library(tidyverse) # for everythinglibrary(tidymodels) # for modelinglibrary(rms) # for more modelinglibrary(Hmisc) # for statslibrary(magrittr) # for numberslibrary(here) # for path managementlibrary(readr)library(openxlsx)library(readr) # for importing datalibrary(broom.mixed) # for converting Bayesian models to tidy tibbleslibrary(dotwhisker) # for visualizing regression resultslibrary(skimr) # for variable summarieslibrary(tableone) # for tableslibrary(gt) # for tableslibrary(gtsummary) # for prettier tableslibrary(patchwork) # for multi-planel figureslibrary(glue) # for working with stringslibrary(RColorBrewer) # for color palettes# Load custom functions# source("C:/Users/kinge/Dropbox/Research/Resources/mtk-custom-functions.R") # Windowssource("/Users/mtk/Library/CloudStorage/Dropbox/Research/Resources/mtk-custom-functions.R") # Mac
This was a retrospective cohort study consisting of patients who underwent operative correction of distal radius fracture malunion at a large, urban, academic institution between 2011 and 2022. Inclusion criteria consisted of any patient who sustained a distal radius fracture complicated by malunion who then underwent operative repair of the malunion with an osteotomy and subsequent fixation. Exclusion criteria consisted of inadequate postoperative follow up, which was defined as less than 1 year or lack of follow up to the point of radiographic evidence of healed osteotomy. Both the electronic medical record system and individual surgeon case logs were searched. The following CPT codes were used to identify potentially eligible patients:
25350 - Osteotomy, radius; distal third)
25365 - Osteotomy; radius and ulna
25391 - Osteoplasty, radius OR ulna; lengthening with autograft
25400 - Repair of nonunion or malunion, radius OR ulna; without graft
25405 - Repair of nonunion or malunion, radius OR ulna; with iliac or other autograft
25415 - Repair of nonunion or malunion, radius AND ulna; without graft
25420 - Repair of nonunion or malunion, radius AND ulna; with iliac or other autograft
For each identified patient, the operative report was used to assess whether the indication and procedure were relevant for this study. Patients undergoing any of the above procedures for an indication other than distal radius malunion were excluded.
2.2 Data Collection
The medical records of all eligible patients were reviewed and relevant data was recorded. This included demographic information (age, sex, BMI, smoking history), details related to the injury and surgery (duration between initial injury and malunion repair, type of osteotomy, type of bone graft), radiographic characteristics (volar tilt, ulnar varianace, radial inclination), and clinical outcomes (wrist range of motion at final follow up, reoperations or revision surgeries). Patients were grouped based on the type of graft used (structural graft consisting of cortical bone versus non-structural graft consisting of cancellous bone). The groups were further stratified based on the type of osteotomy performed (distraction osteotomy versus wedge osteotomy).
Radiographic measurements were performed by two independent orthopedic surgeons. Volar tilt, ulnar variance, and radial inclination were each measured on radiographs obtained preoperatively, immediately postoperative, and at final follow up with evidence of bony consolidation. Intra-class correlation coefficients were used to ensure appropriate agreement and reliability of the radiographic measurements between the two raters.
2.3 Statistic Analysis
Statistical analysis was performed using R (R Foundation for Statistical Computing, Vienna, Austria). The primary outcome of this study was volar tilt subsidence following osteotomy, defined as the change in volar tilt between the immediate postoperative radiographs and the final follow up radiographs after the osteotomy had healed. Secondary outcomes included ulnar variance subsidence and radial inclination subsidence. Comparisons between groups were
A priori sample size estimation was performed based on the primary outcome of volar tilt subsidence. To detect a 3 degree difference in volar tilt angle between the structural graft group and the non-structural graft group with 80% power and alpha = 0.05, assuming a 3 degree standard deviation and roughly 3:1 ratio of patients in non-structural group to patients in the structural group, a total sample size of 44 patients was needed.
A total of 51 patients met inclusion criteria and were included in the analysis. 70.6% of patients were female and the mean age was 51.2 +/- 18.8 years. 38 patients (74.5%) had non-structural bone graft during malunion repair, while 13 patients (25.5%) had structural bone graft during malunion repair. With respect to osteotomy type, 32 patients (62.7%) underwent distraction osteotomy and 19 patients (37.3%) underwent wedge osteotomy. The overall mean duration from injury to operative malunion repair was 18.7 +/- 16.9 weeks. The non-structural graft group had a significantly shorter duration from initial injury to malunion repair (16.6 +/- 17.0 weeks versus 24.7 +/- 15.5 weeks, p = 0.024). Mean follow-up duration was 1.4 +/- 2.2 years from the time of surgery. There was no difference between the non-structural graft group and the structural graft group with respect to baseline demographic characteristics (Table 1).
2 Welch Two Sample t-test; Fisher’s exact test; Wilcoxon rank sum test; Fisher’s Exact Test for Count Data with simulated p-value
(based on 2000 replicates)
Intra-class correlation coefficients (ICC) were used to assess the inter-rater reliability for radiographic measurements. ICCs were calculated using a two-way mixed effects model to evaluate the absolute agreement between raters. Good reliability was observed for radiographic measurement of volar tilt (ICC = 0.901, 95% CI [0.86, 0.93], p = < 0.001), ulnar variance (ICC = 0.805, 95% CI [0.66, 0.88], p = < 0.001), and radial inclination (ICC = 0.853, 95% CI [0.8, 0.89], p = < 0.001).
There was a wide range of preoperative sagittal plane deformity with respect to volar tilt (-41.55 degrees to 28.4 degrees), and patients in the structural graft group were associated with more severe sagittal plane deformity than the non-structural graft group (-20.5 +/- 16.0 degrees versus -7.0 +/- 21.5 degrees, 95% CI [1.5, 25.4], p = 0.028). Despite the more severe dorsal angulation in the structural graft group, there was no difference in the osteotomy gap size or degree of intraoperative volar tilt correction between groups (Table). Importantly, there was no difference in volar tilt subsidence (i.e., change from immediate postoperative to final follow-up) between graft type groups (-0.5 +/- 2.4 degrees versus -0.7 +/- 2.8 degrees, p = 0.658). Given that the structural graft group had more severe preoperative sagittal plane deformity but similar intraoperative correction and subsidence compared to the non-structural graft group, the structural graft group was associated with less optimal volar tilt at the postoperative and final timepoints (Table).
There was no difference in preoperative ulnar variance or radial inclination between the structural graft group and the non-structural graft group. Furthermore, there was no difference in the degree of intraoperative correction achieved or the amount of subsidence between graft type groups when patients who underwent wedge osteotomies and distraction osteotomies were considered together (Table).
Patients were then stratified based on the specific type of osteotomy performed and the radiographic outcomes were again compared between structural grafts and non-structural grafts (Figure 1). For patients who underwent wedge osteotomies, there was no significant difference in volar tilt subsidence, ulnar variance subsidence, or radial inclination subsidence between structural grafts and non-structural grafts (Table). Similarly, for patients who underwent distraction osteotomies, there was no significant difference in volar tilt subsidence, ulnar variance subsidence, or radial inclination subsidence between structural grafts and non-structural grafts.
There was no difference in wrist range of motion between structural and non-structural grafts (Table). Likewise, there was no difference in the reoperation rate between graft type groups. 6 patients in the non-structural graft group (15.8%) underwent reoperation and 3 patients in the structural graft group (23.1%) underwent reoperation (χ2 = 0.018, p = 0.893). There was no difference in reoperation rate when stratifying by osteotomy type (Table). Of the 9 total patients who underwent reoperation following malunion repair, 8 patients underwent removal of symptomatic hardware (2 patients with a concomitant ulnar shortening osteotomy and 1 patient with concomitant EIP to EPL transfer and Darrach procedure).
Clinical Outcomes for Wedge and Distraction Osteotomies
Characteristic
Wedge Osteotomies
Distraction Osteotomies
Non-structural, N = 131
Structural, N = 61
p-value2
Non-structural, N = 251
Structural, N = 71
p-value2
Reoperation
0 (0.0%)
2 (33.3%)
0.088
6 (25.0%)
1 (14.3%)
>0.999
Range of Motion (Degrees)
Flexion
55.0 +/- 17.9
51.0 +/- 12.4
0.616
50.5 +/- 16.7
56.0 +/- 10.8
0.397
Extension
51.7 +/- 19.0
55.0 +/- 20.6
0.772
45.8 +/- 20.3
45.0 +/- 22.4
0.945
Pronation
75.5 +/- 9.8
83.8 +/- 4.8
0.187
74.5 +/- 19.3
77.0 +/- 9.1
>0.999
Supination
78.0 +/- 7.9
76.3 +/- 11.1
0.941
70.5 +/- 17.4
72.0 +/- 10.4
0.861
1 Mean +/- SD; n (%)
2 Welch Two Sample t-test; Wilcoxon rank sum test; Fisher’s exact test
3.5 PROs
Code
# Load PRO datadata_path_pros <-here('Distal-Radius-Malunion_Data_PROs_20231105.xlsx')df_pros <-read.xlsx(data_path_pros,detectDates =TRUE) %>%mutate(across(c(mrn), factor),across(c(date_pro), ymd))# Merge with other dfdf_temp <- df %>%select(mrn, dos, graft, osteotomy)df_pros <- df_pros %>%left_join(df_temp,by ='mrn')# Remove excluded patientsdf_pros <- df_pros %>%filter(!is.na(dos))# Time from surgery to PRO vardf_pros <- df_pros %>%mutate(timepoint = (date_pro - dos) %>%as.numeric()) %>%mutate(timepoint_bin =case_when( timepoint <=0~0, timepoint >0~ timepoint ))# Final follow up PROsdf_pros_final <- df_pros %>%group_by(mrn) %>%# Only get last timepointfilter(timepoint ==max(timepoint)) %>%filter(timepoint >0) %>%# Remove anyone without any PROs completedfilter(!(is.na(promis_intensity) &is.na(promis_interference) &is.na(promis_ue_function) )) %>%ungroup()
A small proportion of patients completed PROMIS questionnaires at the 1 year postoperative timepoint (Table). There was no difference in PROMIS Intensity score, PROMIS Interference score, or PROMIS Upper Extremity Function score between patients with non-structural grafts and those with structural grafts.
