Introduction

Distal radius fractures (DRFs) accounts for approximately 16% of all fractures treated by orthopaedic surgeons [1]. They are the most commonly encountered fracture in the emergency department, and more than 640,000 are reported annually in the United States alone [2]. Although the majority of DRF management strategies involve non-operative methods, a significant trend towards surgical treatment has been noted [3,4,5], largely explained by the introduction of volar locking plate (VLP) fixation which is associated with benefits including stable fixation, short immobilization period, and few complications [6, 7]. However, after plate fixation of the distal radius, adverse outcomes may include infection, tendon irritation, articular violation, or nerve irritation and most of these events require removal of the inserted VLP [8,9,10]. In the absence of complications, many surgeons have insisted that the inserted VLP does not need to be removed [11, 12]. Nevertheless, site-specific removal rates vary greatly from 0 to 100% [13,14,15,16,17]. In some cases, plate removal is performed at the patient’s request or the surgeon’s discretion in the absence of any clinical symptoms [18, 19].

This study aimed to compare characteristics of patients who underwent VLP fixation with and without subsequent implant removal. Second, the rate of implant removal according to the implant position and type was investigated. Finally, we summarized clinical outcome after implant removal, the reasons for removal, and the associated complications.

Materials and methods

Study design, setting, and participants

This retrospective single-centre cohort study was performed in compliance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines [20]. The study design was approved by the local institutional review board. The patient data were collected between January 1, 2008, and December 31, 2017. A highly experienced hand surgeon, according to levels of experience previously defined, conducted all surgeries [21]. During the procedure, the surgeon performed VLP fixation for DRFs using a modified Henry approach to the fractured distal radius in-between the flexor carpi radialis (FCR) and the tendons of the brachioradialis. Regarding the inserted VLP, only one manufacturer (DePuy Synthes Co., Solothurn, Switzerland) was used, but different types of plates were used depending on the fracture patterns per the surgeon’s discretion.

Two separates orthopaedic surgeons collected the data. During the study period, 1455 patients were diagnosed with DRF and treated operatively. Only patients with VLP fixation were included, and those with other treatment modalities, such as closed pinning, external fixation, dorsal plate fixation, and screw-only fixation, were excluded. The patients aged < 18 years were also excluded. During data analysis, the patients with insufficient information due to loss of medical records, those for whom data were not consistently recorded, and those with missing data regarding confirmation of medical comorbidities were excluded. Patients who were lost to follow-up before the bony union was achieved were also excluded.

Implant removal

The same surgeon performed implant removal after administration of anaesthesia (general, brachial plexus block, or bier block). Using a previously reported approach, the inserted plate was extracted under tourniquet inflation. A simple soft dressing was applied postoperatively, and immediate range of motion (ROM) exercises were allowed. Intravenous or oral antibiotics were routinely prescribed for less than 1 week.

Variables, data sources and measurement

The study population was divided into two groups based on subsequent implant removal. Data on patient characteristics, such as age, sex, side of the fractured arm, and comorbidities (diabetes, hypertension, angina or myocardial infarction, nephropathy or dialysis history, hypo- or hyper-thyroidism, and cancer) were collected from medical data base. For the radiologic evaluation, the AO Foundation and Orthopaedic Trauma Association (AO/OTA) classification of the DRF was assessed based on plain X-ray or computed tomography. For plate prominence grading associated with the watershed line in postoperative lateral plain X-ray, the Soong classification system was used [22], and the product name of the inserted plate was obtained. The insurance coverage for the DRF treatment was attempted to be investigated in all included patients.

Depending on groups, we attempted to compare regarding basic characteristics, comorbidity, AO/OTA classification, Soong grade, plate types, and insurance coverage.

Further investigation into the implant removal group was conducted. For functional evaluation, the patients with subsequent implant removal were assessed with the Disabilities of the Arm, Shoulder, and Hand (DASH) score prior to the operation. The DASH score was assessed again after implant removal during the outpatient follow-up before the end of treatment.

The treatment cost for implant removal was also investigated. When possible, the patient was contacted by phone, and further follow-up was conducted by two different orthopaedic surgeons during study investigation period. The reason for implant removal, clinical outcomes, and post-removal complications were investigated.

The reasons for implant removal were investigated from the medical chart review and through follow-up conducted via telephone. The reasons were classified into seven categories, including removal without clinical symptoms, nerve-related issues, tendon-related issues, implant-related issues, problems with osteosynthesis, part of another procedure, and others.

From the telephone phone survey, the personal satisfaction of each patient was assessed through two questions. First, ‘‘How would you describe the overall result of your treatment considering how you felt before implant removal and now, after implant removal??’’. The possible responses were ‘‘very satisfied (5)’’, ‘‘satisfied (4)’’, ‘‘average (3)’’, ‘‘poor (2)’’, and very poor (1)”. The second question was, ‘‘Would you undergo implant removal again in the same situation?’’ The possible answers were ‘‘yes’’ or ‘‘no’’.

Finally, we investigated any complications associated with implant removal and classified them as major and minor complications. The major complications included failure to remove a plate despite an attempt, refracture, tendon and neurovascular damage, decrease in the wrist ROM after implant removal, and deep infection. The minor complications included broken screws that remained at the implant site, superficial infection, post-operative pain, tenosynovitis, neuritis-like transient tingling sensation, and dissatisfaction with the operative scar.

Statistical analysis

The summary statistics are presented as mean ± standard deviation (SD) or numbers and percentages. Pearson’s chi-square test and Fisher’s exact test were used to compare the categorical variables and Student’s t test was used for comparison of continuous variables between groups. Before Student’s t test was performed, a normality test (Shapiro–Wilk test) was performed. The paired Student’s t test or Wilcoxon signed-rank test were performed in cases of paired continuous variables depending on normality tests result. A p value of less than 0.05 was considered statistically significant. With software R (v. 3.1.0; The R Foundation, Vienna, Austria), the statistical evaluation was conducted.

Results

Finally, 806 patients with a total of 814 DRFs were included in the study with a minimum postoperative follow-up duration of 4 months (Fig. 1).

Fig. 1
figure 1

Flow diagram for patient selection

Full size image

Among the 806 patients who underwent VLP fixation for DRFs, 252 (31.3%) patients underwent implant removal. The average time to implant removal from the initial open reduction and internal fixation (ORIF) was 12.1 ± 9.2 months (range 1–170 months). The mean operation time for implant removal was 28.8 ± 9.9 min. The average hospital stay was 2.2 ± 1.9 days and the average medical cost per patient was 616.6 USD (range 0–1594 USD) from admission to discharge.

All implants were manufactured by the same company (Depuy Synthes Co., Solothurn, Switzerland) with five different types of titanium plates used, including the 3.5-mm locking compression plate (LCP) T-plate, 2.4-mm variable angle LCP (VA-LCP) two-column volar distal radius plate, 2.4-mm LCP distal radius system (LDRS)/extra-articular, 2.4-mm LDRS/juxta-articular, and 2.4-mm VA-LCP volar rim distal radius plate.

Comparison between implant removal and retention

The patients who underwent implant removal were significantly younger (50.8 ± 14.0 years) compared to those who retained their implant (62.0 ± 13.4 years, p < 0.001). In the group with the implant removal, the male ratio was 37.3%, which was higher than the male ratio of 30.7% in the retention group, but there was no statistical difference. There were no differences between the two groups in terms of the fractured arm (Table 1).

Table 1 Comparison of patients between implant removal (group 1) and retention (groups 2)
Full size table

When comparing the comorbidities between the two groups, patients who underwent implant removal presented with a significantly lower rate of diabetes (p = 0.001), hypertension (p < 0.001), and any type of cancer (p = 0.004) than those who retained the implant. In the radiologic comparison, there was no difference in the AO/OTA classification between the groups. In terms of plate prominence according to Soong classification, the most common type was G1 in both groups. Although there was no statistical difference (p = 0.075), the G2 type tended to be more common in the removal group (15.0%) compared to the retention group (10.2%) (Table 2).

Table 2 Comparison of AO Foundation and Orthopedic Trauma Association (AO/OTA) classification between implant removal (group 1) and retention (groups 2)
Full size table

Comparison depending on plate type

There was a statistical difference in the implant removal rate depending on plate design (Table 3).

Table 3 Implant removal rates depending on plate types
Full size table

A relatively lower implant removal rate was observed for patients who received the 2.4-mm LDRS/extra-articular plate (21.6% [16 out of 74]) compared to those who received the other types of plates. In contrast, higher removal rates were observed in patients with the LDRS/juxta-articular plate (51.4% [18 out of 35]) and the VA-LCP volar rim plate (100% [2 out of 2]). Given the overall average time to implant removal, 12.1 months, between fracture fixation and implant removal, the average time in patients with the LDRS/juxta-articular plate and VA-LCP volar rim plate was shorter with 249.4 and 244.0 days, respectively.

Insurance coverage

In total, insurance coverage was verified for 185 of the 252 patients (73.4%) with plate removal and 332 of the 560 patients (59.9%) with plate retention. Comparing the insurance coverage for which patients were benefited, there was statistically different between groups (p < 0.001) (Table 4). Among plate removal patients, 46.5% disclosed that they had private health insurance which would cover the entire cost for implant removal compared to 31.0% in the patients with implant retention. In contrast, 53.0% of the patients with plate retention declared that they only had national health insurance coverage compared with 33.5% in the plate removal group. Industrial accident insurance coverage paid for implant removal cost in 11.9% of the total implant removal cases compared to coverage in 2.4% of the patients with plate retention.

Table 4 Insurance coverage for distal radius fracture treatment between implant removal (group 1) and retention (groups 2)
Full size table

Reasons for implant removal

Eighteen reasons for implant removal were identified and assigned to one of the seven defined categories. Overall, 186 (73.8%) patients (73.8%) underwent implant removal without clinical symptoms after bony union. The other reasons reported were carpal tunnel syndrome (n = 13 patients, 5.2%), combination with removal of other implants (n = 13 patients, 5.2%), unclear source of discomfort or pain (n = 12 patients, 4.8%), and foreign body sensation (n = 8 patients, 3.2%) in orders of ratio (Table 5). Among 13 patients with carpal tunnel syndrome, carpal tunnel release was performed together with implant removal in 11 (4.4%) patients.

Table 5 Reasons for implant removal
Full size table

Clinical outcomes

In 207 (82.1%) of the 252 patients who underwent implant removal, the pre- and post-operative DASH scores were obtained. The post-operative DASH scores were assessed on average 5.5 weeks after implant removal (range 2 weeks–15 months). The average pre-operative DASH score of 16.1 (range 0–70.8) improved to 5.1 at the final follow-up after implant removal (range 0–43.3; p < 0.001). Among the 252 patients, 165 (66%) completed a follow-up phone survey. Among 165 respondents, the satisfaction score average was 4.1 ± 0.9 (range 1–5) immediately preceding the implant removal and 4.4 (range 1–5) at the time of the phone survey, which indicates a significant improvement (p < 0.001). Among the 165 patients questioned, 153 (93%) patients answered “yes’’ when asked if they would undergo implant removal again under the same conditions.

Complications after implant removal

During outpatient follow-up, none of the patients experienced major complications. Minor complications were reported for 26 (10%) patients (Table 6). Six patients complained of newly developed postoperative pain with subsidence during outpatient follow-up. Five patients complained of numbness or tingling sensations along the median nerve or around the operative scar. Five patients were rather dissatisfied with their operative scar after the implant removal. Prior to implant removal, 12 patients requested removal due to an unclear source of discomfort or pain, and 3 patients complained of persistent pain even after implant removal (1.2%).

Table 6 Complications with implant removal
Full size table

Discussion

Among 806 patients who underwent VLP fixation for DRFs, 252 patients underwent plate removal during the 10-years study period. When comparing patient characteristics between the removal and retention groups, the patients who underwent implant removal were younger and had lower rates of diabetes, hypertension, and cancer. Although there was no statistical difference, the patients who underwent implant removal tended to have higher rates of grade 2 plate prominence according to Soong classification. There were higher rates of private health insurance coverage in patients who underwent plate removal. In 186 (73.8%) patients, plate removal was performed in the absence of any clinical symptoms. Juxta-articular and volar rim plates were associated with higher removal rates. Overall, the patient satisfaction score improved from 4.1 to 4.4 after implant removal. In total, 93% of the patients in the removal group who participated in the follow-up interview answered that they would choose implant removal again if they were in the same situation. No major complications were reported, but several minor complications were reported in 10% of the patients who underwent plate removal.

Researchers have speculated on the characteristics of the patients who undergo implant removal compared to those who opt for retention. Our results show that the patients who underwent removal were younger, by 10 years, than the patients who did not. In addition, although not statistically significant, men showed a higher rate of removal. Moreover, lower rates of diabetes, hypertension, and cancer were reported among the patients who underwent removal. These results suggest that this may be a more active cohort. Therefore, we posit that the removal rate is higher among patients with higher activity levels, regardless of the clinical symptoms.

There was no fracture pattern difference between the groups according to the AO/OTA classification. In a type C fracture, the distal screw must purchase the distal articular fragment and, therefore, it is usually necessary to place the plate very distally, which can irritate the flexor tendon [23]. Therefore, one would expect a higher rate of implant removal for type C fractures. However, no difference was detected in the current study, which corroborates the results from a previous systematic review [13]. One possible explanation is that the difference could be diluted as so many implants were removed in asymptomatic patients, which may have developed tendon irritation, had the implant been left. Furthermore, the AO classification groups various fracture types together within the C-fractures. Not all these fractures require a very distal position of the plate.

Based on the Soong classification of plate position prominence, more grade 2 cases tended to end in removal. Sellas et al. reported a six times higher plate removal rate among patients with grade 2 prominence compared to that for grade 0 prominence [24].

Regarding the plate type, we found higher removal rates associated with the juxta-articular and volar Rim plates, which were designed to be placed distally to purchase the distal fragment, resulting in Soong classification grade 2 [22] and concern for flexor tendon irritation. If this position is required for fixation, some studies suggest that subsequent removal should be planned after bony union [25, 26]. Asadollash and Keith reported 21 cases of flexor tendon injuries within an average of 9 months from fracture surgery [27]. In the present study, these two types were removed within an average of 8 months. Rather than the patient’s clinical symptoms, the surgeon’s concern and discretion might have affected these outcomes. However, flexor tendon injuries were not reported in patients who retained the juxta-articular plates.

Although there is a small variation depending on the tier of the hospital, the operation costs associated with implant removal in the study setting averaged 179 USD. The cost for hospital stay during implant removal and recovery averaged 616.6 USD. In Korea, the patient is asked to pay 20% of the total cost under the national health insurance service plan. However, about 66.5% of the patients who underwent implant removal had additional health insurance compared with 47% of the patients who retained the implant. There was a significant difference regarding insurance coverage between the groups, especially regarding private health and industrial accident insurance coverage. These findings suggest that patients are more likely to have implant removal if they have no economical burden even in the absence of clinical symptoms. Those that would have had to pay themselves more often retained the implant. Insurance coverage seems to influence the rate of plate removal.

There are some conditions requiring the removal of the inserted volar plate including infection, tendon irritation, articular violation, non-union or malunion, and nerve irritation [8,9,10, 28]. However, some patients opt for plate removal without significant adverse symptoms, despite the lack of evidence regarding the advantages. Yamamoto et al. defined removal without clinical symptoms as routine removal [13, 18, 29]. In our study, 186 (73.8%) patients (73.8%) underwent implant removal without clinical symptoms after bony union.

Studies have reported functional improvement and decreased pain associated with implant removal in patients with DRFs [19, 30, 31]. In our study, the average preoperative DASH score (16.3) improved at the final follow-up (5.1). The majority of patients had implant removed without symptoms, therefore, the DASH improvement could be due to the removal itself. For the other patients who removed plate with specific reasons, another explanation may be that the patients with poor functional ability before implant removal experienced complications associated with VLP fixation, including tendon ruptures, screw joint penetration, malunion and non-union, and carpal tunnel syndrome. Proper management of complications and implant removal could have resulted in improved functional scores. Finally, among 165 respondents, the satisfaction score average was 4.1 ± 0.9 (range 1–5) immediately preceding the implant removal and 4.4 (range 1–5) at the time of the phone survey and 153 (93%) patients answered “yes’’ when asked if they would undergo implant removal again under the same conditions.

It is possible to develop complications as a result of implant removal, and the use of the locking system in fracture management presents technical challenges during removal [8, 19, 32, 33]. One of the main reasons opposing removal in asymptomatic patients is that it could cause problems such as refracture, nerve damage, infection, tendon injury, and development of new pain [12, 18]. After defining the possible complications as major or minor, we did not find any major complications associated with implant removal, and the VLPs were removed without incidence except for two cases where a broken screw remained. In total, only 10% of the patients reported minor complications which were all transient and resolved during the out-patient follow-up.

There are some limitations to the study that must be considered. All surgeries, VLP fixation, and removal were performed by a single hand surgeon. Therefore, implant removal outcomes were based on one surgeon’s experience at a single-centre. In addition, we could not compare clinical outcomes with patients who did not undergo plate removal. The functional outcome measurements were self-reported in the medical chart, and additional data were collected by phone interview during the study. In addition, we encountered difficulties in attempting to assess information in elderly patients who could not participate in the telephone interview.

Our study showed that implant removal is more often performed in a young and healthy cohort. Although the implant removal was conducted without clinical symptoms in the majority of patients, overall patients presented improved functional outcomes with implant removal. The evidence is inconclusive regarding its necessity, however, implant removal after VLP fixation for DRF is not a challenging procedure and is not associated with major complications.