Authors' reply to Pounder and colleagues
Pounder and colleagues suggest that the effect of LIPUS on tibial fracture healing in the TRUST trial  has been obscured by low compliance. As we note in our study, patient compliance was moderate and is a possible explanation for differences in results between TRUST and previous studies. The level of compliance we observed is, however, high enough that, if effects on time to radiographic healing seen in previous studies were present in our patients, we would have seen a substantial (albeit possibly attenuated) impact of the device.
To further inform this issue, we have completed a systematic review of all randomized controlled trials that have explored LIPUS for fracture healing (submitted for publication) and pre-specified compliance as a factor to explain heterogeneity. We considered adherence to be high if patients applied 80% or more of the total LIPUS time prescribed. The majority of trials reported only the surrogate outcome of time to radiographic healing, and our sub-group analysis (Figure, below) shows no evidence for a difference in effect between trials reporting high or moderate treatment compliance (test of interaction p-value = 0.79). What does explain heterogeneity in treatment effect is risk of bias. We categorized trials at high risk of bias if they failed to report concealment of allocation, blinding of patients, caregivers, data collectors or outcome assessors, or reported >20% loss to follow-up. Studies at high risk of bias suggest accelerated radiographic fracture healing (32% reduction in radiographic healing, 95%CI, -39%, -24%) and studies at low risk of bias show no significant effect (2% reduction in radiographic healing, 95%CI, -11%, 9%) (test of interaction p-value <0.001).
Pounder and colleagues advise that “lack of compliance should result in withdrawal from the trial”. We disagree. Such post-randomization exclusions will bias estimates unless those who do not comply are prognostically identical with those who do – a very unlikely circumstance. This is the reason for the consensus among methodologists on the desirability of intention to treat analysis.
Pounder and colleagues further advise that future trials of LIPUS for fracture healing should “…incorporate clinically meaningful and objective endpoints such as incidence of non-union”. We disagree: no patient will present to their surgeon with a complaint of incomplete radiographic union. The issues of importance to patients are whether use of LIPUS for fracture healing can reduce pain and accelerate functional recovery. Our systematic review found no evidence that LIPUS significantly influences either outcome: (1) -6.5 days of tenderness (95%CI, -24.1, 15.1); (2) -2.7 days to full weight-bearing (95%CI, -8.3, 3.1).
Jason W. Busse
Thomas A. Einhorn
James D. Heckman
Paul Tornetta III
Gordon H. Guyatt
On behalf of the TRUST Investigators
1. 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. Low Intensity Pulsed Ultrasound Following Operative Fracture Fixation. BMJ. 2016; 355: i5351
Competing interests: All authors have completed the ICMJE uniform disclosure form at www.icmje.org/coi_disclosure.pdf and declare: TAE, ES, and MB have received consulting fees from Smith & Nephew, the manufacturer of the study device. PT receives royalties from Smith & Nephew. GJDR is a paid consultant for Bioventus LLC, which is 51% owned by Essex Woodlands and 49% by Smith & Nephew. MB is supported, in part, by a Canada research chair, McMaster University.