This retrospective cohort study comparing distal radius fractures treated with a volar plate either using angular stable screws or pegs in the distal fragment showed no difference with regard to loss of reduction at six weeks and 12 months after surgery. Union was achieved in both groups by the time of the 12-month follow-up and implant removal rates were similar.

Duration of surgery was significantly longer for patients who received screws, despite the pegs group having more complex fractures. However, these surgeries were more frequently performed or supervised by junior attendings, while most of the surgeries in the peg group were performed or supervised by a senior staff member. Senior staff members typically treat more complex cases and due to their experience, the duration of a surgery can be reduced. In the study hospital the head of clinic was familiar with the pegs while many junior attendings had not used them before. As the choice of implant was left at the discretion of the treating surgeon, this could explain the unequal distribution. It therefore seems unlikely that the use of pegs truly reduces operating time, instead this was probably due to the experience of the surgeons.

Previous biomechanical studies evaluating the stability of fracture fixation with either fixed angle screws or pegs have come to contradictory results. A study from 2008 examining axial loading of AO type C3 fractures in sawbones that were fixed with various combinations of screws and pegs showed earlier failure if the lunate fragment was solely stabilized with pegs [12]. Similarly, a study from 2010 evaluating AO type A3 fractures in sawbones fixed either using pegs or screws concluded that pegs were mechanically inferior under torsional loading [10]. In 2013 a biomechanical study using cadavers with simulated extraarticular fractures was unable to show a difference when axial loading was tested (this study group did not test torsional loading) [11]. A study from 2006 comparing the failure from axial loading of ten different plates with various fixation methods [23]. Due to the descriptive nature of the study no sound conclusion could be drawn. However, in their study, it seems that locking pegs may possibly be superior, as those plates only failed by means of plate bending with the pegs remaining firmly attached to the plate and undisplaced with the bone. The plates they were compared to included such with locking screws, cortical screws and locking tines, many of which are no longer available on the current market making the comparison difficult and possibly irrelevant today.

To the best of our knowledge only one previous clinical study has compared volar plating of distal radius fractures using angular stable pegs versus screws. Boretto et al. included 27 patients with AO type C2 or C3 fractures, of which 13 patients were treated with locking screws and 14 with locking pegs [13]. Their findings were in line with the findings of this current study. While some displacement occurred in both groups between the early and late postoperative x-rays, there was no difference in the delta value between the groups. While the current study did not examine each delta value, the number of patients that demonstrated secondary displacement was comparable in both groups. The aforementioned study performed an early postoperative and late x-ray. It is not clear when exactly these x-rays were obtained. The current study performed two comparisons, the first was between the intraoperative imaging and the x-ray at the six-week follow-up, the second was between the six-week and the 12-month follow-up. In addition, the current study assessed for displacement in both the ap- and lateral projections, while the previous study only assessed the ap-images. This could have uncovered a previously missed relevant displacement of the volar tilt, which however could not be detected, therefore supporting the conclusion Boretto et al. came to. This study enhances previous findings by including a larger number of patients and incorporating AO type A and B fractures, thereby making the results more generalizable.

Regarding joint penetration, no sound conclusion can be made based on this study. Joint penetration occurred in only one patient, making this complication too rare to provide any statistical value in this cohort.

With regards to plate positioning, there were more Soong 1 plates in the peg group than in the screw group. This could also be an effect of the complex fractures in the peg group, as these often require more distal placement of the plate to be able to support all fragments. Another possible explanation is that the more experienced surgeons that treated the peg patients, were less afraid of placing the plate to distal and risking joint penetration or soft tissue irritation. It seems unlikely that the choice of implant alone would influence the position of the plate. Contrary to previous literature, the Soong classification did not influence the rate of implant removal [24, 25].

The occurrence of CRPS was low in both groups with an overall occurrence of 3.4%. A recent meta-analysis demonstrated CRPS rates in patients with radius fracture to be as high as 13.63% [26]. According to the findings in that study, CRPS occurred more frequently in fractures with great complexity or associated soft tissue damage. Only 2.6% of the patients in the screw group and 6.1% in the peg group had open fractures (total 4 patients), which could explain the low rate of CRPS in this study cohort.

There are some limitations to this study that should be mentioned. Firstly, the retrospective study design prohibited randomization. Randomized control studies are often considered the gold standard when comparing treatment options. However, by leaving the choice of implant to the surgeons’ discretion, each surgeon was likely to choose the implant he/she is most comfortable and experienced with. This way differential expertise bias can be avoided and a sort of pseudorandomization occurs [27]. However, as within the study hospital, the head of clinic was most familiar with the locking pegs, this led to an imbalance in the distribution of fracture types with more complex fractures being treated with pegs. The fact that even though there were more complex fractures in the peg group, there were no differences in radiological outcome between the groups, only emphasizes the finding that pegs and screws are comparable with regard to radiological outcome. The question remains though, whether the outcome would have been the same, if the pegs were used by less experienced surgeons. Another limitation of observational studies is that data is often incomplete for some patients. This is an issue we encountered in our patients. Some required x-ray views were missing from the intraoperative images. As mentioned in the methods, ulnar variance could not be assessed in patients with missing dorso-palmar views. Despite this, these patients were included as all other measurements and outcomes were obtainable. It is possible that some of these patients may have had a loss of reduction based on the ulnar variance. However, excluding these patients from the study could have introduced further bias and reduced the patient cohort so much, that statistical analysis would have become impossible.

A further limitation is variability in the quality of postoperative x-ray images. Not all x-rays were ideal, which could lead to difficulties in measuring the variables accurately. The measurement accuracy is another limitation that should be mentioned. All measurements were performed by a single author. While this eliminates inter-observer variability, it could be possible that a systematic measuring mistake was overseen. As the outcome is based on the difference between two measurements at different times and not absolute numbers, the authors don’t believe this to be a relevant problem. Furthermore, the definitions for “loss of reduction” were chosen arbitrarily. Had different cut-off values been chosen, the results may have been different. Finally, the lack of functional outcome measures leaves it unknown, whether the loss of reduction caused any clinically relevant issues in any of the patients. An attempt was made to include functional outcome, however due to the retrospective design of the study, too much information was missing to be able to perform any evaluation.

In conclusion, this study showed locking pegs and screws to be comparable with regards to short and long-term radiological outcome in distal radius fractures treated with a volar plate. While in this study population surgery duration was shorter when pegs were used, this is most likely influenced by the treating surgeon’s experience level and not by the implant itself. Considering the potential benefits of smooth locking pegs, such as their blunt tips and possibly reduced risk of secondary joint penetration, pegs are a viable alternative to screws.