Acupuncture and sleep disorders in Parkinson’s disease: A systematic evaluation with meta-analysis

Abstract

Background:

Parkinson’s disease (PD) patients commonly suffer from sleep disorders, significantly impacting their quality of life. Western treatments often entail adverse effects, while acupuncture (ACU) presents a safe, nonaddictive alternative.

Methods:

A thorough literature search was performed across PubMed, Cochrane Library, and Embase databases. Eligible studies underwent statistical analysis via RevMan 5.4 software.

Results:

This study synthesized data from 19 randomized controlled trials involving 1300 patients. The ACU cohort showed notable improvement in Parkinson’s disease sleep scale (PDSS) scores (mean difference [MD] = 10.81, 95% confidence interval [CI]: 5.64, 15.98) relative to controls. Subgroup analysis revealed significance for ACU treatments beyond 6 weeks (MD = 15.39, 95% CI: 11.70, 19.09) but not for those 6 weeks or shorter (MD = 3.51, 95% CI: −1.20, 8.23). Notably, electroacupuncture resulted in significant PDSS score enhancements (MD = 12.39, 95% CI: 6.06, 18.71), with sensitivity analysis verifying result stability. However, without electroacupuncture, PDSS score differences were insignificant (MD = 7.83, 95% CI: −2.33, 17.99) and had lower result stability. Additionally, increased ACU session frequency may yield better improvements in PDSS scores. The ACU group also observed Improved Pittsburgh Sleep Quality Index scores (MD = −4.52, 95% CI: −6.36, −2.67). However, no significant variation was identified in Epworth Sleepiness Scale score improvement between groups (MD = −0.90, 95% CI: −3.67, 1.88).

Conclusion:

ACU therapy effectively improves nighttime sleep quality in PD patients. A treatment duration extending beyond 6 weeks is highly recommended. Additionally, increasing the frequency of ACU sessions and incorporating electroacupuncture in the treatment regimen may be essential for optimal results.

1. Introduction

Parkinson’s disease (PD), also known as paralysis agitans, or shaking palsy, is a common neurodegenerative condition that predominantly afflicting middle-aged and elderly individuals. Among the various non-motor symptoms in PD, sleep disorders are a common and challenging issue with complex underlying causes, including disturbances in the central sleep regulation, adverse drug reactions, and interactions between motor and non-motor symptoms. These sleep disorders can manifest as insomnia, excessive daytime sleepiness, restless legs syndrome, rapid eye movement sleep behavior disorder, and sleep-related respiratory disturbances, significantly impacting the patients’ quality of life.^[1] Prior research has suggested that sleep disorders are prodromes of other non-motor symptoms in PD patients.^[2] The quality of nighttime sleep directly influences motor symptoms, attention, and overall functioning.^[3,4] Therefore, effectively treating sleep disorders in PD patients may lead to improvements in both sleep quality and other non-motor and motor symptoms. Western medicine typically relies on sedatives, hypnotics, antianxiety drugs, and antidepressants to manage sleep disorders in PD patients. Nonetheless, these drugs come with side effects such as impaired memory, decreased reaction ability, addiction, relapse following drug withdrawal, exacerbated constipation, and excessive daytime sleepiness.^[5]

In contrast, acupuncture (ACU), rooted in traditional Chinese medical principles, offers an alternative approach for treating diseases by clearing the meridians and regulating the Yin-Yang balance. Extensive research has indicated that ACU can enhance sleep quality by modulating neurotransmitter levels, improving melatonin function, and adjusting the biological clock. This is because ACU influences several chemical factors closely linked to sleep mechanisms, including 5-hydroxytryptamine, norepinephrine, dopamine, gamma-aminobutyric acid, glutamate, and nitric oxide.^[6,7] The benefits of ACU for sleep disorders in PD patients are evident as it poses no adverse events and addiction risks. Additionally, it is a cost-effective treatment option that is suitable for a diverse range of populations.

Recent studies have explored the potential of ACU in treating PD patients. However, many have shortcomings such as small sample sizes, focusing on a single outcome measure, and inconsistent quality. To address this, our study evaluated the efficacy of ACU in managing sleep disorders in PD patients through a comprehensive analysis of relevant randomized controlled trials (RCTs). We utilized Epworth Sleepiness Scale (ESS), Parkinson’s Disease Sleep Scale (PDSS), and Pittsburgh Sleep Quality Index (PSQI) as measurements provide robust evidence for the application of ACU in treating sleep disorders in PD patients.

2. Method

The current study adhered to the guidelines outlined by the guidelines stipulated by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement and was registered prospectively on the International Prospective Register of Systematic Reviews (CRD42022336847).

Patient and public involvement: All analyses are based on previously published studies. Therefore, no public or patient involvement is required.

2.1. Search strategy

A comprehensive, systematic search of databases, including China National Knowledge Infrastructure, VIP, WanFang, Chinese Biomedical Literature Database, PubMed, Cochrane Library, and Embase databases, was conducted up until July 1^st, 2023. Two researchers performed the search independently following a predetermined search strategy, utilizing both controlled terms (such as “Parkinson’s Disease” [Mesh], “Acupuncture” [Mesh], and “Pharmacopuncture” [Title/Abstract]) and uncontrolled terms. Details on the search strategy are provided in Data S1, Supplemental Digital Content, http://links.lww.com/MD/K854, Data S2, Supplemental Digital Content, http://links.lww.com/MD/K855.

2.2. Literature screening

Two investigators independently carried out a thorough literature search adhering to a preestablished strategy. Once duplicates were excluded, titles and abstracts of the remaining articles were reviewed to eliminate nonrelevant studies. Subsequently, the remaining articles underwent a detailed examination to remove any that did not meet the inclusion criteria. Eligible studies were then pooled together, and 2 investigators cross-verified the screening results. In case of any disagreement, a third investigator was invoked to aid in making the final decision.

The inclusion criteria encompassed the following aspects: 1. Study design: randomized controlled trials published both domestically and internationally, examining the efficacy of ACU in treating PD. No restrictions were placed on blinding methods and allocation concealment; 2. Participants: patients had been clinically diagnosed with PD according to the UK Parkinson’s Disease Society Brain Bank Clinical Diagnostic Criteria or comparable accepted diagnostic standards.^[8] No restrictions were imposed regarding gender, age, ethnicity, or disease stage; 3. Interventions: the treatment group received ACU or a combination of ACU and primary treatment, while the control group had no treatment limitations; 4. Outcome measures: PDSS, PSQI, and ESS scores; 5. Language: literature in either English or Chinese was accepted.

Exclusion criteria are as follows: 1. Original trials lacking clear or accepted diagnostic criteria; 2. Duplicates; 3. Non-RCTs, including case reports, self-controlled studies, literature research, and reviews; 4. Original trials with significant methodological flaws, such as data inaccuracies or statistical processing discrepancies; 5. Studies in which outcome data remained unobtainable, despite efforts to contact the original authors.

2.3 . Information extraction

Two investigators performed data extraction independently, incorporating the following information: 1. Core details of the included studies, such as the title, primary author, and year of publication; 2. Baseline characteristics of participants, encompassing sample sizes of both the intervention and control groups, age of patients, and disease duration; 3. Intervention details: treatment modality, frequency, and duration; 4. Essential elements for bias risk evaluation: randomization method, allocation concealment, blinding, attrition, and follow-up; 5. Outcome measures: PDSS, PSQI, and ESS scores. After completion, the 2 investigators’ data extraction outcomes were cross-verified, with discrepancies resolved appropriately.

The PDSS is a patient-reported visual analogue tool consisting of 15 items, each scored on a scale of 1 to 10, with 1 indicating severe symptoms and ten representing symptom-free. The total score ranges from 0 to 150, with lower scores indicating more severe sleep disturbances.^[9]

The ESS is an eight-item self-report measure wherein respondents rate their sleepiness on a scale of 0 (no sleepiness) to 3 (high degree of sleepiness) for each item. The overall score spans from 0 to 24, with higher scores indicative of increased daytime sleepiness.

The PSQI consists of 9 components, with the first 4 requiring descriptive responses and the last 5 presenting multiple-choice questions. Component 5 includes ten sub-questions. After respondents complete all questions, scores are assigned between 0 and 21 based on the scoring guidelines, with higher scores reflecting poorer sleep quality.

2.4. Risk of bias assessment

The methodological quality of the included RCTs was evaluated in line with the criteria outlined in the Cochrane Handbook for Systematic Reviews 5.4.0. These criteria encompassed randomization, allocation concealment, blinding, attrition bias, selective reporting, and other potential biases. The assessment results for each item were categorized as “low” risk of bias (yes), “unclear” risk of bias (unclear), or “high” risk of bias (no).

2.5. Statistical analysis

Statistical analysis was carried out using Revman5.4 software. Treatment effects were measured using the differences in PDSS, PSQI, and ESS scores before and after treatment, with the mean difference (MD) serving as the effect size. In cases where different evaluation tools or methods were used in the included studies, the standard mean difference was utilized as the effect size. Each effect size was accompanied by a 95% confidence interval (CI). The heterogeneity of the included trials was gauged using the I² index. If I² was ≥50% and P < .05, suggesting significant heterogeneity among the trials, a random-effects model was employed to merge the effect sizes, and a sensitivity analysis was conducted. Conversely, if no significant heterogeneity was found, a fixed-effects model was adopted for data analysis. Additionally, publication bias was assessed when the pooled effect size reached or exceeded 10.

3. Results

3.1. Literature search results

Search of databases generated 2713 articles which were subsequently managed using the Endnote software. After eliminating 515 duplicates, titles and abstracts of the remaining studies were screened to exclude reviews, case reports, self-controlled studies, animal experiments, and interventions deemed irrelevant, resulting in the exclusion of an additional 1952 articles. Upon a detailed full-text examination, 218 papers were excluded due to unclear diagnostic criteria, unavailability of outcome measures, or vague treatment method descriptions. Consequently, a total of 19 trials was incorporated into the present study.^[10–28] The literature screening process is depicted in Figure 1.

Figure 1.

The process of literature screening.

3.2. Basic characteristics of included studies and quality assessment

Nineteen trials met the eligibility criteria for this meta-analysis, providing a combined sample size of 1300 participants, with 663 in the experimental groups and 637 in the control groups. These studies were published between 2005 and 2022. The interventions applied in the experimental groups encompassed ACU, acusector, the combination of ACU with the basic treatments administered to the control group, and the combination of acusector with the basic treatments administered to the control group. The control groups received interventions such as drug therapy, madopar, sham ACU, and hyperbaric oxygen (HPO) in conjunction with madopar. Among the studies, eleven reported on PDSS,^{[10,12,15,16,18,20,22–24]} 8 on the PSQI,^{[11,17–19,21,25,26]} and 3 on the ESS.^[12–14] Table 1 displays the basic information of the included studies.

Table 1.

Basic information on included studies.

Author	Year	Egion	Population* (e,c)	Age (mean, SD)	Gender (female, male)	Diseases’s condition†	Duration‡ (year)	Comparisons (I & C)§	Outcomes	The total number of acupuncture sessions	The duration of acupuncture treatment
Aroxa[10]	2017	Brazil	11,11	60.5 ± 11.72	8,14	20 (0.5);10 (1)	NON	ACU + drug versus drug	PDSS	8	8 weeks
Wang[11]	2015	China	28,20	60.85 ± 10.39	26,22	2.0 ± 0.735	2.82 ± 2.64	EA + drug versus drug	PSQI	20	60 days
Kong[13]	2017	Singapore	20,20	64.6 ± 8.4	27,13	NON	5.73 ± 3.79	ACU versus Sham	ESS	10	5 weeks
Kluger[12]	2016	USA	47,47	63.7 ± 11.7	35,59	1,6;1.5,9;2,28;2.5,30;3,16;4,4	NON	ACU versus Sham	PDSS/ESS	12	6 weeks
L.H.Li[14]	2022	China	30,27	61.11 ± 8.22	33,24	2.68,0.39	5.93 ± 1.64	ACU versus Sham	PDSS-2/ESS	26	30 days
Xu[15]	2020	China	33,37	61.85 ± 9.94	34,36	1,17;1.5,16;2,11;2.5,14;3,8;4,4	3.38 ± 2.53	EA + Madopar versus Madopar	PDSS	32	8 weeks
M.Deng[16]	2021	China	53,53	65.37 ± 5.22	50,56	NON	8.60 ± 2.78	ACU + TCM versus TCM	PDSS	20	30 days
Lei.Li[18]	2021	China	50,50	62.03 ± 7.50	47,53	NON	4.11 ± 0.59	EA + Madopar versus Madopar	PDSS/PSQI	56	8 weeks
N.Huang[25]	2014	China	20,20	60 ± 8.47	15,25	NON	3.45 ± 6.98	ACU + Madopar versus Madopar	PSQI	20	30 days
L.Li[21]	2018	China	25,25	69.65 ± 8.66	24,26	NON	NON	ACU + TCM versus TCM	PSQI	60	60 days
Q.J.Dong[20]	2018	China	36,36	59.66 ± 7.11	32,40	1,22;2,39;3,11	3.99 ± 0.94	EA + Madopar versus Madopar	PDSS	30	30 days
Y.Bai[17]	2021	China	29,29	45–78	35,23	NON	0.5–3.2	ACU + alprazolam versus alprazolam	PSQI	14	2weeks
S.Zhou[22]	2016	China	63,64	69.65 ± 4.51	65,62	1,17;2,43;3,60;4,7	8.38 ± 3.76	EA + Madopar versus Madopar	PDSS	56	8 weeks
X.Liang[24]	2014	China	35,35	61.5 ± 23.24	34,38	2.26 ± 1.47	4.36 ± 2.95	ACU versus Madopar	PDSS	72	24 weeks
G.Yan.1[23]	2015	China	20,40	45–82	30,30	2.34 ± 0.78	0.5–11	ACU + Madopar versus Madopar	PDSS	20	30 days
G.Yan.2[23]	2015	China	60,20	42–87	39,41	2.30 ± 0.81	0.08–13	ACU + HPO + Madopar versus HPO + Madopar	PDSS	20	30 days
Y.C.Zhu[[19]	2020	China	27,27	73.9 ± 5.16	33,21	NON	4.35 ± 1.05	ACU + TCM versus TCM	PSQI	7	2 weeks
H.B.Chen[26]	2021	China	31,31	74.43 ± 5.03	34,28	NON	4.30 ± 1.19	ACU + TCM versus TCM	PSQI	7	2 weeks
Z.H.Yu[28]	2019	China	30,30	66.3 ± 5.6	30,30	NON	NON	ACU + TCM versus TCM	PSQI	40	40 days
Nazarova[27]	2002	China	15,15	70.1 ± 6.2	14,16	1.6 ± 0.9	NON	EA + drug versus drug	PDSS	16	8 weeks

Drug used to treat Parkinson’s disease.

ACU = acupuncture; alprazolam = alprazolam; EA = electroacupuncture; HPO = hyperbaric oxygen; Madopar = Madopar; Sham = simulated acupuncture; TCM = traditional Chinese medicine.

^*In the form of (number of treatment group, number of control group).

^†H-y, Hoehn-Yahr Stage, expressed in the form of (stage, mean ± SD) or (stage-n).

^‡Duration is expressed as (years, mean ± SD) or range (years, shortest–longest).

^§I versus C, interventions in the treatment group and the control group.

The 19 studies included in this meta-analysis displayed a moderate overall methodological quality, all adopting a randomized design for group allocation. However, 3 of the them failed to explicitly report details regarding randomization procedures. Among the included studies, 3 utilized sequential allocation based on patient order, one used the drawing of lots method, one employed a coin toss, and one implemented third-party randomization. Computer-generated simple randomization was used in 2 studies, while 8 adopted random number tables. Only 3 studies explicitly mentioned allocation concealment. Blinding of both patients and outcome assessors was implemented in 3 studies. Two studies blindfolded only the outcome assessors, while the remaining studies did not provide information about blinding. Complete outcome data were reported in all studies, with only 5 registering the study in advance. The risk of bias summary and graph are presented in Figure 2A and B, respectively.

Figure 2.

(A) Risk of bias summary; (B) risk of bias graph.

3.3. Meta-analysis results

3.3.1. Effect of ACU on the PDSS scores of patients with PD.

Out of the 19 trials included, L.H. Li^[14] utilized the PDSS-2 for descriptive analysis. The meta-analysis incorporated ten studies that reported PDSS, with G. Yan study^[23] contributing 2 control groups: ACU + Madopar versus Madopar, and ACU + HPO + Madopar versus HPO + Madopar. Given the heterogeneity test result (I² = 60%, P < .05), a random-effects model was used for pooling effect sizes, with MD as the statistical metric. The pooled MD was 10.81 (95% CI: 5.64, 15.98), indicating that the ACU group demonstrated a significant improvement in PDSS score compared to the control group (Z = 4.10, P < .0001). Figure S1, Supplemental Digital Content, http://links.lww.com/MD/K851 displays the PDSS scores. Sensitivity analysis revealed that excluding any single study did not significantly alter the results, reinforcing the stability of the meta-analysis.

The study by L.H. Li^[14] investigated the PDSS-2 scores, revealing no statistical difference in the improvement in PDSS-2 score between the ACU and sham ACU groups on the 30^th-day post-admission. However, the ACU group showed enhanced sleep latency, efficiency, and total sleep time compared to the sham ACU group on the same day, with a statistically significant difference (P < .01).

3.3.2. Effect of ACU on the PSQI scores of patients with PD.

Among the 19 included trials, Y.C. Zhu^[19] and H.B. Chen^[26] solely reported individual subitem scores of posttreatment PSQI and conducted a descriptive analysis of these scores. The meta-analysis of 6 included studies that reported total PSQI scores before and after treatment. Due to the heterogeneity test results (I² = 94%, P < .05), a random-effects model was employed to pool effect sizes, using MD as the statistical metric. The pooled MD was −4.52 (95% CI: −6.36, −2.67), indicating significant improvements in PSQI scores for the treatment group compared to the control group (Z = 4.80, P < .0001). Figure 3A illustrates the PSQI scores. Sensitivity analysis reaffirmed the stability of the meta-analysis, as no substantial changes were observed after excluding any study.

Figure 3.

(A) Forest plot of PSQI scores; (B) forest plot of ESS scores; (C) funnel plot of PDSS scores; (D) forest plot for subgroup analysis of PDSS scores based on course of acupuncture treatment.

Y.C. Zhu^[19] and H.B. Chen^[26] compared the components of PSQI between treatment and control groups, including sleep disorders, sleep efficiency, sleep latency, and sleep duration. The treatment group exhibited significantly better results than the control group (P < .05). Furthermore, Y.C. Zhu^[19] reported significantly improved daytime function scores in the treatment group compared to the control group (P < .05).

3.3.3. Effect of ACU on the ESS scores of patients with PD.

Among the 19 trials, 3 studies reporting ESS scores were integrated into the meta-analysis. Considering the heterogeneity test result (I² = 78%, P < .05), a random-effects model was employed to pool effect sizes, using MD as the statistical metric. The pooled MD was −0.90 (95% CI: −3.67, 1.88), suggesting no significant difference between the groups (Z = 0.63, P = .53 > .05). Figure 3B presents the ESS scores. However, sensitivity analysis showed that excluding Kong study influenced the meta-analysis results, yielding a new MD of −2.24 (95% CI: −4.18, −0.30). This indicates that the treatment group had a significant advantage over the control group in improving ESS scores (Z = 2.26, P = .02 <.05).

3.3.4. Subgroup analysis.

Subgroup analyses were conducted for PDSS scores, considering 3 independent variables: duration of ACU treatment, total number of ACU sessions, and application of electroacupuncture.

Subjects were segregated into 2 cohorts based on the duration of ACU treatment: those receiving treatment for less than or equal to 6 weeks vs. those treated for over 6 weeks. Notably, a significant decrease in heterogeneity was observed, with I² values of 14% and 4%, respectively, leading to the adoption of a fixed-effect model. The combined effect sizes for short-duration and long-duration cohorts were: MD = 3.51 (95% CI: −1.20, 8.23) and MD = 15.39 (95% CI: 11.70, 19.09), respectively. Sensitivity analyses confirmed the robustness of these results. This suggests a nonsignificant difference in PDSS score improvements between the ACU and control groups for treatment durations less than or equal to 6 weeks. Conversely, the ACU group demonstrated significantly improved PDSS scores for treatments over 6 weeks. Detailed results are shown in Figure 3D.

Considering the number of ACU sessions, subgroups were generated using 16 and 20 sessions as cutoffs (Fig. S2A and B, Supplemental Digital Content, http://links.lww.com/MD/K852). For subgroups with a total of sessions exceeding 16 and 20, the ACU group consistently outperformed the control group regarding PDSS score improvements, corroborated by stable sensitivity analysis results. No significant difference was observed between groups in the subgroup with total sessions less than or equal to 16, though sensitivity analysis results were inconsistent. Likewise, the subgroup with less than or equal to 20 sessions initially suggested superior outcomes in the ACU group, but sensitivity analyses revealed inconsistent findings. Detailed results are tabulated in Table 2.

Table 2.

Subgroup analysis of PDSS based on the number of acupuncture sessions, and use of electropuncture.

Subgroup	Heterogeneity (I2)	MD (95%CI)	Sensitivity analysis	Studies	Analysis model
The number ≤ 16	72%	5.93 [−10.04, 21.90]	Unstable	Aroxa, Kluger, Nazarova	Random effects
The number > 16	59%	12.13 [6.39, 17.86]	Stable	G. Yan.1, G. Yan.2, Lei.Li, M. Deng, Q.J. Dong, S. Zhou, X. Liang, Xu	Random effects
The number ≤ 20	44%	8.20 [2.75, 13.64]	Unstable	Aroxa, Kluger, Nazarova, M. Deng, G.Yan.1, G.Yan.2	Fixed effect
The number > 20	73%	13.60 [6.57, 20.63]	Stable	Lei.Li, Q.J. Dong, S. Zhou, X. Liang, Xu	Random effects
Electroacupuncture	63%	12.39 [6.06, 18.71]	Stable	Lei.Li, Nazarova, Q.J. Dong, S. Zhou, Xu	Random effects
Non-electroacupuncture	65%	7.83 [−2.33, 17.99]	Unstable	Aroxa, G. Yan.1, G. Yan.2, Kluger, M. Deng, X. Liang	Random effects

The number = the number of acupuncture sessions.

Regarding the use of electroacupuncture, meta-analysis results indicated that the electroacupuncture subgroup exhibited significantly improved PDSS scores compared to controls, with sensitivity analysis affirming the robustness of these findings (Fig. S3, Supplemental Digital Content, http://links.lww.com/MD/K853). Conversely, no significant difference was observed between groups in the non-electro-acupuncture subgroup, though sensitivity analysis revealed inconsistent results. Detailed results are found in Table 2.

3.3.5. Publication bias assessment.

Publication bias was evaluated through a funnel plot analysis of the studies reporting PDSS scores, as depicted in Figure 3C. The symmetric funnel plot implied the absence of publication bias. On the other hand, fewer than 10 studies reported PSQI or ESS, precluding a comprehensive evaluation of publication bias for these measures.

4. Discussion

4.1. Summary of findings

Our investigation affirms the efficacy of ACU as a therapeutic intervention for sleep disturbances in PD patients. Compared to the control cohort, the ACU-treated cohort demonstrated significant ameliorations in PDSS and PSQI scores, indicative of dependable and consistent outcomes. Conversely, no significant disparity was discerned in the ESS score between the groups. Nevertheless, the sensitivity analysis raised some concerns about the robustness of these results due to the limited number of studies included in the ESS score analysis.

Subgroup analysis of the PDSS scores disclosed an advantage for the experimental cohort in enhancing PDSS scores for those undergoing ACU treatment courses beyond 6 weeks. There was no statistically significant disparity between experimental and control groups when the treatment lasted for 6 weeks or less, and sensitivity analyses confirmed the stability of these findings in both subgroups. When the number of ACU sessions exceeded 16 and 20, results consistently indicated superior PDSS score improvements in the experimental cohort. Sensitivity analysis further substantiated the stability of these findings. However, for those receiving fewer sessions, no statistically significant disparity was observed in PDSS score improvements, albeit with the instability of results as suggested by the sensitivity analysis. In the context of electroacupuncture usage, the experimental group outperformed the control group regarding PDSS score improvement. There was no significant disparity between groups without electroacupuncture, but sensitivity analysis indicated instability of these findings. These subgroup analysis results suggest that a more extended course of ACU treatment might be crucial for enhancing PDSS scores. Furthermore, implementing electroacupuncture and increasing treatment frequency could yield additional benefits.

Based on our research findings, we believe there are significant clinical implications. Our study provides evidence suggesting that ACU can enhance the nocturnal sleep quality of patients with PD, which can be pivotal in elevating their overall quality of life. The results underscore that treatment durations extending beyond 6 weeks are essential, and integrating electro-acupuncture and intensifying weekly treatment frequencies may further optimize therapeutic outcomes. This insight offers valuable guidance for clinicians when designing ACU treatment regimens for PD patients. Additionally, in contrast to pharmaceuticals, ACU presents a secure nondrug therapeutic alternative, devoid of risks associated with addiction or adverse reactions. Furthermore, improved sleep may also contribute to the amelioration of both motor and non-motor symptoms associated with PD.

The risk of bias predominantly arose from the challenge of blinding in the study design, especially regarding practitioner blinding in ACU due to its specific treatment modality. As it stands, there are no effective methodologies to blind the practitioner.

Considerable heterogeneity was observed among the included studies, especially those focusing on PSQI and ESS. We endeavored to analyze this from various perspectives, but the sources of the heterogeneity remained unclear. Aside from the number of treatments, duration of therapy, and use of electroacupuncture, other potential heterogeneity sources include the selection of ACU points, manipulation techniques employed, and sequence of needle insertions. Notably, our assessment detected no publication bias.

4.2. Reflections on PDSS, PSQI, and ESS

PDSS is a comprehensive 15-item scale designed to evaluate sleep disorders in PD patients. It goes beyond assessing overall nocturnal sleep quality and diurnal somnolence, delving into factors that specifically influence sleep quality, such as nocturnal limb restlessness, vivid dreaming, hallucinations, nocturia, and nocturnal limb discomforts including numbness, pain, spasms, and tremors. Previous research has demonstrated ACU’s efficacy in treating conditions like restless legs syndrome,^[29] obstructive sleep apnea,^[30] and increased nocturnal urination,^[31] all of which significantly impact the nocturnal sleep quality of PD patients and constitute vital components of the PDSS. ACU’s ability to alleviate these symptoms could explain its efficacy in enhancing PDSS scores. Our subgroup analysis revealed that the duration and frequency of ACU sessions could influence PDSS score improvement, mirroring previous research outcomes. Other studies investigating ACU’s therapeutic potential for insomnia also advocate for a requisite number of ACU sessions to effectuate effective treatment.^[32] Analogous conclusions have been proposed by RCTs suggesting that an increment in the frequency of ACU treatments engenders superior therapeutic outcomes for insomnia patients.^[33,34] Emerging evidence from functional magnetic resonance imaging (fMRI) studies suggests that ACU can modulate brain activity, indicating that increased treatment frequency could augment physiological outcomes.^[35] Some RCTs focusing on other medical conditions have also suggested that extending treatment duration enhances ACU’s therapeutic effectiveness. For instance, a study by Xiao Xue^[36,37] treating 90 patients with primary dysmenorrhea for 1, 2, and 3 treatment courses reported superior therapeutic outcomes after 2 or 3 courses compared to a single course. A separate study by Eun Hui He^[38] reported enhanced therapeutic outcomes associated with extending treatment duration within 6 weeks, where 42 female patients affected by stress urinary incontinence were treated with electroacupuncture. Moreover, previous meta-analyses have suggested that electroacupuncture is more efficacious than conventional ACU in ameliorating insomnia, closely aligning with our research findings.^[39]

PSQI is a reliable and valid scale routinely employed to assess insomnia.^[40] Although certain PSQI items resemble those of the PDSS, such as vivid dreams, pain, and frequency of nocturia, the scoring methods diverge. While PSQI aggregates the scores of 9 items (5b-5j) to give a final score of 0 to 3, PDSS individually scores similar items (3–13) on a 1 to 10 scale. This differential approach to assessing analogous items somewhat rectifies the inherent limitations in subjective rating scales, rendering the results relatively objective. Previous studies on non-PD insomniacs have confirmed that ACU can effectively improve PSQI scores,^[32,41] aligning with our research results.

The ESS primarily evaluates diurnal somnolence rather than nocturnal sleep quality. It assesses patient sleepiness across 8 everyday scenarios, including watching TV, reading, sitting in public places, traveling in a car for 1 hour, taking a midday break, engaging in conversation, sitting quietly after a meal, and being stuck in a brief traffic jam. All 3 trials included contributing to the meta-analysis of ESS score are control studies that compare ACU with sham ACU treatment. The meta-analysis revealed no statistical difference in ESS score improvement between ACU and sham ACU groups. However, after excluding Kluger study,^[12] the results indicated superior ESS score improvement in the ACU group compared to the sham ACU group. Nonetheless, upon a comprehensive review of all included studies in our meta-analysis, it was found that the heterogeneity caused by Kluger study^[12] cannot be explained by the PICOS principle (participants, intervention, comparison, outcome, and study design). ESS evaluates daytime sleepiness across 8 scenarios, and its scores may be influenced by individual lifestyle, education level, and place of residence. For instance, rural patients with lower educational attainment may not regularly read, travel by car, or frequently experience traffic jams, which could lead to similar scores before and after treatment. However, the enrollment criteria of these 3 studies did not restrict participants based on educational background or residential environment, potentially resulting in subjective scores that could impact our meta-analysis results. A previous study on ACU’s effect on daytime sleepiness in non-PD patients showed that ACU could effectively improve PSQI scores for patients experiencing daytime sleepiness.^[42] Among the 3 studies included in the ESS analysis, L.H. Li^[14] reports positive results, while Kluger^[12] and Kong^[13] report negative ones. These studies utilized varying dosages of L-dopa in both the treatment and control groups. Notably, the study by L.H. Li employed a significantly lower dosage of L-dopa compared to those applied in Kluger and Kong studies. Additionally, long-term oral administration of L-dopa has been identified as a critical factor contributing to daytime sleepiness in PD.^[43]

4.3. Strengths and limitations

This systematic review and meta-analysis pooled various RCTs to create a larger sample size and investigated the efficacy of ACU on sleep disturbances regarding parameters such as PDSS, PSQI, and ESS scores, but its limitations should be noted.

First, some included studies had methodological shortcomings, limited sample sizes, and short follow-up periods. Second, the number of incorporated studies, especially in the meta-analysis of ESS, is relatively low, which may impact the overall findings. Third, the reliance on subjective patient self-reporting or examiner evaluation for PDSS, PSQI, and ESS lacks objective markers such as laboratory diagnostics and imaging examinations. Fourth, this review only included studies published in English or Chinese, potentially limiting the scope of findings. Moreover, the optimal treatment duration for ACU remains unclear due to the limited number of available studies.

Similarly, the investigation identifies that both the application of electroacupuncture and the number of ACU sessions may influence the improvement in PDSS scores. However, these findings necessitate cautious interpretation due to the instability in the subgroup analysis results. Lastly, there is inherent heterogeneity among the included studies, potentially related to variations in acupoint selection and ACU techniques, which may impact the meta-analysis outcomes.

5. Conclusion

ACU may aid in improving the quality of nocturnal sleep for patients with Parkinson’s disease. The duration of the treatment course influences the enhancement of PDSS scores and may also be associated with electroacupuncture and the frequency of ACU sessions. Based on our research, we recommend a treatment duration exceeding 6 weeks for optimal enhancement of PDSS scores. A higher frequency of ACU sessions and the incorporation of electroacupuncture may be beneficial. In light of the limitations of this study, it is recommended that future research investigate rigorous and multicenter RCTs with large-scale samples to further investigate the effectiveness of ACU in treating sleep disorders for PD patients and explore the factors influencing therapeutic outcomes. Furthermore, including objective indicators such as NT-3, IL-1β, cystatin C, and polysomnography in the outcome measures may provide valuable insights.^[44,45]

Acknowledgments

We want to thank the researchers and study participants for their contributions.

Author contributions

Conceptualization: Fei Yan.

Formal analysis: Qiuju Feng.

Investigation: Qiuju Feng.

Methodology: Chen Chen.

Resources: Zongju Huang, Yongliang Chen.

Supervision: Huan Chen.

Writing – original draft: Fei Yan.

Writing – review & editing: Huan Chen.

Supplementary Material

medi-103-e36286-s001.docx ^{(28.9KB, docx)}

medi-103-e36286-s002.docx ^{(59.3KB, docx)}

medi-103-e36286-s003.tif ^{(579.3KB, tif)}

medi-103-e36286-s004.tif ^{(1.3MB, tif)}

medi-103-e36286-s005.tif ^{(671.6KB, tif)}