It was recently noted that fractures of the volar articular margin of the distal part of the radius with volar radiocarpal subluxation (volar shearing, Barton, or AO Type-B3 fractures) are occasionally accompanied by a fracture of the dorsal metaphyseal cortex1. It was suggested that these fractures may be associated with osteoporosis and should be considered to be complete articular, or AO Type-C, fractures1. Failure to identify this fracture line at the time of surgery can lead to dorsal translation and angulation of the distal radial articular surface1-3, particularly when the volar plate is intentionally undercontoured so that it can be used as a reduction aid to push the volar marginal articular fragment back into position.
In this study, we investigated the relative incidence of volar marginal articular fractures with and without a concomitant fracture of the dorsal metaphyseal cortex (i.e., we compared Type-B and Type-C oblique volar marginal articular fractures with volar radiocarpal subluxation) and tested the null hypothesis that there is no difference between these two fractures with regard to wrist function or health status after open reduction and plate-and-screw fixation.
Between 2001 and 2004, 420 patients (423 fractures) were enrolled in a multicenter cohort study of open reduction and plate-and-screw fixation of an isolated fracture of the distal part of the radius (i.e., patients with no other injuries to the ipsilateral limb). Inclusion criteria were an age of eighteen years or greater, open reduction and internal fixation with a plate and screws performed within ten days after the injury, and no prior open reduction and internal fixation. Exclusion criteria included local disorders (e.g., tumors or Paget disease), systemic skeletal diseases (e.g., hyperparathyroidism), motor dysfunction (e.g., a central motor disorder or myasthenia gravis), a history of vascular insufficiency involving the injured limb, polytrauma (an Injury Severity Score4 of >16), or a history of drug or alcohol abuse. After exclusion of forty-four patients with inadequate preoperative radiographs and three with bilateral fractures, 373 fractures were considered for inclusion in the study.
The fractures were classified by the treating surgeon according to the Comprehensive Classification of Fractures (AO classification5). Two of the authors (J.S.S. and D.R.) reclassified all of the fractures. All discrepancies were discussed and resolved with the assistance of another author (J.B.J.). There were 144 Type-A (extra-articular) fractures, sixty-four Type-B (partial articular) fractures (including thirty-seven fractures that would be reclassified as Type C on the basis of an additional dorsal cortical break), and 165 Type-C (or complete articular) fractures.
Among the sixty-four fractures classified as Type B (partial articular), there were two radial styloid fractures (Group B1), five fractures of the dorsal articular margin with dorsal radiocarpal subluxation (Group B2, or a dorsal shearing [Barton] fracture), and fifty-seven fractures of the volar articular margin with volar radiocarpal subluxation (Group B3, or a volar shearing [Barton] fracture).
Thirty-seven (65%) of the fifty-seven Group-B3 fractures were associated with a fracture of the dorsal metaphyseal cortex and therefore might more accurately be considered to be a complete articular, or Type-C, fracture (Figs. 1-A and 1-B). These thirty-seven fractures were reviewed by a group of five experienced hand surgeons and four hand surgery fellows, and there was consensus that they were appropriately considered to be volar shearing (Barton) fractures by virtue of a volar radiocarpal subluxation with an oblique fracture of the volar articular margin of the distal part of the radius.
Type-B Volar Shearing Fractures
The average age of the nine women and eleven men with a partial articular (Type-B) fracture was forty-nine years (range, twenty to eighty years). The causes of the original injury included a fall from a standing height (five patients), a fall from a greater height (one), a sports injury (six), a motor-vehicle collision (three), and unknown (five). Ten injuries (all on the dominant side) involved the right wrist, and ten (one on the dominant side) involved the left wrist. All fractures were closed, and no patient had another injury in the ipsilateral limb. The interval between the injury and the index operative procedure averaged six days (range, zero to ten days). Eleven fractures were treated with a 2.4-mm locking compression plate (LCP; Synthes, Paoli, Pennsylvania); five fractures, with a 3.5-mm locking compression plate (LCP; Synthes); two fractures, with a distal volar radial plate (DVR; Hand Innovations, Miami, Florida); and two fractures, with a T-shaped nonlocking plate (Synthes), which was made of titanium in one and of stainless-steel in the other.
A volar approach through the sheath of the flexor carpi radialis was used in nineteen patients, and a combined volar and dorsal approach was used in one. Additional stabilization was done with a lag screw in one patient. Primary carpal tunnel release was not performed.
There were three complications: one patient had an injury to the palmar cutaneous branch of the median nerve, one had self-limiting tendinitis of the flexor digitorum profundus tendons, and one had chronic regional pain syndrome (type I). The first two patients underwent a planned plate removal, not related to the symptoms, one year after the initial procedure.
Type-C Volar Shearing Fractures
The average age of the twenty-four women and thirteen men with a complete articular (Type-C) fracture was fifty-two years (range, eighteen to eighty years). The causes of the original injury included a fall from a standing height (twelve patients), a direct blow to the wrist while participating in sports (nine), a motor-vehicle collision (eight), and unknown (eight). Eighteen patients had involvement of the right wrist, which was on the dominant side in fourteen of them, and nineteen patients had involvement of the left wrist, which was on the dominant side in sixteen. None of the fractures were associated with a wound or an ipsilateral upper-limb injury. Eight patients had concomitant fractures, none of which involved the extremity with the index fracture.
The interval between the injury and the index procedure averaged five days (range, zero to ten days). Twenty-six fractures were treated with a 2.4-mm locking compression plate; seven fractures, with a 3.5-mm locking compression plate; three fractures, with a distal volar radial plate; and one fracture, with a titanium T-shaped nonlocking plate. A volar operative exposure through the sheath of the flexor carpi radialis was used in all patients6. One patient also had a dorsal exposure and screw fixation. Autogenous iliac crest bone graft was used in four patients. Ancillary stabilization included external fixation in two patients, Kirschner wires in three patients, and a screw separate from the plate in six patients.
There were three complications: dorsal translation and angulation of the articular fracture fragments, a loss of alignment of the volar lunate facet fragment with volar radiocarpal subluxation7 (this patient also had a rupture of the extensor pollicis longus tendon), and injury to the palmar cutaneous branch of the median nerve.
There were no significant differences in demographic or injury characteristics between the patients with an AO Type-C and those with an AO Type-B volar shearing fracture.
Evaluation
Of the fifty-seven patients, forty-seven (sixteen with a Type-C and thirty-one with a Type-B fracture), or 82%, had a six-month evaluation; forty-seven (seventeen and thirty, respectively), or 82%, had a one-year evaluation; and forty-five (fifteen and thirty), or 79%, had a two-year evaluation.
Each patient was evaluated according to the system of Gartland and Werley8 and completed the Disabilities of the Arm, Shoulder and Hand (DASH)9 and Short Form-36 (SF-36)10,11 questionnaires at each follow-up point. The Gartland and Werley score is a physician-based demerit scoring system that combines residual deformity, subjective findings (including pain and disability), the range of motion (as determined with an objective evaluation), and complications including arthritis, nerve complications, and "poor finger function due to cast." The score ranges from 0 points, representing excellent function, to 52 points, representing the poorest function. The DASH questionnaire is a validated health-status questionnaire that results in a score ranging from 0 points, representing no disability, to 100 points, representing maximum disability. The SF-36 is a validated general health-status questionnaire that results in physical and mental component summary scores that are standardized according to population norms.
Pain at rest and pain in motion were evaluated at each time point on a visual analogue scale with 0 representing no pain and 10 representing the worst possible pain.
Radiographic Evaluation
The alignment of the distal part of the radius was measured on posteroanterior and lateral radiographs with use of commercial software (eFilm; Merge Healthcare, Milwaukee, Wisconsin). Ulnar inclination, palmar tilt, ulnar variance, and articular congruity were measured with use of the standardized guidelines described by Kreder et al.12. Because the images were digital and unscaled, it was not possible to measure ulnar variance and articular step-off in millimeters. Instead we used a technique for measuring these variables in reference to the length of the capitate as determined on posteroanterior radiographs with use of the guidelines described by Nattrass et al.13. Ulnar variance and articular step-off were reported as a ratio to the length of the capitate. Radiographic signs of osteoarthritis were rated, according to the system of Knirk and Jupiter, as grade 0 (no arthritis), grade 1 (slight joint-space narrowing), grade 2 (marked joint-space narrowing with osteophyte formation), or grade 3 (bone on bone, with osteophyte and cyst formation)14.
Statistical Analysis
Twenty-two parameters were assessed at each examination and regressed on two indicator variables (dorsal cortical break and follow-up time) and one interaction term (the product of dorsal cortical break and time—used to assess the interaction between the two variables). For each outcome, all data were pooled and analyzed together in one overall linear regression model, with the repeated measurements of each patient taken into account.
The likelihood ratio test was used to study the overall effect of a dorsal cortical break on each outcome parameter: the maximum likelihood estimate of the full regression model (including dorsal cortical break, follow-up time point, and interaction term) was compared with that of a null regression model (time only) to see if adding the dorsal cortical break variable would significantly improve the model. In a similar fashion, the full regression model was compared with a model without the "time" variable to determine the overall time effect on the outcome. Significance was adjusted for multiple testing with use of a Bonferroni correction; a p value of <0.003 (0.05/22) was considered significant. This approach reduces the probability of identifying spurious group differences simply because of the large number of parameters analyzed. For each outcome showing an overall significant effect, the full regression model was used to estimate and test the difference between the two groups (effect size) at each follow-up time.
Statistical adjustment was performed for the difference in sex distribution between the two groups. In addition, the analysis of the DASH scores was adjusted according to the DASH score based on the preinjury disability recorded for each patient at the time of enrollment. The SF-36 score was adjusted for concomitant injuries, and absolute range-of-motion values (in degrees) were adjusted for the mean for the contralateral healthy side across follow-up examinations.
A post-hoc power analysis with use of the scores for pain in motion on a 10-point visual analogue scale at two years indicated a power of 69%. This was based on the average score (and standard deviation) of 0.4 ± 0.6 point in a sample of twenty patients with a Type-B fracture compared with 1.1 ± 1.4 points in a sample of thirty-seven patients with a Type-C fracture and the alpha error level set at 0.05.
Source of Funding
This study was sponsored by the AO Clinical Investigation and Documentation (AOCID) in Dübendorf, Switzerland.
Effect of Dorsal Cortical Fracture
At six months after the surgery, the patients with a Type-B volar shearing fracture reported a mean score for pain in motion of 0.5 point on the 10-point visual analogue scale compared with 2.2 points for the patients with a Type-C fracture (group difference, 1.7 points [95% confidence interval, 0.7 to 2.6 points]; p < 0.001); there was no difference between the groups at twelve or twenty-four months (Table I).
Improvement with Time
There was a small but significant improvement (p < 0.05) in the mean arc of wrist flexion-extension between six months and one year after the surgery in both the patients with a Type-B fracture (from 122° to 127°) and those with a Type-C fracture (from 119° to 129°). The patients with a Type-C fracture also had a significant improvement (p < 0.05) in flexion, radial deviation, pronation-supination, pronation, supination, pain in motion, and grip strength between six months and one year after the surgery (Table I). There were no significant improvements in any of the variables between one year and two years after the initial surgical procedure.
Radiographic Evaluation
There were no significant differences between the groups with regard to volar angulation, radial angulation, or ulnar variance at any follow-up point (Table II). At the two-year follow-up examination, a 1 to 2-mm intra-articular step-off was seen in two patients in the Type-B group and in two patients in the Type-C group (p = 0.28). Grade-1 radiocarpal arthritis was found in seventeen patients (seven with a Type-B fracture and ten with a Type-C fracture), and grade-2 was found in six patients (two with a Type-B fracture and four with a Type-C fracture).
Volar marginal articular (AO Type-B3 [shearing]) fractures are, by far, the most common type of partial articular fractures of the distal part of the radius and accounted for 89% of the partial articular fractures in our series. The characteristic pattern is an oblique fracture line extending from the articular surface to the volar metaphyseal cortex with volar and proximal subluxation of the carpus along with the volar marginal articular fracture fragment. Harness et al.1 reported on a series of eight volar shearing fractures with a concomitant fracture of the dorsal metaphyseal cortex and noted that this fracture line converts the fracture from a Type-B to a Type-C (complete articular) injury. In that series, the dorsal metaphyseal fracture was not recognized intraoperatively in five patients, resulting in a mean dorsal angulation of 9° (range, 0° to 22°) and dorsal translation of the articular fracture fragments. On the basis of the Gartland and Werley score, the result was good for four patients and fair for one despite this deformity1.
Our data from a multicenter international cohort study of plate-and-screw fixation of unstable fractures of the distal part of the radius suggest that, among oblique volar marginal articular fractures of the distal part of the radius with volar radiocarpal subluxation (Barton fractures), the complete articular (Type-C) fractures described by Harness et al.1 predominate (65% [thirty-seven] of fifty-seven patients had Type C and 35% [twenty] had Type B). The occurrence of a dorsal metaphyseal fracture line—which is often subtle and unrecognized—in such a substantial proportion of volar Barton fractures reflects the difficulty in developing a comprehensive and reliable classification system.
When the dorsal fracture is obvious, it is debatable whether the injury should be classified as a volar marginal shearing fracture. In our opinion, the volar marginal articular fracture fragment is the key fragment that defines the injury and directs the treatment. It has long been the consensus that the optimal treatment of volar marginal articular shearing fractures is open reduction and internal fixation with a volar plate and screws. The observed spectrum from no dorsal metaphyseal fracture, to a subtle fracture, to a widely displaced fracture of the dorsal metaphyseal cortex simply reflects the fact that illness and injury often present in a continuum. We suggest that the Type-C volar marginal articular shearing fracture can be considered a transitional lesion between a Type-B shearing fracture and a volarly displaced complete articular fracture without an oblique shearing fracture of the volar fragment. Alternatively, since the latter is so uncommon, this may simply reflect the anatomy of the distal part of the radius with the volar articular margin15 nearly always failing with an unstable, oblique fracture.
We followed a relatively large series of volar shearing fractures prospectively for two years, but the analysis should be interpreted with an awareness of its shortcomings. To start, the distinction between volarly displaced fractures with an oblique shearing fracture mechanism and orientation from those that would be considered to have more of a compression mechanism is somewhat arbitrary and may not be reliable. This paper is based on fractures that a consensus group agreed would be interpreted by most surgeons as a volar Barton (shearing) fracture. In addition, we performed a large number of comparisons in a study with limited statistical power (69%), with the result that some of the findings may have been spurious and others that might have been significant were not found to be so. For instance, early in the recovery process, the differences in pain, supination, radial deviation, and grip strength between patients with a Type-B and those with a Type-C fracture were small and probably clinically irrelevant. However, some of the differences, such as that between the Gartland and Werley scores at two years (1.5 compared with 3.2 points) and that between the DASH scores at two years (3.9 compared with 11.1 points), could be considered clinically relevant and might have been statistically significant if we had not used such stringent criteria to account for multiple comparisons.
These data indicate that volar shearing fractures are usually complete articular fractures. This finding is important for several reasons. First, it changes our understanding of volar shearing and volarly displaced intra-articular fractures and will influence our attempts to classify them. Second, it should help to avoid the pitfalls that might occur both with failure to recognize the dorsal cortical fracture (namely, dorsal translation and angulation of the fracture fragments as documented in several recent series1-3) and with failure to appreciate the unstable volar shearing element of a complete articular fracture, which we have seen lead to inadequate treatment with either a dorsal plate or percutaneous methods. The fact that we found no differences in outcome between complete and partial articular volar shearing fractures should be interpreted in light of the facts that these fractures were treated by experienced surgeons and newer precontoured volar plates were used in all cases. A difference in outcome might have been identified had we been able to study the general treatment of these fractures, although it is encouraging that, with optimal treatment, the additional fracture line appears to be relatively inconsequential. 