This systematic evaluation and meta-analysis was designed and implemented based on the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) guideline (24) and has been registered with Prospero (CRD42023403594).

The inclusion and exclusion criteria for literature screening were predetermined to enable more rigorous literature screening. Based on the PICOS principles of the systematic review, the inclusion criteria for this study included: (1) Patients clinically diagnosed with upper limb tendonitis (include rotator cuff tendonitis, lateral epicondylitis, trigger finger and so on; (2) the experimental group received ESWT or RESWT; (3) the control group received placebo treatment (sham treatment); (4) the outcome indicators of functional activity and analgesic effects were assessed by relevant scales, such as a visual analog scale (VAS), the Constant Murley Scale (CMS) and grip strength (GS); and (5) the experimental design was a RCT.

The exclusion criteria were as follows: (1) retrospective studies, animal studies, single-case reports, protocols, reviews, meta-analyses, poster presentations, or conference abstracts; (2) interventions other than placebo or sham; (3) full-text content not available; (4) missing or duplicated experimental data; and (5) non-English literature.

Searches were performed in the PubMed, Embase, Web of Science, Medline, and Cochrane Library databases from the initial availability date to December 2023 to identify studies for inclusion in the quantitative analysis. The main search terms for this study were “tendonitis,” “rotator cuff tendonitis,” “lateral epicondylitis,” “trigger finger,” “upper limb tendonitis,” and “extracorporeal shock wave.” In addition, we manually searched for other relevant literature, such as studies included in some systematic reviews and meta-analyses, to broaden the search for eligible studies. For instance, the following search strategy was used for PubMed: lateral epicondylitis OR tennis elbow OR rotator cuff tendonitis OR trigger finger OR upper limb tendonitis OR tendinopathy [MeSH terms] AND extracorporeal shockwave therapy [MeSH terms] OR shock wave therapy OR radial extracorporeal shock wave therapies OR HIFU therapy OR extracorporeal high intensity focused ultrasound therapy OR ESWT OR REWST. All the relevant words were examined in advance using the PubMed Subject Glossary. A similar search strategy was used for the other databases.

All studies were imported into the document management system (Endnote x20), and the management function was used to remove duplicate content. Two researchers then read the title and abstract of each study and screened the studies according to the predetermined criteria. For further screening, two researchers downloaded the full texts and read them, removing studies that did not meet our final inclusion and exclusion criteria. If there is a disagreement about the screening process for a particular study, a consensus will be reached through advice provided by the principal investigator.

Two reviewers independently extracted data and items from the literature. During the information extraction process, the two evaluators obtained complete test data by contacting the original authors when they encountered unclear or missing information. If a response was not received after three consecutive emails, the study was defined as having missing data.

The following data were extracted from all eligible studies: (1) General study information: first author of the study, country, year of publication, sample size, age of patients, sex, disease type, and disease duration; (2) Study characteristics: study design, setting, and inclusion and exclusion criteria; (3) Specific therapeutic parameters: pulse count, energy, and sessions; (4) The outcome measures were pain and function scores reported at baseline and 3 and 6 months. When 3-month data were not available, we used the closest data points from 1 to 3 months of follow-up; and (5) Adverse events.

Two researchers independently used the Cochrane Risk of Bias tool (Review Manager 5.40) to assess the quality of risk of bias in the included studies (including the risk of bias graph and risk of bias summary). The assessment included blinding of participants and personnel, outcome assessments, allocation concealment, random sequence generation, incomplete outcome data, selective reporting, and other biases. Risk of bias was graded as high, low, or unclear (25).

The quality of evidence for the outcome indicators was assessed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) system. The assessments included limitations, intermittencies, inconsistencies, and imprecision (26). Evidence levels for each element were categorized as “high,” “moderate,” “low,” or “very low,” with the final strength of recommendation as “strong” and “weak” (27).

The data extracted from the included studies were analyzed for statistical and analytical purposes. Heterogeneity between studies was statistically analyzed using RevMan 5.40. The size of heterogeneity was expressed as I²; I² values of 25%, 50%, and 75% represented no significant heterogeneity, moderate heterogeneity, and large heterogeneity in the combined results, respectively (28). Subgroup and sensitivity analyses were performed when moderate heterogeneity was observed in combined results. Moreover, when I² ≥ 50%, a random-effects model was used, and when I² < 50%, a fixed-effects model was used (29). Subgroup and sensitivity analyses were performed when moderate heterogeneity was observed in the combined results, subgroup analysis and sensitivity analysis were performed (28, 30). The Egger, Begg, and funnel plot methods were also used to test for publication bias. The continuity outcomes were calculated and expressed as 95% confidence intervals (CIs) and mean differences (MDs) to express the effect size.

The outcome indicators of function and pain were assessed using relevant scales such as the VAS, CMS, and GS.

Five databases were searched, and the initial search yielded 1,680 studies. After removing duplicate studies using a literature management software, the results left 1,180 studies. After removing duplicates and selections based on titles and abstracts by the two reviewers, 296 studies were analyzed. After reviewing the full text, 278 studies were excluded, including 20 reviews, six case reports, 15 study protocols, 20 meeting abstracts, 130 non-compliant intervention studies, 80 non-RCTs, and seven studies for which data could not be retrieved. Ultimately, 18 eligible studies were included (Figure 1).

Basic information and intervention parameters of the 18 RCTs were extracted and are summarized in Table 1. There were eight studies about rotator cuff tendonitis, eight studies about lateral epicondylitis, one study about long bicipital tendonitis, and one study about finger tendonitis. A total of 1,351 patients eligible for the studies were included, with 740 patients treated with ESWT and an additional 611 receiving only placebo as controls.

According to the simplest classification about the energy levels of ESWT, with low-energy ESWT having an energy flux density (EFD) of < 0.12 mj/mm² and high energy an EFD between 0.12 and 0.38 mj/mm² (31). Accurate EFD was reported in 13 out of 18 experiments, with five studies for high energy (0.12–0.32 mj/mm²⁾ and 10 studies for low energy (0.01 to 0.12 mj/mm; two of the studies had both high and low-energy groups as well as a placebo group).

The 18 included RCTs were assessed using the Cochrane Risk of Bias Assessment Tool (RevMan 5.40) for risk of bias. Three studies had incomplete outcome data, mainly because of a large or unbalanced number of missing people between the groups, and the deficiency of data significantly affected the effect values. In one study, the diagnoses of the included patients were made based on history and physical examination. This could have led to selection bias. Other studies have shown a low or unclear risk in all risk–bias assessments. Overall, the included RCTs had a low risk of bias (Figures 2, 3). We assessed the level of evidence for the main outcome indicators (VAS, CMS, and GS) of the included studies using GRADE. The results of the Egger and Begg tests showed no publication bias in the included studies, according to the three outcome indicators (P > 0.05). The quality of evidence for the VAS was moderate, owing to high heterogeneity (I² > 80%). Moreover, the quality of evidence for GS was assessed as moderate owing to the smaller sample size and wider CIs. Overall, the GRADE recommendation rating was “strong” for the three outcome indicators (Table 2).

Six studies reported CMS scores that were used to assess the function intensity involving rotator cuff tendonitis (17, 32–36). In these studies, 352 participants participated in ESWT trials, divided into experimental and control groups. Data analysis of the forest plot showed that the experimental group with ESWT as an intervention had a significant improvement in CMS scores in patients with rotator cuff tendonitis at the 3-month follow-up [MD = 7.56, 95% CI: (3.69, 11.43), I² = 37%, P < 0.001; Figure 9]. Subgroup analysis based on the presence of calcification showed that ESWT was effective in improving functional activity in patients with calcific and non-calcific tendonitis at the 3-month follow-up [calcification, MD = 5.76, 95% CI: (0.92, 10.61); I² = 23%, P = 0.02; non- calcification, MD = 12.20, 95% CI: (5.21, 19.18), I² = 48%, P < 0.001; Figure 10]. Additional subgroup analysis based on the EFD of ESWT showed that ESWT was effective in improving functional activity in patients with rotator cuff tendonitis and suggested that the effect was dose dependent [high, MD = 18.47, 95% CI: (13.71, 23.23); I² = 3%, P < 0.001; low, MD = 7.56, 95% CI: (3.69, 11.43), I² = 37%, P < 0.001; test for subgroup differences, MD = 11.91, 95% CI: (8.91, 14.91), I² = 91.8%, P < 0.001; Figure 11].

Four studies reported the GS scores that were used to assess power intensity involving lateral epicondylitis (38, 39, 42, 43). In these studies, 227 participants participated in ESWT trials, divided into experimental and control groups. Data analysis of the forest plot showed that the experimental group with ESWT as an intervention had a non-significant improvement in GS scores in patients with lateral epicondylitis at the 3-month follow-up [MD = 2.38, 95% CI: (−0.43, 5.20), I² = 0%, P = 0.10; Figure 12]. The energy density parameter of ESWT was not mentioned in any of the four studies included in this review. In addition, the evidence level for this outcome indicator was moderate according to the GRADE classification. These findings suggest that additional clinical trials are required to verify the accuracy of our conclusions.

No serious long-term adverse effects were reported in any of the included studies. Some studies reported a temporary increase in pain and local reactions such as swelling, erythema, petechiae, or small hematomas, but these symptoms disappeared at the end of treatment (33, 34, 37, 39, 44). Only a small percentage of the patients could not tolerate the treatment, had worsening symptoms, or withdrew from the trial (38, 39, 48). Some studies have emphasized the use of local anesthesia or oral analgesia during treatment, which can only be used as a factor to increase heterogeneity without affecting the outcome (33, 36, 44).