Low-Intensity Pulsed Ultrasound (LIPUS) for Fracture Healing

• The parameters. The standard device delivers 1.5 MHz ultrasound at a low intensity of about 30 mW/cm² (spatial-average temporal-average), as pulsed bursts (200 microseconds on, at a 1 kHz pulse-repetition frequency), for 20 minutes once daily. The "low-intensity" is the point - it is far below therapeutic/high-intensity ultrasound and is a non-thermal, mechanical stimulus (not the tissue-heating of higher-intensity ultrasound).
• The regulatory history. The FDA approved LIPUS for accelerating healing of fresh fractures in 1994 and for established non-unions in 2000, on the basis of early positive randomised trials (e.g. accelerated tibial and distal-radius union) - which is why it entered practice before the later high-quality trials tempered the fresh-fracture claim.

The topics this sits between are worth reading together. The biology it claims to accelerate is set out in fracture healing, and the problem it is offered for in non-union management, with the commonest site in tibial non-union. The other non-invasive stimulus with a comparable evidence problem is electrical stimulation. The interventions that do have an effect on union - correcting the mechanics and the biology - are covered under bone grafts and, for the single most modifiable host factor, smoking cessation.

• The landmark trial (TRUST), with its numbers. The TRUST trial (Busse et al, BMJ 2016) randomised 501 patients (250 LIPUS, 251 sham) with operatively managed tibial fractures across 43 North American trauma centres, blinded and sham-controlled, followed for one year. LIPUS was no better than sham on function - SF-36 physical component summary mean difference 0.55 (95% CI -0.75 to 1.84, p=0.41) against a minimal important difference of 3 to 5 points, so the confidence interval excludes any clinically important benefit - nor on time to radiographic healing (hazard ratio 1.07, 95% CI 0.86 to 1.34, p=0.55). Safety did not differ. Compliance was only moderate: 73 percent administered at least half the recommended treatments. A linked BMJ Rapid Recommendation (2017) then made a strong recommendation against routine LIPUS for fresh fractures.
• What TRUST deliberately excluded - and why that leaves non-union open. Of 3,105 patients screened, only 599 were eligible. The exclusions included pathological, segmental and bilateral fractures and, critically, any fracture with less than 25 percent cortical contact and more than a 1 cm gap after fixation. That is precisely the biologically compromised fracture in which a healing stimulus would be expected to matter most - so TRUST is a definitive answer for the ordinary, well-reduced, well-fixed tibia, and is silent on the population the non-union argument is actually about.
• The meta-analysis picture. Systematic reviews of LIPUS RCTs find no effect on functional recovery and only low-quality/inconsistent evidence for radiographic surrogates; the earlier small positive trials were at higher risk of bias.
• The bench-to-bedside disconnect. There is robust in-vitro and animal data (accelerated endochondral ossification, upregulated osteogenic genes) but neutral high-quality human trials - explained by surrogate (radiographic) versus patient-important (functional) outcomes, small-trial/publication bias, coupling/dosing/compliance issues, and a ceiling effect (ordinary fractures heal well anyway). Whether the biology helps in non-union (where there is more to gain) remains the open question the ongoing trials address.