Shockwaves for nonunion after fracture surgery: a 2023 narrative review · Lab study · 2023
How might shockwave help a bone that has not healed after surgery?
This review explains how shockwave might restart healing in a bone that has not joined after surgery: lab and animal studies point to tiny cracks, small bleeds and growth signals. In the larger single-group series it lists, the bone healed in about 72% to 80% of treated people, with no untreated group to compare.
Explores how shockwave works- clinical reports listed, with 1 to 172 patients each
- 10
- kinds of device discussed: focused and radial
- 2
- year published
- 2023
Good to know: A narrative review.
Wang H, Shi Y Exp Ther Med, 2023 · Read the paper ↗
The treatment
- Focused regimens in the cited studies
- Rompe 2001: 3,000 impulses at 0.6 mJ/mm² over 50–75 min after local anaesthesia. Elster: median 4,000 impulses at 0.38–0.40 mJ/mm² for 20–60 min under general or local anaesthesia. Cacchio 2009: 4,000 impulses at 0.40–0.70 mJ/mm² (Dornier or Storz lithotripter). Wang 2006: 2,000 impulses, with or without bone-marrow grafting. Wu 2021: three courses of 10-minute sessions, twice a week, for up to 4 weeks; energy not reported.
- Radial regimens in the cited studies
- Zhang 2013: 1,000, 2,000 or 3,000 impulses. Kertzman 2017: 3,000 impulses at 0.18 mJ/mm² per session. Huang 2015: three sessions of 3,000 impulses at "80–120 J", 7 days apart. Jin 2018: 3,000 impulses at 0.54 mJ/mm², with or without autologous growth-factor injection. Session numbers are otherwise not reported.
- Focused versus radial (authors' description)
- Focused devices create a high-pressure spot that reaches deep bone, use high energy and commonly need anaesthesia; radial devices use mid-to-low energy and have recently been used for nonunion of superficial bones such as the navicular and tibia.
Who took part
The review lists 10 clinical reports of bones that had not healed after surgery, plus lab studies in rats, rabbits, mice, sheep and cells.
Side effects
Brief bruising, swelling, small skin bleeds, blisters and pain, without counts. One trial also listed infection, skin ulcers and nerve or vessel damage, without saying in which group.
All the numbers and the full appraisal
What the full paper reports.
No new patient cohort. Table I lists 10 clinical reports of nonunion after fracture surgery or corrective osteotomy, with 1 to 172 patients each. The mechanism section draws on rats, rabbits, mice, sheep, zebra mussels and cultured stem cells, plus skin and muscle studies.
Proposed mechanisms1 reported outcome
Proposed mechanisms
Qualitative finding
Restart of union: energy lost at the soft-tissue–bone interface produces microfractures, small bone fragments (0.1–3.0 mm³) that act like a graft, and a hematoma with IL-1β and IL-6 release. Other proposed routes: higher BMP-2, osteocalcin and TGF-β (rabbit tibia, rat mandible); osteoclast suppression (ovariectomized rats); proliferation and osteogenic differentiation of stem cells (0.4 mJ/mm² doubled proliferation in culture); angiogenesis (mouse limb ischemia and skin wounds, human skin in obesity); and mineralization in zebra mussels at 0.4 mJ/mm².
Analysis details and source
Laboratory, animal and other-tissue findings as summarized by the review; Figure 1 is labelled a hypothetical diagram. Two of the three osteoclast references (lipoteichoic acid, butyrate) and the L-plastin reference do not involve shockwaves, and the muscle-elasticity claim cites stroke and brain-machine-interface papers.
Union after shockwaves alone (single-group and dose series)1 reported outcome
Union after shockwaves alone (single-group and dose series)
Qualitative finding
Rompe 2001: 72.0% (31/43). Elster: 80.2% (138/172). Kertzman 2017: 72.7% (16/22). Zhang 2013: 28.6% (4/14) with 1,000 pulses, 85.7% (12/14) with 2,000 and 78.6% (11/14) with 3,000. One case report: 1/1. From the Zhang series the review states that doses below 2,000 radial pulses were excluded from later trials.
Analysis details and source
Uncontrolled rates show how many treated nonunions united, not how many would have united anyway. Rompe has 42 patients in the text and in Table I's case count, but the union fraction is 31/43. Elster is dated 2009 in Tables I and II but 2010 in the reference list. Kertzman is a retrospective case series by its reference title.
Union with shockwaves versus surgery1 reported outcome
Union with shockwaves versus surgery
Qualitative finding
Cacchio 2009 (randomized, 126 patients): 70.0% (26/37) with a Dornier and 71.0% (27/38) with a Storz device versus 73.0% (28/38) with surgery. Huang 2015 (randomized, 72): 87.2% (31/35) versus 93.9% (29/31) with autogenous bone grafting, described as similar. Wu 2021 (randomized, 65): 97.0% (32/33) versus 75.0% (24/32) with intramedullary nailing.
Analysis details and source
No effect estimates, intervals or tests are reported, so "similar" is not a tested equivalence. The Introduction gives Cacchio as 71% versus 74%. Table I denominators (37, 38 and 38 for Cacchio; 35 and 31 for Huang) are smaller than the enrolled totals, without explanation. For Huang the printed counts and percentages do not match. Wu 2021 is a two-page Chinese report; its prior surgery is open fixation in the text but intramedullary nailing in Table I.
Shockwaves combined with other treatments1 reported outcome
Shockwaves combined with other treatments
Qualitative finding
Wang 2006 (cohort, 42): 84.2% (16/19) with shockwaves plus bone-marrow grafting versus 82.6% (19/23) with shockwaves alone, a difference the review calls not statistically significant. Jin 2018 (randomized, 48): 95.8% (23/24) with shockwaves plus autologous growth-factor injection versus 75.0% (18/24) with shockwaves alone.
Analysis details and source
No test is reported for Jin 2018. Both are Chinese-language reports not read here.
Prognostic factors1 reported outcome
Prognostic factors
Qualitative finding
For union after shockwaves, the review lists a shorter time from fracture to first treatment, a shorter interval between fracture and surgery, good intramedullary stabilization and more treatment sessions. It also states that shockwave treatment versus no treatment was associated with higher union. Other factors (smoking, displacement, bone loss, contrast-ultrasound perfusion) come from revision-surgery studies.
Analysis details and source
No untreated group appears in Table I and no data are shown for the treated-versus-untreated statement. The contrast-ultrasound study is cited as reference 60 (a femoral-neck grafting study) rather than reference 58.
How treatment was delivered
Rompe 2001: 3,000 impulses at 0.6 mJ/mm² over 50–75 min after local anaesthesia. Elster: median 4,000 impulses at 0.38–0.40 mJ/mm² for 20–60 min under general or local anaesthesia. Cacchio 2009: 4,000 impulses at 0.40–0.70 mJ/mm² (Dornier or Storz lithotripter). Wang 2006: 2,000 impulses, with or without bone-marrow grafting. Wu 2021: three courses of 10-minute sessions, twice a week, for up to 4 weeks; energy not reported.
Zhang 2013: 1,000, 2,000 or 3,000 impulses. Kertzman 2017: 3,000 impulses at 0.18 mJ/mm² per session. Huang 2015: three sessions of 3,000 impulses at "80–120 J", 7 days apart. Jin 2018: 3,000 impulses at 0.54 mJ/mm², with or without autologous growth-factor injection. Session numbers are otherwise not reported.
Focused devices create a high-pressure spot that reaches deep bone, use high energy and commonly need anaesthesia; radial devices use mid-to-low energy and have recently been used for nonunion of superficial bones such as the navicular and tibia.
Table II lists adverse events by type only, with no counts: transient hematoma, local edema, petechial hemorrhage, blisters and pain; Wu 2021 lists infection, skin ulceration and peripheral vessel and nerve damage, and the review does not say whether these events occurred in the shockwave or the nailing group; two reports did not report harms and two reported none. The review calls the safety profile generally acceptable.
Author-stated: the optimal dose is unknown (too much causes excessive damage, too little falls short), effects on other tissues and organs need study, equipment design deserves further work, and further high-quality studies are needed.
Extractor: narrative review without search, selection criteria, risk-of-bias appraisal or pooling; trials, single-group series and a case report are placed side by side, and union definitions and follow-up are mostly not given. Equal-looking union rates without a tested comparison cannot show that shockwaves and surgery perform alike.
Internal inconsistencies: Rompe 42 patients in the text and in Table I's case count versus a union fraction of 31/43; Cacchio 71% versus 74% (Introduction) and 70–71% versus 73% (Section 3); Huang counts 31/35 and 29/31 printed with 87.2% and 93.9%; Wu prior surgery open fixation (text) versus intramedullary nailing (Table I); Elster dated 2009 in Tables I and II versus a 2010 reference; contrast-ultrasound study cited to the wrong reference.
Extractor: the summary calls shockwaves effective and tolerable, but the harms in Table II are listed without counts (including infection and nerve damage in one trial, without saying whether in the shockwave or the nailing group), and the mechanism section mixes bone studies with mussels, skin, limb ischemia and substances other than shockwaves. Funding: none received. Competing interests: none declared.
Findings checked against the full source by a separate automated reviewer.
Limitations noted by the reviewers
- Author-stated: the optimal dose is unknown (too much causes excessive damage, too little falls short), effects on other tissues and organs need study, equipment design deserves further work, and further high-quality studies are needed.
- Extractor: narrative review without search, selection criteria, risk-of-bias appraisal or pooling; trials, single-group series and a case report are placed side by side, and union definitions and follow-up are mostly not given. Equal-looking union rates without a tested comparison cannot show that shockwaves and surgery perform alike.
- Internal inconsistencies: Rompe 42 patients in the text and in Table I's case count versus a union fraction of 31/43; Cacchio 71% versus 74% (Introduction) and 70–71% versus 73% (Section 3); Huang counts 31/35 and 29/31 printed with 87.2% and 93.9%; Wu prior surgery open fixation (text) versus intramedullary nailing (Table I); Elster dated 2009 in Tables I and II versus a 2010 reference; contrast-ultrasound study cited to the wrong reference.
- Extractor: the summary calls shockwaves effective and tolerable, but the harms in Table II are listed without counts (including infection and nerve damage in one trial, without saying whether in the shockwave or the nailing group), and the mechanism section mixes bone studies with mussels, skin, limb ischemia and substances other than shockwaves. Funding: none received. Competing interests: none declared.
Source-linked critical reading
Narrative review
Laboratory and animal studies suggest shockwaves may restart healing by causing tiny cracks, bleeding and growth signals at the nonunion site. The clinical studies listed here report union in most treated patients, with rates that varied widely, but most are small, their designs differ, and the review does not test how much of that healing came from the treatment.
How might shockwave treatment restart healing in fractures that fail to unite after surgery, and what do published studies report on union, safety and prognostic factors?
Narrative review with no described search, selection criteria or risk-of-bias appraisal. It summarizes proposed mechanisms from animal, cell and other-tissue studies, then tabulates union rates from 10 clinical reports (four called randomized trials, five cohorts and one case report; several in Chinese) without pooling them.
No new patient cohort. Table I lists 10 clinical reports of nonunion after fracture surgery or corrective osteotomy, with 1 to 172 patients each. The mechanism section draws on rats, rabbits, mice, sheep, zebra mussels and cultured stem cells, plus skin and muscle studies.
Regimens differ across studies. Focused: 3,000 pulses at 0.6 mJ/mm² (Rompe), a median of 4,000 at 0.38–0.40 mJ/mm² (Elster), 4,000 at 0.40–0.70 mJ/mm² (Cacchio), 2,000 pulses (Wang), usually with local or general anaesthesia. Radial: 1,000, 2,000 or 3,000 pulses (Zhang), 3,000 at 0.18 mJ/mm² (Kertzman), 3,000 at 0.54 mJ/mm² (Jin), and three sessions of 3,000 pulses at "80–120 J" a week apart (Huang). For Wu 2021 only timing is given (three courses of 10 minutes, twice weekly, for up to 4 weeks), with no energy.
Surgery in three randomized trials (nail, plate or combined fixation for Cacchio, per the Introduction; autogenous bone grafting; intramedullary nailing); 1,000 and 2,000 versus 3,000 pulses in one dose series; shockwaves alone versus shockwaves plus bone-marrow grafting or plus autologous growth-factor injection. Four of the 10 reports have no comparison group, and none has an untreated group.
Fracture union rate as printed in Table I (timing given only for some studies, e.g. 9 months for Rompe and 6 months for Kertzman; union definitions not stated), adverse events listed by type in Table II, and prognostic factors described in the text.
Single-group union rates in Table I are 72.0% (Rompe), 80.2% (Elster) and 72.7% (Kertzman). In a dose comparison (Zhang), 28.6%, 85.7% and 78.6% with 1,000, 2,000 and 3,000 radial pulses. Against surgery: 70.0% and 71.0% versus 73.0% (Cacchio), 87.2% versus 93.9% with bone grafting (Huang), 97.0% versus 75.0% with nailing (Wu). Add-ons: 84.2% versus 82.6% with bone marrow (not significant) and 95.8% versus 75.0% with growth factors (Jin). No pooled estimates, intervals or tests are given for the surgical comparisons. The mechanisms (callus microfractures, growth factors, osteoclast suppression, stem cells, angiogenesis, biomineralization) rest on laboratory and animal work, and Figure 1 is labelled hypothetical.
Table II lists adverse events by type only, with no counts: transient hematoma, local edema, petechial hemorrhage, blisters and pain; Wu 2021 lists infection, skin ulceration and peripheral vessel and nerve damage, and the review does not say whether these events occurred in the shockwave or the nailing group; two reports did not report harms and two reported none. The review calls the safety profile generally acceptable.
No search or appraisal: small single-group series and Chinese-language reports carry the same weight as the larger randomized trial, and union definitions are not given. The numbers disagree in places: Rompe has 42 patients in the text and in Table I's case count, but the union fraction is 31/43; Cacchio is 71% versus 74% in the Introduction but 70–71% versus 73% in Section 3, and the Table I denominators (37, 38 and 38) fall short of the 126 enrolled; for Huang the printed counts (31/35, 29/31) do not match the printed percentages (87.2%, 93.9%); Wu is described as after open fixation in the text but intramedullary nailing in the table. Several mechanism citations concern mussels, skin, limb ischemia or substances other than shockwaves. No funding; no competing interests declared.
An overview of the biological hypotheses and of union rates published after shockwave treatment for nonunion following surgery. It cannot say whether shockwaves match revision surgery for a given nonunion, because the rates come from varied, mostly small studies without pooling or quality appraisal.
A systematic review that appraises each trial with consistent union definitions and timing, and adequately sized randomized trials against revision surgery that count adverse events per patient.
Read the original article.
Extracorporeal shock wave treatment for post‑surgical fracture nonunion: Insight into its mechanism, efficacy, safety and prognostic factors (Review).
Wang H, Shi Y. · Exp Ther Med · 2023
Open the publication record or available full-text files.
Checked against the full paper · Clinical review: Dr. Marcus Yu Bin Pai, · How we review →