 ms, 20 mA) was more effective than MA (and superficial acupuncture, RN) in enhancing survival of ischaemic musculocutaneous flaps in the rat (all treatments were for 60 minutes).³²² EA may more effectively enhance peripheral nerve regrowth than MA, even in the presence of neurotmesis.³²³

EA is more effective than MA in reducing canine gastric acidity and enhancing gastric bicarbonate and sodium levels.³²⁴^,³²⁵ Likewise in humans, EA (5 Hz, BL-21, ST-36, Ren-12) reduces sham-feeding-enhanced gastric acidity more than MA.³²⁶ Auricular EA (20 Hz, moderate motor level) enhanced bile secretion in rats more effectively than MA at the same points.³²⁷

EA may be more effective than MA in reducing muscle spasticity³²⁸ and experimentally induced itch.³²⁹

The general conclusion has been drawn that EA is more effective than MA in stimulating nerves, so should be used if this is the aim of treatment.³³⁰

Studies where EA and MA are equally effective

Both intense EA and MA may excite primary C-afferent fibres (and so activate diffuse noxious inhibitory control, DNIC).³³¹ At a less noxious level, both ALTENS and ~CEA (intramuscular electrical stimulation via wire electrodes) consistently induce marked and long-lasting elevations of ipsilateral muscular PT.³³² In one very small study (N = 4) involving an experienced acupuncturist, neither EA (3 Hz, 0.2 ms, for 60 minutes, motor level from ‘uncomfortable’ to ‘distinctly painful’) nor MA (RN) at bilateral points selected as suitable for thyroidectomy (GB-21, LI-4) or thoracotomy (SJ-8, LI-14), or at sham points had any appreciable effect on PT, either during stimulation or at 15 or 30 minutes afterwards.³³³^,³³⁴

MA and 8 Hz EA at Du-26 may both rapidly enhance rat brain oxygenation.³³⁵^,³³⁶ In one randomised controlled trial (RCT), both EA and MA at bilateral GB-30 had equivalent effects on enzyme levels in the rat posterior pituitary;³³⁷ similar results were found for pituitary intermediate lobe enzymes, using 2 Hz manual needle rotation for 30 minutes (!).³³⁸

EA and MA at body points both increased EMG amplitude in patients with ischaemic cerebrovascular disease (whereas scalp acupuncture did not).³³⁹

LF low-intensity EA was found to be equivalent to MA in one study of AA in rabbits,³⁴⁰ and MA and EA resulted in very similar levels of analgesia in mice (as measured by the phenylquinone-induced writhing test); hypophysectomy did not alter EAA.³⁴¹

Studies where MA is more effective than EA

Deqi sensations of distension and heaviness, or soreness (conveyed by group III and IV fibres, respectively) may be elicited more readily with MA than with EA.³⁴²

In dogs, Du-26 MA may more effectively increase cardiac output, stroke volume, heart rate and mean arterial pressure than EA³⁴³ (although ‘electrocautery’ was superior to both³⁴⁴). Even acupressure at Du-26 may greatly increase cardiac output and arterial pressure, but again EA has little such effect (though more than RN).³⁴⁵

In one detailed study,³⁴⁶ plasma progesterone decreased less with EA than MA.

In rabbits pretreated with high doses of dexamethasone (to induce Kidney xu syndrome), MA (reduction) at ST-36 decreased body weight, but EA did not. There were no other evident differences between EA and MA, whether reinforcing or reduction was used; unfortunately the EA parameters used were not disclosed.³⁴⁷

In rabbits, MA reduced reaction to noxious tooth pulp testing more than EA.³⁴⁸ In an early rabbit study by Han’s Beijing group, although 16 Hz BL-60 EAA ~ MAA ~ finger pressure analgesia, the after-effect of MA was greatest, probably owing to local tissue damage as a result of vigorous needling.³⁴⁹ One interesting study protocol added manipulation (rotation or lifting/thrusting, 1 minute out of every 5, or 12 seconds in every minute) during EA (20 minutes) and found that TFL in rats was considerably enhanced, particularly by the shorter, more frequent manipulation.³⁵⁰

MA may initiate temperature changes and EA not³⁵¹ (although other studies have found no difference between the two³⁵²). EA may induce local short-term cooling before recovery and generalised warming as occurs with MA.³⁵³^,³⁵⁴ MA (‘mildly painful’) may result in a more consistent vasodilation than occurs with non-painful LF EA.³⁵⁵

In young pigs with overt E. coli diarrhoea, EA (at ST-36, Du-1 and caudal baihui) led to slower initial recovery than MA at the same and additional points together with moxibustion and bleeding at other points (rather an uneven comparison!).³⁵⁶ Similarly, EA may reduce fever induced by E. coli in rabbits less than MA.³⁵⁷

In healthy subjects, MA significantly increases salivary flow, whereas non-painful LF EA does not (and EA decreases salivary flow already stimulated by chewing, whereas MA has no such effect); this is possibly because of opposite effects on sympathetic tone.³⁵⁸

Studies where EA is more effective than moxibustion

Brief LF EA (2 mA) may be more effective than brief moxibustion in prolonging latency of peripheral antidromic action potentials (APs), with EA mediated predominantly by substance P (SP), moxibustion by somatostatin (SS).³⁵⁹ Also, EA has been shown more effective than moxibustion in its inhibition of vibration-induced finger reflex.³⁶⁰

Focus on the needle

Apart from the obvious method of stimulating two separate AcPs with needles connected to an EA device, the circuit completed through the body (Fig. 6.2a), groups of needles can be connected, for instance along one or two meridians³⁶¹ or in a localised area of pain,³⁶² and stimulated from the same output (Fig. 6.2b). As an alternative method to treating two separate points, for strong local treatment two needles can be inserted close to and either side of an area of pain, so that although they enter the skin maybe 4 cm apart their tips are much closer together (0.5–1.5 cm) (Fig. 6.2c).³⁶³^,³⁶⁴^,³⁶⁵ Chronically implanted needlelike electrodes to stimulate otherwise inaccessible regions of the peripheral nervous system have been advocated in neurosurgery.³⁶⁶


Figure 6.2 Connecting needles. (a) Needles are usually connected in pairs to an EA device (the circuit completed through the body). (b) However, groups of needles can be connected, positioned either along a meridian, or in a particular area. (c) Focused needling is also possible, with the needle tips close together (0.5–1.5 cm, with the needles angled towards each other).

To TENS or not to TENS?

Clearly, in many respects EA and TENS are very similar in their neurophysiology, especially if used at the same locations. Some authorities believe they are virtually interchangeable when used at AcPs,³⁶⁷^,³⁶⁸^,³⁶⁹ although others emphasise correctly, if a little primly, that the term ‘electroacupuncture’ should not be used to describe treatments where needles do not puncture the skin.³⁷⁰

Much of the research on which TENS is based has been ‘poached’ from EA studies, especially the neurochemical ones. However, there are differences. For one thing, although surface stimulation can activate nerves even 4 cm beneath the skin,³⁷¹ needle stimulation obviates problems of skin impedance (see SubCh. 5.1) so that much less current is needed to trigger action potentials (APs) in deeper afferent nerves than with surface electrodes (25 mA, with 1-millisecond pulse durations³⁷²). With the latter, at least with isolated pulses, there may be an initial transient current surge (due to the capacitative element of the impedance); this does not occur with needles.³⁷³ The deeper afferents contain a smaller proportion of thin unmyelinated fibres than superficial ones; such superficial C fibres may be activated by the higher currents needed with TENS.³⁷⁴ For these several reasons, EA may in some ways be less uncomfortable than TENS³⁷⁵ (skin/subcutaneous tissue PT tends to be lower than muscle PT, both being greater than fascial or periosteal PT³⁷⁶).

However, in one case study comparing EA and TENS, although the pain relief obtained was similar, more urinary corticosteroids were found after EA, which possibly indicates that it was experienced as more stressful.³⁷⁷ (Patients may well be less apprehensive about non-invasive stimulation, even in China!³⁷⁸^,³⁷⁹) The possibility that strong deep stimulation is endorphinergic and strong TENS less likely to be (see SubCh. 6.4, ‘gentle, shallow needling’) may also have some bearing on this.

Studies where EA is more effective than TENS

In a British study, local EA produced significant elevation of PT in response to cold immersion of the hand, whereas 100 Hz TENS did not. 8 Hz TENS also raised PT, but with considerable variation between individuals.³⁸⁰ In an important Japanese study of EA and TENS (both 100 Hz, 0.1 millisecond biphasic), only EA through needles insulated except at the tip significantly elevated PT in muscle and periosteum.³⁸¹ Correspondingly, EA may reduce SEP amplitude more than TENS.³⁸²

Both EAA and TENS analgesia were significant in one study on low-intensity LF stimulation, but latency to EAA was shorter.³⁸³

Yoshiaki Omura reported that EA was more effective for wound repair than TENS, and also in reducing muscle spasticity when used at GB-21 (though TEAS was superior to MA).³⁸⁴

In rats with experimental diabetes, EA resulted in significantly lower plasma glucose, with attenuation of polyphagia, polydipsia and polyuria, raised PT and normalisation of motor nerve conduction after 4 weeks of treatment (20 minutes every 2–3 days, bilaterally at ST-36 and BL-23). In TENS-treated animals, plasma glucose levels reduced only slightly, motor nerve conduction normalised after 6 six weeks and PT decreased after induction of diabetes using streptozotocin i.p.³⁸⁵

Studies where EA and TENS are equally effective

One major study by Han Jisheng’s Beijing group found that (1) both ST-36/SP-6 EA and TENS in rats produced analgesia of slow onset and offset at 2, 15 or 100 Hz (the time course being similar for both EA and TENS in each case), (2) systemic naloxone almost completely antagonised 2 Hz, partially antagonised 15 Hz and failed to affect 100 Hz analgesia resulting from both methods, and (3) tolerance to one form of stimulation at a given frequency reduced analgesia induced by the other form at the same frequency. The authors concluded that there is no significant difference between EA and TENS, in terms of either antinociception or the neural mechanisms involved.³⁸⁶^,³⁸⁷^,³⁸⁸ Richard Chapman also concluded that TENS and EA are pretty much equivalent.³⁸⁹^,³⁹⁰

Low-intensity LF EA and TENS at ST-36 both resulted in analgesia and enhanced cellular immune function.³⁹¹

In one Japanese study already mentioned, EA and TENS (both 100 Hz, 0.1 millisecond biphasic) significantly elevated PT in skin and fascia, while EA also (but not significantly) elevated muscle PT. The size, shape and material of the TENS electrodes (13 mm diameter rubber or metal cones, or soft rubber electrodes, 50 × 150 mm) did not greatly affect results.³⁹² In further Japanese studies, both high-intensity TENS and EA (at 1 Hz, for 5 minutes) led to comparable increases in local skin blood flow, temperature and cold tolerance.³⁹³^,³⁹⁴

One Chinese study found no significant difference between the use of EA or TEAS in the treatment of psychosis, although sleep improved more with the former. Deqi and PSM were reported by patients with both treatments.³⁹⁵

EA (5 Hz at BL-21, ST-36, Ren-12) and TENS (3 Hz) both significantly reduced gastric acid secretion in sham-fed humans, compared with MA or sham acupuncture.³⁹⁶

Meg Patterson found empirically that needles were unnecessary in her form of CES, and that pads were as effective.³⁹⁷

Studies where TENS is more effective than EA

Han’s group, using a protocol of 2 hours of treatment every other day for 10 days, noted that tolerance to HF EA was greater than to TENS of the same parameters (100 Hz), especially by the fifth session (with reductions in both level and duration of EAA).³⁹⁸ Other Chinese researchers have found the analgesic effect of DD TENS greater than that of DD EA using similar frequencies.³⁹⁹

In one study on stressful stimulation, 20 minutes of noxious TENS (100 Hz, 0.1 ms, motor level, cathode to volar wrist, anode to upper arm⁴⁰⁰) resulted in localised reduction in experimental pain, but the same duration of noxious EA of ‘classical’ AcPs (88 Hz at maximum acceptable intensity, to LI-4 and points at anterior and posterior axial crease, radial and medial elbow crease⁴⁰¹) did not produce any analgesia.⁴⁰² Similarly, in a much quoted early study (small, and not placebo controlled), Andersson’s Lund group found that TENS increased dental PT more than EA despite almost identical subjective sensation and very similar onset and decline of effect.⁴⁰³^,⁴⁰⁴ They considered that the larger electrode interface enabled higher currents to be used, resulting in recruitment of additional afferent fibres without excessive stimulation of nociceptive ones.⁴⁰⁵

One Japanese group found TENS more effective than EA for improving peripheral flap circulation and survival in the rat, possibly because stimulation was applied over a larger area in the former (the effects of intensity were unclear).⁴⁰⁶

Studies where MA is more effective than TENS

In a study comparing MA, familiar to the subjects, with ALTENS (200 ms, gradually increased to elicit strong but non-painful muscle contractions at 15–40 mA), to which they were naive, dental PT increased significantly only with the former.⁴⁰⁷

MA at Du-26 in dogs was superior to TENS in increasing cardiovascular parameters under halothane anaesthesia.⁴⁰⁸ And in one small human study, temperature changes with HF TENS (albeit at 30–40 mA) were less than those found when testing MA.⁴⁰⁹

Studies where TENS is more effective than MA

Both 2 Hz 40 mA and 100 Hz 20 mA TEAS at LI-4 increased the amplitude of the H-reflex evoked by soleus muscle stimulation in healthy volunteers, whereas MA at LI-4 for the same length of time (15 minutes) had no effect. Results suggest that TEAS enhances the excitability of the motoneuron pool in the spinal cord, but MA (at least, with the needle twisted initially to obtain deqi and then retained for 15 minutes) does not.⁴¹⁰

Other comparisons

EA (1–3 Hz, 0.4 ms, 10–12 mA at P-6 and (?) banmen (M-UE-13)) yielded a progressive, long-lasting and naloxone-sensitive decrease of the nociceptive blink reflex, whereas segmental TENS (80-100 Hz, 0.2 ms, 2-4 mA both at the wrist and at two points over the supraorbital nerve) had a more rapid effect that ceased immediately the painful stimulation was removed and was not affected by naloxone.⁴¹¹

Localised and general temperature changes induced by EA and TENS are not the same, although both may be sympathetically mediated (see SubCh. 6.3, ‘simultaneous sympathetic and parasympathetic activation’).

In a study of low-intensity LF stimulation, naloxone partially antagonised EAA but not TENS analgesia, and whereas the former was almost completely blocked by deep (but not intradermal) injection of the local anaesthetic agent procaine, the latter was partially blocked by either deep or intradermal injection.⁴¹²

Electrical stimulation of the median nerve produced a less detectable fMRI image from the contralateral sensorimotor cortex than when an acupuncture roller was used along the corresponding meridian (the image with rolling coincided with that obtained during hand and finger motor tasks, which was larger when these were more complex).⁴¹³

In one TENS study, smaller (1.5 cm diameter) electrodes resulted in more local temperature increase/vasodilation than larger (4 cm) ones (electrode temperature did not rise); however, other parameters were covaried, such as frequency (3 Hz or 100 Hz) and intensity (1.5 or 3 times the sensory threshold of 5–9 mA), which makes interpretation difficult.⁴¹⁴ Also, although electrode size, shape or material did not affect PT increases in the Japanese study mentioned above,⁴¹⁵ or quality of sensation in one Italian report,⁴¹⁶ others have suggested that effective stimulation may be experienced as unpleasant if electrodes are less than about 6 cm² in area⁴¹⁷ or 3 cm in diameter – at least using some waveforms.⁴¹⁸ It certainly seems easier to obtain the electrical equivalent of deqi with such electrodes than with very small ones,⁴¹⁹ presumably because more nerves can be activated (Ch. 4, ‘easier to activate more nerves’). With pads that are larger still, however, current density may be too low to activate deep muscle afferents.⁴²⁰

Other forms of electrical point or meridian stimulation

A small-diameter handheld probe (pTENS) was found to have slightly less of an enhancing effect on PT than EA in rats (although only 2 V was used in EA, and 7 V with the probe). Optimal duration was 10–20 minutes for both treatments.⁴²¹ Audiofrequency pTENS has been used in a number of studies to induce PSM,⁴²² associated with changes in temperature.⁴²³

A LF piezoelectric stimulator was found to be effective in reducing flexion reflex to pain elicited by sural nerve stimulation, whether applied segmentally or extrasegmentally.⁴²⁴ Also, a handheld electrode (roller?) has been used to massage and stimulate along the Gall Bladder meridian to induce gall bladder contractions and enhance bile flow in dogs (the return surface electrode being positioned at an appropriate AcP).⁴²⁵

Other methods of stimulating acupoints and meridians

Heat and cold

Traditionally, moxibustion is a major part of acupuncture practice. There are many studies on its effects, for instance on rat TFL (comparing moxibustion at different AcPs and temperatures, needling with moxa, and different durations of heating⁴²⁶^,⁴²⁷), on gastrointestinal electrical activity,⁴²⁸ on rat renal function and BP (using BL-15 and BL-27),⁴²⁹ on immune function and ageing (mouse thymus and pituitary activity,⁴³⁰ local immunocyte influx in rats and guinea pigs⁴³¹ and spleen T-cell activity in mice⁴³²), on how different methods and dosages alter blood histamine levels⁴³³ or affect other neurotransmitters,⁴³⁴ on how it affects the EEG, enhancing alpha production,⁴³⁵ on intestinal activity⁴³⁶ and so on. The difference between heat applied through the needle (activating both muscular and cutaneous afferents) and radiant heat (activating predominantly cutaneous afferents) has also been explored.⁴³⁷

As an alternative to burning herbs, some ingenious electrical methods of heating have been developed. A method of ‘electrocautery’ of Du-26 using high temperatures (~80°C) for 10 minutes via a 3-millimetre probe significantly increased cardiac activity and BP in a number of studies on anaesthetised dogs under various circumstances,⁴³⁸^,⁴³⁹^,⁴⁴⁰^,⁴⁴¹^,⁴⁴² with more effect than EA, MA,⁴⁴³ TENS or finger pressure.⁴⁴⁴^,⁴⁴⁵ Simple incandescent lamps have been used in many reports,⁴⁴⁶ or specifically infrared emitters.⁴⁴⁷ And many devices combine electrical heating with the use of herbs.⁴⁴⁸ One such even adds the benefit of an ‘electric wind’, claiming that a gentle low-temperature breeze for 10–20 minutes may strengthen wei qi (resistance to external disease), whereas a faster hotter breeze for 3–5 minutes (longer if directed at a more muscular area) may be used to treat febrile disease.⁴⁴⁹

The increase in PT that occurs with non-noxious heat (40°C) may be mediated by oxytocin (Ox) increases in CSF or plasma, or both.⁴⁵⁰

One study comparing moxibustion, incandescent heat and microwave concluded that there were no obvious differences between them (whether for xu or shi subjects), but that moxibustion instruments are more convenient to use than traditional methods.⁴⁵¹ Another report suggested that combining EA with electrical warming of the needle might increase its therapeutic effect.⁴⁵² Electrically heated needles have been used to stimulate immune function⁴⁵³ and in the local hyperthermia treatment of tumours (in mice, for instance⁴⁵⁴). Moist heat and shortwave diathermy (see Ch. 4, ‘27.12 MHz’) both relieved pain at trigger points (TrPs), with more effect from the latter at those that were only moderately sensitive initially.⁴⁵⁵

Not many studies have investigated the effects of cold (cryotherapy) at AcPs. Cooling the needle handle to induce small electric currents was suggested by a French doctor, Maurice Mussat, in the early 1970s.⁴⁵⁶ Recently, in an unpublished pilot study by British physiotherapist Lester Ward, a week of daily cold needling was found to reduce high levels of salivary cortisol possibly more effectively than using needles at room temperature.⁴⁵⁷ In a Chinese study, ‘freezing acupuncture’ was likewise found to increase T-lymphocyte counts significantly more than ordinary acupuncture in Kidney xu outpatients (most with chronic nephritis).⁴⁵⁸ Freezing acupuncture has been investigated for some respiratory conditions.⁴⁵⁹

Ice massage to LI-4 decreased acute dental pain significantly more than tactile massage (alone or accompanied by explicit positive suggestions) in a well-known study by Ronald Melzack and associates.⁴⁶⁰ Jeffrey Mannheimer and Gerald Lampe, in their classic tome on TENS,⁴⁶¹ also refer to cooling AcPs. However, despite the existence of some Japanese equipment for cryotherapy of small regions, I know of only one study on a device for the relief of pain by cooling AcPs directly.⁴⁶² In contrast, intermittent intense cold (with ice massage or a cooling spray) is a standard method for inactivating TrPs.⁴⁶³

Microwaves

There are some mentions of ‘microwave acupuncture’ (MWA) in the literature, although it is not always clear at first glance whether this is thermal or athermal in effect. MWA, for example, may (like EA) induce a temperature drop along a meridian from the treated point.⁴⁶⁴ Also, one comparative study found MWA analgesia to be greater than either MAA or that from HeNe laser acupuncture. 5HT was implicated in all three forms of analgesia.⁴⁶⁵

Millimetre waves (Extremely high frequency, EHF)

This form of non-thermal stimulation was introduced in an earlier chapter (see Ch. 4, ‘Millimetre waves (extremely high frequency, EHF)’). One of its most frequent applications is in the treatment of peptic ulcer. For this condition it is used at AcPs (particularly ST-36⁴⁶⁶). A standard protocol would be 8–10 daily sessions, 20–25 minutes long, for example.⁴⁶⁷ Biopsies taken from 10 patients with duodenal ulcer showed an apparent acceleration of ulcer repair with EHF treatment.⁴⁶⁸ In rats (particularly those under stress), 60 GHz ST-36 EHF reduced gastric hypersecretion,⁴⁶⁹ and had both protective and remedial effects on experimental peptic ulcer (with 5- or 10-, but not 15-minute irradiation).⁴⁷⁰

In another human uncontrolled peptic ulcer study, EHF enhanced immune function,⁴⁷¹ as it did in other conditions of compromised immunity when AcPs were irradiated.⁴⁷²^,⁴⁷³ In one report,⁴⁷⁴ immune enhancement after five 20–30 minute treatments was dependent on the initial immune status of the subject, and could be reversed by overtreatment (> 15 sessions). EHF also resulted in a reversible increase of the sensitivity to certain antibiotics of some staphylococci at AcPs (although without any obvious direct effect on them).⁴⁷⁵

EHF at AcPs has been used in the treatment of experimental tumours in mice⁴⁷⁶ and rats.⁴⁷⁷ (Frequencies in the 35.9–42.3 GHz range, scanned at a 12.5 Hz rate, were more effective than in two higher ranges.)

The effect of auricular EHF modulated at different frequencies on hypothalamic activity in humans has already been mentioned (see SubCh. 6.1, ‘local effect of EHF radiation on the brain’). Similar results for increased hypothalamic (and somatosensory cortex) activity in rats with SJ-20 55–65 GHz modulated at 26 Hz have also been reported,⁴⁷⁸ as well as effects (with different modulating frequencies) in rabbits.⁴⁷⁹ 60 GHz ST-36 EHF may inhibit hypothalamo-pituitary neurosecretion and thyroid hyperfunction in rats, particularly those subject to stress.⁴⁸⁰

In humans, EHF (for instance, at GB-20, BL-10, SI-16 and SJ-13) is considered to reduce autonomic hyperactivity (whether sympathetic or parasympathetic).⁴⁸¹

52–78 GHz EHF directed to a particular combination of body points (Ren-2, 6 and 18, Du-20, and bilateral BL-22, KI-3, P-6, LI-18) appeared to increase levels of plasma dehydroepiandrosterone (DHEA) in one American study using 3 minutes of daily irradiation in two 5-day courses followed by less frequent treatment.⁴⁸² DHEA increases were more prominent if progesterone cream was also applied, and the Liss cranial stimulator (see above, ‘optimum parameters’) used too (!).⁴⁸³

In healthy volunteers, 52–78 GHz at Du-20 and HE-7, P-6, LI-4 and auricular shenmen (all on the left side) led to a general reduction in EEG power, particularly in the alpha (8–13 Hz) and beta1 (13–18 Hz) bands. Most decreases were greater on the left, particularly in the theta (4–8 Hz) band. Together, these effects are suggestive of greater relaxation. Changes were more marked in those unfamiliar with the treatment. However, control points were not used.⁴⁸⁴ In patients with chronic amputation stump pain, EHF to AcPs (classical or ashi points) reduced slow wave power in the EEG.⁴⁸⁵ In rabbits, low-intensity (10 mW/cm²) EHF at 55–75 GHz reduced hypothalamic EEG power in some narrow frequency bands, and enhanced it in others, with more effect at ear points than directly over the scalp. Exposure in both areas reduced power centred on 15.9 Hz, and increased that at 25.6 Hz.⁴⁸⁶

In studies on dental pain threshold in rabbits, equivalent results to those obtained with acupuncture were found with very-low-power EHF (1–20 mW); during this treatment breathing became more regular.⁴⁸⁷ Although LI-4 EHF (modulated at 1–10 Hz) was less effective than MA, it was more so than EA in this situation.⁴⁸⁸

Using skin potential (SP) measurement at AcPs (one on each of the Small Intestine, Kidney, Liver, Stomach and Spleen meridians on the left side) and non-AcPs, the return electrode being a needle inserted subdermally in the occipital region, EHF (70 GHz) and LF electrical stimulation (1 Hz, 50 µA) of the stomach meridian point (presumably ST-36) were found to have synergistic effects in dogs.⁴⁸⁹

EHF to the auricular ‘Heart’ point may reduce heart rate in rabbits.⁴⁹⁰

Athermal levels of EHF may have heating or cooling effects at AcPs outside the treated area.⁴⁹¹ Subjective sensations of heat or cold, accompanied by corresponding changes in pulse rate, may occur too, as well as spreading sensations (PSM) of vibration or paraesthesia following a more local reaction, and even visual (colour) sensations.⁴⁹²^,⁴⁹³^,⁴⁹⁴ Intestinal peristalsis may also increase.⁴⁹⁵ Such ‘sensor reactions’ have been used to as a guide to optimal EHF wavelength,⁴⁹⁶^,⁴⁹⁷ but may be less obvious in complicated cases⁴⁹⁸ and tend to become less pronounced with repeated treatment.⁴⁹⁹ The Omura ‘bi-digital O-ring test’ has also been used to select appropriate treatment parameters, which may well vary for the same patient.⁵⁰⁰

Locally, low-intensity (0.05 mW/cm²) 42.0 GHz EHF to P-8 leads to degranulation of dermal mast cells,⁵⁰¹ as occurs too after EA (SubCh. 6.1, ‘degranulation of mast cells’).

One authority on the use of EHF considers that reinforcement treatment should not be longer than 2–5 minutes, and reduction should be carried out for 15–25 minutes.⁵⁰²

In a very small pilot study, skin reflectivity to 61 GHz EHF at the AcPs P7, P-8 and P-9 was greater than that in neighbouring areas,⁵⁰³ and increased hydration at AcPs (in line with Hiroshi Motoyama’s findings (see SubCh. 5.2, ‘Motoyama Hiroshi’s acupuncture meridian measurements’) has been adduced as a reason why EHF may be better absorbed there than at non-AcPs.⁵⁰⁴

Light

When the backs of rabbits (but not their eyes) were exposed to ordinary white light (but not yellow filtered light), continuous or intermittent (9 Hz), serum acetylcholinesterase significantly increased in the treated rabbits, but not in control animals.⁵⁰⁵ Blood glucose levels were affected quite differently by the continuous and intermittent light: they increased with the latter, but only well after irradiation.⁵⁰⁶ Intermittent light was also found to increase plasma dopamine.⁵⁰⁷ As with EHF, weak (athermal) infrared irradiation of the auricular ‘Heart’ AcP may alter both EEG and ECG characteristics.⁵⁰⁸

Low-intensity laser and polarised light

Lasers and other forms of light with an energy density of around 40 J/cm² or more will have a heating effect and may thus be comparable to moxibustion in some respects. Low-intensity laser (or light) therapy (LILT) is quite different, as discussed above in general terms (see Ch. 4, ‘Low-intensity lasers and polarised light’). LILT in the form of ‘laser acupuncture’ (LA) has in particular been promulgated as a pain-free and non-traumatic alternative to needles⁵⁰⁹ or even TENS.⁵¹⁰ Invasive LA is also used, the light being conducted via an optic fibre threaded through a hollow needle, so combining the benefits of LILT and acupuncture (or even EA⁵¹¹). In terms of the very different mechanisms involved, the term ‘laser acupuncture’ should perhaps be restricted to the invasive form,⁵¹² but here ‘LA’ will be used to cover LILT as applied cutaneously to AcPs (if needles are used, this is stated explicitly).

There are, as yet, few studies of the effects of LA within the brain, but in one pilot study diode LA (880 nm, 1 J/cm²) at Du-14 and TrPs in the trapezius muscle was found to alter regional cerebral blood flow significantly in various parts of the brain associated with nociception (the thalamus, caudate nucleus and prefrontal cortex).⁵¹³

The effects of LILT on pain are still controversial (see Ch. 4, ‘More controversial than the use of LILT’). In particular, the opioid effects of LA are still not agreed. For instance, blood levels of βEP and LE may be normalised using HeNe LA.⁵¹⁴ In keeping with this result, i.v. naloxone was found to reverse PT increases in rabbits in one GaAs ST-36 LA study. Here AcP irradiation had very similar effects to deep fibular nerve irradiation, suggesting, in contradiction to Thomas Lundeberg,⁵¹⁵ that LILT may indeed owe its effects to nerve stimulation.⁵¹⁶ However, in rats the increase in PT obtained with 4 Hz HeNe LA at Du-1 was not reversed by a low dose of naloxone (2 mg/kg) given beforehand. In this study, PT changes were assessed by TFL and hot plate (HP) methods, and it also transpired that a high dose of naloxone (20 mg/kg) did actually reverse the LA effect for HP but not for TFL.⁵¹⁷ Naloxone was also unable to block TFL increases in a further study, using 60 Hz pulsed LA.⁵¹⁸ However, tail flick is a local spinal reflex (unlike HP), not centrally mediated, so really this is hardly surprising.

A RCT of auricular HeNe LA in healthy subjects demonstrated a significant increase in PT (to painful electrical stimulation of the ipsilateral wrist) in healthy subjects, with no such change in the sham treatment group,⁵¹⁹ despite a low-dosage protocol that was subsequently criticised by David Baxter (as was use of PT as the sole outcome measure).⁵²⁰ Baxter also expressed concern about the dosage used in a SEP study of HeNe LA (0.7 mW or 1.26 J at unilateral LI-4 and P-6 for half an hour) in which effects faded after an hour,⁵²¹ suggesting that results might be due to expectation and habituation rather than LA per se. However, the author of this last study cites another Chinese study in which SEP effects were similar with both LA and ‘needle acupuncture’ (presumably MA).⁵²²

In one Chinese study, though, the effect of ST-36 HeNe LA on pain tolerance (rather than PT) was found to be a useful pretest for the epidural dosage required for gastrectomy when the epidural was used in combination with AA.⁵²³

Some reports give more clearcut results on LA and pain. In 10 out of 14 horses, infrared LA at nine AcPs (300 mW, 904 nm, 360 Hz, for 2 minutes per point) significantly alleviated signs of chronic back pain and greatly improved performance (with continuing benefits 1 year post treatment).⁵²⁴ In rats with experimental hindpaw arthritis, HeNe BL-60 LA (3 mW, 10 minutes daily for 5 days) significantly reduced ankle swelling and pain scores more than sham or no treatment (but without any effect on PT except immediately after treatment).⁵²⁵ Also, LA significantly increased PT in rabbits, dogs, pigs, sheep and goats, the effect being potentiated by administering a combination of L-tryptophan, D-phenylalanine, D-leucine, acetylcholine and insulin (!).⁵²⁶ This is in line with findings that 5HT, OPs, ACh and other neurotransmitters may mediate LILT (see Ch. 4, ‘LILT may significantly alter brain levels’), and the effect of hypoglycaemia on EAA (see below, ‘hypoglycaemia may potentiate ALS’).

LA may have cardiovascular effects. Thus HeNe laser irradiation of P-6 in cats (using 2.7 J/point) has been shown to reduce HR, with other effects on the ECG as well.⁵²⁷ HeNe LA has similar effects in cows, together with a reduction in haemoglobin levels (the same changes, and some additional ones, were observed if the uterine cervix was irradiated, rather than AcPs⁵²⁸). However, in rats HeNe LA (or associated stress?!) may increase HR (and physical activity in general, with decreased AChE particularly in young rats).⁵²⁹ LA also influences ventricular function in humans (3–5 mW HeNe at P-6).⁵³⁰ However, although P-3 LA speeded recovery in rabbits with acute myocardial ischaemia (AMI), it did not reduce myocardial injury if used after severing afferent nerves from that point, which indicates the involvement of the afferent pathway in LA.⁵³¹

LILT is well known for its effects on tissue repair. HeNe LA, in combination with the bioflavonoid quercetin, has been shown to reduce postoperative inflammation of injured nerve.⁵³² HeNe LA also improved microcirculation and enhanced muscle regeneration in rats with experimental myopathy due to denervation,⁵³³ particularly when combined with physical loading.⁵³⁴ Myopathy due to forced inactivation (including an increase of slow muscle fibres) was inhibited by LA.⁵³⁵ As with EA, VIP and CGRP have been invoked to explain the tissue repair effects of LA.⁵³⁶

LA may have particularly useful effects on the immune system. It appears, for example, to enhance some cell-mediated and humoral immune characteristics in rats⁵³⁷^,⁵³⁸ and mice, as well as non-specific antiviral resistance (to acute experimental influenza) in the latter.⁵³⁹^,⁵⁴⁰

Bilateral 2 mW HeNe ST-36 laser irradiation (10 minutes daily, 14 treatments in all) may beneficially affect a number of immune parameters.⁵⁴¹ However, the effect of using ‘photobustion’, an incandescent light flashing at 0.5 Hz, to heat points to 40–43°C daily for 10 minutes, was not significantly different. Direct irradiation of the thymus area in rats exposed for the first time may induce a stress response. Rather as with EA (see SubCh. 6.4, ‘greater increases in serum cortisol’ and ‘relaxation only in subsequent sessions’), repeated (or longer) treatment may reverse this, but with little effect on immune parameters.⁵⁴² However, some effects of LA such as those on chronic inflammation may well be cumulative.⁵⁴³

Laser cautery, like the use of strongly heated needles, may elevate leukocyte counts and serum immunoglobulins.⁵⁴⁴ Invasive LA has been used to irradiate the blood directly as a means to enhancing immune function.⁵⁴⁵

LI-20 HeNe LA (2.5 mW) has been used in the treatment of experimental allergic rhinitis, resulting in temperature increases of up to 5°C at the contralateral LI-20 point.⁵⁴⁶ Rather as with EHF (see above, ‘Such ‘sensor reactions’’), sensor reactions of warmth, tingling or sharp pricking may indicate that treatment has had an effect and need not be continued.⁵⁴⁷ Such deqi, if one may call it that, is perhaps similar to that obtained with MA, but usually much less marked.⁵⁴⁸ However, in one informal report of 450 treatments using the LASERneedle system (see Ch. 11), deqi was obtained in over 90% of cases and was found to be greater than with MA, particularly in acute cases.⁵⁴⁹

In one rabbit study, however, ST-36 HeNe LA did not inhibit anaphylactic reactions, although HeNe LA did appear to have some regulative effect on blood histamine levels.⁵⁵⁰

LA affects gastroelectric activity in rabbits⁵⁵¹. Also, Du-1 LA has been used for dysentery in young lambs, with rapid reduction of hyperperistalsis and borborygmus.⁵⁵²

Polarised light (together with MA), which is associated with partial recovery of sensation, has been used on scalp and back acupoints in a dog with loss of sensation following a road traffic accident.⁵⁵³

Comparative studies of ‘laser acupuncture’

Some careful studies have been carried out comparing laser and acupuncture for pain. In one on rats by Thomas Lundeberg’s group,⁵⁵⁴ using HeNe and GaAs lasers at the base of the tail, both power (1.56 and 0.07 mW) and energy (~1.4 J and 0.06 J per point over 15 minutes) were probably too low to be effective. So although 80 Hz local EA did prolong TFL, the conclusion that LA analgesia might be due to placebo is unwarranted. In a German study, single-blind MA (bilaterally at LI-4) increased PT to painful heat stimuli more than MA at a control point, and also more than double-blind HeNe laser irradiation for 1 minute at each point (bilaterally at LI-4 and jianqian, M-UE-48).⁵⁵⁵ However, this study has been criticised by David Baxter as depending solely on PT as the outcome measurement, and (like studies by Lundeberg’s group) because the laser dosage used was very low.⁵⁵⁶

Ukrainian researchers (using non-specified LILT parameters) found no analgesic effect with LA.⁵⁵⁷ Yet Chinese researchers found that 30 minutes of either EA or LA increased amplitude of the ‘N80’ component of SEP in response to median nerve stimulation, although it look longer for the N80 to return to pretreatment level with EA (20 minutes) than with LA (10 minutes). Importantly, they found that there was no effect on N80 amplitude in normal subjects⁵⁵⁸ – something that might help to explain the negative results of some LILT nerve conduction studies (see Ch. 4, ‘enhanced slow components’). In a Hong Kong study, both LF EA (20 minutes, intermittent) and LA (0.6 J/point, 120 seconds per point, ‘closed contact’ method) at LI-4 and ST-36 decreased late component (cognitive) SEP amplitude in response to painful dental stimulation. Results using EA were marginally superior to those using LA.⁵⁵⁹

In a study of chronic back pain in horses, LA (904 nm, 300 mW, 360 Hz, for 2 minutes/point), MA (even needling) and ‘injection acupuncture’ (with isotonic saline) all appeared to be equally useful.⁵⁶⁰ Similarly, TENS, EA and LA all gave comparable results in one human case study.⁵⁶¹

HeNe laser at ST-36 resulted in similar levels of analgesia and enhanced cellular immune function to those obtained with TENS or EA (low-intensity LF) at the same AcP.⁵⁶² A further comparison, between HeNe ST-36 LA and ‘photobustion’, was mentioned above.⁵⁶³

When compared with MA, HeNe LA was found to elevate PT significantly more than MA in rabbits, with (infrared) CO₂ being more effective than HeNe LA in dogs.⁵⁶⁴ However, in rats, lower power HeNe LA resulted in greater analgesia in one study (6 mW as opposed to 15 mW), LA at both powers being less effective than MA and MWA.⁵⁶⁵

Yoshiaki Omura reports that 15 minutes of 10 mW GB-21 LA was less effective in reducing muscle spasm (cramp) than EA, TENS or even MA.⁵⁶⁶ Litscher and Wang, in a small pilot study, found that the effects of LA (785 nm, 19.6 mW) on cerebral oxygenation were less than those of manual needling for the same length of time (20 seconds). With both modalities, effects were greater at the acupoints used than at a sham point.⁵⁶⁷

LA and VLF low-intensity EA at P-6 have been compared for their effects on cardiac function. LA appeared to have greater effects in this study.⁵⁶⁸

In a single case report in which various stimulation modalities were used, urinary glucocorticoids increased more when LA was pulsed at 2 Hz than at 64 Hz.⁵⁶⁹ In a more detailed study, 1 mW HeNe laser at low SR tail points (~Du-1) in rats for 15 seconds resulted in rapid onset short duration (24-hour!) hypoalgesia when pulsed at 4 Hz, delayed and longer lasting (48-hour) effects on TFL at 60 Hz, and no effect at 200 Hz.⁵⁷⁰ (Compare the Ulster study on 16 Hz versus 70 Hz mentioned above, see Ch. 4, ‘slower in onset yet longer lasting’.) As David Baxter comments, this appears to be a genuine frequency effect (as power differences at the different frequencies were taken into account). Less clear is a result reported by Jean-Claude Darras: 28 mW HeNe LA tended to affect the rate of flow of a radioactive tracer through the connective tissue more when pulsed at 24 Hz than 48 Hz (but the method of pulsing was undefined).⁵⁷¹

LILT was found to be less effective than both 3 Hz EA and 5 Hz TENS in reducing sham-feeding-induced gastric acidity in humans.⁵⁷²

Preliminary data from an Australian group indicate that near-IR laser (780 nm, 3 mw) is preferentially absorbed when pulsed at LF (50/50 duty cycle) by the AcP LI-4 compared with a non-AcP (both located by measuring SR). Above 10 Hz, no difference was found.⁵⁷³ Although this study is important in that it draws attention to a possible link between low SR (thin corneal layer? greater hydration?) and LILT absorption, and may shed some light on variation in response using different pulse frequencies, it is not entirely clear whether the result is in fact due to the different frequencies used or to cumulative effects with longer pulse durations.

Low-intensity pulsed electromagnetic fields

Very low (pTesla) PEMF applied to the head may increase release of 5HT or enhance 5HTergic transmission.⁵⁷⁴ Stronger fields (on their own, or in combination with CES) may induce drowsiness.⁵⁷⁵ In rats with depressed humoral and cellular immune function (following lesions to the anterior hypothalamus or caudate nucleus), magnetic stimulation may to some extent restore immune function.⁵⁷⁶ In a small pilot study, Saul Liss noted trends to increased blood levels of 5HT, βEP and ACTH, with decreased cortisol, after treatment with the Medicur device (see Ch. 11).⁵⁷⁷

While PEMF has been applied to TrPs in clinical practice,⁵⁷⁸ there appear to be few experimental studies on PEMF within the context of acupuncture.

Permanent magnets

Small magnets (750 or 1500 G) applied to AcPs may enhance leukocyte activity, phagocytosis and a non-specific immune response.⁵⁷⁹

Magnets applied to the skin, or even light finger pressure in the presence of a small (2 G) magnetic field, may reverse, deviate or weaken PSM induced by LILT.⁵⁸⁰^,⁵⁸¹ In rabbits, a 1300 G magnet (‘North pole’) applied over muscle may restore muscle twitch depressed by long-term tetanic stimulation (10 Hz, for 1 hour), possibly by re-establishing local microcirculation via cholinergic mechanisms (i.e. inhibition of AChE). This might well have an effect on local muscle pain as well.⁵⁸²

The use of magnets in conjunction with Yoshiaki Omura’s bidigital O-ring test is mentioned below (see below, ‘magnets are positioned’). Jacques and Andrée Pontigny have suggested that magnets may be beneficial only if the subject is moving relative to the geomagnetic field; wearing them at night may thus not be useful.⁵⁸³

Vibration

Pain relief with 100 Hz vibration is not affected by (low-dose) i.v. naloxone.⁵⁸⁴ However, non-noxious 100 Hz vibration may be mediated (at least in rats) partly through activation of oxytocinergic mechanisms.⁵⁸⁵ 100 Hz vibration may also induce vagally mediated increases in plasma gastrin and cholecystokinin (CCK) levels, as may 40°C warming or 2 Hz EA, with SS changing only in response to the latter, that is intramuscular stimulation.⁵⁸⁶ Furthermore, adenosine appears to mediate HF (100 Hz) vibration-induced analgesia,⁵⁸⁷ as well as low-intensity HF TENS analgesia or EAA (see SubCh. 6.2, ‘two or three cups of coffee’, and SubCh. 6.4, ‘mediated by adenosine’).

100 Hz vibration and TENS, as well as 2 Hz EA, have all been found superior to placebo in the treatment of myalgic pain.⁵⁸⁸ Like TENS and EA, 100 Hz vibration does not affect thermal sensitivity to non-noxious heat, even if relieving clinical pain in the same subjects.⁵⁸⁹

When used intrasegmentally at or proximal to a site of experimentally induced pruritus, vibration at 10 Hz, 100 Hz or 200 Hz (and 2 Hz or 100 Hz TENS) reduced subjective itch intensity, with 100 Hz being the most effective especially locally. Itch duration also decreased most after prior 100 Hz vibration, which had the most rapid effect as well.⁵⁹⁰ However, 100 Hz vibration and TENS had no effect on experimentally induced itch when applied extrasegmentally, although 2 Hz TENS did.⁵⁹¹ In contrast, some researchers have noted that vibration, at 60 Hz for instance, does have contralateral effects on itch⁵⁹² or, at 20 Hz, 50 Hz, 100 Hz or 200 Hz, on pain.⁵⁹³^,⁵⁹⁴^,⁵⁹⁵

Clearly, 100 Hz vibration stimulating large-diameter cutaneous afferents shares quite a number of characteristics with HF TENS or (gentle) EA, even though these may activate Aβ or even Aδ. Generally, vibration at > 50 Hz takes effect rapidly and that at < 10 Hz more slowly.⁵⁹⁶^,⁵⁹⁷ The effects of 20 Hz vibration, however, are mixed: as with LF EA, the increase in PT is slower in onset than with HF stimulation but, as with HF EA, more rapid in returning to pretreatment level.

Importantly, with 20 Hz or HF vibration the duration of effect on clinical pain may be much greater than that on experimental PT. Increases in PT occur in phase with enhanced local (and possibly diminished peripheral) blood flow.⁵⁹⁸

LF vibration of the acupuncture needle is sometimes used in MA to good analgesic effect.⁵⁹⁹ In rabbits, twitch response to stimulation at TPs was greater with vibration (tapping) than simple ‘dry needling’ (though less than with the mechanical equivalent of snapping palpation).⁶⁰⁰

Sound

Although sound has been applied to AcPs in clinical practice,⁶⁰¹^,⁶⁰²^,⁶⁰³ there are few experimental studies in this area. Some research has been carried out on the ‘Qi Gong machine’ (see Ch. 10). In a more traditional vein, Susan Tomkins, an acupuncturist trained in the Five Element approach, has looked at the effects of 329.6 Hz applied to Earth points on the different meridians.⁶⁰⁴

Despite the paucity of research, some devices for audiofrequency electrical stimulation have been utilised.⁶⁰⁵^,⁶⁰⁶ Clinically, the combination of sonic with electrical stimulation may markedly affect body temperature.⁶⁰⁷ As already mentioned (see SubCh. 5.2, ‘may also have bioluminescent’), meridians may have their own acoustic characteristics.

Ultrasound

Very-low-intensity ultrasound (0.05–0.2 W/cm²) has been used to treat auricular AcPs (even at 0.4 W/cm², repeated ultrasound stimulation could lead to electrical ‘hyperactivity’ at the points).⁶⁰⁸ In one controlled study, 20 minutes of ultrasound applied to ST-36 was found to increase somatic and gastralgia PT in rabbits, the effect peaking 25–30 minutes post treatment.⁶⁰⁹ Use of a special applicator to ‘focus’ ultrasound deep within the tissues at ST-36 and ST-41 also improved microcirculation in the human leg.⁶¹⁰ Ultrasound has been used too to detect TrPs⁶¹¹ and to investigate differences in sensation at body and auricular AcPs.⁶¹² In one fascinating study, Joie Jones (a colleague of Zang-Hee Cho at UC Irvine) found that athermal 50–80 W/cm² ultrasound, at various frequencies (centred on 5 MHz) and applied through a 2 cm diameter head mounted in a water-filled flexible chamber over BL-67, resulted in fMRI records of occipital lobe activity indistinguishable from those obtained with BL67 MA (see too SubCh. 6.1, ‘activation of the visual cortex’).⁶¹³ Optimal results were obtained when the ultrasonic pulse was focused at the same depth as the needle tip had been; if focused more deeply or on the skin surface, no cortical activity was observed.

Possible adverse effects of stimulation

Electrical stimulation

EA has endorphinergic effects, with the implication that awareness of painful stimuli may be dulled, possibly resulting in injury. However, as Anders Ekblom and Per Hansson found, while EA may reduce clinical pain intensity, thermal sensitivity (for instance) is not affected.⁶¹⁴ This has also been reported for SPES⁶¹⁵ and some other forms of electrostimulation.⁶¹⁶

However, during TENS treatment, and for a short while after, sensitivity to electrodiagnostic testing (the rheobase, see Ch. 4) may be elevated.⁶¹⁷ With EA, if a nerve is slightly injured by an acupuncture needle the threshold to subsequent stimulation may be reduced, so that EA may be experienced as more painful than usual until the nerve has healed.⁶¹⁸ In general though, even with repeated treatments, there do not appear to be any long-term adverse effects on nerve conduction characteristics with peripheral electrical stimulation.⁶¹⁹

If chronic pain is the result of severely impaired descending inhibitory pathways, then EA or TENS may exacerbate rather than relieve it.⁶²⁰ This aggravation might provide a clue that such impairment is present, and could well be less likely to occur when using methods such as CES that do not directly affect primary afferent synapses in the dorsal horn.

Remembering the Arndt–Schulz law, excessive levels of stimulation (whether in terms of amplitude, or pulse or treatment duration) may be counterproductive⁶²¹ – not least because they may trigger a stress response (see SubCh. 6.4, ‘stress mechanisms became involved’). Stressful stimulation may adversely increase plasma/blood viscosity and erythrocyte aggregation rates, for instance.⁶²² Strong (or prolonged) stimulation should be avoided at particular AcPs such as Du-3.⁶²³ Somebody who is already stressed may experience even low-level stimulation as stressful, with possible adverse responses.⁶²⁴

Strong electrical muscle stimulation (as opposed to comparable voluntary contractions) may temporarily damage muscle fibre membrane.⁶²⁵

As with any treatment that leads to muscle movement, motor level EA or TENS may affect tissue uptake or distribution of locally or systemically administered substances. While this effect may be usefully harnessed, it is not always desirable.

Spinal cord stimulation may reduce heart rate via sympathetic inhibition.⁶²⁶ This should be remembered when using dumai (Governor Vessel) or huatuojiaji points in yang xu patients. HR may also be affected by DC stimulation of HE-9 with Ren-17 (see SubCh. 6.4, ‘adverse, if temporary, cardiac effects’).

Vagal stimulation in rats may increase the proportion of degranulating mast cells in the trachea and bronchi (more so on the stimulated side), mediated possibly by polymodal neurons of the jugular ganglion.⁶²⁷ If a similar pathway exists in humans, auricular stimulation (especially bilateral) might be expected to cause bronchoconstriction.

Caution is generally advised in stimulating points on the throat, in part because of possible baroreceptor effects via the carotid sinus, leading to HR and BP reduction, and in part because such treatment might lead to airway restriction. However, using TENS (1 ms, 2.5–4 mA) both unilaterally and bilaterally over the tracheoesophageal groove led to increasing glottic aperture at lower frequencies (maximal at 30 Hz) and closure only at higher frequencies (total closure at 100 Hz), with no effects on respiration or heart function.⁶²⁸

Certain stimulation frequencies have been considered to be in some way undesirable. For example, if HF EA or TENS releases spinal cord dynorphins, and these inhibit vasopressin and LHRH (SubCh. 6.3, ‘Gonadotropin-releasing hormone (LHRH or GnRH)’), it is possible (if unlikely) that such stimulation might inhibit or delay ovulation (see SubCh. 5.1, ‘dynorphin may inhibit release’).

LF stimulation that leads to 5HT-mediated peripheral vasodilation can also sometimes result in headache.⁶²⁹

Other recommendations on unsafe frequencies come from the CES literature, sometimes with very little except scattered observations to support them. Meg Patterson, for example, advocated avoiding 8–12 Hz in her method of CES (see below), Ifor Capel’s group found that 7 Hz was ‘rather deleterious’ with SPES,⁶³⁰ whereas Mike Heffernan, using a LF form of CES, has suggested that > 14 Hz might elicit ion cyclotron resonances (see Ch. 7), with unforeseeable consequences.⁶³¹

As a generalisation, symmetrical biphasic waveforms produce more effective stimulation and fewer side-effects than others.⁶³²

Although DC EA⁶³³ and EA hyperthermia have both been used in the treatment of tumours (see too Ch. 4, ‘Rudolf Arndt, a psychiatrist’), they are very much experimental procedures, not a part of normal clinical practice. If poorly understood, it is very possible that they might enhance tumour growth or metastasis.

However, there is no indication that very-low-intensity stimulation (such as SPES) can adversely affect cancer,⁶³⁴ and one report on the use of EAA (MAA?) with epidural anaesthesia in a small number of patients (14) found no evidence of any lymphocyte mutagenic or carcinogenic effects.⁶³⁵ Nevertheless, metastatic activity was dramatically increased in one study of rats receiving 100 Hz PAG stimulation analgesia after injection with carcinosarcoma (ascites) cells.⁶³⁶ In view of some similarities between this form of analgesia and EAA, it would seem sensible to err on the side of caution when treating patients known to have cancer.

Of course, in general clinical practice needles should not be used with DC stimulation because of the effects of electrolysis⁶³⁷ (see Ch. 3, ‘thoroughly explored’). Even without overt electrolysis, needles may undergo chemical reactions at the metal surface, with the possible adsorption of residual byproducts.⁶³⁸ Thus they should not be reused, certainly without proper cleansing and preferably not at all.

There is a greater risk of adverse skin reactions when using TENS electrodes with a small surface area for extended treatments (and more so with premodulated interferential currents or long trains of pulses at high intensity).⁶³⁹ Although allergic responses to electrode gels also occur with extended treatment,⁶⁴⁰^,⁶⁴¹^,⁶⁴² advances in manufacture mean that they are less common than they used to be.

An in vitro study on the hippocampus from young rats (< 2 weeks) shows LF stimulation may have double the effect found in that from older ones (> 5 weeks).⁶⁴³ As the effects being studied (reversible long-term depression and potentiation of neuronal activity) are very probably also involved in the dorsal horn synaptic mechanism of EA and TENS, this is a pointer to caution in the treatment of children. Furthermore, if there is any substance to some of the theories of EEG resonance described in the next chapter, the effects of LF repetitive stimulation on children whose brains do not yet show adult patterns of activity may be quite different from those found in adults. For instance, in children there is a more ‘introverted’ dominance of theta rather than of alpha, as in adults.⁶⁴⁴ Until such differences are clarified, caution is again indicated.

LILT

Although dosage, at least for tissue repair, should be ~4 J/cm² (Ch. 4, ‘~4 J/cm²’), HeNe even at 0.3–0.5 J/cm² was observed to cause considerable damage to the mitochondria of fruitfly (Drosophila auraria) larvae, but not at < 0.3 J/cm².⁶⁴⁵ Whether this is significant in terms of human mitochondrial damage is another question.

Caution should be observed when using LILT on a patient who is taking St John’s Wort (SubCh. 6.2, ‘cautiously in conjunction with LILT’).

The vexed question of epilepsy

Strong visual stimulation may induce seizures at certain low frequencies (12–18 Hz),⁶⁴⁶ though not at 4 Hz for instance,⁶⁴⁷ unless long pulse durations are used.⁶⁴⁸ This effect is probably also wavelength dependent,⁶⁴⁹ white or blue light being safer than green or red, for instance.⁶⁵⁰^,⁶⁵¹ However, the risk of triggering epilepsy is minimal with LILT, for which in any case safeguards should be in place to prevent direct exposure of the eyes during clinical use.

In patients with physically obvious pre-existing exaggerated responses (as in myoclonus) or abnormal EEG patterns, prolonged nerve stimulation at particular frequencies (in one case at 10–14 Hz, for example) may lead to generalised convulsions. Fortunately such patients are not commonly encountered, and are able to recognise something different and uncomfortable in the overall subjective sensation elicited by stimulation before convulsions occur.⁶⁵² Early reports of this epileptogenic effect of LF stimulation were used by Meg Patterson to support her use of HF CES,⁶⁵³ but since then it has been found quite unlikely that even direct visual stimulation at < 15 Hz⁶⁵⁴ or magnetic stimulation at < 17 Hz⁶⁵⁵ will cause problems, while the extrapolation to somatosensory stimulation may be quite unjustified for the vast majority of patients.

For one thing, epileptogenic ‘kindling’ (see Ch. 7) is more likely to occur at higher frequencies, even when the brain is directly stimulated.⁶⁵⁶^,⁶⁵⁷ Although LF (3 Hz) amygdala stimulation has been used to induce kindling in rats,⁶⁵⁸^,⁶⁵⁹ for example, elsewhere within the brain LF (1 Hz DC or 4 Hz biphasic) may ‘quench’ such amygdala-kindled seizures.⁶⁶⁰^,⁶⁶¹^,⁶⁶²

Secondly, generalised seizures tend to occur only in patients with a personal or family history of epilepsy,⁶⁶³ and kindling is unlikely to occur in those without such a history.⁶⁶⁴ However, any sign of rhythmic myoclonic jerks occurring in time with the applied stimulation (or at a harmonic of that frequency), particularly in such a patient, should be viewed as a contraindication to continued treatment (mild local muscle twitching in response to EA or TENS occurs commonly, and should not be confused with myoclonus).

Thirdly, EA has been found to reduce experimental epilepsy and convulsions, and their effects in the brain, in a number of animal studies,⁶⁶⁵^,⁶⁶⁶^,⁶⁶⁷^,⁶⁶⁸^,⁶⁶⁹^,⁶⁷⁰^,⁶⁷¹^,⁶⁷²^,⁶⁷³^,⁶⁷⁴^,⁶⁷⁵^,⁶⁷⁶^,⁶⁷⁷^,⁶⁷⁸^,⁶⁷⁹^,⁶⁸⁰ although in one detailed report on work with cats and mice the results were not so clearcut.⁶⁸¹ MA or EA (2 Hz trains of 50–70 ms, internal frequency 950 Hz, 0.2 ms, intensity not indicated) was applied to 24 acupoints used clinically for epilepsy, in various combinations. In another study using MA at the same points (some distally on the limbs, some on or near the head, such as GB-20, Du-14 and Du-26), acupuncture clearly increased EEG spike activity in most animals, particularly if used during seizure discharges.⁶⁸² Prior MA has also been found to reduce induced epileptiform EEG in rats.⁶⁸³ Rats prone to seizures (with a functional defect in central CCK mechanisms) may in fact respond better to EA than other rat types in some respects.⁶⁸⁴^,⁶⁸⁵

EA at ST-36 has been used to inhibit cAMP-induced epileptiform waves (in amplitude, frequency and duration),⁶⁸⁶ and at Du-26 and Du-24 to prevent hyperbaric oxygen-induced convulsions.⁶⁸⁷ Lumbar points on the dumai (Governor Vessel) seem to be particularly effective, tending to inhibit both amplitude and frequency of epileptiform hippocampal activity, for instance, especially at higher intensities, with frequency (1 Hz or 5 Hz) seemingly a less important factor.⁶⁸⁸^,⁶⁸⁹

However, MA at non-acupoints that do not effectively inhibit epileptiform activity may in fact potentiate it (see SubCh. 6.1, ‘potentiated epileptiform activity’), and repeated treatment may reduce antiepileptic effectiveness (possibly through enkephalin tolerance).⁶⁹⁰ Furthermore, strong EA applied to head points may occasionally exacerbate penicillin-induced epileptiform bursts⁶⁹¹ and MA, as mentioned, may also increase penicillin-induced seizure activity.⁶⁹²

The antiepileptic effects of EA may be mediated by 5HT.⁶⁹³^,⁶⁹⁴ GABA has also been implicated in Du-24/26 attenuation of hyperbaric convulsions⁶⁹⁵ and, with endorphins and CCK, in penicillin-induced epileptiform discharges.⁶⁹⁶^,⁶⁹⁷^,⁶⁹⁸^,⁶⁹⁹ Hippocampal βEP and LE elevated by ECT or penicillin microinjection are reduced by protective EA (134 Hz, 7–10 mA) at Du-8 and Du-16,⁷⁰⁰^,⁷⁰¹ probably as a result of it down-regulating their synthesis (in this study EA at the same points was at 100 Hz, 6 mA, so still an intense stimulation).⁷⁰² However, Enk may have both pro- and antiepileptic activity.⁷⁰³ Dynorphin levels, in contrast, which are reduced after penicillin-induced seizure, may be enhanced by EA.⁷⁰⁴ The anticonvulsant effect of EA may also be mediated by decreases in some NO synthases.⁷⁰⁵

Clinically, auricular acupuncture (with press needles) has been used for canine epilepsy,⁷⁰⁶ and EA may have useful antiepileptic effects in humans⁷⁰⁷ (see SubCh. 9.1). SPES too may have some antiepileptic effects.⁷⁰⁸

Nevertheless, as epilepsy may involve repetitive discharge with reduced variability in some parts of the brain that are also involved in acupuncture pathways (see Ch. 4), the use of regularly repetitive strong stimulation should be avoided particularly at points near or on the head. As has been shown repeatedly, high-intensity ‘electronarcosis’ like ECT may well induce convulsions.⁷⁰⁹ HF EA at 7–10 mA has been used to attenuate such convulsions in animals⁷¹⁰ (decreasing β and increasing δ band EEG activity⁷¹¹), but also, at even higher intensity, to induce convulsion as a less damaging alternative to ECT.⁷¹² Strong EA used for prolonged periods in surgery has likewise been found to trigger seizures.⁷¹³

Even 10–20 G static magnetic fields applied to the head may elicit epileptiform activity in the hippocampus in epileptic patients,⁷¹⁴ although some magnetic fields – including static ones applied preemptively⁷¹⁵ – may also attenuate artificially induced epilepsy.⁷¹⁶

A number of drugs may induce epileptiform activity in the brain. These include chlorpromazine and lithium (markedly), thioxanthenes, butyrophenones and amitriptyline (moderately), and imipramine, monoamine oxidase inhibitors (MAOI)s, thioridazine, meprobamate, chlordiazepoxide and methylphenidate (slightly).⁷¹⁷ This is more likely to occur during the first weeks of treatment or during withdrawal (as may occur too with drugs such as haloperidol that lessen seizure frequency if taken for long periods⁷¹⁸). EA should be used with particular caution during these periods. Even during alcohol withdrawal, although spontaneous epileptiform activity is rare, a ‘photomyogenic’ response may be triggered by photic stimulation. Bearing in mind the similarities between photic and somatosensory stimulation (Ch. 7, ‘cooperativity between neurons’), use of EA at this time should be undertaken with caution. Cocaine use can lead to seizures,⁷¹⁹ and amphetamine may sensitise to subsequent kindling,⁷²⁰ so the possibility of adverse effects from EA in past users should not be overlooked.

Prior exposure to some pesticides may also predispose to seizures.⁷²¹^,⁷²²^,⁷²³

However, regular exercise may inhibit the development of kindling.⁷²⁴

Factors having possible adverse effects on stimulation

The state of the patient

Because acupuncture can sometimes lower temperature,⁷²⁵ it has been suggested that care be exercised when using strong needle manipulation in yang xu patients.⁷²⁶ However, with comfortable (non-stressful) EA, which in general is warming (SubCh. 6.3, ‘generalised temperature increase’), this caution may not apply.

Chronic or acute stress may interfere with 5HT metabolism, and so with many forms of stimulation.⁷²⁷ Anxious patients may be less able to tolerate electrical stimulation,⁷²⁸ and anxiety in general is found to reduce EA effectiveness.⁷²⁹^,⁷³⁰^,⁷³¹ Paradoxically though, it may sometimes result in greater PT elevation.⁷³²^,⁷³³^,⁷³⁴

Patients with low initial PT may respond better to EA than those whose PT is already high.⁷³⁵ However, in those with neurogenic pain (e.g. radiculopathy), LF electrical stimulation of the involved nerve can sometimes cause the pathological pain to be reproduced, if temporarily.⁷³⁶ As mentioned above, exacerbation of chronic pain may occur if descending inhibitory pathways are impaired in some way.

In diabetic rats, the analgesic potency of morphine is reduced,⁷³⁷ probably due to changes in morphine’s kinetics rather than opiate receptor characteristics,⁷³⁸ although direct PAG stimulation does appear to be mediated by different opioid receptor interactions than in normal rats.⁷³⁹ Thus ALS, like morphine,⁷⁴⁰ may be less effective at ‘low doses’ when it comes to clinical practice. However, diabetes does not appear to alter the potency of analgesia produced by delta- and kappa-opioid agonists,⁷⁴¹ so that HF electrostimulation is less likely to be affected. Interestingly, it is possible that hypoglycaemia may potentiate ALS, as it does the antinociceptive effects of aspirin and morphine (at least, when insulin-induced).⁷⁴² A study was also mentioned above in which insulin potentiated LA (above, ‘acetylcholine and insulin’). One implication of these results is that EAA and also LA effects may be reduced in diabetic human patients.

Parkinson’s disease, which decreases cerebral 5HT, might reduce the effectiveness of EA and related methods.⁷⁴³

Patients with recent cerebral haemorrhage should be treated cautiously, owing to the vasodilatory effect of acupuncture.⁷⁴⁴

In rats, opioid receptors may decline during ageing, which may also increase hormonal responses to stress.⁷⁴⁵ It may be necessary to take both these factors into consideration when treating elderly patients with EA. Decline in pineal function with ageing may also reduce opioid-mediated analgesia.⁷⁴⁶ However, ST-36 EA (2 Hz, 1.5 V) was found to facilitate sensory transmission to S1 (in rats where it had been strongly inhibited by β-amyloid topically applied to S1 as a model for the neuronal alterations in early Alzheimer’s disease), although surprisingly it did not modulate this transmission in normal rats.⁷⁴⁷

Adequate nutrition is a prerequisite for SPES⁷⁴⁸ and presumably to some extent for EA too: without sufficient precursor intake the neurotransmitter mechanisms involved may not be able to function efficiently, as very readably described in the book Natural Highs by Patrick Holford and Hyla Cass.⁷⁴⁹ Vitamin B6, for instance, may facilitate some GABA- or 5HT- mediated EA effects.⁷⁵⁰^,⁷⁵¹ In some circumstances, as mentioned (above, ‘hypoglycaemia may potentiate ALS’), hypoglycaemia may potentiate the effects of ALS. Furthermore, hypoglycaemia (e.g. due to fasting) may exaggerate EEG anomalies and so potentially also anomalous responses to EA (particularly in those with a personal or family history of epilepsy).⁷⁵² However, food intake shortly before acupuncture treatment may affect both breathing rhythm and peripheral blood flow,⁷⁵³ which may be misleading, for instance, to a TCM practitioner.

Drug interactions

(This section can usefully be read in conjunction with Table A3.1 in Appendix 3)

Any strong stimulation to the head in a patient on anticoagulant drugs should be undertaken only with great caution, because of its potentially vasodilatory effect (this is an extrapolation from findings with ECT, electroconvulsive therapy).⁷⁵⁴ This may even apply to those taking such seemingly innocuous drugs as aspirin, or alternatives such as gingko biloba or garlic.

Caffeine (and theophylline), which is found in coffee, tea and cola drinks, may reduce the effectiveness of HF stimulation, whether EA, TENS or vibration, through antagonistic effects on adenosine⁷⁵⁵ (SubCh. 6.2, ‘two or three cups of coffee’). Caffeine may also affect peripheral blood flow and breathing rhythm,⁷⁵⁶ as well as increase circulating βEP.⁷⁵⁷ Foods or beverages containing these stimulants should therefore be avoided for some hours prior to HF (and probably LF) EA treatments.⁷⁵⁸ However, caffeine does not adversely affect SPES.⁷⁵⁹

Any drug that enhances spontaneous EEG rhythms within a particular frequency range may potentiate the effect of strong EA or TENS at similar frequencies.⁷⁶⁰

Monosodium glutamate (MSG) may adversely affect SPES if the blood/brain barrier is leaky,⁷⁶¹ and also EA⁷⁶²^,⁷⁶³ (presumably in the same circumstances). Because it selectively affects the hypothalamic NArc (see SubCh. 5.1), MSG is more likely to reduce the effectiveness of CEA than TLEA.⁷⁶⁴

Whereas chronic L-DOPA (for Parkinson’s disease) may deplete 5HT (and so interfere with EA), the antidepressant sertraline, which inhibits 5HT reuptake and is often given for Parkinson’s, may improve response to EA and allied methods.⁷⁶⁵

Antihistamines may counteract SPES⁷⁶⁶ and possibly EA (SubCh. 6.2, ‘peripheral histamine synthesis may initially decrease’). SPES may also interact with other drugs (e.g. cortisol-like agents, calcium channel blockers, ACE inhibitors⁷⁶⁷) as mentioned in the CES section above (‘calcium channel blockers’ and ‘cortisol-like agents’).

Benzodiazepines do not interfere with SPES.⁷⁶⁸

SPES may paradoxically potentiate the effects of drugs such as cocaine, opiate derivatives or alcohol (if used before intake).⁷⁶⁹ Such an effect has not been reported for EA, although there are anecdotal reports that acupuncture too may ‘enhance’ some illicit drug effects.⁷⁷⁰

Similarly, certain drugs may synergise with EHF, leading to overdose effects.⁷⁷¹

Prenatal exposure to alcohol modifies central opioid analgesia systems in adult rats.⁷⁷² It is possible that patients so exposed may exhibit potentiated EAA.

Other treatments and environmental factors

Prior surgical intervention may reduce the effect of EHF.⁷⁷³ This has been observed for other modalities as well, including EA, TENS and LA (see SubCh. 9.1). Whether patients who have undergone surgery but still require treatment are more likely to be unresponsive to any further intervention, or predominantly to those that depend on intact endogenous mechanisms for their effectiveness, requires further research.

It has been suggested that prolonged prior exposure to electricity might reduce the effectiveness of MENS or CES.⁷⁷⁴ This is in line with early reports by French acupuncturists that prior electrotherapy appeared to reduce the effectiveness of traditional acupuncture.⁷⁷⁵ To my knowledge, these impressions have not been substantiated, however. Also, although the ‘electromagnetic smog’ in which many of us now live may well have its own adverse effects on our health⁷⁷⁶^,⁷⁷⁷ (Ch. 4, ‘electromagnetic smog’), I do not know of any clear evidence that this interferes directly with any form of acupuncture treatment.

However, when more subtle forms of stimulation such as SPES are used in close proximity to other devices (such as power lines, VDUs, mobile phones or even hearing aids), the strong signals from these may indeed disrupt the effects of treatment.⁷⁷⁸ Conversely, any electrical stimulator such as TENS or EA may interfere with other sensitive electronic apparatus. Their use for patients with cardiac pacemakers is thus generally considered to be contraindicated in case the output of the latter is affected (see Ch. 12). TENS (or epidural stimulation) may also create artifacts in ECG readings⁷⁷⁹^,⁷⁸⁰ that might be misinterpreted as a malfunctioning pacemaker,⁷⁸¹ although there are technical solutions to this problem, at least for HF TENS.⁷⁸² EA, of course, is unlikely to be overlooked in the same way!

Despite the accepted view that electrical stimulation should not be used in pacemaker patients, in one case report careful monitoring failed to observe any effect on the ECG in a patient using TENS at sites non-local to the (temporary) pacemaker (two anterior electrodes near the midaxillary line and two dorsal and paraspinal, both pairs positioned in the left fifth and tenth intercostal areas).⁷⁸³ In another study of 51 patients with 20 different types of permanent pacemakers, no episodes of interference, inhibition or reprogramming of the devices occurred with TENS at four different sites (lumbar area, cervical spine, left leg and lower arm area ipsilateral to the pacemaker), although locations parallel to the line of the pacemaker electrodes were not attempted.⁷⁸⁴ Thus, if electrodes are positioned carefully, interactions are not inevitable.

Experimental measurements

Electrodermal measurements at AcPs and TrPs in response to treatment: in search of the objective

Preamble

A number of EA systems enable the practitioner to locate points that give unusual electrical readings (generally of skin conductance, SC), treat the points and then take further measurements to demonstrate that readings have changed. It is generally assumed that electrical properties of AcPs tend to normalise with treatment⁷⁸⁵ (even dental treatment⁷⁸⁶). This may be interpreted purely as a local reflex phenomenon (as in relief of muscle tension and pain), or as corresponding in some way to a normalisation of energetic function with possible eventual repercussions on organ function. Leaving aside issues of practitioner or patient expectation, self-delusion and the considerable leap of faith involved in relating electrical measurements at traditionally defined AcPs to the state of internal organs whose functions may be very different from those described in traditional acupuncture models (see Ch. 10), as well as the practical difficulties involved (see SubCh. 5.2), there is something very alluring about methods that give gratifyingly instant and persuasive feedback on treatment effectiveness.

Such methods may not use electrical measurement at all. Pressure at TrPs is commonly used before and after treatment to check and demonstrate response. Heat (from a lighted moxa stick) has been used to locate sensitive points, treat them and demonstrate ‘cure’ when strong sensation is no longer found.⁷⁸⁷ Muscle testing is used in applied kinesiology (AK).⁷⁸⁸^,⁷⁸⁹ One form of muscle testing, Yoshio Omura’s ‘bidigital O-ring test’ is popular with some acupuncturists⁷⁹⁰^,⁷⁹¹^,⁷⁹²^,⁷⁹³ (see SubCh. 9.12) and so on. Traditionally, of course, pulse taking is just such a form of ‘measurement’; Paul Nogier’s ACR (auriculocardiac reflex, sometimes known as VAS, or vascular autonomic signal)⁷⁹⁴^,⁷⁹⁵ is a modernised version.

All these methods involve interaction between subject/patient and practitioner, and occur within a particular context. Although the focus may be on what the tests shows, their results are subject to many influences that cannot be consciously accessed at the time.⁷⁹⁶ In particular, they all involve muscle activity by the practitioner, and if AK has taught us anything it is that muscle strength can be affected by very many things, from the weather or stray electromagnetic (EM) fields to the state of mind of both parties or what they had for breakfast. If these can be filtered out then the test may show what it purports to.

Sadly, though, the holy grail of objective measurement remains elusive. This has been demonstrated even with trained or experienced practitioners for pressure sensitivity at TrPs⁷⁹⁷ (which anyway may not change in response to temporary suppression of myofascial pain⁷⁹⁸), for AK⁷⁹⁹^,⁸⁰⁰ (although some interesting results have been obtained with handheld dynamometers,⁸⁰¹^,⁸⁰² and with AK for certain muscles⁸⁰³) and for traditional pulse taking,⁸⁰⁴^,⁸⁰⁵^,⁸⁰⁶^,⁸⁰⁷ except when considering a very limited range of parameters.⁸⁰⁸^,⁸⁰⁹^,⁸¹⁰ Also, any acupuncturist who has used the Akabane test (testing points for sensitivity with a moxa stick) knows how tricky that can be.

However, this is not to belittle any of these methods. Most diagnostic procedures do not provide definitive information about the true state of a patient but, being imperfect, only probabilities. All require interpretation as well. Skilled acupuncturists from different schools may read the pulse in very different ways and use very different treatment protocols as a result. Sometimes patients may get better, it seems, almost regardless. Similarly with the other methods mentioned: in the hands of a skilled practitioner and within the total context of that particular healing encounter they can be most useful. Taken out of context, comparing results even from those trained in the same tradition, and more so when comparing results from different teaching lineages, such methods may seem almost laughable to the sceptic looking at them as isolated techniques.⁸¹¹

However, as careful studies on TrP examination have shown, both training and experience are required for interrater reliability.⁸¹²^,⁸¹³ For TrP palpation at least, there are now a few somewhat more positive studies on the reliability of pressure threshold and tolerance measurements,⁸¹⁴ although twitch response, palpable band palpation and referred pain have not always been found clinically useful in determining TrPs, at least in low back pain.⁸¹⁵ David Simons has written an overview of the issues involved.⁸¹⁶

Given all the difficulties involved in these hands-on methods, an interest in electrical measurement is not surprising. At first glance it does seem to offer something both more immediate and rather more valid and repeatable. Electrical impedance measurements have even been used, for instance, in attempts to show that the ACR itself is a reliable and objective phenomenon.⁸¹⁷^,⁸¹⁸ There have also been numerous attempts to objectify pulse taking using pressure and other transducers of various sorts,⁸¹⁹^,⁸²⁰^,⁸²¹^,⁸²² as well as devices designed to simulate different pulse qualities to assist students in learning them. Inevitably perhaps, these systems fail to do justice to the nuances involved.⁸²³

Back to the point

If a point is stimulated, its characteristics will inevitably change. Both needling (see Fig. 5.9A) and surface stimulation (see Ch. 4, ‘a characteristic pricking or burning pain’) will affect SC, and needling will cause skin potential (SP) changes.⁸²⁴ Electrical stimulation in any form will presumably perturb local current and SP patterns, at least temporarily, with possible further repercussions (for instance, Darren Starwynn claims that greater changes following treatment are more likely to signal subsequent symptomatic improvement⁸²⁵).

Pressure at a point will affect the piezoelectric characteristics of underlying tissue. Ultrasound treatment (of the inner ear) may result in increased SC at nearby (auricular) AcPs.⁸²⁶ Vibration likewise will change SP measures.⁸²⁷ More drastically, Russian surgeons using external braces to treat deformities or severe bone lesions discovered in the 1970s that if points of low SC were used as sites for the pins that fixed the brace to the bone then internal organ pathology occurred in debilitated patients, whereas it did not if such points were avoided.⁸²⁸

All in all, it is hardly surprising, as reported in many EA studies,⁸²⁹^,⁸³⁰^,⁸³¹ that electrical measurements at treated acupoints change after treatment (perhaps more than at non-points⁸³²).

A less obvious result is that stimulation using laser may also change electrical characteristics at AcPs, including those of EMG,⁸³³ SP and SC, along with corresponding changes in organ function.⁸³⁴^,⁸³⁵^,⁸³⁶ (It has been suggested that a photoconductive effect may be set in motion when AcPs are illuminated.⁸³⁷) In one randomised double-blind study, for example, three successive treatments using HeNe laser at trigger points increased their SR.⁸³⁸ Unfortunately, as Peter Baldry points out,⁸³⁹ there was no mention of whether these changes correlated with subjective changes in pain. In another such study, infrared laser was used to irradiate TrPs in healthy subjects, using a dose of 1.5 J/point, in a double-blind placebo-controlled crossover trial. Significant differences in TrP PT were found between laser and placebo, with increases in PT also at non-treated TrPs.⁸⁴⁰

Such changes at untreated points are perhaps of more interest than the more predictable changes at the points that have been stimulated (see SubCh. 5.2, ‘points both on the same and on other meridians’). For instance, EA at body points (ST-36, BL-20 through to BL-21) that enhanced healing of experimental gastric ulcer in rats was accompanied by improved microcirculation and yet reduced SC at auricular points (whether this was a generalised auricular response, or at specific points only, was not stated).⁸⁴¹ In contrast, when unspecified muscular regions of the body were stimulated or injured in rabbits, auricular SC noticeably increased proportionally with the strength of stimulation.⁸⁴² Such studies on auricular changes in response to treatment often do not state whether measurements were carried out solely at AcPs where changes would be expected or at other unrelated points as well.⁸⁴³

The authors of one study on phantom leg sensation (not pain) in an amputee reported that contralateral HF TENS (at 100–120 V) reduced paraesthesiae and concurrently changed stump SC, whereas auricular pTENS (4 Hz, 20–24 V) or placebo TENS produced only insignificant changes.⁸⁴⁴

One intriguing Lithuanian study found that 1 Hz or 28 GHz stimulation of points on the Stomach, Spleen, Liver and Small Intestine meridians affected electrical potentials in stomach and small intestine organs, while stimulation to these organs was also reflected in SP changes at the AcPs. Small Intestine and Stomach meridian points were most responsive.⁸⁴⁵ The authors of one early American report observed that following cholecystectomy (but not herniorrhaphy), points such as GB-24, GB-25 and GB-26 (on the right side) became more easily detectable electrically, but were not convinced that this had anything to do with visceral condition.⁸⁴⁶ In a less invasive demonstration of the relationship between internal organs and associated AcPs, Steven Rosenblatt demonstrated that when, on the one hand, subjects used audio biofeedback to alter their own heart rate, an increased HR was accompanied by increased SC at HE-7, and decreased HR by lower SC. On the other, when they used the audio signal to change SC at this point, the HR would also change correspondingly. The order of HR and SC biofeedback did not affect results, and no relationship between HR and SC could be demonstrated when LU-9 or a nearby point not on any meridian was used instead of HE-7.⁸⁴⁷

However, such organ correlations are not found in all conditions. Under EAA for ophthalmic (cataract) surgery, for example, SR at xi (Accumulation) points of all meridians except the Kidney decreased more than with conventional local anaesthesia (during which it increased at some points), the difference presumably being due to sympathetic activation under insufficient EAA.⁸⁴⁸ And under general anaesthesia for gastric surgery, SR gradually declined on all meridians except the Kidney and Bladder (with no particular changes on Stomach).⁸⁴⁹

Changes in electrodermal activity (EDA) may occur in response to rather more subtle stimulation too. As mentioned above (see SubCh. 5.2, ‘under hypnosis’), potential gradients in the body alter under hypnosis. Viktor Adamenko, using electrodes of different metals,⁸⁵⁰ reported that both SC and the potential difference between linked AcPs may change in a hypnotised patient according to the proximity of the hypnotist, and even with their unvoiced mental suggestions from a distance!⁸⁵¹

Of course, there are also studies that show no demonstrable effect on SR from EA (although in one very small study (see above, ‘In one very small study (N = 4)’), no changes in PT were found either, with frankly uncomfortable motor level 3 Hz 0.2 ms EA at SJ-8, GB-21 and LI-4).⁸⁵²^,⁸⁵³ There are also reports that the effectiveness of EA may depend on the prior bioelectric characteristics of associated AcPs.⁸⁵⁴

Electroacupuncture according to Voll (EAV)

Despite the great influence that Reinhold Voll’s EAV (SubCh. 5.2, ‘Reinhold Voll’s EAV’) has had throughout the world on the development of numerous EA-based diagnostic systems, there are surprisingly few experimental studies on its validity or repeatability. Voll himself never conducted any, so far as I have been able to ascertain through quite extensive enquiries,⁸⁵⁵^,⁸⁵⁶ although he was well aware that such research was needed.⁸⁵⁷ As for remedy testing, one of the two lynchpins of EAV (see SubCh. 5.2), one author wrote in 1990: ‘that no formal studies have been conducted in the last 35 years is a cause for concern’.⁸⁵⁸

A small Swiss study on EAV methodology carried out by A Comunetti with colleagues from the pharmaceutical company Hoffmann-La Roche⁸⁵⁹ found that conductivity measurements (at jing (Well) points LU-11, LI-1, P-9) are relatively independent of the degree of skin humidity under the probe, and also of the applied voltage (provided it does not exceed about 1 V), whereas both conductance and its change over time (indicator drop, ID) are strongly affected by the electrode material used. Used on dry skin, the polarity of the probe may affect readings: with chemically active electrodes (silver and brass) a positive probe may help to overcome epidermal resistance.

Systematic measurements on points with the highest local SC give ‘reasonably good reproducibility’ of SC and ID readings if repeated within about 20 minutes, but vary considerably outside that period. They also vary greatly between individuals. The authors, who gave no statistical analysis in their report, concluded that ‘to diagnose under the usually less rigorous conditions in medical practice seems to us a difficult task in which the physician’s intuition is predominant’. As one experienced EAV practitioner was forced to conclude, ‘when dealing with the complexity of the human then possibly truth is ultimately subjective rather than an external’.⁸⁶⁰

Another experimental study from a major journal located in my literature search merely describes how, when EAV was used to measure SC and ID at the 24 jing points in subjects before and after performing their accustomed qigong exercises, SC and ID readings changed in various ways. As some subjects performed sitting and others moving qigong, it is not surprising that the changes were not consistent over the different subjects when this test was carried out blinded; however, they had been when subjects were unblinded and in a conference situation,⁸⁶¹^,⁸⁶² with SC increasing and IDs decreasing as might be expected.

In another unblinded study from Russia, exposure to negative ions was found to normalise EAV readings in 87% of subjects and worsen them in 13% (surprisingly, in no subjects did they remain unchanged!).⁸⁶³

A serious attempt to put EAV on an objective footing was made by Orm Bergold. He first investigated the effects of different electrode materials on ID, then developed a method to maintain constant probe pressure on the AcP and observed the effects of measurement duration and intervals between measurements on readings, as well as their inherent variability. Bergold went on to demonstrate that the results of Voll’s remedy test are an artifact of the measurement situation, pointing out that some EAV users actually ‘confirm’ their results using pendulum diagnosis, and that Voll himself told him that he did not need to carry out the remedy test because he could sense what patients lacked and which remedy would be useful to them (based on their medical history and his own extensive experience). Bergold also commented that the test may still give results even when leads from the remedy well or handheld electrode are not connected to the measurement device.⁸⁶⁴ As Keith Scott-Mumby has kindly put it, ‘the body itself acts as the most sensitive quantum detector we have’.⁸⁶⁵

Voll’s ‘quadrant’ conductance system (see Fig. 10.13d) was assessed by a group of 18 German doctors in the mid 1970s. Results depended on the age of the patient and on weather conditions, but did not correlate with blood pressure. Remembering that this test does not make use of handheld probes on AcPs, and so is less influenced by pressure artifacts, it is perhaps not surprising that readings were fairly consistent between doctors and their assistants.⁸⁶⁶

Dutch EAV studies

Rather more meaty are two detailed Dutch reports, by Roel van Wijk and Fred Wiegant of Utrecht University.⁸⁶⁷^,⁸⁶⁸ In the first of these they demonstrated that, using a Dutch device based on Voll’s original instrument, SC readings on the fingers are greatest at the jing points (or the equivalent non-traditional points at the other angles of the nails), on both fingers and toes. They therefore used these points almost exclusively, and in their first study made the following observations:

The point of greatest SC may shift location within a given area, occasionally leading to widely differing readings on repeated measurements (despite absence of visible skin damage).
SC varies considerably between different points, with a standard deviation (SD) of approximately 10% of the average value.
Provided the monitored downward force on the probe tip is kept within a ‘critical zone’ of 120–140 grams, SC variability at a given point is always < 5% (the type or amount of water used to moisten the skin beforehand and the size of area moistened produce unimportant variations).
SC varies with the length of time that pressure is exerted on a point, as well as the time interval between repeated measurements at the same point.
Even with only a short application of the probe, SC changes.
This change is generally an ID in the short term, and persists more if greater pressures are used.
Even after the probe is removed, SC may continue to change (whether or not subsequent measurements are carried out), but may increase, decrease or remain unchanged; these changes are dependent on the pressure, but not on the electrical current used (a similar conclusion had been reached many years before by Abraham Noordegraaf and Dennis Silage,⁸⁶⁹ see SubCh. 5.2).
Such changes may persist (e.g. initially decreasing for 30 seconds and then increasing) for at least 15 minutes (or 25 minutes if there are no further measurements).
Thus, if two experienced testers measure SC at a series of points within a short period, their findings do not agree (correlation is non-significant), although there is significant correlation of their average SC readings.

Van Wijk and Wiegant also found that:

When particular homeopathic remedies in sealed tubes are placed in contact with a brass plate electrically connected to an electrode held by the subject, the pressure-induced ID at selected nail points may be disinhibited.
This does not occur if control tubes, similarly labelled, are used (in a possibly double- but not triple-blind manner), or tubes containing placebo substances.
For a given remedy, this increase occurs in a similar manner at the nail point and another related EAV point nearby, but may not occur at other unrelated points.

Because of the complexity of the procedure, van Wijk and Wiegant recommend that only those conversant with this type of measurement should engage in this type of research.

In a second study, on homeopathic remedy testing, authored by van Wijk alone (although assisted by Wiegant and others), results were as follows:

Statistically significant differences in SC are measured when verum and placebo remedies are tested blind.
However, under blind conditions, serious errors are also found (false positives from placebo tubes).
With repeated measurements, ID is reduced or may disappear completely (increasing the risk of subsequent false positive results).
However, repeated measurements using one test remedy show little variation, greater changes occurring on substitution of a different remedy (although others have found that repeating the identical test clearly can affect readings⁸⁷⁰).
The results of the medicine test are linked to the way pressure is applied at the measurement point used: differences in pressure lead to different readings.
However, once the measurement probe has been applied at the optimum pressure (~ 120 g), further increase in pressure does not tend to precipitate a rise in SC (as Niboyet found, the point may become unresponsive; see SubCh. 5.2, ‘impossible to regain’).
Some visual contact between subject and tester is necessary.
If readings on a subject are ‘normalised’ in response to specific remedies by one tester (correcting abnormal baseline measurements to a reading of ‘50’ on the meter of the EAV device), a second tester will not obtain the same results when measuring the same points in response to the same remedies. In other words, results are specific to the interaction between subject and tester.

Van Wijk concludes that the changes in SC found in EAV do not indicate altered conditions in the subject, but are the result of pressure changes on the probe that must be attributed to the tester (as other researchers have also suggested⁸⁷¹^,⁸⁷²). Sealed tubes containing remedies somehow influence the (in)voluntary (un)conscious fine tuning of the tester’s muscle tone during the interaction with the test subject. Thus, as William Tiller has also concluded,⁸⁷³ the tester is using himself as an intermediary in selecting remedies within the context of the total situation, and any errors found in artificial blind experiments cannot be extrapolated to the very different clinical situation.

A British EAV study

Using a protocol based on the Dutch studies, together with input from a skilled EAV practitioner (Ken Andrews) and Fred Wiegant (although ignoring his advice that only experienced testers should carry out EAV research!), Riccardo Cuminetti and the present author performed a small pilot study on two subjects (A and B) using a premarket prototype EAV measurement device, the EBM2, which Riccardo Cuminetti had designed in collaboration with Ken Andrews. Because measurement is time consuming and tiring at least when beginning this kind of work, we decided to concentrate on hand points only, using jing points and also Voll’s ‘control measurement points’ (CMPs) proximal to these.

Our primary objectives were to determine whether SC readings at AcPs were repeatable, and also sensitive to the effects of a non-specific 20-minute treatment using a CES device: the Liss Bipolar Cranial Stimulator (SBL 202-M) (see above, ‘optimum parameters’). We carried out two sets of measurements in succession before the treatment, and then two more afterwards. I (DM) acted as the tester. Unfortunately we did not at the time have access to a constant pressure probe, and (possibly mistakenly) maintained pressure on the point for a standard 5 seconds with each measurement. The tester was not blinded to the results of current or previous measurements. Other details of the protocol have been presented elsewhere.⁸⁷⁴

Our results, based on analysis by acupuncturist and research statistician Mark Bovey (of the Acupuncture Research and Resources Centre, then at Exeter University), were as follows:

Only peak SC measurements were used. In a pretest, using Hiroshi Motoyama’s AMI electrodes affixed to the points and allowed to stabilise for some minutes, and just touching the moistened measurement probe to them with minimal pressure, virtually no IDs could be elicited. This indicates that the ID is very likely a pressure artifact, as found by the Dutch group. Even when the probe was used directly on the skin, IDs did not vary sufficiently to be able to apply normal statistical methods. They were thus excluded from subsequent analysis.
Peak SC jing point or CMP readings do not in general vary hugely across meridians within a subject. However, readings on the same points in A and B were quite different. The most likely explanation for this was that A (over 50) and B (aged 13) had very different skin characteristics. High readings on the Lung jing point in both A and B could be accounted for as both tended to bite or pick at the skin around the nails, particularly at LU-11 (Voll places the jing point of the Lung meridian on the ulnar side of the nail, not the radial, as classically located). In subject B, as healing occurred visibly at the right Lung jing point over the course of the study, SC readings dropped dramatically (from ~80 in the first session to ~40 in the third).
There are considerable differences between averaged jing point and CMP readings in the same subject (again possibly owing to differences in skin condition). From such a small sample, it is not possible to state whether measurements are more repeatable at the set of jing points or at the CMPs.
Using light pressure on the probe to locate points initially, it is clear that they can vary in position within quite a short time, as the Dutch group found. However, when using heavier pressure to measure SC, results are rather different. Although there are naturally variations in readings over time as in any functional measurements of a living biological system, indicated here by large variations across sessions (sometimes upwards of 24%!), the EBM2 readings at both jing points and CMPs exhibit considerable and significant repeatability within sessions. (Remember, though, that the tester was not blinded to the results of previous measurements.)

We incorporated safeguards worked out by the Dutch group to minimise effects of repeat readings on SC measurements, and yet still found short-term ‘stability coefficients’ of 0.90 or over, with the electrode removed between readings (a result reported usually only when electrodes remain in situ between measurements⁸⁷⁵).

Large variations across sessions were probably to some extent due to measurement errors, attributable to inexperience, fatigue and inattention.

It is unclear whether treatment with the Liss device altered readings in any consistent way in these subjects. Although there was a statistically significant decrease of averaged maximum SC after treatment, this could be a result of repeated measurement (van Wijk) or learning on the part of the tester. Certainly the very minor differences between pre- and post-treatment readings on three jing points (Triple Burner, Pericardium and Large Intestine) in subject A virtually rule out a treatment effect on these channels in this subject. It may be that readings on foot points would show more responsiveness to this particular intervention, or it may be that using the device was not an appropriate intervention to elicit a rapid response in essentially healthy people.

On the basis of this very small but very time-consuming study using a very stunted version of EAV and rather minimal statistics, my chief conclusion at the time was that this is not a method that gives useful results in the hands of an untrained tester. It requires a great deal of concentration and in practice would, I suspect, reduce the richness and depth of interaction with patients.

Although diagnosis is of course important and has its own therapeutic value, except in extreme situations it should not, in my view, take the foreground in the healing encounter between patient and practitioner, at least within the context of acupuncture practice. Also, if indeed it can take well over a year to become proficient in this method,⁸⁷⁶ before taking it up I would need to be convinced that it offers real benefits beyond those of traditional acupuncture (for the practitioner as well as for the patient).

The short-term stability of readings is impressive, but more studies using experienced practitioners are needed to show its accuracy in the longer term. As van Wijk comments, in EAV ‘there is a clear tendency to regard the range of knowledge as a closed system, an approach in which diagnosis and therapy are mutually supportive. This assumption is delusive and may cause one to be blind to one’s limitations.’⁸⁷⁷ Clinical decision making should certainly not depend on such methods alone.⁸⁷⁸ They can only be an ancillary aid, part of a more comprehensive approach to the patient.

An American EAV study

Carlo Calebrese and colleagues from Bastyr University have carried out an unpublished preliminary study using the Computron EAV device.⁸⁷⁹ They assessed reliability and validity of measurements by three different operators, all highly skilled, on six subjects, using three identical machines. Readings were carried out by the same operator on each subject twice, using two different machines. The six subjects included three without major health problems and three each of who had a well-defined but superficially inapparent disease condition. Operators were unaware of this, and were indeed screened from direct visual contact with subjects by a hanging curtain. No verbal communication was permitted between any participants. Ten points on each hand and 10 on each foot were measured (40 measurements in all), selected to provide relatively stable measurements and also an overall health evaluation.

From their careful statistical analysis, the authors concluded that there were no apparent patterns of points distinguishing patients or distinguishing the normal from the ill patients, and that more variation occurs between operators than between subjects (or machines), with operator variance possibly nine times larger than patient variance. Thus, to compare results on different patients would require that they were tested by the same operator. However, their statistical method did not permit any conclusions on whether differences between individual point readings might be useful diagnostically.

In summary, they could see no evidence for the clinical reliability of the Computron instrument.

Other electrodermal measurement systems

Richard Croon’s elektroneural Diagnostik (see SubCh. 5.2) has been criticised for many of the same reasons as EAV. Sturm and Ludwigs, the authors of one such critique,⁸⁸⁰ found for example the following:

Electrical (resistive and capacitative) differences between reactive points and the surrounding skin were not so great as those Croon reported.
Results were dependent on probe pressure, angle and how long it was held on the point, thus on who carried out the measurements and their individual technique.
Thus readings are not easily reproducible (also varying with time of day and season).
There were no differences between measurements on corpses and living people, which suggests that electrical characteristics are unrelated to inner organ function.
Measurements did not depend on blood flow either.
However, results were dependent on the number and activity of sweat pores at the points measured.
No other anatomical substrate could be found for Croon’s Reaktionsstellungen.
Measurements on healthy subjects and patients with specific conditions did not support Croon’s hypothesis that point measurements correlated in any way with disease processes, except possibly purely local inflammation.
In particular, they found no evidence for the notion that abnormal readings could provide early indications of functional disorders.

A variant of EAV, also developed in Germany, is the Prognos meridian measurement system, in which a very small direct current (1.1 µA) is applied very briefly (~223 ms) via a constant pressure probe (~4.6 mm diameter) to determine skin resistance.⁸⁸¹ In one small study, reliability of four repeated skin resistance readings at the jing Well points for all 24 traditional meridians was found to vary between 0.44 and 0.82 for single meridian readings. Overall reliability was 0.72, with lowest reliability at the yang meridian end points. Blinding of the operator and subjects, who all worked in the same clinic, is not mentioned.⁸⁸² In a further study on the Prognos system, this time carried out in the USA, mean reliability of a single measurement carried out in 31 volunteers was 0.758 for a standard measurement protocol of four sequential sweeps of the 24 jing Well points, increasing if the points were marked with ink and the measurements carried out in quick succession. In contrast to the earlier study, yang meridian measurements were more reliable than those at yin meridian jing points.⁸⁸³ However, at body points other than the jing Well points, it seems that even with this well designed device there may still be pressure artifacts from variations in tissue ‘firmness’.⁸⁸⁴ These studies represent an important and much needed step forward in the exploration of electrical acupoint measurement.

Measuring temperature

Temperature and SC increases may occur together in response to acupuncture when qi arrives in an area of ‘obstruction’ where pain (bi) is experienced.⁸⁸⁵ Changes in both have been used as indicators of recovery, for instance in patients with gastro-oesophageal reflux when treated with laser in combination with EA.⁸⁸⁶

As already mentioned (see SubCh. 5.2), temperature measurements at AcPs have been used diagnostically, although whether AcPs are in fact relevant to the results obtained has been strongly questioned. AcP temperature increases have also been found following MA.⁸⁸⁷ Thus infrared thermography may be a useful and relatively objective tool in assessing changes at AcPs in response to both MA and EA.⁸⁸⁸

Simultaneous temperature increases and decreases at different points in the ear have been reported in response to opening or closing the eyes, or even to a brief flash of bright light directed at the ear.⁸⁸⁹

Electrical imaging methods

Techniques such as Kirlian photography (see Ch. 10) use a high-frequency corona discharge to produce images that supposedly represent the body’s bioelectric field. Theoretically, if sensitive enough, they should actually reveal the AcP, whose SP and SC are greater than those of the surrounding skin (see SubCh. 5.2, ‘electrical properties of acupuncture meridians and points’), or meridians, which may exhibit some kind of bioluminescence (SubCh. 5.2, ‘may also have bioluminescent’). However, Kirlian-type effects do not appear to correlate with SC changes.⁸⁹⁰

Unfortunately, experimental studies on these techniques, although very interesting, give very little useful information about acupuncture.⁸⁹¹^,⁸⁹² However, changes in Kirlian and related images have been reported in response to needling some AcPs, but not always,⁸⁹³^,⁸⁹⁴ and changes in ‘bioelectroluminescence’ at AcPs have also been observed in response to intraoral laser stimulation.⁸⁹⁵

One particular disadvantage of these methods is that images are never the same when using HF discharges, even if repeated immediately. Indeed, it may be almost impossible to differentiate between genuine image and artifact.⁸⁹⁶ As one author concluded, there is no difference between Kirlian images of healthy and ill patients, and most variability is explainable in terms of physical factors not psychophysiology (although anxiety or arousal, or even hypnosis, may affect the images obtained⁸⁹⁷^,⁸⁹⁸).

In particular, with fingertip imaging undue importance cannot be attached to the pattern of the corona discharges obtained.⁸⁹⁹ As Rosemary Steel (formerly vice-president of the International Kirlian Research Association, IKRA) has stated, Kirlian is not so much about ‘diagnosis’ as a ‘visual corroboration of process’.⁹⁰⁰

Even with these limitations, Kirlian photography remains fascinating. For example, a study by Gordon Gadsby confirmed that, using black and white film (responsive more to blue-violet rather than red light), denser fingertip images correlate with high electrical resistance measured between the hands, and fainter ones with low resistance, although there were no such correlations with Kirlian photographs of the feet and foot-to-foot resistance. Age and gender contributed significantly to the variations found.⁹⁰¹ (It is tempting to speculate that the greater thermal sensitivity of the foot jing points in men may have some bearing on this,⁹⁰² although skin temperature/vasodilatation does not correlate with Kirlian-type effects.⁹⁰³)

In keeping with bioelectrical measurements (see SubCh. 5.2, ‘Burr’s work’ and SubCh. 5.2, ‘not necessarily for long’), it may be possible to detect ovulation using Kirlian photography,⁹⁰⁴ as well as fingertip images even after death that correspond in some way to the mode of dying or the chronicity of illness prior to death.⁹⁰⁵ There are also claims that Kirlian images of fingers and toes change in response to chiropractic adjustment,⁹⁰⁶ as well as allopathic drugs and homeopathic remedies,⁹⁰⁷ and that – as with SC readings (see SubCh. 5.2, ‘greater left-right asymmetries’) – extreme left–right differences (one side ‘in excess’, the other ‘deficient’) may indicate a chronic condition.⁹⁰⁸

Julian Kenyon, in a minimally reported study⁹⁰⁹ comparing the repeatability of measurements carried out at the same time on consecutive days with three different systems – the Motoyama AMI, Vega Segmental Electrogram and Knapp Plasmaprint (a variant on Kirlian photography) – found coefficients of variation (CVs) respectively of around 9%, 30% and 45%. Taking a CV of 15% or less as indicating good repeatability, this is not very favourable for the second two methods (as for EAV in the hands of an untrained tester, as described above, ‘measurement errors’).

With ‘electronography’, a system developed by Ioan Dumitrescu and colleagues in Romania, a single impulse of known characteristics is used rather than the complex random discharge of Kirlian photography. Dumitrescu suggested that this method captures images of the thin layer of ions immediately next to the body, has claimed that these are reproducible and (although without giving detailed results) that electronography detects nearly 90% of points tender to palpation. The electrodermal points detected may be distributed in a line ‘which can be considered as corresponding to the meridian’.⁹¹⁰ Such lines are evident in some Kirlian photographs too.⁹¹¹^,⁹¹² However, in the latter case I suspect they may indicate nothing more than the particular part of the body nearest to the plate electrode when the discharge occurred.

A more homespun method, using a simple chemical reagent film on the skin beneath a negative copper electrode and a DC of not more than 0.2 mA for 1–5 minutes, has been claimed to reveal high SC points up to 2 mm in diameter.⁹¹³

A rather different approach, without clinical application as yet, involves measuring biophoton emission rate from the fingers, with no applied external field. In one study, more biophotons with wavelengths between 300 and 650 nm were emitted from the fingernails than from the pads of the fingers, with individuals varying as to which finger gave out the strongest emissions.⁹¹⁴

Section summary

Kirlian photography (see Ch. 10) produces images that supposedly represent the body’s bioelectric field. As such, they should reveal both acupoints and meridians, if these can be defined electrically. Unfortunately, the beautiful images of Kirlian photography are more likely to be the product of physical artifact than psychophysiological changes. As a result, experimental Kirlian research gives very little useful information about acupuncture, although there may be correlations between hand-to-hand electrical skin resistance and the images obtained, and MA may (sometimes) affect these images.

Some parallels between Kirlian photography and electrical measurement methods are described. Julian Kenyon has found one variant of the Kirlian method far less consistent than the Vega Segmental Electrogram (itself a variant of Voll’s original quadrant conductance measurement method), with by far the most repeatable results from the Motoyama AMI device (see SubCh. 5.2).

With ‘electronography’, a system developed by Ioan Dumitrescu and colleagues in Romania, a single impulse of known characteristics is used rather than the complex random discharge of Kirlian photography. It is claimed to give more reproducible results, and to detect nearly 90% of points tender to palpation.

Biophoton emission rates from different parts of the body have been measured without the application of external electric fields.

Measurements of magnetic fields

Biomagnetic fields (2–4 mG, 4–10 Hz) at least 1000 times stronger than those of normal people have been detected from the palms of those able to ‘project’ qi, despite the absence of any measurable bioelectrical current that might have induced them.⁹¹⁵ They have also been detected from the hands of ‘therapeutic touch’ practitioners (peaking at 7–8 Hz).⁹¹⁶

In Russian studies, paramagnetic changes in human blood have been detected in response to microcurrent stimulation of AcPs,⁹¹⁷ while magnetic fields detected around the arm during stimulation of various AcPs appeared to indicate that the electrical signal was possibly not being transmitted along nerves but along other pathways.⁹¹⁸ (However, the latter study has been criticised as omitting to take a number of physiological and biophysical factors into account.⁹¹⁹)

Box 6.5.1 details the effects of magnets on muscle.

Box 6.5.1 A note on muscle and magnets

Manaka found that with the North pole of a magnet (Ch. 3, ‘conventional labels are used’) directed towards the skin, pressure pain at an ashi point is reduced, whereas the South pole increases it.⁹²⁰ Correspondingly, Yoshiaki Omura has suggested that when using the bidigital O-ring test, a magnet placed on the part of the body being tested will affect results. He finds that usually the North pole increases muscle strength, and the South pole weakens it,⁹²¹ with opposite results if magnets are positioned over functionally impaired organs. Ambient magnetic fields may also affect the tester’s response⁹²² and so magnets have been used to amplify this in some very different systems, such as the VEGAtest and a number of dowsing methods.⁹²³^,⁹²⁴

Whether there is some physiological explanation for this (Ch. 4, ‘magnetic fields appear to influence’), some interaction with the tester’s muscle,⁹²⁵ with the tester’s own magnetic field, or even with small currents in the arm(s) holding the measurement device, or whether the effect is ‘just psychological’, the fact that these parallels exist is tribute to the archetypal attraction of the mysterious magnet with its bipolar characteristics.

In conclusion

There is clearly a vast amount of experimental research on EA and associated modalities, much of it of sufficiently good quality to be able to draw conclusions that can be tested in clinical practice. Inevitably, however, given the range of topics, animals, approaches, settings and researchers involved, there are contradictions and disagreements, and a need for further research to explore where these lead. Even with these difficulties (or opportunities), it is now clear beyond reasonable doubt that EA, and indeed some of the other non-traditional methods of acupoint stimulation, do have potentially useful effects that are not simply the result of placebo, distraction or poor study design. The next chapter explores some of the mechanisms that have been invoked to explain how these methods may work.