Low intensity pulsed ultrasound for bone healing: systematic review of randomized controlled trialsBMJ 2017; 356 doi: https://doi.org/10.1136/bmj.j656 (Published 22 February 2017) Cite this as: BMJ 2017;356:j656
All rapid responses
Response to Dr. Farrar re: low intensity pulsed ultrasound for bone healing: systematic review of randomized controlled trials
Dr. Farrar questions the results of our systematic review that found low intensity pulsed ultrasound is not effective,(1) based on two hypotheses: 1) LIPUS, although ineffective in fresh fractures, may work in non-unions; and 2) the effectiveness of LIPUS is dose-dependent.
Our systematic review included three trials that enrolled patients with non-union of a fracture; all were at high risk of bias and none reported patient-important outcomes. We found no trustworthy evidence for patients with non-union, but did find moderate to high quality evidence that LIPUS does not improve outcomes for fresh fractures. The current evidence is insufficient to support the use of LIPUS in patients with non-union (this is discussed in detail in the linked Rapid Recommendation: http://www.bmj.com/content/356/bmj.j576)
Compliance in the TRUST trial was only moderate and likely reflected the real use of LIPUS in outpatients with tibial fracture.(2) Most patients (73%) administered LIPUS for ≥50% of recommended treatments: 45% of patients had ≥75% compliance and 28% had ≥50% but <75% compliance.
If the effect of LIPUS was dose-dependent, imperfect compliance in the TRUST trial could have attenuated the effect but not eliminated it. Since the TRUST trial found no effect on any outcome, then if LIPUS was effective only in those who used it perfectly it must be harmful in those with imperfect compliance. Reporting treatment effects only among highly compliant patients is inappropriate, as the balance in prognostic factors achieved by randomization is lost. A more appropriate approach is to explore for a subgroup effect between trials based on patient compliance.
Our systematic review found no evidence for subgroup effect based on compliance (interaction p=0.99 for radiographic healing). We did, however, find a credible subgroup effect based on risk of bias: only trials with high risk of bias showed a benefit of LIPUS. Although we cannot exclude a subgroup effect of LIPUS on non-union, it would be unlikely given the absence of any patient-important benefits on other types of fracture. A more plausible interpretation of the body of evidence is that LIPUS has no effect on bone healing.
Stefan Schandelmaier, Reed A Siemieniuk, Thomas Agoritsas, Per O Vandvik, Gordon H Guyatt, Jason W Busse
1. Schandelmaier S, Kaushal A, Lytvyn L, Heels-Ansdell D, Siemieniuk RA, Agoritsas T, Guyatt GH, Vandvik PO, Couban R, Mollon B, Busse JW. Low intensity pulsed ultrasound for bone healing: a systematic review of randomized controlled trials. BMJ. 2017; 356: j656
2. TRUST Investigators writing group, Busse JW, Bhandari M, Einhorn TA, Schemitsch E, Heckman JD, Tornetta III P, Leung K-S, Heels-Ansdell D, Makosso-Kallyth S, Rocca GJD, Jones CB, Guyatt GH. Re-evaluation of low Intensity Pulsed Ultrasound Following Operative Fracture Fixation. BMJ. 2016; 355: i5351
Competing interests: JWB and GHG were co-authors of the TRUST trial, which was supported in part by an industry grant from Smith & Nephew, a manufacturer of LIPUS devices.
Re: Low intensity pulsed ultrasound for bone healing: systematic review of randomized controlled trials
I am a regular user of pulsed ultrasound for treatment of non-union of fractures and have observed excellent outcomes with this technique and observe success rates of 90-100%. I achieve considerable cost savings for my Trust in avoiding surgery in these patients.
In 2 acute fractures of the distal 1/3 of the shaft of the tibia, which I assessed as high risk for delayed or non-union, I achieved 100% union with definite acceleration of healing observed in both cases.
My patients are highly motivated and are encouraged to utilise the device regularly. As a result I see 95-100% compliance with use of the devices, as measured by checking the 'tick' chart on the device.
In the TRUST trial compliance was 43%. It is therefore no surprise that the trial concluded the device, on an intention to treat basis, did not accelerate healing in this group of patients. This is manifestly unfair to patients who utilise the devices well with high compliance, and calls into doubt the claim of the trial that ultrasound fails to accelerate bone healing. The trial team should therefore be asked to report the outcomes of the subset of patients with high compliance, in their results. This would then start to answer the actual question about the value of ultrasound in accelerating fracture healing.
Competing interests: No competing interests