Assessment of methodological quality
Assessment of each study’s methodological quality was performed using the Quality Index tool developed by Downs and Black [
21]. This tool has been shown to have high internal consistency (KR-20 = 0.89), good test-retest reliability (
r = 0.88) and good inter-rater reliability (
r = 0.75). The Quality Index tool consists of 27 items, and allows for assessment of internal and external validity, reporting and power.
For this systematic review, an
a priori decision was made to remove two items where they did not apply to the respective studies identified. Firstly, Item 25 was removed for non-randomised controlled studies, as it has been shown that case mix adjustment cannot reduce the extent of bias in non-randomised trials [
22]. Secondly, Item 27 was removed for all studies, as a minimally important difference has not been established for measuring the outcomes of ESWT in the treatment of ulceration.
We chose to present the quality assessment results as percentage scores, which is typical of previous studies using the Quality Index tool [
20,
23-
24]. Furthermore, after obtaining data from the included studies, the findings were combined using a narrative rather than a quantitative approach, owing to study heterogeneity.
Trial characteristics
There were three randomised control trials, one quasi-experimental study and one case-series study identified, the characteristics of which are presented in Table
3. Moretti
et al. [
26] conducted a randomised controlled study in participants with neuropathic diabetic foot ulcers, to evaluate standard care and ESWT (intervention group) against standard care alone (control group). Both groups received standard care consisting of therapeutic footwear, debridement and dressings, although none of these variables were described in detail. Furthermore, the treatment of infection was undertaken where necessary and was not considered as a reason to exclude participants.
Table 3
Characteristics of included studies
| Rate of re-epithelialization | RCT | (a) 15 | 56.2 ± 4.9 | 40 | UD | Neuropathic (diabetic) plantar foot ulceration ≥6 months duration; area >1 cm2 and diameter between 0.5 and 5 cm | 3 sessions (every 72 hours); 100 pulses per 1 cm2; EFD 0.03 mJ/mm2 |
15 | 56.8 ± 7.5 | 53 | UD |
| Exudate, | Quasi-experimental | (a) 30 | 58.5 | 43 | 5.3 | Venous ulcers; diabetic ulcers; unresponsive to conservative care for ≥3 months duration | 4 to 10 sessions; 100 impulses per 1 cm2; EFD 0.037 mJ/mm2; frequency of 4 Hz or 240 impulses/min |
Granulation and | 10 | 66.6 | 40 | 5.2 |
Fibrin/necrotic tissue |
| Healing rates | RCT | (a) 34 | 58.6 ± 12.6 | UD | 22.7 ± 20.9 | Diabetic foot ulcer >3 months duration | 3 treatments; repeat course performed in cases with incomplete healing; 300 plus 100 pulses per 1 cm2; EFD 0.11 mJ/cm2 |
Histopathological analysis | (a) 36 | 63.4 ± 10.3 | UD | 19.0 ± 19.5 |
| Healing rates | RCT | (a) 39 | 60.5 ± 14.0 | UD | 6* (3 to16) | Diabetic foot ulcer >3 months duration | 6 treatments; Ulcer size dependent treatment; minimum 500 pulses; EFD 0.27 mJ/cm2 |
Histopathological analysis | (a) 38 | 62.5 ± 14.0 | UD | 6* (6 to10) |
| Safety and feasibility of ESWT | Case-series | (a) 31 | 61.0 | 48 | UD | Complicated, non-healing, acute and chronic venous and arterial ulcers | Mean sessions 1.9 to 3.7; 100 impulses per 1 cm2; EFD 0.1 mJ/mm2 |
Saggini
et al. [
27] conducted a quasi-experimental study and investigated 40 participants with chronic posttraumatic, venous and diabetic ulcers (17 venous ulcers, 7 diabetic). The case group (n = 30) received ESWT; between ESWT treatment sessions, the participants continued previous conservative treatment, although it is unclear what this treatment consisted of. There were 10 control participants who received regular conservative dressings, although the authors did not explain what conservative care involved.
Wang
et al. [
28] compared ESWT with hyperbaric oxygen therapy (HBO) in chronic diabetic foot ulcers. Seventy-four participants were randomly divided into two groups according to the dates that they were referred into the study. Thirty-six participants with 38 ulcers received ESWT, whereas 38 participants with 38 ulcers received HBO therapy. A repeat course of ESWT was performed in cases with incomplete healing from the first course of treatment. Participants resumed the same wound care technique at home after treatment including offloading on the affected leg, wound cleansing with sterile normal saline solution, and application of silver sulfadiazine cream. Participants in the HBO group received the same wound care as the ESWT group. In their second study comparing ESWT with HBO therapy in chronic diabetic foot ulcers, Wang
et al. [
29] randomly divided participants according to computer generated block labels. Forty-three participants were assigned to the ESWT group and 45 participants were assigned to the HBO group. After ESWT, participants resumed their initial wound care protocol including offloading on the affected foot, wound cleansing with sterile normal saline solution and application of silver sulfadiazine cream. Participants in the HBO group received the same wound care protocol after treatment as the ESWT group.
In a study by Schaden
et al. [
17], the safety and feasibility of ESWT was assessed in 208 participants with a variety of wounds that included 31 ulcers of a neurovascular origin. The intervention applied to all participants was a combination of debridement, ESWT and moist wound dressings.
Evidence for the effectiveness of ESWT in the treatment of lower limb ulceration
Table
4 provides a description of the mean differences of ulcer healing between groups. Moretti
et al. [
26] found that after 20 weeks of treatment, 53.33% of the ESWT group had complete wound closure compared with 33.33% of the control group, and healing times were 60.8 and 82.2 days respectively (
p < 0.001).
Table 4
Mean differences in ulcer healing between groups of included studies
| Wound closure: 53.33% | 33.33% | < 0.001 | UD |
healing time: 60.8 days (SD 4.7 days) | 82.2 days (SD 4.7 days) | <0.001 | 4.43, Huge effect |
| UD | UD | UD | UD |
| 31% completely healed | 22% completely healed | <0.001 | UD |
58% improved | 50% improved | <0.001 | UD |
11% remained unchanged. | 28% remained unchanged | <0.001 | UD |
≥50% improved in 89% of participants | ≥50% improved in 72% of participants | <0.001 | UD |
| Completely healed 57% | Completely healed 25% | =0.003 | UD |
≥50% improved in 32% of participants | ≥50% improved in 15% of participants | =0.071 | UD |
unchanged ulcers in 11% | unchanged ulcers in 60% | <0.001 | UD |
| NA | NA | NA | NA |
Saggini
et al. [
27] treated 32 ulcers with ESWT and reported that 16 ulcers healed completely within six sessions of ESWT. In those ulcers that did not completely heal, statistical significance (p < 0.01) was reported with regard to decrease in ulcer size, after four to six sessions of ESWT. There was no evidence in this study of a difference between the two groups regarding ulcer healing or change in ulcer size.
Wang
et al. [
28] found that in the ESWT group, 31% of ulcers completely healed, 58% improved and 11% remained unchanged. In the HBO group 22% completely healed, 50% improved, and 28% remained unchanged. These differences were significant at
p = 0.001. Furthermore, greater than 50% improvement of the ulcer was observed in 89% of participants in the ESWT group and 72% of participants in the HBO group (
p < 0.001). In their second study comparing ESWT and HBO, Wang
et al. [
29] found: completely healed ulcers in 57% and 25% (
p = 0.003); ≥ 50% improved ulcers in 32% and 15% (
p = 0.071), and unchanged ulcers in 11% and 60% (
p < 0.001) respectively.
Schaden
et al. [
17] found that venous stasis ulcers demonstrated the worst healing rates (36% versus 66% for all other ulcers,
p = 0.001). Furthermore, arterial insufficiency ulcers did not completely heal in 33% of cases, the second worst healing rate of all ulcer types. The primary outcome assessed in their study was the safety and feasibility of using ESWT on wounds, the authors concluding that ESWT is a safe and effective treatment.
Discussion
The aim of this systematic review was to investigate the effectiveness of ESWT for the treatment of lower limb ulcers. We evaluated five studies in this review, and identified a trend to suggest that ESWT may be effective in improving wound healing and decreasing wound size. Furthermore, ESWT may also be a safe treatment option with few complications associated with its use, however; we found average study quality for the studies identified. External validity across studies was rated most poorly, due to deficient definitions of the source population and methods of patient selection, and poor identification of confounding factors. It is difficult therefore, to generalise the findings of the studies to the populations from which the study participants were derived. Furthermore, it is unknown whether participants were representative of the population from which they were recruited. As such, all five studies performed poorly on the external validity questions, scoring a mean of only 27% for questions 11 to 13 on the quality Index tool.
All studies also rated poorly on the internal validity component of the Quality Index (questions 14 to 26). For example, Moretti
et al. [
26] conducted a randomised controlled study in participants with neuropathic diabetic foot ulcers, to evaluate standard care and ESWT (intervention group) against standard care alone (control group). While it is unclear what standard care comprised of in both groups, it appears that both groups received therapeutic footwear, debridement and dressings, and treatment of infection where present. Subsequently, treatment effects in the intervention group could have been influenced by the variation in the standard care regime, especially where antibiotics were used to treat infection. In the study by Saggini
et al. [
27], the authors described the control group as receiving usual conservative dressings but did not clearly describe the standard care received by the intervention group. The lack of standardization is of particular concern considering the non-blinded design of the studies. Therefore, the conclusions made by the authors of these studies should be considered in light of these methodological flaws.
The internal validity of the studies identified may also have been threatened due to a loss of participants during the study. While Moretti
et al. [
26] and Saggini
et al. [
27] suggested that all participants in their study completed the respective trials; Wang
et al. [
28] described a loss of four participants during their study, with no reference to how these losses were accounted for in their final analysis. In their second study, Wang
et al. [
29] described a loss of 11 participants during their trial, again with no discussion or statistical analysis to account for this loss. As there was no reference to an intention-to-treat analysis in the studies by Wang
et al. [
28,
29], group characteristics may have changed during the trial, resulting in over-estimation of the treatment effect; only the study by Schaden
et al. [
17] made an attempt to adjust their statistical analysis with intention-to-treat analysis. Further threats to internal validity might have occurred as no attempt was made to blind those responsible for measuring the outcomes in any of the five studies. Furthermore, the details of randomization in the study by Moretti
et al. [
26] were not made clear in their study. Consequently, differences in baseline characteristics may have influenced the effects of the intervention.
The classification of ulceration varied across the five studies identified in this review. Moretti
et al. [
26] defined their inclusion of neuropathic diabetic foot ulceration as occuring below the malleoli for a period of at least 6 months with an area wider than 1 cm
2. Although Moretti
et al. [
26] made an attempt to define peripheral neuropathy and standard care, ulcer classification and measurement of change in ulcer size was not undertaken using a recognised or validated measure. Saggini
et al. [
27] included venous and diabetic ulcers in those participants with a history of chronic ulcers for more than three months. While Saggini
et al. [
27] defined a recognised measure of wound exudate, the measurement of change in ulcer size was not appropriately described in their methods. The two studies by Wang
et al. [
28,
29] included participants with recurrent chronic diabetic ulcers of the foot for more than three months duration. In these two studies [
28,
29], and in the study by Schaden
et al. [
17], there was no description of a recognised method used to determine the change in ulcer size following the intervention. Consequently, the variation in ulcer classification and the poor definitions of change in ulcer size used across studies renders these results susceptible to bias.
The ESWT protocol varied between studies resulting in study heterogeneity, making comparison difficult. Specifically, there were differences in the duration, frequency and strength of ESWT application identified between studies. Furthermore, these differences were also noticeable within studies. For example, a repeat course of treatment was performed in cases with incomplete healing from the first course of treatment in the study by Wang
et al. [
28]. Moreover, the ESWT treatment dosage for each participant in the studies by Wang
et al. [
29] and Schaden
et al. [
17] was dependent on the size of the participant’s ulcer, rather than a pre-determined intervention dose. In the study by Saggini
et al. [
26], participants’ received anywhere between four and 10 sessions of ESWT; essentially, in these three studies, the likelihood of the results being due to the actual intervention cannot be determined.
This systematic review has identified a number of important implications for future research. Firstly, to reduce bias it is essential that when evaluating the effectiveness of ESWT for the treatment of lower limb ulceration, that rigorous randomised controlled trial (RCT) methods are used [
30,
31]. Second, it is necessary that outcome measures used are reliable and valid and include both specific and generic measures [
32]. This should also include detailed information about the criteria used to identify the presence of lower limb ulceration as there is substantial variability in the criteria used. Third, acknowledgement and adjustment for confounding variables should be included in future trials and if necessary, stratification of analyses should be made on the basis of the type of ulceration. This will ensure that such trials include sufficient information for the methods to be critiqued and allow comparisons to be made with similar investigations. Finally, the establishment of an optimal ESWT regimen remains to be established and should be the focus of future research.
The existing evidence that supports the use of ESWT for treatment of lower limb ulceration therefore needs to be viewed in light of some limitations. Firstly, there were only two studies (one of which was an RCT) that investigated the effect of ESWT versus standard treatment, and there were small participant numbers in the studies identified. Secondly, this review identified significant methodological heterogeneity between studies. For example; one of the studies in this review included smokers and also assessed ulceration associated with multiple comorbidities [
27], whereas the other three studies did not. Third, the outcome measures assessed across studies was inconsistent; although all studies assessed the change in ulcer size or ulcer healing rates as their primary outcome, the definitions used to determine these changes varied between studies. There were some limitations to our review design; there was no pooling of data for meta-analysis and no statistical measure of heterogeneity performed. There were however, two strengths of this review; the use of a validated quality assessment tool [
20], and the systematic approach used.