Interventions for Treating Tennis Elbow: An Overview of Systematic Reviews in the Past Ten Years

purpose: Tennis elbow is a common orthopedic disease, and there are many ways to treat it. This overview aimed to summarize the evidence of different treatments for tennis elbows, so as to provide the best guidance for clinical treatment. Methods: Use computer to search CNKI, WanFang database, WeiPu database, CBM database, PubMed, Cochrane Library and Embase from the time of establishment to May 31, 2019.Te Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) and latest Assessment of Multiple Systematic Reviews 2 (AMSTAR2) checklists were used to assess reporting characteristics and methodological quality, respectively. Results: A total of 37 references were included. Methodological quality and reporting quality were unsatisfactory. Methodological quality was generally low and many key items were not reported. Some research reports are of high quality, but there is no trial registration and protocol written in advance, which may lead to some bias in the research process. The most frequent problems included non-registration of study protocol, absence of a list of excluded studies, and unclear acknowledgment of conicts of interests. The different types of interventions included have been shown to relieve pain, improve quality of life, and restore elbow function, but there has been a lack of comparative studies. Conclusion: The reporting and methodological quality of systematic reviews and meta-analysis studies were sub-optimal, which demands further improvement. Comparative studies of different types of interventions are needed to determine unclear. they can alleviate pain,restore elbow joint function and improve quality of life to varying degrees.

and exclusion criteria formulated by this study, two researchers(YJ.YT) independently read the titles and abstracts of the literature found and conducted the preliminary screening, excluding the studies that obviously did not meet the inclusion criteria, and then read the full text to select the studies that met the standards for data extraction.If the same study is reported by different literature for many time, the most detailed was selected to be included in the study.Data extraction contents included author, year, publishing unit, the literature evaluation tool, literature type, sample size, intervention measures, included a research type, outcome index, etc.Both researchers independently screened the literature and extracted data, and then cross-checked them.If there were any inconsistent studies, the third researcher decided whether or not to include them or not after discussion. 2.6.SRs Methodological Quality Assessment. Methodological quality evaluation is used to evaluate internal authenticity, while reporting quality evaluation is used to evaluate external authenticity.AMSTAR2 [17](A Measure Tool to Assess Systematic Reviews 2) a methodological quality assessment tool for SRs.AMSTAR2 is updated on the basis of AMSTAR, which is applicable not only to the systematic review of randomized controlled studies, but also to the systematic review of non-randomized intervention studies.AMSTAR2 has 16 entries, with 7 key entries as a recommendation level(Table1)."Yes", "no" and "partially yes", according to the compliance. A general rule was applied for rating overall con dence in the results of review: SR with no or one noncritical weakness was rated as high;with more than one noncritical weakness was rated as moderate;with one critical aw with or without noncritical weaknesses was rated as low;with more than one critical aw with or without noncritical weaknesses was rated as critically low. In addition, PRISMA [18] was a reporting quality assessment tool for evaluating SRs, with 27 entries.Rated as "yes" and "no" according to whether they are in conformity or not. All articles that meet the requirements were evaluated by two reviewers using the above two assessment tools, and the two reviewers discussed and decide upon the differences. A total of 234 articles were found in the literature retrieval.And 73 were repeated.After reading the questions and abstracts, one article was published repeatedly, 67 of which did not meet the inclusion and exclusion criteria, and 47 of which were published 10 years ago.After looking at the full text, there were 6 conference papers, 2 without full text, and 1 article with incomplete data.37 SR papers were listed in this overview( Figure 1).
3.2Description of the included reviews.
The main characteristics of the included SRs were presented in table 2.Among the included studies, there were 15 in China and 22 in English, among which 2 in China were actually the same study.The one with the most complete description was chosen for evaluation.Twenty-nine of the studies included were randomized controlled trials, and 21 were assessed for bias using Cochrane's risk bias assessment tool.Seven was assessed using the Jadad scale, and three were assessed using the PEDro score.Ten literature was reviewed systematically, and 27 were meta-analyzed.Interventions, 12 articles used acupuncture, 4 using the shock wave therapy, 14 with non-surgical therapy, including platelet rich plasma, botulinum toxin injection, autologous blood injection, corticosteroid injections, etc.3 article adopting physical therapy, 4 by operation therapy.See Table 2 for full details.
3.3.Assessment of methodological quality and quality of evidence 3.3.1.Methodological quality Table 3 shows the methodological quality assessment results of the included reviews.According to the latest AMSTAR2 evaluation criteria, since all SRs had more than one critical weakness (Items 2, 4, 7, 9, 11, 13, and 15), their quality was considered as critically low.
They all employed the PICO approach (population, intervention, control group, and outcome) as an organizing framework for establishing study questions.Most of the included studies adopted the elementary rule of PICO, including 5 non-random, cohort and case-control studies, and the rest were included in RCT.Only 1 article [54] provides protocol registration, and the rest do not provide protocol registration or publication before commencement of the review (AMSTAR2 Item 2).Two SRs [38,44] described the reasons for the type of included study, while none of the others described the reasons for the type of included study(AMSTAR2 Item 3).Only one review [48] had conducted a comprehensive literature search.One review [53] only retrieves one database, which is not inclusive enough(AMSTAR2 Item 4).In two reviews [41,44] authors had not performed study selection and data extraction in duplicate.None of the reviews provided a complete list of potential related studies and reasons for excluding each(AMSTAR2 Item 7).Thirty seven reviews partly provided characteristic information of their included studies.Three SRs [35,41,53] did not evaluate the risk of the included studies, while the rest of them evaluated the quality of the included literature with different scales.Three SRs [23,32,54] had reported funding sources for the included studies and whether there is a con ict of interest between the included studies, while the rest are not reported.Twelve reviews [22, 27-28, 29-30, 38-39, 41, 43, 45-47] applied meta-analytical methods appropriately, explaining factors for xed or random effects model selection and methods used for heterogeneity investigation.Eleven reviews [32,35,37,42,44,50,[52][53][54][55] were only described qualitatively without meta-analysis (AMSTAR2 Item 11).Two reviews [29,51] described the potential impact of the quality of the included studies on the results of evidence synthesis, while the rest were not described.Nineteen reviews [19, 23, 25-26, 28, 30-31, 33, 34, 36, 38-40, 42-43, 45-52, 54] had discussed the impact of quality assessment in results interpretation of the review(AMSTAR2 Item 13).Seventeen reviews [22, 24, 27-28, 35-36, 38-39, 41, 43, 45-47, 51, 53-55] had explored possibilities of heterogeneity and discussed its in uence on the results conclusions and clinical recommendations.Twelve reviews [19-20, 22-23, 25-28, 33, 39, 48-49] only investigated publication bias, but did not describe the causes of publication bias and its impact on results(AMSTAR2 Item 15).Twelve reviews [20, 23, 25-26, 28, 30-31, 37-39, 41, 52] didn't describe funding sources or con icts of interest.

Reporting Quality.
The results of the assessment on the reporting of included reviews were presented in Table 4 and Figure 2.Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) were applied to assess report quality of included SRs.In all the review topics, it was clearly stated that they were systematic evaluation, meta-analysis, or both.All of them adopted the abstract of structural formula, described the disease burden and clinical symptoms, and proposed clear research problems to be resolved.One review [54] reported the trial registration number, and 48.75% provided a complete retrieval method for at least one database.And 86.94% reported risk bias assessments.8.11% reported whether there were other methods for meta-analysis, 91.89% reported the literature ow chart, 70.27% summarized the main ndings of the study, 75.68% discussed the limitations and risks of bias in the outcome level, and 72.97% reported the sources of funding and the role of funder in completing the systematic review.

3.4.1.VAS
Twenty-ve SRs included VAS as an outcome measure.Of these SRs were conducted meta-analyses.The intervention methods are:Seven SRs using acupuncture and moxibustion treatment of tennis elbow, four SRs with shock wave therapy, two piece of SRs used botulinum toxin injection treatment, one SR used autologous blood treatment, one SR with Platelet rich plasma therapy treatment with 1 piece of corticosteroid injection, 2 SRs adopted physical therapy, 1 piece of SR with Saline Injections of treatment, For the SR, low-level laser therapy (LLLT) was used in 1 paper, Percutaneous Needle Tenotomy was used in 1 paper, non-surgical therapy was used in 1 paper, and surgical therapy was used in 3 plates.All the results showed that it could effectively relieve pain.

Effectiveness of different interventions
Eight SRs(89 RCTs, 6992 participants) included effective rate as an outcome measure.The intervention measures include re needle, warm acupuncture, oating needle, small needle knife, massage, acupuncture, extra corporeal shock wave therapy, etc.Most of them are acupuncture and moxibustion, and the results all show that acupuncture and moxibustion can improve the effective rate of treating tennis elbow and reduce the adverse reactions.

Adverse Events.
Only ve studies reported adverse events, The intervention measures with small needle knife, acupuncture, botulinum toxin injections, autologous blood injection, platelet rich plasma, etc., adverse reactions are fainting during acupuncture treatment, extensor weakness, pain at the injection site,and paresthesia.

Discussion
Summary of Main Findings.
Overviews is a comprehensive research method that comprehensively collects systematic reviews related to the treatment,etiology,diagnosis and prognosis of the same disease or health problem for reevaluation [56] .This overview has provided a summary of the effects on pain improvement and functional relief among patients with tennis elbow treated bydifferent interventions in thirty-seven eligible SRs.This overview included RCT,cohort studies,and Case series.This overview included Twenty-ve meta-analyses and twelve narrative SRs,The majority of SRs were considered as relatively high reporting quality and critically low methodological quality,by using the PRISMA and AMSTAR2 tools,respectively.

4.2.Suggestions for Better Methodological and Reporting Quality.
In this overview,all 37 SRs employed PICO approach in organizing research question and describing inclusion criteria but only one review has provided a documented protocol or register information.It is noted that obtaining an open register of a SR in advance is quite essential for conducting a SR.It can help facilitate processing transparency and avoiding post hoc decision bias in methodology [57] ,The register item has also been recorded in both the PRISMA and AMSTAR2 checklists,indicating its great importance.In summary,system reviewers should register in advance before performing a system review.
Most systematic reviews only included randomized controlled trials and used the cochrane risk bias tool for bias assessment,with a small number of unreported tools for bias assessment.A small number of systematic reviews included cohort studies and case series observations,but no corresponding scales were used for bias assessment,which may affect the authenticity and reliability of the results due to the low quality of the included studies.Based onAMSTAR2 users guide,The author of the system review should also report whether other types of analysis have been carried out during the meta-analysis,so as not to cause selective reporting results and affect reliability.The authors should also report on the source of funding for the included study,and if it is corporate funding,there may be an interest relationship that affects the reliability of the results.
In the systematic review,at least two databases should be searched,and one of the included studies only searched one database,which had great methodological problems.Most studies do not provide search strategies,and detailed search strategies are conducive to the repeated implementation of the study.Most reviews processed the study selection and data extraction in duplicate, to decide by discussion andintroduction of a third author while the remaining two did not mention this in the article.A few studies assessed publication bias,but did not elaborate on the causes of publication bias.PRISMA guideline is mainly responsible for reporting quality assessment.Most literature reports are of poor quality,Common items with low scores included structured summary,protocol and registration,search strategy,and funding information.It is hoped that in the future systematic reviewers can learn PRISMA and AMSTAR2 in advance when conducting systematic reviewings,which can reduce many methodological and reporting errors and improve the quality of research.

Strengths
Firstly,it is the latest overview on Different interventions for the treatment of tennis elbow,which can provide new evidence reference for clinical practice.Based on the current results and conclusion of high quality SR,the overview suggested that there are many clinical measures to treat tennis elbow,including acupuncture therapy,drug injection,surgical therapy,physical therapy,etc.,all of which can relieve pain to different degrees.However,acupuncture treatment of tennis elbow has good effect,less side effects and stable long-term effect.Although drug injection has a good short-term effect,it is easy to relapse and has a long course of treatment.Physical therapy has the advantages of good curative effect,low cost and few complications.Although the effect of surgical treatment is relatively fast,it is prone to complications,which could be useful for decision-making for tennis elbow treatment in clinic.Secondly,AMSTAR2 and PRISMA,the latest methodological quality assessment tools,are combined to provide a new direction for how to evaluate and improve the quality of systematic reviews.

Limitations
This overview has several potential limitations.First,The types of studies included in this review include not only RCT,but also cohort studies and case series observations,and the risk bias assessment tools used by some systematic reviewers were inappropriate.Second, there are too many types of interventions, only descriptive analysis was performed, and no comprehensivequantitative analysis was performed.Third,The quality of the research methodology included is not high and the quality of the evidence of the results is not satisfactory.Fourth,Most research and research processes were incompletely reported,the quality of reports was low,and few documents were implemented in accordance with PRISMA's reporting speci cations.Finally,There are too many interventions and relatively few literatures comparing different types of interventions, so it is necessary to further study the comparison between different types of interventions.

4.5.Opportunities for Future Research
By analyzing and pointing out these insu ciencies in these published systematic reviews,We found that the most common problems were that no trial registration was conducted,no plan was written in advance,no inclusion exclusion list was provided,and no funding source for the included study was reported.Bias and publication bias may be selected during the trial,resulting in poor study quality.It is proposed that the study should be implemented in strict accordance with AMSTAR2 and PRISMA in the future.

Conclusion
In recent years,there have been more and more treatments for the treatment of tennis elbow,but most of them are of low quality and cannot provide valid evidence.Because there are too many research methods,there are not many related studies on the comparison between different types,and it is impossible to draw conclusions on which treatment measures are more effective,but they can alleviate pain,restore elbow joint function and improve quality of life to varying degrees.   Table 3 Methodological quality of the included reviews assessed by AMSTAR2 Author  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15 16 Ranking of quality Hu Qiulan [19] Hu Lan [20] Chen Yan [28] Hao ZhengAn [30][31] Sims,S.E.G. [32] Table 4 Reporting quality of the included reviews assessed by PRISMA Author  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25 Hu Figure 1 Literature retrieval ow chart