Using meta-regression approach to explore the dose-response association between acupuncture sessions and acupuncture effects on chronic prostatitis/chronic pelvic pain syndrome

Introduction

Category III prostatitis, also known as chronic prostatitis/chronic pelvic pain syndrome (CP/CPPS) presents as chronic pain in the region of the pelvis, and often associated to ejaculatory pain, erectile dysfunction, and psychosocial symptoms, lasting for at least 3 to 6 months (1). As the most common type of prostatitis, CP/CPPS affects approximately 90% to 95% of men worldwide (2,3). Owing to its long-lasting symptoms that are tough to treat and quantify, indeed both patient and physician frustration is high in dealing with this disorder.

The benefits of acupuncture on CP/CPPS have been well established according to clinical trials (4-8), and guideline also supported its effectiveness for CP/CPPS (5,9,10). However, the existence of a possible dose-response association between acupuncture and its effects has not been well investigated. Currently, the acupuncture sessions of published studies ranged from 6 to 24 times, optimal dosing for acupuncture is not fully understood and clinical guidelines including acupuncture do not describe minimum treatment session that would constitute a full course of treatment for given condition (7).

As which is particularly important in pharmacology and epidemiology, dose-response relationship should be considered in the acupuncture clinical practice and research. White and colleagues realized question of “dose-response” for acupuncture study decade ago (11), suggesting that there is an urgent need for studies to guide decisions about what is an “adequate” dose of acupuncture for different conditions. To date, few studies examined factors associated with the number of acupuncture sessions completed, considered the question of adequate acupuncture sessions and applied criteria. Research into what constitutes an adequate dose of acupuncture has long been neglected and uncertainty exists regarding the dose-response relationship between treatment sessions and the effect of acupuncture (11). This study also demonstrated that there appear to be relationships between treatment timing, treatment frequency, and mode of needle stimulation, albeit these conclusions were drawn by qualitative approaches instead of quantitative analyses (11).

To address the aforementioned shortfalls and develop evidence-based recommendations, we introduced meta-regression approach into dose-response association between acupuncture sessions and acupuncture effects on CP/CPPS.

Methods

This meta-analysis was reported according to the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) statement and recommendations for improving the methodological quality of dose-response meta-analysis (12,13).

Literature search

Two reviewers independently searched PubMed, EMBASE, and Cochrane CENTRAL Register for Clinical Trials for English-language publications from their inception through December 8, 2017 and updated at May 20, 2018 using the search terms: “chronic prostatitis” OR “chronic pelvic pain syndrome” OR “CP/CPPS” AND “acupuncture” (see Table S1, which describe the search terms and strategies). In addition, a manual search was performed for relevant systematic reviews and guidelines. Both two authors have over 3 years’ experience in evidence-based practice. Another author (C Xu) checked the excluded records to ensure there was no mis-operation.

Table S1 Search strategies
Full table

Study selection

Trials were included if (I) patients diagnosed with CP/CPPS or chronic nonbacterial prostatitis; (II) efficacy data were available for acupuncture participants for at least one time point other than baseline and endpoint; (III) the National Institutes of Health Chronic Prostatitis Symptom Index (NIH-CPSI) was utilized as outcome measurement of CP/CPPS (14). Trials were excluded if (I) patients with acute bacterial prostatitis, benign prostatic hyperplasia, prostate cancer or other prostate diseases; (II) non-invasive acupuncture was used (laser acupuncture, acupressure et al.); (III) secondary analysis using the published clinical sample; (IV) adjunctive therapy was provided to the acupuncture group.

We defined the total acupuncture sessions as the total counts of acupuncture therapy; we use the acupuncture sessions as primary intervene because compared with other variable it significantly varies with different studies; the primary outcome, NIH-CPSI, refers to the total score corresponding to each time point.

Data extraction and risk of bias assessment

Two authors extracted data from selected studies using an electronic standardized data extraction form. Any disagreements were solved by consensus. Study characteristics, patient characteristics, acupuncture sessions, acupuncture types, acupoints selection, and the NIH-CPSI score (range from 0 to 43). Included trials provided NIH-CPSI for at least two time points (baseline, first visit point). Additionally, if trials reported outcomes in a figure rather than in a table, GetData Graph Digitizer version 2.26 was used to extract data from the figures (software available at http://getdata-graph-digitizer.com).

Two authors evaluated the quality of included studies independently. We used the Cochrane Collaboration tool to evaluate the risk of bias for the RCTs. For assessing CSSs, a tool developed by Moga et al. was used, which consists of 20 items and can systematically evaluate the CSSs (15,16).

Statistical analysis

To build the relationship of acupuncture treatment sessions and NIH-CPSI score, we conducted dose-response meta-analyses using the approach of robust-error meta-regression (REMR) (17). Based on the “one-stage” framework, REMR method treated each included study as a cluster, and fitted the meta-regression of treating sessions against the mean NIH-CPSI score within a whole data set. Additionally, we weighted each dose-specific effect for the pooling by using the inverse variance method, which, at the same time, can balance heteroscedasticity in REMR model and ensure the unbiasedness of parameter estimation.

Restricted cubic spline (RCS) was used for approximating non-linear relationship. This is reasonable since the data points of dose-specific effect are sufficient (at least four data points between two adjacent knots) to fit an RCS function (18). We set three fixed knots to place splines inserting values, with both tails of the curve restricted to linear. The step would generate two splines, in which, the first spline represents the initial “dose” and the second one would be the non-linear function (spline transformation) of dose (18). We forced the regression coefficient of the second spline to zero to test whether the trend correspond to non-linear, with the P<0.1 indicating the evidence of non-linear.

The mean value of NIH-CPSI score and the mean difference of NIH-CPSI score were used as effect estimators while the treatment sessions as “dose” in this meta-regression analysis. We illustrated the dose-response results in following two methods: initially, we synthesized the means of NIH-CPSI score to demonstrate association between the different acupuncture sessions and dose-specific NIH-CPSI score; then, we synthesized the change of means of NIH-CPSI score, with the mean difference equals zero as the reference.

Additionally, we undertook sensitivity analyses by omitting CSSs and studies with sample size less than ten participants. The subgroup analysis was classified by different acupuncture types (manual acupuncture vs. electroacupuncture). Because for one-stage dose-response meta-analysis, there were currently no valid measurement methods for assessing heterogeneity, we used the statistic (1-R²) as a rough estimation of heterogeneity. Analyses were performed using the Stata 13.0 software (Stata Corp., College Station, TX, USA), with a two-side test of α=0.05 as the significant level.

Results

Characteristics of included studies

A flowchart describing the selection of eligible trials is presented in Figure 1. After removal of duplicates, 141 citations for eligibility with 57 full-texts were used. Finally, a total of ten studies (six RCTs and four CSSs) were included in qualitative review and meta-analysis (6,8,19-26). All studies were published in English. Table 1 summarizes the parameters of included studies. In general, the quality items of included studies were generally acceptable. Tables 2,3 present the quality assessment of included studies.

Figure 1 Study flow chart.

Table 1 Characteristics of selected studies
Full table

Table 2 Risk of bias assessment for included randomized controlled trials
Full table

Table 3 Quality assessment for assessing included case series studies (16)
Full table

Acupuncture sessions and the NIH-CPSI score

Ten studies (six RCTs and four CSSs) were included in the meta-regression model and the results indicated that more acupuncture sessions were associated with greater therapeutic effects. A J-shaped association between acupuncture sessions and NIH-CPSI score was presented: after 6 acupuncture sessions, the NIH-CPSI decreased from 26.1 (95% CI: 25.3–27.0) to 18.5 (95% CI: 11.6–25.4), with a between-session difference of −7.6 (95% CI: −14.6 to −0.7). After 18 acupuncture sessions, the NIH-CPSI decreased to 17.8 (95% CI: 15.9–19.7), with a difference of −8.3 (95% CI: −10.4 to −6.3), and a 6-point decline of NIH-CPSI score have been used as the optimal threshold for minimal clinically important difference (27). After 24 acupuncture sessions, the NIH-CPSI increased to 19.0 (95% CI: 15.3–22.7), with a difference of −7.1 (95% CI: −10.9 to −3.4). Moderate heterogeneity was observed between included studies (1-R² =54.7%). Figures 2,3 presented the results of NIH-CPSI score and changes of NIH-CPSI score, respectively.

Figure 2 Dose-response relationship between acupuncture sessions and NIH-CPSI score.

Figure 3 Dose-response relationship between acupuncture sessions and changes of NIH-CPSI score.

In sensitivity analysis, when CSSs were omitted from the analysis, there remained a similar dose threshold in treatment sessions providing clinical effectiveness. Six RCTs were included in the meta-analysis showed the association between acupuncture sessions and NIH-CPSI score (6,8,22-24,26). The results from sensitivity analyses are congruent with the primary results. After 18 acupuncture sessions, the NIH-CPSI decreased from 26.3 (95% CI: 25.6–27.0) to 18.2 (95% CI: 16.3–20.2), with a difference of −8.1 (95% CI: −9.5 to −6.7). Sensitivity analysis by omitting studies with small sample size found no important changes compared with remaining studies (6,8,22,24-26). After 18 sessions, the NIH-CPSI changed −8.3 (95% CI: −10.1 to −6.5). Subgroup analysis by classified studies using electroacupuncture found no significant difference compared with studies using manual acupuncture.

Discussion

Summary of main findings

To the best of our knowledge, this is the first study to explore the association between acupuncture sessions and acupuncture effects using dose-response methodology. We reported a dose-response association between acupuncture sessions and NIH-CPSI score, at least six acupuncture sessions might be recommended, and prolonged sessions of acupuncture may have clinically different effects.

Compared with existing studies

There is an urgent need for studies to guide decisions about what is an adequate dose of acupuncture for different conditions. In 1990s, a team from Dutch epidemiologists firstly considered the adequacy of acupuncture treatment in a series of systematic reviews, which suggested that a better understanding of adequate acupuncture treatment dose would support communication with patients and inform health policy (28,29). Despite with incomplete knowledge of acupuncture dosing, these studies have made more accessible to patients and stakeholders. In 2008, White and colleagues recommended that acupuncture systematic reviews should provide a subgroup analysis including only the studies that meet criteria for “adequate” acupuncture (12). Specifically, Ezzo and colleagues demonstrated that trails using ≥6 sessions of acupuncture for osteoarthritis were more likely to be positive than those using <6 sessions (30). However, the naïve subgroup analysis could not systematically exam dose-response relationship between variables and identify the existed cut-off value. Armour and colleagues investigated the relationship between components of acupuncture dose and menstrual pain (31). The results showed that there appear to be relationships between treatment timing and mode of needle stimulation, and menstrual pain outcomes; needle location, number of needles used and frequency of treatment showed dose-response relationships with menstrual pain outcomes. However, it was a narrative review without statistical model and quantitative analysis. Schwehr et al. surveyed the insurance system data belongs to National Health Interview Survey (NHIS), demonstrating that most people who use acupuncture did not receive a full treatment course, which may affect the effectiveness of treatment (32). Although a short course of acupuncture treatment may be sufficient and effective in some cases, especially for acute pain. However, for other cases such chronic pain and internal disorders with long-term conditions, which are likely to require a series of treatments for optimal results.

Implications for practice and research

Increasing patients suffering from CP/CPPS seek acupuncture for help. However, the association between acupuncture sessions and its effects remain unclear, which limited the application of acupuncture regardless available evidence found that acupuncture could relieve the symptoms of CP/CPPS effectively. Every treatment involves a therapeutic relationship, and similar to drugs, the dose of acupuncture therapy should also be considered in clinical practice.

Limitations

Several limitations should be noted. First, the quantity of included studies is small, only ten studies with 329 participants were used for data synthesis. Second, to ensure a better trend approximating on the dose-response relationship, it is expected to include more dose categories with each study. However, most of the studies reported two time points such as baseline and endpoint which may influence the robustness of the results. Third, although the acupuncture time in different studies was similar, the “dose” in this study specific to the treatment sessions, and the treatment was not adjusted in our results. Fourth, owing to the characteristics of “one-stage” approach model, currently there is no valid method to estimate the heterogeneity. Instead, we reported the “1-R²” of the model to quantitatively measure the heterogeneity, which may not precisely reflect the “true” heterogeneity.

Conclusions

The treatment session plays a vital role in acupuncture effects and prolonged acupuncture sessions (≥6 sessions) may have clinically relevant benefits for CP/CPPS. Further, large-scale, well-designed clinical studies reporting results from longitudinal data are strongly needed to explore the association between acupuncture treatment sessions and acupuncture effects.

Acknowledgements

Funding: This study has been supported by China Academy of Chinese Medical Sciences (grant No. ZZ10-012). The funder has no role in design of protocol, data collection, analysis, and interpretation.

Footnote

Conflicts of Interest: The authors have no conflicts of interest to declare.

References

1. Krieger JN, Nyberg L Jr, Nickel JC. NIH consensus definition and classification of prostatitis. JAMA 1999;282:236-7. [Crossref] [PubMed]
2. Clemens J. Male chronic pelvic pain syndrome: Prevalence, risk factors, treatment patterns, and socioeconomic impact. Curr Prostate Rep 2008;6:81-5. [Crossref]
3. Habermacher GM, Chason JT, Schaeffer AJ. Prostatitis/chronic pelvic pain syndrome. Annu Rev Med 2006;57:195-206. [Crossref] [PubMed]
4. Qin Z, Wu J, Tian J, et al. Network Meta-Analysis of the Efficacy of Acupuncture, Alpha-blockers and Antibiotics on Chronic Prostatitis/Chronic Pelvic Pain Syndrome. Sci Rep 2016;6:35737. [Crossref] [PubMed]
5. Qin Z, Wu J, Zhou J, et al. Systematic Review of Acupuncture for Chronic Prostatitis/Chronic Pelvic Pain Syndrome. Medicine (Baltimore) 2016;95:e3095. [Crossref] [PubMed]
6. Qin Z, Zang Z, Zhou K, et al. Acupuncture for Chronic Prostatitis/Chronic Pelvic Pain Syndrome: A Randomized, Sham Acupuncture Controlled Trial. J Urol 2018;200:815-22. [Crossref] [PubMed]
7. Engeler D, Baranowski AP, Borovicka J, et al. EAU Guidelines on Chronic Pelvic Pain. 2017. Available online: http://uroweb.org/wp-content/uploads/23-Chronic-Pelvic-Pain_2017_web.pdf
8. Lee SW, Liong ML, Yuen KH, et al. Acupuncture versus sham acupuncture for chronic prostatitis/chronic pelvic pain. Am J Med 2008;121:79.e1-7. [Crossref] [PubMed]
9. Liu BP, Wang YT, Chen SD. Effect of acupuncture on clinical symptoms and laboratory indicators for chronic prostatitis/chronic pelvic pain syndrome: a systematic review and meta-analysis. Int Urol Nephrol 2016;48:1977-91. [Crossref] [PubMed]
10. Posadzki P, Zhang J, Lee MS, et al. Acupuncture for chronic nonbacterial prostatitis/chronic pelvic pain syndrome: a systematic review. J Androl 2012;33:15-21. [Crossref] [PubMed]
11. White A, Cummings M, Barlas P, et al. Defining an adequate dose of acupuncture using a neurophysiological approach--a narrative review of the literature. Acupunct Med 2008;26:111-20. [Crossref] [PubMed]
12. Xu C, Liu Y, Jia PL, et al. The methodological quality of dose-response meta-analyses needed substantial improvement: A cross-sectional survey and proposed recommendations. J Clin Epidemiol 2019;107:1-11. [Crossref] [PubMed]
13. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. J Clin Epidemiol 2009;62:e1-34. [Crossref] [PubMed]
14. Litwin MS, McNaughton-Collins M, Fowler FJ Jr, et al. The National Institutes of Health chronic prostatitis symptom index: development and validation of a new outcome measure. Chronic Prostatitis Collaborative Research Network. J Urol 1999;162:369-75. [Crossref] [PubMed]
15. Moga C, Guo B, Schopflocher D, et al. Development of a quality appraisal tool for case series studies using a modified Delphi technique. Edmonton: Institute of Health Economics, 2012. Available online: http://www.ihe.ca/advanced-search/development-of-a-quality-appraisal-tool-for-case-series-studies-using-a-modified-delphi-technique
16. Guo B, Moga C, Harstall C, et al. A principal component analysis is conducted for a case series quality appraisal checklist. J Clin Epidemiol 2016;69:199-207.e2. [Crossref] [PubMed]
17. Xu C, Doi SAR. The robust error meta-regression method for dose-response meta-analysis. Int J Evid Based Healthc 2018;16:138-44. [PubMed]
18. Wold S. Spline functions in data analysis. Technometrics 1974;16:1-11. [Crossref]
19. Capodice JL, Jin Z, Bemis DL, et al. A pilot study on acupuncture for lower urinary tract symptoms related to chronic prostatitis/chronic pelvic pain. Chin Med 2007;2:1. [Crossref] [PubMed]
20. Chen RC, Nickel JC. Acupuncture for chronic prostatitis/chronic pelvic pain syndrome. Curr Urol Rep 2004;5:305-8. [Crossref] [PubMed]
21. Honjo H, Kamoi K, Naya Y, et al. Effects of acupuncture for chronic pelvic pain syndrome with intrapelvic venous congestion: preliminary results. Int J Urol 2004;11:607-12. [Crossref] [PubMed]
22. Küçük EV, Suceken FY, Bindayi A, et al. Effectiveness of acupuncture on chronic prostatitis-chronic pelvic pain syndrome category IIIB patients: a prospective, randomized, nonblinded, clinical trial. Urology 2015;85:636-40. [Crossref] [PubMed]
23. Lee SH, Lee BC. Electroacupuncture relieves pain in men with chronic prostatitis/chronic pelvic pain syndrome: three-arm randomized trial. Urology 2009;73:1036-41. [Crossref] [PubMed]
24. Sahin S, Bicer M, Eren GA, et al. Acupuncture relieves symptoms in chronic prostatitis/chronic pelvic pain syndrome: a randomized, sham-controlled trial. Prostate Cancer Prostatic Dis 2015;18:249-54. [Crossref] [PubMed]
25. Tugcu V, Tas S, Eren G, et al. Effectiveness of acupuncture in patients with category IIIB chronic pelvic pain syndrome: a report of 97 patients. Pain Med 2010;11:518-23. [Crossref] [PubMed]
26. Zhou M, Yang M, Chen L, et al. The effectiveness of long-needle acupuncture at acupoints BL30 and BL35 for CP/CPPS: a randomized controlled pilot study. BMC Complement Altern Med 2017;17:263. [Crossref] [PubMed]
27. Propert KJ, Litwin MS, Wang Y, et al. Responsiveness of the National Institutes of Health Chronic Prostatitis Symptom Index (NIH-CPSI). Qual Life Res 2006;15:299-305. [Crossref] [PubMed]
28. Kleijnen J, ter Riet G, Knipschild P. Acupuncture and asthma: a review of controlled trials. Thorax 1991;46:799-802. [Crossref] [PubMed]
29. ter Riet G, Kleijnen J, Knipschild P. Acupuncture and chronic pain: a criteria-based meta-analysis. J Clin Epidemiol 1990;43:1191-9. [Crossref] [PubMed]
30. Ezzo J, Berman B, Hadhazy VA, et al. Is acupuncture effective for the treatment of chronic pain? A systematic review. Pain 2000;86:217-25. [Crossref] [PubMed]
31. Armour M, Smith CA. Treating primary dysmenorrhoea with acupuncture: a narrative review of the relationship between acupuncture 'dose' and menstrual pain outcomes. Acupunct Med 2016;34:416-24. [Crossref] [PubMed]
32. Schwehr NA, Shippee ND, Johnson PJ. Acupuncture 'dose' (number of treatments) and insurance benefits in the USA. Acupunct Med 2018;36:88-95. [Crossref] [PubMed]