Extracorporeal shockwave therapy in the management of knee osteoarthritis: A systematic review of dose-response meta-analysis

Abstract

Introduction

The aim of the research was to conduct a systematic review of dose-response meta-analysis, examining the correlation between the energy of extracorporeal shockwave therapy (ESWT) and clinical outcomes for patients with knee osteoarthritis (OA).

Methods

We conducted a systematic review of three online databases - PubMed, Embase, and Cochrane Library - to collect relevant articles from their inception to May 2023. We included the articles that investigated the efficacy of ESWT in treating knee OA and were designed using randomized controlled trials (RCTs). The main outcomes were measured using the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) for primary outcomes and the Visual Analog Scale (VAS) for secondary outcomes. We conducted a dose-response meta-analysis to examine the correlation between the energy of ESWT and clinical outcomes.

Results

Our study included a total of 7 RCTs (comprising 450 subjects) that met the inclusion criteria. We found a negative linear relationship between the doses of ESWT and WOMAC scores at the 1-week (p = 0.0398) and 4-week (p = 0.001) follow-up periods, but not at the 12-week follow-up period (p = 0.202). Furthermore, at the 4-week (p = 0.0477) and 12-week (p < 0.001) follow-up periods, a negative linear connection was found between the ESWT dose and VAS scores, but not at the 1-week follow-up period (p = 0.2268).

Conclusions

Our results demonstrate a dose-response connection between the energy of ESWT and clinical outcomes. This suggests that using higher energy levels of ESWT to treat knee OA could lead to greater improvement in clinical outcomes compared to using lower energy levels.

1. Introduction

Knee osteoarthritis (OA) is a common disease, particularly among the elderly, that can cause functional disability due to impaired joint function and pain. Approximately 50% of adults between 45 and 85 years old have experienced knee OA,¹ with women being twice as likely to have the disease as men.² Joint replacement surgery is the ultimate therapy for end-stage knee OA. However, several therapies have been introduced to treat early-stage knee OA and prevent joint replacement surgery, such as hyaluronic acid injection,³ platelet-rich plasma injection,⁴ radiofrequency ablation,⁵ extracorporeal shockwave therapy (ESWT),⁶ nonsteroidal anti-inflammatory drugs,⁷ and kinesio taping.⁸

Several meta-analyses have performed the effect of ESWT in treating knee osteoarthritis (OA).9, 10, 11 Two of these reports found that ESWT significantly improved pain and physical function, with effects lasting up to 6–12 months.^10,11 Another meta-analysis reported that the ESWT group had better pain relief and physical function than the physical therapy group at the 1-month follow-up visit.⁹ Minor complications were the only observed adverse events after receiving ESWT for knee OA.¹¹ Although ESWT is an effective therapy for knee OA based on current meta-analyses, statistical heterogeneity was detected in most of these analyses.9, 10, 11

To achieve maximum efficacy of ESWT for knee OA, different energies of ESWT have been used.9, 10, 11 As far as I know, only one randomized controlled trial (RCT) has compared the efficacy of two energies of ESWT for treating knee OA.¹² The results demonstrated that high energy led to obvious improvement in pain and physical function for knee OA compared to low energy.¹² This suggests that the energy of ESWT is positively related to improvement of knee OA. It is clinically important to use an optimal energy of ESWT for treating knee OA. However, the relationship between the energy of ESWT and its efficacy for knee OA is still unclear.

The objective of this study was to evaluate the relationship between the efficacy and energy of ESWT in treating knee OA using dose-response meta-analysis. The findings may offer medical professionals’ useful guidance on selecting the ideal energy level for ESWT in the treatment of knee OA.

2. Materials and methods

2.1. Literature search

The PRISMA standards were followed in the conduct of this study, and it was registered under the registration number CRD42023427568 on the International PROSPERO. We searched three electronic databases, including PubMed, Embase, and Cochrane Library, for articles related to the efficacy of ESWT for treating knee OA from inception to May 2023.

The search terms used were as follows: (“shockwave” OR “shock wave” OR “Extracorporeal Shockwave Therapy [MeSH Term]") AND “knee” AND (“osteoarthritis” OR “arthritis” OR “arthritic” OR “Osteoarthritis [MeSH Term]"). Two authors (H.S.C. and C.L.S.) independently searched for related articles.

Firstly, we identified articles from the three databases, and removed duplicates using EndNote software. Then, we screened the remaining articles by title/abstract, and retained possibly relevant articles. Finally, we screened these articles by full-text, and determined relevant articles. The two authors had discussions to settle any differences and come to a conclusion. Additionally, we manually screened the reference lists of the included articles for possible relevant articles.

2.2. Inclusion and exclusion criteria

The study's included articles fulfilled the subsequent requirements: (1) patients with primary knee OA; (2) patients who were treated with ESWT; (3) a control group that received sham ESWT or a lower energy of ESWT; (4) an RCT study design; and (5) clinical outcomes related to pain and physical function. We excluded articles that did not meet the above criteria, including: (1) non-original articles such as notes, letters, comments, conference abstracts, and reviews; (2) articles that were not published in English; and (3) animal studies.

2.3. Data extraction

The following major characteristics were extracted for the systematic review: initial author, publication year, patient age on average, sample size, sex ratio, therapy, process of treatment, type of clinical outcomes, and follow-up visits. These data were extracted independently by the two authors, and any disagreements were discussed until consensus was achieved. If the sham ESWT group did not provide a value for the energy flux index, a value of zero was adopted for that group.

The WOMAC questionnaire is a self-administered tool widely used to assess clinical outcomes of knee OA. It has 24 items total, broken down into 3 subscales: physical function, stiffness, and pain. Several clinical outcomes have been adopted to assess the effect of therapies for knee OA. The VAS is a subjective measure used to assess pain levels. It uses a 10-cm line, with 0 and 10 cm representing “no pain” and “worst pain,” respectively.¹³ Therefore, WOMAC is considered the primary outcome, and VAS is secondary. If the article presents mean and standard deviation values as figures, these values are measured from those figures.

2.4. Quality assessment

The quality of each article was assessed independently by two authors according to the Cochrane Collaboration's criteria for systematic reviews. Three levels of risk were used to evaluate quality: high risk, unclear risk, and low risk. Any disagreements were discussed by the authors until a consensus was reached. Review Manager Version 5.41 (Cochrane Collaboration, Oxford, England) was used to perform the quality assessment.

2.5. Statistical analysis

A two-stage random-effects meta-analysis was conducted using a linear model to investigate the relationship between the dose of ESWT and clinical outcomes (VAS or WOMAC). The R package “dosresmeta” was used to perform the dose-response meta-analysis. The I² statistic was used to assess statistical heterogeneity among articles. I² values less than 50% indicate no heterogeneity, while values greater than or equal to 50% indicate heterogeneity. The model was tested using the Wald test, and a p-value <0.05 indicates statistically significant. R software (version 4.1.0, http://rproject.org/) was used to conduct all statistical analyses.

3. Results

3.1. Literature search

First, 276 papers in all were found across three databases. After removing duplicates, 182 articles remained and were screened by analyzing their titles and abstracts. Subsequently, 18 articles were selected for full-text analysis (Fig. 1). Out of these, 7 articles met the inclusion criteria, while the remaining 11 articles were excluded for reasons such as not having a reference dose of ESWT, being combined with other therapies, not being ESWT, or other reasons.

Fig. 1.

Progress of systematic review for identifying related articles.

3.2. Major characteristics of included articles

Table 1 records the main characteristics of the articles analyzed, which were published from 2013 to 2021. Seven RCTs in total were included, comprising 450 patients with a mean age ranging from 49.7 to 75.5 years. The energy flux density of ESWT varied between 0 and 0.25 (mJ/mm²), and the follow-up periods ranged from 1 week to 12 weeks. All RCTs adopted sham ESWT as a control group, with the exception of one RCT which used lower energy as the control.¹²

Table 1.

Major characteristics of included articles.

Author	Group	Patients	K-L	Sample size	Mean age	Sex ratio (M/F)	Shock number	Energy flux density (mJ/mm2)	Treatment duration	Clinical outcomes	Follow-up periods
Zhao 2013	ESWT	knee OA	II ∼ III	34	59.9 ± 11.3	14/20	4000	0.25	once per week for 4 weeks	VAS and WOMAC	1, 4, and 12 weeks
	sham ESWT			36	61.8 ± 9.8	11/25	0	0
Kim 2015	low-energy ESTW	knee OA	II ∼ III	30	65.1 ± 6.3	26/4	1000	0.04	once a week for 3 weeks	VAS and WOMAC	1, 4, and 12 weeks
	high-energy ESTW			30	63.5 ± 5.4	27/3	1000	0.093
Cho 2016	ESWT	chronic stroke patients with knee OA	2.0 ± 1.1	9	75.5 ± 7.7	8/1	1000	0.05	once a week for 6 weeks	VAS	1 week
	sham ESWT		1.8 ± 1.1	9	72.7 ± 5.9	7/2	1000	0
Imamura 2017	ESWT	knee OA	II ∼ VI	52	70.0 ± 6.5	NA	2000	0.1–0.16	once a week for three weeks	VAS and WOMAC	1 and 12 weeks
	sham ESWT			53	72.4 ± 6.5	NA	2000	0
Zhong 2019	ESWT	knee OA	II ∼ III	32	62.5 ± 8.2	11/21	2000	0.105	once per week for 4 weeks	VAS and WOMAC	5 and 12 weeks
	sham ESWT			31	63.2 ± 7.7	11/19	2000	0
Hammam 2020	low-energy ESWT with strengthening exercise	knee OA	II	15	50.4 ± 3.4	6/9	2000	0.02	once per week for 4 weeks	VAS	NA
	median-energy ESWT with strengthening exercise			15	49.9 ± 2.6	5/10	2000	0.178
	sham ESWT with strengthening exercise			15	49.7 ± 3.1	8/7	2000	0
Zhang 2021	ESWT_group1	knee OA	II ∼ III	19	60.8 ± 8.4	11/8	2000	0.12	once a week for 4 weeks	VAS and WOMAC	4 weeks
	ESWT_group2			19	62.7 ± 7.5	12/7	4000	0.12
	ESWT_group3			19	58.2 ± 9.5	10/9	2000	0.24
	ESWT_group4			18	63.7 ± 6.9	12/6	4000	0.24
	sham ESWT			14	61.5 ± 5.4	8/6	1000	0.02

K-L: Kellgren-Lawrence. ESWT: extracorporeal shockwave therapy. OA: knee osteoarthritis. VAS: visual analog scale. WOMAC: Western Ontario and McMaster Universities Arthritis Index.

3.3. Risk of bias

Fig. 2 displays the outcomes of the risk of bias assessment. All RCTs reported how to generate a random number, except for one. Four RCTs used allocation concealment during random assignment, but the other three RCTs did not mention this. Four RCTs had a double-blind study design. Three RCTs reported that their outcome assessors were blinded to the assigned group during outcome assessment. The rate of loss to follow-up was less than 20% for all RCTs. The major outcomes were reported for all RCTs.

Fig. 2.

Risk of bias for each article (upper figure) and a summary for all articles (lower figure).

3.4. Dose-response meta-analysis

This dose-response meta-analysis mainly used the total score of WOMAC and VAS, as they were the most commonly used outcomes in the included RCTs. However, the clinical outcomes observed after receiving ESWT decreased over time. To establish an accurate relationship between clinical outcomes and doses of ESWT, a dose-response meta-analysis was performed based on different follow-up periods.

3.4.1. WOMAC score

At the 1-week follow-up, only two randomized controlled trials (RCTs) reported WOMAC scores after receiving ESWT. A dose-response meta-analysis was performed, which revealed a negative relationship between doses of ESWT and WOMAC scores when using a linear model (β = −13.91, 95% CI = −27.18 to −0.65, p-value = 0.0398, I² = 97.2%; see Fig. 2A). At the 4-week follow-up, four RCTs reported WOMAC scores after receiving ESWT. A negative correlation was observed between doses of ESWT and WOMAC scores (β = −16.90, 95% CI = −26.93 to −6.86, p-value = 0.001, I² = 96.7%; see Fig. 2B). At the 12-week follow-up, two RCTs reported WOMAC scores after receiving ESWT. However, a significant linear association between doses of ESWT and WOMAC scores was not found (β = −14.08, 95% CI = −35.73 to −7.57, p-value = 0.202, I² = 93.8%; see Fig. 2C).

3.4.2. VAS score

At the 1-week follow-up, four RCTs reported the VAS score after receiving ESWT. A dose-response meta-analysis was performed, but a linear connection between the doses of ESWT and VAS scores was not observed (β = −10.89, 95%CI = −28.54–6.77, p-value = 0.2268, I² = 95.2%; see Fig. 3A). At the 4-week follow-up, three RCTs reported the VAS score after receiving ESWT. A negative correlation between the doses of ESWT and VAS score was observed (β = −15.36, 95%CI = −30.56∼-0.16, p-value = 0.0477, I² = 93.0%; see Fig. 3B). At the 12-week follow-up, two RCTs reported the VAS score after receiving ESWT. A negative linear association between the doses of ESWT and VAS scores was obtained (β = −14.93, 95%CI = −19.81∼ −10.04, p-value <0.001, I² = 19.8%; see Fig. 3C) (see Fig. 4).

Fig. 3.

Dose-response meta-analysis between the energy of ESWT and total score of WOMAC at 1-week (A), 4-week (B), and 12-week (C) follow-up periods.

Fig. 4.

Dose-response meta-analysis between the energy of ESWT and VAS score at 1-week (A), 4-week (B), and 12-week (C) follow-up periods.

4. Discussion

ESWT has recently become a widely used treatment for knee OA, with evidence showing that it is effective. However, the relationship between ESWT dose and clinical outcomes for knee OA is still unclear. To clarify this relationship, a dose-response meta-analysis was conducted. The analysis demonstrated that ESWT doses were negatively correlated with the total WOMAC score at 1- and 4-week follow-up periods, except for at the 12-week follow-up period. Furthermore, a negative correlation between ESWT doses and VAS scores was observed at 1-, 4-, and 12-week follow-up periods, respectively. These results suggest that higher doses of ESWT could lead to better improvement in clinical outcomes for knee OA. It should be noted, however, that publication bias of the included RCTs was not assessed due to the small number of included articles.

In our study, we observed a negative correlation between the dose of ESWT and VAS or WOMAC score at all follow-up periods except for the WOMAC score at the 12-week follow-up. Our pooled dose-response meta-analysis showed that the effect of ESWT doses on VAS scores decreased over time. Another meta-analysis demonstrated that clinical outcomes for patients with knee OA after receiving ESWT would also decrease with time.¹¹ In contrast, the WOMAC score after receiving ESWT would increase with the follow-up period. Finally, there was no evidence of a dose-response association during the extended follow-up period. Therefore, a dose-response relationship was not observed for the WOMAC score at the 12-week follow-up.

It is important to determine the association between energy flux density of ESWT and clinical outcomes for patients with knee OA who receive ESWT. One RCT compared the efficacy of two energy flux densities of ESWT for treating knee OA.¹² The results showed that higher energy (0.093 mJ/mm²) led to better improvement in pain relief and function restoration than lower energy (0.040 mJ/mm²), and both energies were effective for knee OA when compared with the baseline.¹² Another RCT also showed that higher energy (0.178 mJ/mm²) led to better improvement in clinical outcomes than lower energy (0.002 mJ/mm²), and both energies were effective for knee OA when compared with sham ESWT.¹⁴ These results suggest that there may be a dose-response relationship between ESWT energies and clinical outcomes for knee OA. Our pooled dose-response meta-analysis demonstrated a dose-response relationship between ESWT energies and clinical outcomes (VAS or total score of WOMAC), which was observed using a linear model. A higher energy of ESWT could lead to better improvement for knee OA. Therefore, physicians may consider adopting the maximum energy of ESWT that patients can tolerate the pain induced from ESWT for treating knee OA.

In this study, we conducted a dose-response meta-analysis to demonstrate the dose-response relationship between ESWT energies and clinical outcomes in patients with knee OA. The articles included in the meta-analysis were RCTs, providing our results with a high level of evidence. However, the study has several limitations. Although the included RCTs were of high quality, only a limited number of studies (n = 7) were included in the meta-analysis, which prevented us from performing a nonlinear model. Additionally, we could only adopt two clinical outcomes (total score of WOMAC and VAS) in our analyses. The WOMAC total score consists of pain, stiffness, and physical function, but our included RCTs did not separate these outcomes. Therefore, we could only assess a dose-response relationship for pain relief (VAS score). Our results also demonstrated statistical heterogeneity across all of the pooled dose-response meta-analyses, but we could not investigate the possible factors contributing to heterogeneity using subgroup analysis due to the limited number of included RCTs. For example, the number of shocks has been reported to affect the efficacy of ESWT in treating knee OA.¹⁵ Future studies with more articles are needed to investigate the effect of the number of shocks on clinical outcomes in knee OA.

5. Conclusions

In this study, we performed a dose-response meta-analysis to investigate the association between the energy levels of ESWT and clinical outcomes for patients with knee OA. Our results demonstrate a dose-response relationship between the energy of ESWT and clinical outcomes, as measured by the total score of WOMAC and VAS score, when using a linear model. This indicates that using a higher energy of ESWT for treating knee OA could result in better clinical outcomes than using a lower energy. Our results suggest that physicians could use the maximum energy of ESWT that patients can tolerate the pain induced by ESWT for treating knee OA. However, due to the limited number of included articles, a nonlinear model could not be used to investigate the dose-response relationship and an optimal energy of ESWT could not be determined in this study. Therefore, more articles are required to investigate the dose-response relationship between ESWT and clinical outcomes of knee OA in the future.

Funding

No funding.

Data availability

Not applicable.

Conflict of interests

None.

Ethics approval

Approval from the institutional ethics committee was not required for this review article.

Consent to participate

For this review article, there were no participants enrolled. As a result, informed permission was not necessary.

Funding statement

We do not receive any funding regarding this research.

Guardian/patient's consent

Not application.

CRediT authorship contribution statement

Tzu-Yin Chen: Writing – original draft. Shih-Hsiang Chou: Conceptualization, Writing – review & editing. Chia-Lung Shih: Conceptualization, Methodology, Validation, Writing – review & editing.