Clinical Trial Comparing Volar, Dorsal, and Custom Thermoplastic Splinting in the Treatment of Acute Mallet Finger

In most cases of acute mallet finger, or avulsion of the extensor digitorum communis terminal tendon from the distal phalanx, splinting is the treatment strategy employed. These are considered Doyle I injuries- closed injuries with or without a small (<20%) avulsion fragment. Absolute and relative indications for surgical intervention include an open injury, or a bony mallet involving greater than 20% of the articular surface of the distal phalanx with volar subluxation of the distal phalanx relative to the proximal phalanx. The type of splint used to treat Doyle I mallet injuries continues to be controversial. Custom splints have recently been advocated though no strong comparative evidence is available.

Objectives. The primary objective of this investigation was to show superiority of the custom thermoplastic splint over both the dorsal and volar non-custom splints. The null hypothesis was that a radiographic lag difference greater than 5 degrees would not be observed between groups. The primary outcome measure is the residual extensor lag difference (compared to the contralateral equivalent digit) at 12 weeks post-splinting. Week-12 was chosen as a compromise between minimizing drop-out and the most distant timepoint reasonably obtainable.

A novel outcome lag measurement was developed for this study, using the contralateral normal digit as an internal control for establishing residual extensor lag. The residual extensor lag difference was calculated with a single radiograph obtained with the beam directed laterally at the affected digit and the contralateral equivalent digit (Figure 2). For example, if the injured mallet finger was the right index finger, a single lateral radiograph of the right and left index fingers on the same film was obtained. The patient was instructed to maximally extend the both digits. This allowed for accurate determination of extensor lag and comparison with the unaffected side . Adequacy of the lateral view was assessed by considering the superimposition of the phalangeal condyles of the affected digit. If they were not not superimposed, the radiograph was repeated.

A pilot study was performed to establish the most reliable technique for measuring residual lag radiographically. Three measurement techniques were trialled by individuals with three levels of training: resident, fellow, and consultant staff. A clearly defined technique was employed using digital images only (Inteleviewer, Intelerad, Montreal, Canada). The radiograph was enlarged on the viewing monitor so that the middle and distal phalanges maximally filled the screen. The techniques used bisecting lines for the middle and distal phalanges referencing either the volar cortex, dorsal cortex, or longitudinal axis of the phalanges (Figure 3). Twenty-six lateral finger radiographs were blinded and assessed, and then repeat measurements were obtained three weeks later. Inter-item correlation matrices and intraobserver Pearson's correlations showed extremely high correlation values ranging from 0.993 to 0.999 for the longitudinal axis technique. This pilot study established the longitudinal axis technique as the most reliable in measuring residual radiographic lag.

Several secondary objectives were investigated. These included correlations between residual lag and other factors, including patient age, Michigan Hand Outcome Questionnaire scores (MHQ) and the development of complications; incidence of complications in each splinting group; comparing radiographic to clinical extensor lag differences; and evaluating the course of residual lag after splinting has been discontinued. Exploratory analyses were performed as hypothesis generators only, including sub-group analysis.