Exploring amplitude in transcutaneous electroacupuncture stimulation - - PowerPoint PPT Presentation

Exploring amplitude in transcutaneous electroacupuncture stimulation (TEAS)

David F Mayor MA, BAc, MBAcC, Hon member AACP Acupuncture practitioner and Visiting Fellow, School of Health and Social Work, University of Hertfordshire, UK

AACP Leeds Conference,

Principal Met Hotel, Leeds 13 October 2018

1

2

This presentation is based on data, gathered during 2015-16, during an experimental study into the effects of TEAS on the brain brain (using electroencephalography, EEG) and heart heart (assessing heart and pulse rate variability, HRV, PRV). With minor adjustments,

• ur findings are likely to be applicable

to electroacupuncture (EA) as well as TEAS.

Background – the study data

2

Contents of this presentation

slides

• 1. Background

[2,4-8]

• 2. Study objectives

[9]

• 3. Materials and methods

[9-20]

• 4. Main findings – amplitude & frequency

[21-30]

• 5. Factors affecting amplitude (& freq.)

[31-39]

• 6. Subjective responses

[40-43]

• 7. Effects on mood

[44-52]

• 8. Factors affecting mood

[53]

• 9. Some conclusions

[54-60]

• 10. Breaking news ... !

[61]

3

4

Background – EA and TEAS parameters

EA and TEAS parameters fall into two main families: ‘low frequency (LF), high intensity’(ALS) ‘high frequency (HF), low intensity’ (TLS). Sometimes the label ALS is restricted to ‘burst’ stimulation (brief LF trains of HF pulses). In addition, there is a middle range, of around 10-20 pps.

LF: Low frequency, 2-6 pps HF: High frequency, 50-200 pps ALS: Acupuncture-like TLS: TENS-like pps: Pulses per second

5

Equinox: alternating monophasic pulses PPS vs frequency (cycles per second, Hz)

PPS = 2 x Hz

6

Background – ‘Strong but Comfortable’ (SbC)

Literature review: Literature review: Most authorities recommend that TENS,TEAS and EA should be experienced as ‘strong but strong but comfortable comfortable’ (SbC), or at the most as ‘non-painful’ and tolerable

• tolerable. In addition, for LF (or burst) stimulation, ‘SbC’

muscle twitches should be elicited. In TENS/TEAS, A-β afferent fibres are mostly involved. EA is also likely to activate A-δ fibres. In either case, if stimulation is strong, C (or type IV) nociceptive (pain) fibres will start to fire, and stress mechanisms may become involved. Stimulation at close to sensory threshold (ST) is unlikely to be clinically effective.

7

STIMULATION SENSATION SHOULD FEEL ‘STRONG BUT COMFORTABLE’

None Some Strong Painful Max possible

So aim for a sensation between ‘Strong’ and ‘Painful’

Background – what we told our participants

ST: Sensory Threshold SbC

a

8

Background – data from an online database

www.electroacupunctureknowlege.com is a comprehensive open-access online database of information from more than 8,000 EA, TENS/TEAS and related clinical studies published before 2004. Studies including the words ‘comfort*’ or ‘toler* under treatment parameters. Parameters were described in terms of ‘comfort comfort*’ in relatively more TENS TENS than EA/TEAS studies, and in terms of ‘toler toler*’ in relatively more EA/TE EA/TEAS than TENS studies.1

Modality All studies ‘comfort*’ % total ‘Toler*’ % total EA 3880 148 3.8% 907 23.4% TEAS 450 22 4.9% 89 19.8% TENS 1170 107 9.1% 126 10.8%

9

Study objectives

To investigate amplitude settings for sensory threshold (ST) and ‘strong but comfortable’ (SbC) responses in our study participants – what affects these, and what do they affect?

Materials and methods

66 participants were recruited. They attended for four sessions, a week or more apart, in which TEAS was applied for 20 minutes to each hand between the acupuncture points LI4 (hegu) and the ulnar border (JR Worsley’s location for SI3, houxi).

TEAS Stimulation at 2.5 pps

11

Study protocol

In each session, TEAS was delivered at a SbC intensity and at either 2.5, 10 or 80 pps. In the fourth session, for sham sham TEAS, amplitude was set at zero and frequency at 160 pps. The order of sessions in which each frequency was used was varied in a balanced manner. Four participants dropped out after their first session, and a fifth after the third session. 2.5 pps 65 sessions 10 pps 64 sessions 80 pps 61 sessions Sham 61 sessions Total Total 251 sessions

12

Online initial questionnaire

8 x 5-minute slots Baseline Stim1 Stim2 Stim3 Stim4 Post1 Post2 Post3

Physiological measurements (EEG, HRV, PRV, temp, head movement)

Online initial questionnaire Online debriefing questionnaire

Session – time line Room temp Room temp Room temp Room temp

13

Online initial questionnaire

8 x 5-minute slots Baseline Stim1 Stim2 Stim3 Stim4 Post1 Post2 Post3

Physiological measurements (EEG, HRV, PRV, temp, head movement)

Online initial questionnaire Online debriefing questionnaire

Session – time line e.g. past EA or TENS lifestyle … which session: most/least change most/least intense most/least pleasant

14

Online initial questionnaire

8 x 5-minute slots Baseline Stim1 Stim2 Stim3 Stim4 Post1 Post2 Post3

Physiological measurements (EEG, HRV, PRV, temp, head movement)

Online initial questionnaire Online debriefing questionnaire

Initial session questionnaires Post session scales PSS NRS‐M (month) Sleep/tired Rate intensity/pleasantness

15

NRS for Mood (NRS-M) Used for ‘how you have felt in general over the past month’ and ‘how you feel right now’

16

Online initial questionnaire

8 x 5-minute slots Baseline Stim1 Stim2 Stim3 Stim4 Post1 Post2 Post3

Physiological measurements (EEG, HRV, PRV, temp, head movement)

ST 1SbC 1 (R  L)

Online initial questionnaire Online debriefing questionnaire

ST 2SbC 2 (R  L) ST: Sensory threshold SbC: Strong but Comfortable

17

Online initial questionnaire

Outcome measures

8 x 5-minute slots Baseline Stim1 Stim2 Stim3 Stim4 Post1 Post2 Post3

Physiological measurements (EEG, HRV, PRV, temp, head movement) Post session scales NRS-M 1 BRUMS 1 NRS-M 2 NRS-M 3 NRS-M 4 BRUMS 2 NRS-M NRS-M Numerical rating scale for mood BRUM BRUMS Brunel Mood Scale Chan Change ge: : NRS 3 – NRS 1 Chan Change ge: : BRUMS 2 – BRUMS 1

Online initial questionnaire Online debriefing questionnaire

Initial session questionnaires

18

Brunel Mood Scale (BRUMS) Used for ‘how you feel right now’

19

Online initial questionnaire

Outcome measures

8 x 5-minute slots Baseline Stim1 Stim2 Stim3 Stim4 Post1 Post2 Post3

Physiological measurements (EEG, HRV, PRV, temp, head movement) Post session scales NRS-M 1 BRUMS 1 NRS-M 2 NRS-M 3 NRS-M 4 BRUMS 2 NRS-M NRS-M Numerical rating scale for mood BRUM BRUMS Brunel Mood Scale Chan Change ge: : NRS 3 – NRS 1 Chan Change ge: : BRUMS 2 – BRUMS 1

Online initial questionnaire Online debriefing questionnaire

Initial session questionnaires

RT RT RT RT

20

Statistics

Almost all our data was not ‘normally distributed’2 Therefore ‘nonparametric’ methods of analysis were used: Simple change counts, and medians and quartiles rather than averages and standard deviations.

Nonparametric methods used for differen differences ces

• Wilcoxon and Sign tests (for pairs of repeated measures)
• Mann-Whitney U test (for pairs of unrelated samples)
• Friedman test (to compare >2 repeated measures)
• Kruskal-Wallis test (to compare >2 unrelated samples)
• Binomial test (to compare counts

counts of pairs: M/F, or +/-) Nonparametric method used for similarities similarities

• Correlation using Spearman’s ρ

Limitati Limitations

• ns: Multilevel modelling methods and

post hoc corrections were not used.

Level (0‐10) 2.5 pps (N=63) 10 pps (N=64) 80 pps (N=61) Amplitude median range median range median range L ST 1 2.9 3.1 2.3 3.2 2.0 3.9 L SbC 1 4.1 7.9 3.6 8.1 3.0 6.1 R ST 1 2.9 4.3 2.7 2.9 2.1 2.5 R SbC 1 4.5 7.6 3.9 6.3 3.1 5.5 L ST 2 3.1 4.8 2.9 3.8 2.5 3.0 L SbC 2 5.0 8.5 4.0 7.8 3.2 6.5 R ST 2 3.1 4.2 2.9 3.2 2.9 3.7 R SbC 2 5.1 8.4 4.0 7.1 3.6 6.1

21

Main findings

• 1. Stimulation amplitudes vary with frequency

frequency Median amplitudes were higher at 2.5 pps than at 10 pps, and these in turn were higher than at 80 pps, for both SbC and (most) ‘sensory threshold’ (ST) levels.

2.5 pp 2.5 pps (N=63) 63) 10 pps pps (N=64 =64) 80 pp 80 pps (N=61 =61) Level Level median range median range median range L ST L ST 1 2.9 3.1 2.3 3.2 2.0 3.9 L Sb L SbC 1 4.1 7.9 3.6 8.1 3.0 6.1 R ST R ST 1 2.9 4.3 2.7 2.9 2.1 2.5 R Sb R SbC 1 4.5 7.6 3.9 6.3 3.1 5.5 L ST L ST 2 3.1 4.8 2.9 3.8 2.5 3.0 L Sb L SbC 2 5.0 8.5 4.0 7.8 3.2 6.5 R ST R ST 2 3.1 4.2 2.9 3.2 2.9 3.7 R Sb R SbC 2 5.1 8.4 4.0 7.1 3.6 6.1

L 1 L 2 R 1 R 2

22

1.1. Stimulation amplitudes at different frequencies These differences were significant (disregarding sham TEAS).3 Comparing medians medians for pairs of frequencies4

• a. Effect Size r = Z/√N

Note the increasing ‘effect size’ r as the difference between stimulation frequencies is increased.

2.5 2.5 – 10 pps 10 pps 10 – 10 – 80 pps 80 pps 2.5 2.5 – 80 pps 80 pps Level Level p value e.s. ra p value e.s. ra p value e.s. ra L ST 1 L ST 1 <0.001 mod <0.001 strong <0.001 v strong L SbC 1 L SbC 1 0.001 mod <0.001 strong <0.001 strong R ST 1 R ST 1 0.003 weak 0.001 mod <0.001 strong R SbC 1 R SbC 1 0.003 weak <0.001 strong <0.001 strong L ST 2 L ST 2 0.002 mod 0.001 mod <0.001 strong L SbC 2 L SbC 2 <0.001 mod <0.001 v strong <0.001 strong R ST 2 R ST 2 n.s. weak n.s. v weak <0.001 mod R SbC 2 R SbC 2 <0.001 mod <0.001 strong <0.001 strong

23

1.2. Stimulation amplitudes at different frequencies Counts Counts of signed differences for pairs of frequencies

** p<0.001; * p<0.050; n.s. not significant5

Note the increasing positive-to-negative ratios as the difference between stimulation frequencies is increased.

2.5 2.5 – 10 pps 10 pps 10 – 10 – 80 pps 80 pps 2.5 2.5 – 80 pps 80 pps Level Level Positive Negative tied Positive Negative tied Positive Negative tied L ST 1 L ST 1 43** 10 8 47** 7 6 55** 3 3 L SbC L SbC 1 39* 19 3 52** 6 2 54** 3 4 R ST 1 R ST 1 41** 14 6 38* 13 9 46** 11 4 R SbC R SbC 1 40* 17 4 44** 12 4 53** 6 2 L ST 2 L ST 2 41* 16 4 40** 13 6 47** 7 6 L SbC 2 L SbC 2 39* 15 7 53** 3 3 56** 3 1 R ST 2 R ST 2 38* 19 4 28ns 23 8 41** 13 6 R SbC R SbC 2 47** 11 3 48** 8 3 53** 7

24

1.3. Relationship between stimulation frequency and ST: A little experiment (N=2) ST was estimated at 9 different frequencies, from 160 pps down to 0.625 pps, in 50% steps. After a brief pause, the process was reversed, from LF to HF. SbC was not estimated, because of possible carry-over effects. At each frequency, ST was first assessed on the right hand and then, after the right output had been returned to zero,

• n the left hand, and so forth.

25

1.3. Relationship between stimulation frequency and ST (N=2) Averaged results for the 2 hands (i.e. 4 sets of measurements) suggest a possible (exponential?) relationship between stimulation frequency and ST:

ST vs stimulation frequency for two individuals, A & B.

Note difference between av STs for A ( ) and B ( ) Av A 3.4 Av B 3.1

26

• 2. There is a clear increase between ST and SbC levels.

This occurred for both measures (1 & 2) on each hand (L & R).

Significance Significance of differences6

• a. Effect Size r = Z/√N

Note that effect size is high throughout. Counts Counts of differences

** p<0.0017 Level Level ST ST  SbC SbC (2.5 pps) 2.5 pps) ST T  SbC (10 pps) SbC (10 pps) ST T  SbC (80 pps) SbC (80 pps) p value e.s. ra p value e.s. ra p value e.s. ra SbC – SbC – ST (L1) T (L1) <0.001 v strong <0.001 v strong <0.001 v strong SbC – SbC – ST (L2) T (L2) <0.001 v strong <0.001 v strong <0.001 v strong SbC – SbC – ST (R1) T (R1) <0.001 v strong <0.001 v strong <0.001 v strong SbC – SbC – ST (R2) T (R2) <0.001 v strong <0.001 v strong <0.001 v strong Level Level ST ST  SbC (2.5 SbC (2.5 pps) pps) ST T  SbC SbC (10 pps) 10 pps) ST T  SbC (80 pps) SbC (80 pps) Positive Negative tied Positive Negative tied Positive Negative tied SbC SbC – ST (L1) ST (L1) 58** 5 61** 3 59** 2 SbC – SbC – ST (L2) T (L2) 60** 1 2 61** 3 58** 1 1 SbC – SbC – ST (R1) T (R1) 60** 2 1 62** 1 1 60** 1 SbC – SbC – ST (R2) T (R2) 61** 1 1 60** 2 2 58** 1 1

27

• 2. There is a clear increase between ST and SbC levels.

Median differences In general, ST to SbC differences decrease with frequency. Thus, practitioners using TEAS (or EA) may need to increase amplitude increase amplitude more cautiously at higher more cautiously at higher stimulation frequencies stimulation frequencies. Three participants were particularly sensitive to this difference, preferring that stimulation not be increased at all

• r even decreased once ST was reached.

28

• 3. There is a clear increase

increase in ST and SbC levels during each session (ST1ST2, SbC1SbC2)

Significance Significance of differences8

a. Effect Size r = Z/√N

Counts Counts of differences

** p<0.0019

Note increasing effect sizes and greater positive-to-negative ratios at higher frequencies. However, only 4 participants showed consistent increases (positive changes) in all 4 levels in all of their 4 sessions.

Level Level 12 (2.5 2 (2.5 pps) pps) 12 (10 2 (10 pps) pps) 12 (80 2 (80 pps) pps) p value e.s. ra p value e.s. ra p value e.s. ra L ST L ST <0.001 mod <0.001 strong <0.001 strong L SbC L SbC <0.001 strong <0.001 mod <0.001 strong R ST R ST <0.001 strong <0.001 strong <0.001 strong R SbC R SbC <0.001 strong <0.001 strong <0.001 strong Level Level 12 (2.5 2 (2.5 pps) pps) 12 (10 2 (10 pps) pps) 12 (80 2 (80 pps) pps) Positive Negative tied Positive Negative tied Positive Negative tied L ST L ST 43** 8 12 45** 7 12 49** 6 6 L SbC L SbC 46** 10 7 47** 9 8 48** 7 6 R ST R ST 45** 7 11 48** 4 12 53** 4 4 R SbC R SbC 49** 8 6 43** 7 14 46** 8 7

29

• 3. Increase in ST and SbC levels during each session

(ST1ST2, SbC1SbC2)

Median values At 2.5 pps, ST appears to increase least, and SbC to increase most.

2.5 pps 10 pps 80 pps

30

3.1 Do ST and SbC levels increase between sessions?

Median values There is an apparent increase for SbC at 2.5 pps, but this is not significant.10

2.5 pps

31

• 4. Many factors could affect ST and SbC levels

For example: 4.1 Age 4.2 Gender 4.3 Prior experience of EA or TENS 4.4 Perceived stress over the past month 4.5 Other recent mood states 4.6 Fatigue 4.7 Room temperature

32

4.1 Age

Age (greater or less than the median age of the complete sample) had a significant effect on ST but not SbC ST but not SbC levels.11

Median age: 43.5

33

4.1 Age

As might be expected, younger people retained mo more se re sensory se nsory sensitivity nsitivity when compared to older people – but there was no difference in SbC level.12

a. Effect Size r = Z/√N

Note that effects of age (r) are weak or very weak, being greatest for STs at 10 pps. Correlations Correlations between age and level were also stronger for STs than SbCs, mostly strongest at 10 pps (moderate e.s.), weakest at 80 pps (weak e.s.).

2.5 2.5 pps pps 10 pps 10 pps 80 pps 80 pps Level Level </> mdn P value e.s. ra </> mdn P value e.s. ra </> mdn P value e.s. ra L ST 1 L ST 1 2.8/3.1 0.020 0.020 weak 2.1/2.6 0.004 0.004 weak 2.0/2.1 0.044 0.044 weak L SbC L SbC 1 4.0/4.5 n.s. v weak 3.4/3.9 n.s. v weak 2.9/3.1 n.s. v weak R ST 1 R ST 1 2.9/3.0 n.s. weak 2.4/2.8 0.010 0.010 weak 2.2/2.1 n.s. 0.19 R SbC R SbC 1 4.5/4.2 n.s. v weak 3.8/3.9 n.s. v weak 3.1/3.2 n.s. v weak L ST 2 L ST 2 2.9/3.3 0.013 0.013 weak 2.7/3.0 0.008 0.008 weak 2.3/2.5 n.s. weak L SbC 2 L SbC 2 5.0/5.0 n.s. v weak 3.9/4.1 n.s. v weak 3.2/3.3 n.s. v weak R ST 2 R ST 2 3.0/3.2 0.023 0.023 weak 2.9/3.0 0.014 0.014 weak 2.8/2.9 0.027 0.027 weak R SbC R SbC 2 5.1/5.1 n.s. v weak 4.0/4.1 n.s. v weak 3.4/3.8 n.s. v weak

34

4.2 Gender

As for age, gender has an effect more on ST than SbC – although not exclusively.

a. Effect Size r = Z/√N Significance of differences by gender: *** p<0.001, ** p<0.01, * p<0.050; n.s. not significant.13

Note that levels are consis consisten tently lower for women than for men ly lower for women than for men, at all frequencies, and are lower at higher frequencies (cf slides 20 & 26).

** * ** * ** ** * * * * ** *** ** *** ** * ** * ** n.s. n.s. n.s. n.s. n.s.

35

4.3 Prior experience of EA or TENS

L SbC was higher in those with prior EA or TENS experience at all frequencies (significantly so at 2.5 pps).14 Differences in R SbC between those with and without prior EA or TENS were only small, and at 2.5 pps R SbC 1 was greater for those without such prior experience. However, median ST or SbC levels were higher among those with prior EA or TENS experience in 16 out of 24 comparisons, the same in 3, and only lower in 5 (all 5 for ST1 and SbC 1).15 Thus those with Thus those with pri prior experi r experience of EA

• f EA or
• r TENS are

TENS are likely to accept likely to accept higher levels of stimul higher levels of stimulation ation (e (especially if specially if this needs t is needs to be be increased increased after 10 minutes). after 10 minutes).

36

4.4 Perceived stress over the past month

Cohen’s Perceived Stress Scale (PSS) was used in the first and last sessions of the study to assess stress experienced over the previous month. In session 1, PSS was negatively correlated with L SbC1 at 10 pps.16 Nonsignificant negative associations with PSS were found in sessions 1 and 4 for most ST and SbC levels at 10 and 80 pps, but at 2.5 pps most were positive. This suggests a This suggests a tenden tendency y for recen for recent stress stress to reduce to reduce tol toleran ance ce of TEAS

• f TEAS at 10 and 80 pps,

at 10 and 80 pps, but possi but possibl bly to enha y to enhance tol e toleran ance at 2.5 pps. ce at 2.5 pps. This may warrant further exploration.

37

4.5 Other mood states - over the past month and current However, any correlations between NRS mood states

• ver the past month and ST or SbC levels

were either weak or not significant.17 No patterns No patterns of positive or negative correlations with ‘good’ or ‘bad’ moods were observed, either in general

• r at particular frequencies.

For current moods, in session 2, there were weakly significant correlations between ‘good’ mood states such as feeling comfortable, relaxed or in a good mood overall and SbC levels, but only at 2.5 pps.17 Again, no pattern no patterns of positive or negative correlations with other moods were observed.

38

4.6 Fatigue

At the start of each session, participants were asked ‘Do you feel tired today?’, and could answer ‘Yes’, ‘No’ or ‘Don’t know’.

Fatigue had a clinically relevant impact on SbC for 2.5 pps 2.5 pps TEAS, but not 10 pps or 80 pps.18

Significance of differences by gender: *** p<0.001, ** p<0.01, * p<0.050; n.s. not significant.

ST levels at all frequencies were often highe higher when r when not fatigue not fatigued, even if not usually significantly so.

n.s.n.s. n.s. ** n.s. n.s. n.s. n.s. n.s. n.s. n.s. * n.s. n.s. n.s. n.s. n.s. n.s.n.s. n.s.n.s. ** ** **

39

4.7 Room temperature We attempted to keep room temperature constant, but our study was conducted in an old lab in which this was difficult. We also allowed at least 15 minutes for people to acclimatise, but some people took longer than others to do so. For all stimulation frequencies, all correlations between room temperature and ST or SbC levels were negative (except for L SbC 2 at 10pps), and a number of these were significant.19 Thus if room temperature is low, Thus if room temperature is low, it is it is likely that ST and SbC likely that ST and SbC will be high. ill be high.

40

5.1 Subjective response after each session At the end of each session, participants were asked to rate how ‘intense’ and ‘pleasant’ they found the stimulation. Correlations between subjective intensity and objective amplitude as indicated by dial settings were significant only for some STs at 2.5 pps (weak positive correlations) and 80 pps (weak negative correlations!).20 Even the non-significant correlations were negative at 80 pps. Younger (but not older) participants found sham stimulation significantly more pleasant than either 2.521 or 8022 pps TEAS. Overall, , women ra men rated TEAS a ted TEAS as m more pleasant re pleasant (at all frequencies) than men than men, but not significantly so. Younger women (but not men) found 10 pps more pleasant than 2.5 pps TEAS.22

41

5.2 Subjective response from online debriefing questionnaire

Participants were asked to keep notes of their experiences during the study, and then completed the debriefing questionnaire (usually < 1 week after session 4).

‘Most intense’ stimulation was reported most commonly at 2.5 pps (followed by 80 pps), and ‘least pleasant’ at 80 pps (followed by 2.5 pps). 28 participants (43.1%) reported most change most change from sessions in which they found the stimulation most inten most intense, and 30 (46.2%) experienced least least change change during or after sessions in which they found the stimulation least inten least intense. Both these conditions were true for 18 respondents.

pps pps Most Most change change Least Least change change Most Most intense intense Least Least intense intense Most Most pleasant pleasant Least Least pleasant pleasant sham sham 7 19 6 37 15 7 2.5 2.5 13 9 22 4 14 14 10 10 16 13 13 8 12 10 80 80 11 5 18 5 11 18 DK DK 17 19 6 11 12 16

42

5.2 Subjective response from online debriefing questionnaire

• 30 (46.2%) found the most intense

most intense stimulation least pleasant least pleasant

• 10 (15.4%) found the most intense

most intense stimulation most pleasant. most pleasant.

• 17 (26.2%) found the least intense

least intense stimulation most pleasant most pleasant

• 8 (12.3%) found the least intense

least intense stimulation least pleasant least pleasant.

• 9 found both the most intense

most intense stimulation least pleasant least pleasant and the least intense least intense stimulation the most pleasant most pleasant

• 3 found both the most intense

st intense stimulation most pleasan most pleasant t and the least intense least intense stimulation least pleasant least pleasant.

43

5.3 Comparing subjective response during and after the study

Participants’ scores of intensity and pleasantness of the stimulation during sessions and at debriefing were compared:

a. Counts include ambivalent scores during sessions.

More information was forgotten about the sham stimulation than about the active interventions by the time the online debriefing questionnaire was completed.

Sham Sham 2.5 2.5 pps pps 10 pps 10 pps 80 pps 80 pps When When recorded recorded During session Debrief agrees During session Debrief agrees During session Debrief agrees During session Debrief agrees Greates Greatest intensi intensity ty 2 1 25 12a 21 7 16 7 Least Least intensi intensity ty 56 34a 3 1 3 Most Most pleasant pleasant 23 8 15 5 14 3 11 3 Least Least pleasant pleasant 18 6a 11 3 11 2 25 6 Discrep screpancies cies 50 34 35 39

44

6 6 The effects of stimul The effects of stimulation on outcome (mood) ation on outcome (mood)

6.1 Stimulation frequency

There were no significant differences no significant differences in mood changes among the 4 different stimulation frequencies (2.5, 10 and 80 pps, and sham), as measured by NRS-M1/NRS-M3 or BRUMS1/BRUMS2 differences.23 Comparing medians for pairs of frequencies, there were some significant differences between mood, but effect sizes were all ‘small’ rather than ‘medium’, except for feeling relaxed, for 10 pps vs 80 pps, with 37 participants feeling more relaxed at 10 pps, and 23 at 80 pps.24 Overall mood also showed more change at 10 pps - and a marginally better balance of increases and decreases in mood - than at 2.5 pps25 (35 vs 10 participants) or 80 pps24 (35 vs 25 participants). There were no differences in BRUMS.

These results suggest possible possible slight benef slight beneficia cial effects for effects for 10 pps 10 pps stimulation, but they are hardly very convincing.

45

6.1 Stimulation frequency (continued) In addition, 10 participants (15.2%) showed similar mood changes

• mostly ‘bad’ (?) – irrespective of stimulation frequency

irrespective of stimulation frequency:

Part rticip icipan ant t ID 2. 2.5 5 pps 10 p 10 pps 80 p 80 pps 0 p 0 pps BA BA confused ↑

• verall mood ↓
• verall mood ↓

confused ↑

• verall mood ↓

confused ↑ BH BH anxious ↑ fatigued ↑ anxious ↑ fatigued ↑ anxious ↑ fatigued ↑ fatigued ↑ BI BI lively ↑ lively ↑ lively ↑ BM BM lively ↓ lively ↓ lively ↓ BT BT fatigued↓ fatigued↓ gloomy ↓ fatigued↓ gloomy ↓ gloomy ↓ BU BU comfortable↓

• verall mood ↓

comfortable↓ lively ↓ fatigued ↑

• verall mood ↓

lively ↓ fatigued ↑

• verall mood ↓

comfortable↓ lively ↓ fatigued ↑

• verall mood ↓

CH CH comfortable↓ comfortable↓ comfortable↓ CJ CJ confused ↑ fatigued ↑ confused ↑ fatigued ↑ confused ↑ fatigued ↑ CS CS lively ↓ lively ↓ lively ↓ DA DA comfortable↓ lively ↓ lively ↓ comfortable↓ comfortable↓ lively ↓

46

6.1.1 Stimulation frequency – individuality of response

It is important to remember that acupuncture-related treatments ‘should take account of individual patient variation not just group responses’ (Lundeberg & Lund 2009). For some participants, particular mood scores increased, for others they decreased, and for yet others they were unchanged. The ratios of negative to positive differences were often significant26 More participants experienced a decreas decrease in feeling anxious, lively, confused, gloomy and overall mood than an increase, and more experienced an increase increase in fatigue than a decrease. Consistent changes in feeling lively or fatigued occurred in over half the participants, virtually irrespec irrespective of tive of stimulation frequency stimulation frequency.

Lundeberg T, Lund I. Treatment recommendations should take account of individual patient variation not just group responses. Acupuncture in Medicine. 2009;27(1):31-2

NRS-M NRS-M changes changes Anxious Anxious Comfortable Comfortable Relaxed Relaxed Lively Lively Confused Confused Fati atigued gued Gloomy Gloomy Overal verall mood mood 2.5 pps 2.5 pps 0.019 n.s. n.s. <0.001 n.s. n.s. n.s. n.s. 10 pps 10 pps 0.033 n.s. n.s. <0.001 0.008 n.s. n.s. n.s. 80 pps 80 pps n.s. n.s. n.s. <0.001 n.s. 0.008 n.s. 0.024 sham sham <0.001 n.s. n.s. <0.001 0.024 n.s. 0.004 n.s. counts counts

• >
• > +

+

• ≈ +
• +
• ≈ +

+

• > +

> +

• > +

> +

• < +

< +

• > +

> +

• > +

> +

47

6.1.1 Stimulation frequency – individuality of response (continued) Counts of individual participants showing three or more mood changes in the same direction for the four stimulation frequencies (2.5, 10 and 80 pps, and sham): BRUMS showed itself to be less sensitive to chan less sensitive to change ge than NRS-M, with over 40 of the 66 participants indicating no Pre-to-Post changes in Anger, Depression, Confusion and Tension. Results for Fatigue and Vigour were in line with those for NRS-M fatigued and lively, while those for BRUMS total mood disturbance (TMD) were in keeping with those for NRS-M overall good mood. Stimulation freque frequenc ncy had no y had no eff effect on BRUMS ct on BRUMS changes.

NRS-M NRS-M changes changes Anxious Anxious Comfortable Comfortable Relaxed Relaxed Livel Lively Confused Confused Fati atigued gued Gloomy loomy Overal verall mood mood increase increases 3 13 17 3 4 23 5 12 decr decrease eases 20 23 13 41 14 12 18 21 p value p value27

27

p<0.001 n.s. n.s. p<0.001 p=0.033 n.s. p=0.011 n.s.

48

6.2.1 Stimulation amplitude –‘high’ and ‘low’ relative to the median

For each of the active stimulation frequencies, medians of the four SbCs for each participant (L and R SbC 1 and 2) were calculated, and then the median of the resulting median SbCs. Those participants with median SbC ≥ the group median for that stimulation frequency were considered as ‘high SbC’ types, and those with median SbC < the group median for that stimulation frequency were considered as ‘low SbC’ types. At 2.5 pps, significantly more ‘high SbC’ (15) than ‘low SbC’ (5) participants recorded a decrease in feeling relaxed.28 However, at 10 pps there were more decreases in feeling relaxed for the low SbC participants (13) than for the high (8), and at 80 pps numbers of decreases in relaxation were virtually equal for high (15) and low (14) SbC participants. Otherwise, using this definition of ‘high’ and ‘low’ types, SbC SbC appears ppears to to have had little effect on have had little effect on changes changes in mood in mood using the NRS-M and BRUMS as outcome measures.

49

6.2.2 Stimulation amplitude –‘high’ and ‘low’ quartiles

A more extreme ‘high SbC’ type could be defined as having SbC in the upper quartile of all four amplitude settings (L and R SbC 1 and 2) for at least two out of three of the active stimulation frequencies, and correspondingly for the ‘low SbC’ type (in the lower quartile). Using this definition, 15 participants were ‘high SbC’ types, preferring strong stimulation, and 21 were ‘low SbC’ types, preferring it gentle. ‘High SbC’ types felt significantly more gloomy following 80 pps stimulation than sham stimulation (but ‘low SbC’ types did not).29 In contrast, ‘Low SbC’ types recorded significantly better overall mood after 80 pps than 10 pps, and after 10 pps and sham than after 2.5 pps TEAS.29 BRUMS TMD (total mood disturbance) was correspondingly lower after 10 pps than 2.5 pps.29

50

6.3 Stimulation amplitude – ST to SbC increases (see Section 2)

• At 2.5 pps, the median ST-to-SbC difference (SSD) correlated negatively

with changes in feeling relaxed and lively, suggesting that a greater greater SSD resulted in feeling less less relaxed and lively after treatment (cf Section 6.2.1).

• At 10 pps, SSD correlated negatively with ‘absolute’ change

in overall mood, suggesting that a greater greater SSD resulted in smaller smaller increases or decreases in overall good mood.

• At 80 pps, SSD correlated negatively with ‘absolute’ change

in feeling comfortable, suggesting that a greater greater SSD resulted in smaller smaller increases or decreases in feeling comfortable. However, all these correlations are weak,30 so may not be found in other situations.

51

6.4 ‘To twitch, or not to twitch’ Literature review showed that for LF (or burst) stimulation, ‘SbC’ muscle

twitches should be elicited (see Background).

52

6.4 ‘To twitch, or not to twitch’ (continued)

More participants reported feeling muscle twitching initially during stimulation at 2.5 pps and 10 pps than at 80 pps. The amount of twitch was moderately correlated with SbC levels at 2.5 and 10 pps,31 but less so at 10 than 2.5 pps, for all four SbC levels (L, R, 1, 2). At 80 pps, correlations were mostly weak.32 The presence of twitch also correlated with SbC levels, but less strongly, and again most strongly for 2.5 pps and least strongly for 80 pps. Neither the presen Neither the presence or ce or amount of amount of twitch had any significant effect twitch had any significant effect

• n changes
• n changes in mood

in mood. 6 participants did not report twitch for 2 out of 3 active interventions. They were all ‘low SbC’ types or explicitly found high intensity stimulation unpleasant.

n twitches itches 1 2 3 4 N > 0 > 0 average verage > 0 > 0 2.5 pps 2.5 pps 3 3 5 8 44 60 3.6 10 pps 10 pps 14 6 6 9 38 59 3.4 80 pps 80 pps 11 4 14 6 26 50 3.1

53

7 7 Factors affecting outcome (mood) Factors affecting outcome (mood)

7.1 Age

Older (but not younger) participants reported feeling significantly more comfortable,33 more vigorous34 and in a better overall mood33 after 10 than 2.5 pps. They also reported feeling more relaxed and in a better overall mood after 10 than 80 pps,34 and more lively after 2.5 pps33 or sham TEAS34 than after 80 pps. 7.2 Gender Men reported feeling more gloomy after 80 than 10 pps33 or sham TEAS,34 and more lively after 2.5 than 10 pps33 or sham TEAS.34 However, although they also felt more Vigour after 2.5 than 10 pps,33 they felt more vigorous after sham than 2.5 pps.34 Statistics need to be used carefully!

54

Some conclusions - Some conclusions - generalities eneralities

• It takes a lot of work to organise and run a research study
• A literature review is essential before you start [5]
• ‘SbC’ stimulation seems to be the norm for EA, TEAS and TENS

(preferably with muscle twitching at LF) [5-6]

• www.electroacupunctureknowledge.com

www.electroacupunctureknowledge.com may be useful to check

• ut past clinical research [7]
• Consult with a statistician when designing a study and deciding

how to analyse results [19]

[Numbers in square brackets refer to previous slides]

55

Some conclusions - Some conclusions - Stimulation timulation

• For most people, stimulation amplitude is higher at 2.5 pps than

at 10 pps, and this in turn higher than at 80 pps, for both SbC and ST) levels [20-22]

• Stimulation amplitude increases between ST and SbC levels [25-26]
• The ST-to-SbC difference decreases at higher frequencies [26]
• There is an increase in both ST and SbC after 10 minutes of

stimulation, but few people allow this every time [27-28]

• When treatment is not consistently repeated, ST and SbC levels

may not increase over a series of treatment sessions [29]

• Many factors may affect findings in any study [30]

56

Some conclusions – Some conclusions – Facto actors affecting findings rs affecting findings

• Age

Age increases ST in older people, but is unlikely to affect SbC [32]

• Gender

Gender: ST may be lower in women than in men [33]

• Those with prior experience

prior experience of EA or TENS

• f EA or TENS are likely to accept

higher levels of stimulation [34]

• Rece

Recent stress stress may reduce tolerance of TEAS at 10 and 80 pps, and even enhance tolerance at 2.5 pps, but this requires verification [35]

• Otherwise, current and

current and recent moods recent moods did not appear to be consistently related to ST or SbC in this study [36]

• Fati

Fatigue gue may reduce ST and SbC (the latter especially at 2.5 pps) [37]

• If room temperature

room temperature is low, ST and SbC are likely to be high [38]

57

Some conclusions – Some conclusions – Subjective responses ubjective responses

• Subjective intensity and objective amplitude may not correlate

strongly [39]

• Age and gender may not have clear group effects on how intense
• r pleasant people find TEAS [39]
• ‘Most intense’ stimulation was usually stated as 2.5 pps (followed

by 80 pps), and ‘least pleasant’ at 80 pps (followed by 2.5 pps) [40]

• Almost half our participants reported most cha

most change ge from sessions in which they found stimulation most intense most intense, and least change least change from sessions in which they found the stimulation least intense least intense [40]

• More participants found most intense

most intense stimulation stimulation least pleasant least pleasant (or least intense stimulati least intense stimulation most pleasant

• n most pleasant) than vice versa [41]
• By the time of debriefing, more information was forgotten about

sham than verum stimulation [42]

58

Some conclusions – Some conclusions – effects of ffects of stimulation on outcome (mood) stimulation on outcome (mood)

• Mood changes were rarely and weakly associated with stimulation

frequency; [43] Stimulation frequency had no effect on BRUMS [46]

• There may be slight beneficial effects for 10 pps stimulation,

but this is a very tentative finding [43]

• 10 participants (15.2%) showed similar mood changes irrespective

irrespective

• f stimulation frequency
• f stimulation frequency [44]
• Acupuncture-related treatment responses are likely to be highly

highly individual individual – in particular, mood scores may increase, decrease, or remain unchanged [45]

• More participants experienced a decrea

decrease se in feeling anxious, lively, confused, gloomy and overall mood than an increase, and more experienced an increase increase in fatigue than a decrease, virtually irrespective of stimulation frequency [45-46]

59

Some conclusions – Some conclusions – stimulation amplitude timulation amplitude

• Stimulation amplitude (defined as high or low relative to the

median) had little effect on NRS-M or BRUMS mood changes [47]

• If ‘high’ and ‘low’ SbC types were defined by upper and lower

quartiles (over all active frequencies), there were marginally more ‘low’ than ‘high’ SbC types in this study, and they possibly responded differently to the different stimulation frequencies [48]

• This finding merits further investigation.
• At 2.5 pps, a greater median ST-to-SbC difference (SSD) may

result in feeling less relaxed and lively after treatment [49]

• The presence and amount of muscle twitch correlated with SbC

level, especially at 2.5 pps [51]

• Neither the presence or amount of twitch had any significant

effect on changes in mood [51]

60

Some conclusions – Some conclusions – more on factors and

• re on factors and generalities

generalities

• Both age and gender may have some effect on mood changes in

response to stimulation at different frequencies, [52] but this requires further study

• Our outcome measures were not sufficiently sensitive to

differentiate between changes in mood in response to treatment at different frequencies (or sham)

• It is likely that many of the outcome changes found were simply

the result sitting still in a quiet room for 40 minutes

• We hope that physiological measurements will reveal more about

differences in the effects of 2.5, 10 and 80 pps, and sham TEAS.

61

Breaking news ... Breaking news ... !

• We now have data on changes in finger temperature and blood

flow in response to TEAS.

• Preliminary results indicate significant correlations between ST1

and these changes – the higher the initial threshold, and the greater the ST1-to-ST2 difference, the more temperature and blood flow increased.

• However, this was only significant for 80 pps stimulation.
• We will soon have data on changes in Heart Rate Variability (HRV)

and cortical activity (EEG) as well.

• So watch this space ... !

62

Acknowledgements

This research was conducted in collaboration with Tony Steffert, Aiste Noreikaite and Lidia Zaleczna. Our academic supervisor was Tim Watson, professor of physiotherapy. It would not have been possible without the support of many members of staff at the University of Hertfordshire, the study participants, and my wife. And finally a big big than thank you you to the AACP for both helping to fund this research, and for inviting me to give this presentation.

63

Statistics

Slide 8 (1) χ2 =122.259, p<0.001 Slide 20 (2) Assessed using Shapiro-Wilk test and calculating skewness and kurtosis Slide 22 (3) Friedman test: χ2 between 15.86, for R ST 2, and 74.92, for L SbC, p<0.001 Slide 22 (4) Wilcoxon signed ranks test Slide 23 (5) Binomial test, with test proportion 0.50 Slide 26 (6) Wilcoxon signed ranks test Slide 26 (7) Binomial test, with test proportion 0.50 Slide 28 (8) Wilcoxon signed ranks test Slide 28 (9) Binomial test, with test proportion 0.50 Slide 30 (10) Friedman test; Kruskal-Wallis and Mann-Whitney U tests for data split by pps Slide 32 (11) Mann-Whitney U test Slide 33 (12) Mann-Whitney U test Slide 34 (13) Mann-Whitney U test Slide 35 (14) Mann-Whitney U test (p=0.036) Slide 35 (15) Binomial test, with test proportions derived from data Slide 36 (16) Spearman’s ρ=-2.22 (p=0.041), a weak effect Slide 37 (17) Correlations, using Spearman’s ρ Slide 38 (18) Mann-Whitney U test Slide 39 (19) Correlations, using Spearman’s ρ Slide 40 (20) Correlations, using Spearman’s ρ Slide 40 (21) Wilcoxon signed ranks test and Sign test Slide 40 (22) Wilcoxon signed ranks test

64

Statistics (continued)

Slide 44 (23) Friedman test Slide 44 (24) Wilcoxon signed ranks test Slide 44 (25) Sign test Slide 46 (26) Binomial test, with test proportion 0.50 Slide 47 (27) Binomial test, with test proportion 0.50 Slide 48 (28) p=0.041 using the Binomial test, with test proportion 0.50 Slide 49 (29) Wilcoxon signed ranks test and Sign test Slide 50 (30) Correlations, using Spearman’s ρ Slide 52 (31) Correlations, using Spearman’s ρ (0.4<rho<0.6) Slide 52 (32) Correlations, using Spearman’s ρ (0.2<rho<0.4) Slide 53 (33) Wilcoxon signed ranks test and Sign test Slide 53 (34) Wilcoxon signed ranks test