Fracture healing, delayed union and nonunion · Single-group study · 2024
Can shockwave help a thigh or shin bone that has not healed?
In this study of 91 people whose thigh or shin bone had not healed, about 63% of the bones healed within 12 months after a single shockwave session. Healing was less likely when the gap in the bone was wider, the bone ends showed little healing activity, or the person had had more than one earlier operation.
About 6 in 10 healed after one session- people with a thigh or shin bone that had not healed
- 91
- shockwave session
- 1
- months of follow-up (at most)
- 12
Good to know: A look back at records of people who all had shockwave; there was no comparison group.
Wu KT, Cheng JH, Jhan SW, et al. International Journal of Surgery (London, England), 2024 · Read the paper ↗
How this study worked
Conceptual animation: a focused front converges, then spreads. Not to scale. Tap the field to send a pulse.
The treatment
Shockwave
Shockwave, 1 session
- Sessions
- 1
- Pulses
- 3,000 per area · in each of two planes (45° and 60°)
- Strength
- 0.62mJ/mm² maximum in the treatment zone; 28 kV
- Numbing
- Not mentioned · done under general anaesthesia
- Device
- OssaTron (SANUWAVE)
- Everyone also got
- Partial weight-bearing with crutches for 4–6 weeks
Compared with: No comparison group
Settings are what each study used, not a prescription.
Who took part
91 people whose thigh-bone (femur) or shin-bone (tibia) fracture had not healed.
Side effects
The authors report that there were no complications such as bruising or bleeding under the skin.
All the numbers and the full appraisal
Every reported result
The results, in numbers.
Ninety-one patients with femur or tibia nonunion treated with ESWT; the clinical text and Table 1 report 57 unions, while Table 2 classifies 56 as union and 35 as nonunion. This source inconsistency is preserved.
Radiographic union definition1 reported outcome
Bony union after ESWT
Radiographic union definition Patients with union / patients
Observed subgroup counts; these are not randomized treatment arms.
The article reports union in 57/91 (62.6%). There was no control group; this is a descriptive rate, not an estimate of benefit versus another treatment.
Analysis details and source
Table 2 assigns 56 patients to union and 35 to nonunion (also total 91), inconsistent with the reported 57 unions.
Radiographic union1 reported outcome
Union by nonunion morphology
Radiographic union Patients with union / subgroup
Observed subgroup counts; these are not randomized treatment arms.
Reported union proportions were 30/46 (65.2%) hypertrophic, 21/27 (77.8%) oligotrophic and 6/18 (33.3%) atrophic; Table 1 reports p=.017. These are treated-cohort subgroups, not randomized comparisons.
Analysis details and source
The 57 events sum across subgroups and conflict with Table 2’s union n=56. Table 2 also lists 19 atrophic patients (6 union, 13 nonunion), whereas Table 1 gives 6/18 and the Results text 18 atrophic nonunions.
Multivariable logistic regression3 reported outcomes
Number of prior surgeries and nonunion
Multivariable logistic regression Odds ratio
Adjusted association of surgery number with nonunion
91 participants in this estimate
Multivariable model
Reported estimate and confidence interval. Logarithmic ratio scale.
The source reports an independent association with nonunion after ESWT. Because all patients received ESWT and the design is retrospective, this is prognostic association, not evidence that ESWT causes or prevents nonunion.
Analysis details and source
The abstract/results frame these as risk factors for failure; regression coding details are limited. Preserve reported OR and CI. Table 2 lists the number of surgeries among the continuous variables (mean 1.18 ± 0.39 in the union group vs 1.46 ± 0.56 in the nonunion group), so this OR refers to the number of surgeries; the "more than one surgery" threshold comes from the separate ROC analysis (Table 3, cutoff value 2).
Maximum fracture gap and nonunion
Multivariable logistic regression Odds ratio
Adjusted association per reported fracture-gap model (mm)
91 participants in this estimate
Multivariable model
Reported estimate and confidence interval. Logarithmic ratio scale.
The source reports an independent association with nonunion after ESWT. Because all patients received ESWT and the design is retrospective, this is prognostic association, not evidence that ESWT causes or prevents nonunion.
Analysis details and source
The abstract/results frame these as risk factors for failure; regression coding details are limited. Preserve reported OR and CI.
Atrophic versus trophic nonunion and nonunion outcome
Multivariable logistic regression Odds ratio
Adjusted association for nonunion type
91 participants in this estimate
Multivariable model
Reported estimate and confidence interval. Logarithmic ratio scale.
The source reports an independent association with nonunion after ESWT. Because all patients received ESWT and the design is retrospective, this is prognostic association, not evidence that ESWT causes or prevents nonunion.
Analysis details and source
The abstract/results frame these as risk factors for failure; regression coding details are limited. Preserve reported OR and CI. Table 2 enters type of nonunion as Atrophic (6 union, 13 nonunion; 19 patients) against a reference row printed as "Hypertrophic" (50 union, 22 nonunion; 72 patients). As printed, 72 matches the trophic group (46 hypertrophic + 27 oligotrophic = 73 in Table 1) rather than the 46 hypertrophic nonunions, and 19 atrophic differs from the 18 in Table 1 and the Results text.
ROC curve1 reported outcome
Discrimination/cutoffs for ESWT failure predictors
ROC curve AUC and cutoffs
Maximum fracture gap AUC 0.743 (95% CI 0.590–0.862), cutoff 3.94 mm (sensitivity .46, specificity .83, p<.001). Number of surgeries AUC 0.628 (95% CI 0.569–0.790), cutoff >1 surgery (sensitivity .50, specificity .85, p=.01).
Analysis details and source
Predictive performance in this cohort only; no external validation.
How treatment was delivered
OssaTron (SANUWAVE); single treatment under general anesthesia with fluoroscopic localization. Two planes at 45° and 60° to long axis, 3,000 impulses per plane, 28 kV, maximum treatment-zone EFD 0.62 mJ/mm². Partial weight bearing with crutches for 4–6 weeks.
In this cohort, the authors report no complications such as ecchymosis or hematoma formation. The report does not provide counts or systematic ascertainment detail for other adverse events.
Retrospective study without a comparator; only radiographic union outcomes were measured, with no function assessment.
Sample size is limited and some relevant factors (e.g. NSAID use, osteoporosis, postmenopausal status) were not analyzed; no external ROC validation.
The overall union count conflicts: clinical text and Table 1 report 57/91 (62.6%), whereas Table 2 lists union n=56 and nonunion n=35. Preserve both source statements without reconciling them.
Findings checked against the full source by a separate automated reviewer.
Limitations noted by the reviewers
- Retrospective study without a comparator; only radiographic union outcomes were measured, with no function assessment.
- Sample size is limited and some relevant factors (e.g. NSAID use, osteoporosis, postmenopausal status) were not analyzed; no external ROC validation.
- The overall union count conflicts: clinical text and Table 1 report 57/91 (62.6%), whereas Table 2 lists union n=56 and nonunion n=35. Preserve both source statements without reconciling them.
Source-linked critical reading
Retrospective treated-cohort prognostic study
Technical source record
Full XML, treatment method, three raw tables, ROC figure caption, all results/discussion and declarations reviewed.
The study reports union after ESWT in about 63%, but its totals conflict and it had no comparison group. Some fracture features were linked with failure and need confirmation.
What proportion of long-bone nonunions unite after ESWT, and which baseline factors associate with failure?
Retrospective single-cohort study; 91 femur/tibia nonunions treated with ESWT.
Patients with femoral or tibial nonunion; overall union counts conflict between text/Table 1 (57/91) and Table 2 (56 union, 35 nonunion).
Single OssaTron session under general anesthesia and fluoroscopy; two planes, 3,000 impulses each, 28 kV, maximum EFD 0.62 mJ/mm²; partial weight bearing 4–6 weeks.
No untreated or alternative-treatment comparison.
Union reported in 57/91 (62.6%) by text/Table 1; subtype results: hypertrophic 30/46, oligotrophic 21/27, atrophic 6/18.
Adjusted associations with failure/nonunion: number of prior surgeries OR 3.7 (95% CI 1.257–10.886), gap OR 1.985 (1.250–3.153), atrophic type OR 4.423 (1.296–15.096). These are prognostic associations within ESWT-treated patients.
No bruising/hematoma reported; broader adverse-event ascertainment not described.
Retrospective, uncontrolled, radiographic outcomes only, limited sample and unmeasured factors; union totals conflict; ROC cutoffs not externally validated.
Results may help identify poorer prognosis after ESWT but do not show comparative treatment efficacy.
Prospective cohorts/trials should reconcile union definitions and denominators, measure function, validate thresholds and include comparison groups.
Read the original article.
Prognostic factors of extracorporeal shockwave therapy in the treatment of nonunion in long bones: a retrospective study.
Wu KT, Cheng JH, Jhan SW, Chen PC, Wang CJ, Chou WY. · International journal of surgery (London, England) · 2024
Open the publication record or available full-text files.
Checked against the full paper · Clinical review: Dr. Marcus Yu Bin Pai, · How we review →