Acupuncture has been described as a philosophy-based, not science-based, medicine [
The Western medical community now accepts that acupuncture works, at least for the relief of pain, but it still struggles to translate the mechanism of acupuncture into the Western paradigm. In recent decades, the increasing popularity of Chinese medicine in general and acupuncture in particular among the public around the world has motivated the scientific community to undertake intensive investigation of acupuncture’s efficacy, as well as the physiological basis behind it. After decades if not centuries of treating individual problems, the Western medical community is coming to understand that human health is a holistic phenomenon reflecting balance—or imbalance—in the bio-psycho-social-spiritual aspects of life. While this balance reflects complex factors, there are parameters we can measure—there are physical indications—of how an individual responds to a disease state in order to maintain the bio-psycho-social-spiritual balance within our body. Heart rate variability (HRV) is a measure of this kind.
HRV can be measured by time domain analysis or frequency domain analysis as a power spectrum produced by a selected chain of heart to heart beat. Using power spectral analysis [
Diagram of the power spectral density heart rate variability (HRV) (adapted from [
According to [
Summary of the normal values of HRV.
Component | Outcome parameter | Clinical indication |
---|---|---|
Very low frequency (VLF) | Frequency (0.0033–0.04 Hz) | Indicator of overall activity of various slow sympathetic activities |
Low frequency (LF) | Frequency (0.04–0.15 Hz) | Indicator of sympathetic activities with parasympathetic influence when the respiration rate is lower than 7 per minute |
High frequency (HF) | Frequency (0.15–0.4 Hz) | Indicator of parasympathetic activities |
LF/HF ratio | 1.5–2.0 | Indicator of sympathetic activities; a higher number means increased sympathetic activities or reduced parasympathetic activities |
sNormalized LF and HF (LF norm, HF norm) | Frequency (Hz) | The monotonic nonlinear transformation of LF/HF ratio. The LF/HF ratio, LF norm, and HF norm indicate the same aspects of sympathovagal balance of autonomic nervous system |
HRV measures the balance of our autonomic nervous system which reflects physiological, hormonal, and emotional balance. The autonomic nervous system is composed of two branches, namely, sympathetic nervous system and parasympathetic nervous system. Sympathetic nervous system is responsible for flight and stress situation while parasympathetic nervous system is dominant when relaxed [
The autonomic nervous system dynamically controls the response of the body to a range of external and internal stimuli, providing physiological stability [
HRV is an established tool in cardiology research and is being increasingly used for a range of clinical applications [
Both timedomain heart rate variability and respiration rate in various positions—for example, lying down in a laboratory, cycling in a laboratory, and sleeping in an ambulatory surrounding—showed high intraclass correlation coefficients [
Acupuncture is generally documented to be effective in treating chronic and acute pain conditions by inserting needles into specific “acupuncture points” (acupoints) on the patient’s body [
Previous studies have showed a relationship between acupuncture and heart rate variability. Wang et al. reported that acupuncture enhances cardiac vagal activity and suppresses sympathetic activities in healthy humans [
Several studies reported the effect of acupuncture on HRV using specified acupoints. For instance, acupuncture on Neiguan (PC6) and Gongsun (SP4) decreased sympathetic activity and balanced the autonomic nervous system for those who were under stress [
To understand the mechanism of acupuncture, we need to understand how clinical scientific studies were done and how acupuncture established its effects for some specific health conditions. A well-designed trial, therefore, isolates the specific treatment variable to see what effect it has. To this end, a systematic review investigating (a) the effect of acupuncture on HRV and (b) the potential for using HRV as an indicator of the therapeutic effect of acupuncture was conducted in this study.
It was a systematic review with meta-analysis.
In this research, randomized control trials (RCTs) were included. Table
Inclusion criteria of the RCTs.
Component | Criteria |
---|---|
Participant | Adults (aged ≥18) were treated with needle acupuncture with or without electric stimulation |
Intervention | Trials employed acupuncture as the sole treatment |
Comparator | Trials compared needle acupuncture with any type of sham acupuncture or no treatment (control) |
Outcome measures | Trials used spectral analysis of HRV as the outcome measure |
There was no language restriction and only papers with fulltext were included for analysis. Studies were excluded if trials testing forms of acupuncture other than needle acupuncture, for example, laser acupuncture or moxibustion; and, if there was only one acupuncture session reported in the studies.
The following keywords were used in the database search: acupuncture OR meridian OR acupoint AND heart rate variability or HRV or heart rate. For Chinese databases, keywords in Chinese used were
Databases searched were CINAHL (1937-), EMbase (1980-), PsycInfo (1597-), Cochrane Central Register of Controlled Trials (1898-), Ovid Medline (1950-), Korean Studies Info Services System, and China Academic Journals Fulltext Database (CAJ) (1915-).
The initial search yielded 94, 848, 30, 155, 89, 3, and 1 paper from CINAHL, EMbase, PsycInfo, Cochrane Central Register of Controlled Trials, Ovid Medline, Korean Studies Info Services System, and CAJ, respectively. Then, the researcher screened the title and the abstract to determine the relevancy. As a result, 1,220 articles were found and comprised the initial data extraction.
The researcher designed a data extraction form, and the two independent reviewers completed the data extraction according to the topics predetermined in the form (Table
Summary of extracted RCTs.
Authors (year) |
Design sample size Conditions | Intervention (regimen) | Heart rate variability | Main results of intergroup differences | Risk of bias |
---|---|---|---|---|---|
Kwak et al. [ |
Parallel; 42 healthy students with examination stress | (A) AT ( |
(1) HRT (beats/min); |
(1) A→, B→, C |
U, U, U, U, U |
|
|||||
Huang et al. [ |
Parallel; 111 healthy subjects | (A) AT (P6, 20 min on both forearms, |
(1) Mean RR interval (ms); |
(1) A↑, B↑, C→: |
U, N, N, N, N |
|
|||||
Chang et al. [ |
Crossover, 15 healthy subjects | (A) EA (2 Hz, 30 min, |
(1) LF/HF; |
(1) A↑, B→: NS; |
U, N, Y, U, U |
|
|||||
Chang et al. [ |
Crossover, 15 healthy subjects | (A) EA (2 Hz at Zusanli, 30 min, |
(1) LF/HF; |
(1) A↑ during and after EA, B→, C→, D→, E→: NS; |
U, N, Y, U, U |
|
|||||
Hsu et al. [ |
Crossover, 10 healthy subjects | (A) EA (2 Hz at BL15, 10 min, |
(1) LF norm; |
(1)–(4) A |
U, N, U, U, U |
|
|||||
Hsu et al. [ |
Crossover, 10 healthy subjects | (A) AT (scalp-Sishencong, 10 min, |
(1) LF norm; |
(1)–(4) A |
U, N, U, U, U |
|
|||||
Kim et al. [ |
Crossover, 38 female subjects with regular menstrual cycle | (A) AT (LI4 and SP6, 15 min, |
(1) LF; |
(1) A↑, B↑: NS; |
U, N, Y, U, U |
|
|||||
Streitberger et al. [ |
Crossover, 20 healthy subjects | (A) AT (LI4, 5 min and 15 sec manual stimulation, |
(1) LF; |
(1) and (2) A→, B→: NS; |
Y, N, Y, Y, U |
|
|||||
Li et al. [ |
Parallel, 29 male healthy subjects (after 3 hr driving workout) | (A) AT (15 min, |
(1) LF/HF; |
(1)-(2) A |
U, N, Y, U, U |
|
|||||
Chang et al. [ |
Parallel, 12 male healthy subjects with no neurological diseases | (A) AT (PC6, 3 sessions with one-week washout, 30 min, |
(1) VLF; |
NS for all cases | U, N, U, U, U |
|
|||||
Kurono et al. [ |
Crossover, 14 male healthy subjects | (A) AT (epifascial stimulation at CV17 and CV16, needle inserted for 1s only, |
(1) LF; |
(1)-(2) A↑ in CV17, but not CV16; |
U, N, Y, U, U |
|
|||||
Liu et al. [ |
1st phase (acute); Crossover, 27 patients with functional dyspepsia (Rome II) | (A) TEA (25 Hz, PC6 and ST36, 30 min, 2 sessions, |
(1) HF; |
(1) A↑ during 1st 30 min Tx, B→, C→: |
Y, N, U, U, U |
|
|||||
Liu et al. [ |
2nd phase (chronic); Crossover, 27 patients with functional dyspepsia (Rome II) | (A) TEA for 2 weeks (twice daily, 30 min), 1 week-wash-out, sham TEA for 2 weeks; |
(1) HF; |
(1) A↑ after 2 weeks, B→, C→: |
Y, N, Y, Y, U |
|
|||||
Shi et al. [ |
Crossover, 20 patients with coronary heart disease | (A) AT (30 min, |
(1) LF; |
(1) A |
U, N, U, U, U |
AT: acupuncture treatment; EA: electroacupuncture; TEA: transcutaneous electroacupuncture; Co: control; Tx: treatment.
Risk of bias (1) sequence generation; (2) incomplete data; (3) patient-blinded; (4) assessor blinded; (5) allocation concealment performed.
Cochrane criteria were used to assess the risk of bias of the studies [
Hard copies of all the articles with full text were printed and studied by two independent reviewers. The two independent reviewers agreed on the assessment of risk of bias. The results of the two reviewers were the same, and the results are shown in Table
Chi-squared test and Higgins
Using the keyword search, the researcher identified 1220 articles initially. Upon reading the abstracts, 1133 articles were excluded because they were not related to acupuncture or HRV. The remaining 87 articles in full text were then reviewed by the two reviewers. Seventy-three articles were excluded because of non-RCTs and incomplete outcome measures. Fourteen studies were included in the analysis (Figure
Process of studies selection.
For risk of bias evaluation, the researcher used the following criteria: sequence generation, incomplete data, patient-blinded, assessor-blinded, and allocation concealment (Table
Results of risk of bias analysis.
Fourteen studies meeting the selection criteria were included in the study. It is noted that some of the included studies had more than one group, that is, acupuncture, sham, and control. In view of the fact that all the tests in the groups had the same outcome measures, the researcher entered the outcome measures of the respective group as one study ID in the Revman 5.0. Six outcome measures were input for analysis, namely, HF, LF and LF/HF ratio, HF norm, and LF norm. Thus, there may be more than one study ID with the same author and year entry in the subsequent meta-analysis for one included study.
Seven studies were included in this part of the analysis, corresponding to 16 study IDs, as explained in the previous section (Figure
Forest plot of the effects of acupuncture on HF for healthy subjects.
Six studies were evaluated in this section of the study, while there were 13 study IDs in this analysis (Figure
Forest plot of the effects of acupuncture on LF for healthy subjects.
Nine studies were evaluated in this section of the study; for these 9 studies, there were 13 study IDs (Figure
Forest plot of the effects of acupuncture on LF/HF ratio for healthy subjects.
Five studies were evaluated in this section of the study, corresponding to 8 study IDs (Figure
Forest plot of the effects of acupuncture on HF norm for healthy subjects.
Five studies were included while there were 8 study IDs in this analysis (Figure
Forest plot of the effects of acupuncture on LF norm for healthy subjects.
Two studies were included while there were 3 study IDs in this analysis (Figure
Forest plot of the effects of acupuncture at ST36 on HF for healthy subjects.
Two studies were included while there were 4 study IDs in this analysis (Figure
Forest plot of the effects of acupuncture on HF for nonhealthy subjects.
One study was included while there were 2 study IDs in this analysis (Figure
Forest plot of the effects of acupuncture on LF for nonhealthy subjects.
Two studies were included while there were 4 study IDs in this analysis (Figure
Forest plot of the effects of acupuncture on LF/HF ratio for nonhealthy subjects.
Two studies were included while there were 4 study IDs in this analysis (Figure
Forest plot of the effects of acupuncture at PC6 on HF for nonhealthy subjects.
Two studies were included while there were 4 study IDs in this analysis (Figure
Forest plot of the effects of acupuncture at PC6 on LF/HF for nonhealthy subjects.
Two studies were included while there were 2 study IDs in this analysis (Figure
Forest plot of the effects of acupuncture at LI4 on HF for all healthy and nonhealthy subjects.
Four studies were included while there were 6 study IDs in this analysis (Figure
Forest plot of the effects of acupuncture at LI4 on LF/HF ratio for all healthy and nonhealthy subjects.
Two studies were included while there were 4 study IDs in this analysis (Figure
Forest plot of the effects of acupuncture at LI4 on HF norm for all healthy and nonhealthy subjects.
Two studies were included while there were 4 study IDs in this analysis (Figure
Forest plot of the effects of acupuncture at LI4 on LF norm for all healthy and nonhealthy subjects.
Three studies were included while there were 6 study IDs in this analysis (Figure
Forest plot of the effects of acupuncture at PC6 on HF for all healthy and nonhealthy subjects.
Two studies were included while there were 4 study IDs in this analysis (Figure
Forest plot of the effects of acupuncture at PC6 on LF for all healthy and nonhealthy subjects.
Six studies were included while there were 11 study IDs in this analysis (Figure
Forest plot of the effects of acupuncture at PC6 on LF/HF ratio for all healthy and nonhealthy subjects.
Three studies were included while there were 5 study IDs in this analysis (Figure
Forest plot of the effects of acupuncture at PC6 on HF norm for all healthy and nonhealthy subjects.
Three studies were included while there were 5 study IDs in this analysis (Figure
Forest plot of the effects of acupuncture at PC6 on LF norm for all healthy and nonhealthy subjects.
Three studies were included while there were 5 study IDs in this analysis (Figure
Forest plot of the effects of acupuncture at ST36 on HF for all healthy and nonhealthy subjects.
Two studies were included while there were 3 study IDs in this analysis (Figure
Forest plot of the effects of acupuncture at ST36 on LF for all healthy and nonhealthy subjects.
Three studies were included while there were 6 study IDs in this analysis (Figure
Forest plot of the effects of acupuncture at ST36 on LF/HF ratio for all healthy and non-healthy subjects.
Table
A summary of the heterogeneity for random effect.
Outcomes | No. of studies | Heterogeneity Chi-sq test | Higgins | ||||||
---|---|---|---|---|---|---|---|---|---|
All studies | Healthy | Non-healthy | All Studies | Healthy | Non-healthy | All studies | Healthy | Non-healthy | |
HF | 9 | 7 | 2 | 41.31 ( |
16.99 ( |
24.31 ( |
64% | 35%# | 88% |
LF | 7 | 6 | 1 | 51.93 ( |
47.44 ( |
0.15 ( |
77% | 79% | 0% |
LF/HF ratio | 11 | 9 | 2 | 977.50 ( |
913.49 ( |
0.75 ( |
98% | 99% | 0% |
HF norm | 5 | 5 | — | 32.35 ( |
32.35 ( |
— | 78% | 78% | NA |
LF norm | 5 | 5 | — | 15.39 ( |
15.39 ( |
— | 55% | 55% | NA |
LI4 on HF | 2 | — | — | 0.00 ( |
— | — | 0% | — | — |
LI4 on LF/HF ratio | 4 | — | — | 73.89 ( |
— | — | 93% | — | — |
LI4 on HF norm | 2 | — | — | 26.62 ( |
— | — | 89% | — | — |
LI4 on LF norm | 2 | — | — | 13.43 ( |
— | — | 78% | — | — |
PC6 on HF | 3 | — | 2 | 24.71 ( |
— | 24.31 ( |
80% | — | 0% |
PC6 on LF | 2 | — | — | 6.33 ( |
— | — | 53% | — | — |
PC6 on LF/HF ratio | 6 | — | 2 | 85.78 ( |
— | 0.75 ( |
88% | — | 0% |
PC6 on HF norm | 3 | — | — | 28.68 ( |
— | — | 86% | — | — |
PC6 on LF norm | 3 | — | — | 15.22 ( |
— | — | 74% | — | — |
ST36 on HF | 3 | 2 | — | 33.34 ( |
0.49 ( |
— | 88% | 0% | — |
ST36 on LF | 2 | — | — | 7.39 ( |
— | — | 88% | — | — |
ST36 on LF/HF ratio | 3 | — | — | 108.59 ( |
— | — | 95% | — | — |
A summary of the effect of acupuncture in HRV outcome measures in various subject groups and by different acupoints.
Outcomes | Overall effect, |
Mean difference | ||||
---|---|---|---|---|---|---|
All studies | Healthy | Non-healthy | All Studies | Healthy | Non-healthy | |
HF | 4.69 ( |
1.39 ( |
5.25 ( |
0.08 [0.05, 0.10]* | 0.07 [−0.03, 0.17]# | 0.08 [0.05, 0.10]* |
LF | 0.03 ( |
0.72 ( |
2.04 ( |
0.01 [−0.38, 0.40] | 0.15 [−0.26, 0.56] | −0.98 [−1.93, −0.04]* |
LF/HF ratio | 2.18 ( |
1.26 ( |
15.82 ( |
−0.33 [−0.63, −0.03]* | −0.26 [−0.66, 0.14] | −0.50 [−0.56, −0.44]* |
HF norm | 5.00 ( |
5.00 ( |
— | 0.51 [0.31, 0.70]* | 0.51 [0.31, 0.70]* | NA |
LF norm | 4.03 ( |
4.03 ( |
— | −0.30 [−0.44, −0.15]* | −0.30 [−0.44, −0.15]* | NA |
LI4 on HF | 15.82 ( |
— | — | 0.21 [−0.05, 0.47] | — | — |
LI4 on LF/HF ratio | 0.04 ( |
— | — | −0.01 [−0.29, 0.27] | — | — |
LI4 on HF norm | 0.24 ( |
— | — | 1.20 [−8.64, 11.04] | — | — |
LI4 on LF norm | 0.36 ( |
— | — | −1.45 [9.32, 6.41] | — | — |
PC6 on HF | 5.47 ( |
— | 5.25 ( |
0.07 [0.05, 0.10]* | — | 0.08 [0.05, 0.10]* |
PC6 on LF | 0.57 (0.57) | — | — | −0.05 [−0.24, 0.13] | — | — |
PC6 on LF/HF ratio | 2.23 ( |
— | 15.82 ( |
−0.27 [−0.51, −0.03]* | — | −0.50 [−0.56, −0.44]* |
PC6 on HF norm | 0.20 ( |
— | — | 0.80 [−6.99, 8.58] | — | — |
PC6 on LF norm | 0.15 ( |
— | — | −0.45 [−6.42, 5.52] | — | — |
ST36 on HF | 4.47 ( |
3.00 ( |
— | 0.07 [0.04, 0.11]* | −99.51 [−164.56, −34.45]* | — |
ST36 on LF | 0.96 ( |
— | — | −136.15 [−414.89, 142.59] | — | — |
ST36 on LF/HF ratio | 0.61 ( |
— | — | −0.09 [−0.36, 0.19] | — | — |
To date, there has been one published systematic review which studied the effect of acupuncture on HRV [
Since this systematic review adopted the same inclusion and exclusion criteria as Lee and colleague’s study [
The findings in this systematic review reveal two phenomena. First, acupuncture has reduction effect on HRV. Acupuncture decreases the LF and LF/HF ratio among nonhealthy subjects. However, the effect of acupuncture on healthy subjects is inconclusive because there were no data of nonhealthy subjects for combined analysis.
Second, acupuncture does not have any effect on the HF component of HRV. Acupuncture also has reduction effect on the HRV of sham/control. Sham/control decreases the HF for non-health subjects and HF norm for healthy subjects. Again, the effect of sham/control on healthy subjects is inconclusive because there were no data of nonhealthy subjects for combined analysis.
We find that changes in LF can change the values of LF/HF ratio or LF norm and HF norm. This is because LF or HF is a function of the LF/HF ratio or LF norm and HF norm. Therefore, it is worthwhile to look at the decreasing modulating effect of acupuncture on LF for nonhealthy subjects. LF refers to the parasympathetic influence on the balance between sympathetic and parasympathetic activities when one is in a relaxed state, like slowing heart rate, decreasing blood pressure, stimulating the gastrointestinal tract, eliminating waste, restoring energy, and building tissues. This modulating effect by acupuncture is important because, in general, parasympathetic activities represent conservative and restorative functions.
Effects of ST36 and PC6 acupoints on HRV were demonstrated that ST36 is an important acupoint of the Stomach Meridian which is documented to be related to general wellness and good
Evidence from this systematic review partially supports the possible effect of acupuncture in modulating the low frequency component of heart rate variability. This may represent a mechanistic pathway for global physiological regulation, which is congruent with East Asian medical theory. It is important to highlight the significance of LF component of HRV. This modulation effect by acupuncture is important because, in general, parasympathetic activities represent conservative and restorative physiological activities. However, we do not have enough published work in this area to determine if HRV can be used as an indicator for the therapeutic effect of acupuncture because the dynamic change between LF and HF in maintaining the optimal status is yet to be answered in this systematic review.
We need more RCTs with high quality in this area so as to provide a direction for our evidence-based practice. We need larger sample sizes in the studies because this allows better randomization to work. The control for heterogeneity is an issue which can be overcome. We need to have studies in which both subjects and assessors can be blinded in order to minimize the psychological biases. Current studies were subject to psychological biases which would in turn affect the quality of the systematic review.
In this thesis, a systematic review with meta-analysis was conducted on 14 RCTs. All the studies on related HRV outcome measures showed moderate to high heterogeneity. Results showed a decreasing modulating effect of acupuncture on LF and LF/HF ratio for nonhealthy subjects and on LF norm for healthy subjects. However, the overall effect was in favour of the sham/control, not acupuncture group, for HF in nonhealthy subjects and for HF norm in healthy subjects. Evidence from this systematic review suggests that acupuncture modulates the low frequency component of HRV but not the HF component. However, we do not have enough published work in this area to determine if HRV can be an indicator for the therapeutic effect of acupuncture.
The authors have no conflict of interests whatsoever to declare.