Do different doses of acupuncture matter on autonomic nervous activity and symptom management in dysmenorrhea? A randomized controlled trial.

Background The study examined whether the acupuncture dose (number of acupoints stimulated) impacted the ecacy of acupuncture on dysmenorrhea and the relationship with autonomic nervous system regulation. Methods This three-arm randomized controlled study included a high-dose acupuncture (12 acupoints, N = 23), low-dose acupuncture (6 acupoints, N = 30) and control (N = 30) arm. The treatment course was three months. We set heart rate variation (HRV) and analgesics dependence as the primary outcome measurements; Visual Analog Scale (VAS) score, Verbal Multidimensional Scoring System (VMSS) and the 12-Item Short Form Health Survey (SF-12) quality of life questionnaires were set as secondary outcomes. SPSS version 24 was used for data analysis. Results Low-dose acupuncture was superior to high-dose in analgesics dependence (p value: low/high/control: p = 0.043/p = 0.056/p = 0.376); symptom relief (VMSS: low/high/control: p < 0.001/p = 0.007/p = 0.109); and physical quality of life (low/high/control: p < 0.001/p = 0.01/p = 0.007). The groups did not differ in HRV parameters (p > 0.05). In intergroup analysis, more signicant changes were noted in the high-dose than in the low-dose group. The scattered nature of the signicant changes implies that acupuncture may have a short-term effect on HRV parameters which does not correlate with the acupuncture dose. reduce analgesic 12-Item Short Form Health Survey; HRV:Heart rate variation; HF:High frequency; HPO:Hypothalamus-pituitary-ovary; LF:Low frequency; NSAIDs:Non-Steroidal Anti-Inammatory Drugs; OCT:Oral contraceptive pills; R-MSSD:root-mean square of differences of successive RR intervals; SDNN:standard deviation of adjacent peak-to-peak (NN) intervals; TP:Total power; VMSS:Verbal Multidimensional Scoring System; Analog Scale; VLF:Very low

More than two-thirds of women with dysmenorrhea take analgesics (5) and nearly 80% have adopted some complementary therapy, such as herbal medicine, acupuncture, moxibustion or aromatherapy. As NSAIDs may cause gastrointestinal discomfort or dyspepsia and hormone therapy can lead to other concerns, some women now adopt acupuncture as a safer therapeutic alternative (6,7). Women who adopt acupuncture as an alternative therapy usually wish to withdraw the medication use or improve their pain and other symptoms in order improve their quality of life. (8) As acupuncture has been widely used in dysmenorrhea, we observe that people expect greater e cacy with the use of more needles of stimulation; however, the relationship of acupoint selection and optimal treatment is not established (9).
No study has yet compared the e cacy of different schedules of acupoints in treating dysmenorrhea, and the acupoints used differ in different trials. We therefore wanted to evaluate whether the number (or dose) of acupoints would affect the outcomes of acupuncture in dysmenorrhea.
Furthermore, in terms of the mechanism of acupuncture, the symptoms related to dysmenorrhea such as anxiety, insomnia and diarrhea are non-speci c and related to autonomic dysfunction, namely, vagosympathetic tone imbalance (10). A previous study also noted that women with dysmenorrhea had decrease heart rate variation (HRV) and autonomic dysfunction compared with the control group (11).
Research con rmed that dysmenorrhea a icted pain and lowered high frequency (HF), indicting lower parasympathetic tone and increasing sympathetic tone (12), compatible with disharmony in the yin/yang. As acupuncture has proven e cacy in adjusting autonomic dysfunction (13,14) and treating dysmenorrhea (7,15), we sought to understand whether the effect of acupuncture in dysmenorrhea was related to adjustments in the autonomic nervous system. Lastly, we wanted to know if acupuncture could help women with dysmenorrhea take fewer NSAIDs or improve their quality of life.
According to the theory of Traditional Chinese Medicine, acupuncture regulates the energy of the meridians and coordinates the balance of yin/yang (16,17). Western medicine interprets the mechanism of pain control of acupuncture as an effect of neurophysiology (18) and the release of neuropeptides such as endorphin (19). Additionally, in our previous study, we noted that the theory of yin/yang (meridian) is compatible with the autonomic nerve theory (20). Therefore, in this clinical trial, we aim to evaluate: 1) the effect of different acupuncture doses (by number of acupoints) in dysmenorrhea; 2) the relationship of the mechanism of acupuncture with the adjustment of autonomic nerve activity in dysmenorrhea; and 3) the effect of acupuncture on analgesic dependence in women with dysmenorrhea.

Subjects and Allocation
We selected the participants according to the inclusion criteria: 1) a diagnosis of primary dysmenorrhea according to the Primary Dysmenorrhea Consensus Guideline(21); 2) age from 16 to 35 years without history of delivery; 3) normal menstrual cycle (28+7 days and duration 7+3 days); and 4) menstrual pain scoring more than 40 on a 100-point Visual Analog Scale (VAS) during the baseline menstrual period according to the patient-kept dairy. The main exclusion criteria were: 1) secondary dysmenorrhea related to endometriosis, endometrial polyps, pelvic in ammatory disease or uterine myoma; and 2) other gynecological problems. Patients enrolled in the acupuncture group were allowed to take NSAIDs at their usual usage and record the dosage in a painkiller diary which was collected.
In total, 83 patients were included in the acupuncture (high-dose: 12 acupoints: N=23; low-dose: 6 acupoints, N=30) and control (N=30) groups from 01, June, 2018 to 31, Dec. 2019. Patients were allowed to receive acupuncture or medication as their intention, yet those in the intervention groups were randomized to different doses (number of acupoints) of acupuncture by using random allocation software 2.0. Sequentially numbered containers were performed, all of the containers were tamper-proof, equal in weight, and similar in appearance. The study was performed in Traditional Medicine and Gynecological Departments, Taipei City Hospital from March 2019 to Dec. 2019. The trial was approved by the hospital Institutional/Independent Review Board (approval no. TCHIRB 10701111). The trial is registered at ClinicalTrials.gov, with number of NCT03881319. Figure 1 provides a ow chart of subject recruitment.

Intervention
In the acupuncture group, patients received acupuncture twice in each luteal phase, for three consecutive menstrual cycles, represented as Acupuncture 1-1; 1-2; 2-1, etc. The acupoints selected by following the Traditional Chinese Medicine and meridian theory in the low-dose group were: PC6 ( ), SP4 ( ), SP6 ( ), SP10 ( ), RN4 ( ) and RN6 ( ). Those selected for the high-dose group were the acupoints for the lowdose group, plus: RN3 ( ), LI4 ( ), ST36 ( ), KI3 ( ), LR3( ) and EX-CA1 ( ) ( Figure 2). De-qi sensation (22) was manipulated with 20 mins of stimulation in each treatment session, using sterile disposable acupuncture needles. In the control (medication) group, participants took only NSAIDs with or without OCT. Patients in the acupuncture groups were allowed to take NSAIDs if they still felt pain and were asked to record the dosage used. The study included at least three consecutive menstrual cycles of treatment.

Outcome measures
To investigate the effect of acupuncture on autonomic function, we used HRV and analgesics dosage as the primary outcomes to assess the relationship and effect of different doses (number of acupoints stimulated) on autonomic nervous status and analgesic dependence. The HRV parameters included the time domain and frequency domain, before and after each course of intervention in both groups. To compare treatment e cacy by group, we used the VAS score to measure pain; the verbal multidimensional scoring system (VMSS), which asks participants to grade their ability to work, their systemic symptoms and their analgesics taking; and the 12-Item Short Form Health Survey (SF-12) quality of life questionnaire as secondary outcomes measurements.

Statistical analysis
In this trial, the estimated sample size was 34 in each group which determined by G power software, with z score:1.96, margin of error(e=0.1) , and the statistical analysis was performed using SPSS version 24.0 (IBM Corp., Armonk, NY). The data analysis was based on the intention-to-treat population. Means and standard deviations were determined for all data. We applied one-way ANOVA to compare the effect of acupuncture in the acupuncture and control groups, while the chi-squared test was used to analyze the categorical variables. The level of signi cance was established as P <=0.05; 95% con dence interval).

Results
In total, 98 patients were enrolled in the trial, 83 completed the trial: 23 in the high-dose group, 30 in the low-dose group and 30 in the control group. During the trial, no participants had any side effect from the intervention or treatment. Table 1 reveals no signi cant difference in baseline characteristics or HRV parameters between groups. MCS: mental health composite score.
*p < 0.05, ** p < 0.01, *** p < 0.001 The effect of acupuncture dosage on heart rate variation (HRV) parameters Table 2 compares the change in HRV parameters before and after treatment in the acupuncture and control groups. No signi cant difference was noted before and after the patients in control group who took 3 months of medication for their dysmenorrhea (p > 0.05). A similar phenomenon was noted in the acupuncture groups, with no signi cant change noted in patients' HRV parameters in the high-dose group, yet some meaningful improvement was found in the low-dose acupuncture group in both time and frequency domains (i.e., the percent of differences of adjacent RR intervals > 50 ms, p = 0.04; the percentage of low frequency bands, p = 0.032; and the percentage of high frequency bands, p = 0.032). *p < 0.05, ** p < 0.01, *** p < 0.001 Despite the lack of obvious difference in between-group analysis, the intergroup-analysis of the acupuncture groups (Table 3) showed that patients had some improvement in autonomic nervous activity after certain acupuncture treatments. For example, in the high-dose acupuncture group, the Standard Deviation of Normal to Normal (SDNN, represents overall heart rate variability index) signi cantly changed after acupuncture session 1-1 ( rst menstrual cycle, rst acupuncture session, p = 0.015); session 2 − 1 (p = 0.016); session 2-2 (p = 0.025); and session 3 − 1 (p = 0.031). HF (represent parasympathetic tone) signi cantly increased after sessions 2-2 (p = 0.042) and 3 − 1 (p = 0.025). Total power (TP), which represent autonomic activity, signi cant improved after sessions 1-1 (p = 0.013), 2 − 1 (p = 0.019) and 2-2 (p = 0.035). However, the signi cance of these changes was less in the low-dose acupuncture group. Nevertheless, the signi cant change distribution implied that acupuncture might have a short-term effect on some HRV parameters, but there was no obvious tendency of superiority for either high-or low-dose acupuncture (Table 3).  Furthermore, high-dose acupuncture (p < 0.001), low-dose acupuncture (p < 0.001) and medication (p < 0.001) all had a signi cant effect on relief of pain, as re ected in the VAS scores. The difference between the three groups not reach signi cance (p = 0.145). In terms of quality of life as measured by the SF-12, low-dose acupuncture improved participants' physical component score (PCS) much more meaningfully then did high-dose acupuncture (low-dose: p < 0.001; high-dose: p = 0.001; control: p = 0.007). In the mental component score (MCS), high-dose acupuncture was superior to the other two groups, which implies a placebo effect (high-dose: p < 0.001; low-dose: p = 0.025; control: p = 0.01) ( Table 4). VMSS: Verbal multidimensional scoring system.
Grade 0-1: menstruation is not painful or mild pain.
Grade 2-3: Menstruation is moderate pain or severe pain.
*p < 0.05, ** p < 0.01, *** p < 0.001 The effect of different doses of acupuncture on dysmenorrhea symptoms In terms of severity of dysmenorrhea scores, patients in both acupuncture groups has signi cant improvement in VMSS scores. Low-dose acupuncture was more e cient in relieving symptoms than high-dose acupuncture (high-dose: p = 0.007; low-dose: p < 0.001; control: p = 0.109) ( Table 4).

Discussion
Dosage effect This is not the rst study to con rm the e cacy of acupuncture in reliving primary dysmenorrhea (23,24).
However, few studies have evaluated the e cacy of different doses of acupuncture or the use of more acupoints. Acupuncture has been widely applied to diverse conditions: pain, arthritis, neuropathy, insomnia, and even some cancer-related side effects (25,26). However, scholars and clinical physicians have not yet reached consensus on the optimal acupuncture prescription, including the number of acupoints, the number of acupuncture sessions, or even the location of acupoints or the optimal total dose. A study by Sun et al. (2019) pointed out that a higher dosage of acupuncture might provide better treatment outcomes in terms of relief of pain and dysfunction in patients with knee osteoarthritis (27), but others consider that more acupuncture sessions will only relieve symptoms in those with chronic pain (28). In primary dysmenorrhea, only one previous study has demonstrated that the timing of the acupuncture treatment and the mode of needle stimulation were germane to menstrual pain outcomes, but even this study failed to identify the optimum dose parameters for this condition (29). From our study, we noted that, when the parameters of acupuncture timing and frequency were xed, acupuncture with fewer acupoints (lower dose) was not inferior to higher dose, and even better in terms of dependence on analgesics (6 acupoints: 0.043; 12 acupoints: 0.056); dysmenorrhea symptom control (VMSS: 6 acupoints: p<0.001; 12 acupoints: p=0.007); and quality of life (SF-12 PCS: 6 acupoints: p<0.001; 12 acupoints: p=0.001). Therefore, the clinical physician should understand that clinical e cacy is not depend on the number of acupoints and should convey this concept to patients, to reduce the waste of medical resources (9). In quality of life assessment, we found that subjects had more improvement in mental quality of life than in physiological quality of life, so that the placebo effect should be considered a relevant psychological factor. In this regard, a study claimed that the correlation coe cients between de-qi and the therapeutic e cacy of acupuncture were greater than those between psychological factors and therapeutic e cacy in dysmenorrhea (30). Therefore, the mechanisms driving acupuncture e cacy are complex and extend beyond the simplistic "more is better" concept of treatment.

Relationship to autonomic activity adjustment
Some studies indicate that de-qi is the most important factor in making acupuncture effective (31). A study in 2017 revealed that patients with dysmenorrhea who experienced actual de-qi sensation had more pain relief and a greater analgesic effect (32). An interesting study revealed that de-qi elicited a signi cant response to acupuncture in speci c brain regions, indicating that de-qi acupuncture triggers the neuroendocrine network to alleviate pain, but the mechanisms are not clear (22). In our study, we used HRV as an outcome measure, to check whether the acupuncture effect was correlated to autonomic change. HRV has long been recognized as a noninvasive indicator of autonomic nervous system activity (20,33). Indeed, HRV coherence is related to many physiological bene ts, including short-term improvement in cardiovascular and respiratory function, and an increase in vagal afferent activity, which converses pain signals and sympathetic out ow, and thereby promotes temporal synchronization of the body (34,35). Accordingly, the improvement in autonomic dysfunction corresponds with the balance of yin/yang and implies the adjustment of neuroendocrine in the body microenvironment (20). In this study, we noted that acupuncture has only a short-term effect in adjusting the autonomic dysfunction within the acupuncture group; the bene t did not meaningfully outweigh that received by the medication (control) group (p>0.05). Nevertheless, inter-group analysis revealed that subjects in the high-dose group had greater HRV change, a result which can be interpreted two ways. First, the effect of acupuncture on the autonomic system is short-term rather than long-term. Second, the stimulation of more acupoints may lead to more change in HRV activity, but not greater e cacy in pain relief or symptom control. The mechanism by which acupuncture relieves dysmenorrhea may not directly derive from a change in autonomic nervous activity. We therefore wonder whether the effect of acupuncture in dysmenorrhea is correlated with alterations in serum cytokines, such prostaglandin E and prostaglandin F2 alpha, as has been revealed in previous literature (36). Further study to evaluate the role of serum cytokines is encouraged.
Potential mechanisms of pain reduction The mechanism of pain associated with dysmenorrhea can be addressed in two aspects: the neuroendocrine regulated and the neuro-physiologic related. As our study indicates that subjects who adopt acupuncture could use less analgesics, we suppose that acupuncture may have some neuro-endocrine effect (37,38). In this regard, a previous study noted that acupuncture could have both an antiin ammatory and an analgesic effect, which might be associated with its inhibition of spinal p38MAPK activation (39) or its regulation of serum prostaglandin E2 or F2 alpha (40). Because patients who used acupuncture required less NSAIDs, acupuncture may have a similar anti-in ammatory effect as NSAIDs, but without the side effects of gastrointestinal discomfort or ulcer stimulation.
On the other hand, some studies have noted that the shift in the estradiol/progesterone ratio in favor of estradiol seems to be the pathogenic principle behind dysmenorrhea (41,42). We therefore suppose that acupuncture might somehow affect the related hormone balance. In animal models, acupuncture had been noted to regulate neuroendocrine activity and the related receptor expression of the hypothalamuspituitary-ovary (HPO) axis (38,43). Clinically, some gynecologist also use oral contraceptives or progestins (OCT) for primary dysmenorrhea (44). If acupuncture could adjust the HPO axis and thereby relieve the related symptoms, then patients can avoid taking OCTs which might have unwanted side effects such as increased risk of endometriosis, blood clot, stroke or cancer (45).
Holistic effect of acupuncture on dysmenorrhea The results indicate that acupuncture could reduce subjects' dependence on analgesics and improve their VAS and VMSS scores, which means increasing their ability to work and allowing them to resume more normal activity with fewer symptoms. Compared to the alternatives, acupuncture has fewer side effects and is more cost-effective, especially considering that the number of acupoints is not linearly related to outcomes.

Limitations
We noted some bias and limitations regarding the study. The rst is that the size of the study could me more to be more powerful. Second, there is as yet no consensus on the optimal acupuncture schedule for any individual disease. We designed the schedule based on the Traditional Chinese Medicine meridian theory, but others might select different acupoints and devise a different schedule when treating the same condition.

Conclusions
We conclude that acupuncture can mitigate dysmenorrhea in terms of providing pain relief, reducing analgesic dependence, improving symptoms and improving quality of life. Furthermore, treatment e cacy is not directly proportional to the dose or number of acupoints stimulated. A well-designed acupuncture schedule with de-qi maneuver is probably more important than the number of needles. We thus concluded that the effect of acupuncture in dysmenorrhea is not direct correlated with the autonomic nervous system adjustment. Acupuncture might have a short-term effect in changing the HRV parameters, yet the true mechanism for acupuncture in dysmenorrhea still warrants further investigation.

Availability of data and materials
The datasets used for the current study are available upon reasonable request from the corresponding author. Figure 1 Flow chart of study Figure 1 Flow chart of study Figure 2 Selected acupoints in the study high dose acupuncture group: All acupoints low dose acupuncture group gray-colored mark: PC6, SP4, SP6, SP10, RN4,RN6