1. Introduction
Distal femur fractures represent a relatively small percentage of all fractures (0.5%), but the incidence is increasing, now occurring in up to 37 people per 100,000.1–4 Fractures of the distal femur are commonly treated with a lateral locked plate (LLP), particularly in the elderly population. Nonunion remains a significant clinical challenge after LLP fixation, and several large case series have reported nonunion rates ranging between 10% and 22%.5–8 Surgical treatment of distal femur nonunion often results in significant morbidity to the patient and significant cost to the health care system. While the diagnosis of nonunion is often straightforward, no clear consensus on clinical or radiographic parameters exists. Recent literature has advocated for the use of the Radiographic Union Scale for Tibia (RUST) and modified RUST scores in metadiaphyseal fractures including the distal femur.9
Several independent risk factors for nonunion after LLP fixation of distal femur fractures have been identified in previous studies including age, obesity, diabetes, smoking status, open fracture, infection, and degree of comminution.6–8 Rodriguez et al8 recently highlighted additional variables that may affect nonunion rates, such as plate material (stainless steel vs. titanium) and construct rigidity score.8 Relatively few studies, however, have investigated the impact of different immediate postoperative weight-bearing regimens on nonunion rates. Many authors recommend non–weight-bearing (NWB), typically because of concerns related to poor bone quality and marginal construct stability.10 Other investigators, however, have suggested that immediate weight-bearing as tolerated (WBAT) is safe and not associated with early postoperative fixation failure or other complications.11,12
The primary purpose of this study, therefore, was to investigate immediate postoperative weight-bearing status, along with other potential risk factors for nonunion after locked plating of distal femur fractures. We also sought to investigate the association between radiographic union scores and clinically determined fracture union in distal femur fractures.
2. Methods
2.1. Patient Selection and Variables of Interest
Approval for this study was obtained from the Health Sciences Institutional Review Board. The study was deemed exempt by the Institutional Review Board, and the research was conducted in accordance with the Declaration of the World Medical Association. All adult distal femur fractures treated operatively over a 10-year period (2007–2017), at a single academic, Level-1 trauma center were identified using billing records. The departmental billing database was queried using Current Procedural Terminology (CPT) codes for plate fixation of distal femur fractures (CPT codes 27,507 and 27,511) in patients aged 18–89 years. Exclusion criteria included patient younger than 18 years or older than 89 years, fractures with intra-articular (transcondylar) extension, pathologic fractures, diaphyseal fractures, fractures treated with intramedullary nailing, patients requiring restricted weight-bearing of the contralateral leg, or those with clinical/radiographic follow-up <12 weeks.
Standard demographic variables were extracted from the database including sex, age at time of surgery, race/ethnicity, body mass index (BMI), American Society of Anesthesiologists score, and tobacco use. Obesity was classified as BMI of 30.0 or greater. Perioperative and postoperative clinical variables of interest included fracture type (open vs. closed), periprosthetic fracture (yes/no), immediate postoperative weight-bearing status (weight-bearing all tolerated/full weight-bearing, touch-down weight-bearing (TDWB) or partial weight-bearing, and non weight-bearing (NWB)), time interval from surgery until full weight-bearing was permitted, reoperation, reason for reoperation, and clinical fracture union. The primary outcome for the study was nonunion of the distal femur. Consistent with previous studies,7,8,13 nonunion was defined as the incidence of any additional surgical procedure performed after the index surgery to promote or assist with bone healing. Patients who did not ever require secondary surgery for bone healing were defined as having achieved clinical union.
2.2. Radiographic Analysis
All postoperative radiographs were reviewed for each patient. Immediate postoperative radiographs were reviewed, in conjunction with the operative report, to determine construct characteristics including plate material, plate length, number of screws proximal to the fracture, total screw density, and proximal screw density. A rigidity score was then calculated according to the method of Rodriguez et al,13 resulting in a score from 0 to 5, which represents the range from low to high construct rigidity. Radiographs from all follow-up visits were reviewed to assess for evidence of implant failure (screw and/or plate breakage) and fixation failure.
To determine RUST and modified RUST scores, 3 independent reviewers (Author 1, Author 2, and Senior Author) evaluated anteroposterior (AP) and lateral radiographs at final follow-up or at the time of nonunion diagnosis. The reviewers evaluated each cortex for the presence and character of fracture callus and calculated RUST and modified RUST scores (ranging from 4 to 12 and 4 to 16, respectively) for each patient.9,14 For the modified RUST score, each of the cortices on the AP and lateral radiographs were scored as previously described by Litrenta et al9 (1 = no callus, 2 = callus present, 3 = bridging callus, 4 = remodeled, fracture line not visible). For the RUST score, as described by Whelan et al,14 the middle 2 categories are collapsed, and any callus present results in a score of 2, whereas remodeled cortices earn a score of 3.
2.3. Statistical Analysis
Standard descriptive statistics were used to characterize the study population. Means and standard deviations were used for continuous variables, and frequencies and percentages were used for categorical variables. Fisher exact tests were used to compare union versus nonunion in categorical variables, and t-tests or Wilcoxon rank sum tests, as appropriate, were used for comparisons between continuous variables.
Interclass correlation coefficients (ICC: 2,1) were calculated to evaluate inter-rater reliability between observers.15 Further classification analysis was used to establish radiographic union threshold values that maximized sensitivity and specificity while minimizing false positives and negatives for both RUST and modified RUST scores.
3. Results
Our database query initially yielded 258 patients with distal femur fractures. After application of our inclusion and exclusion criteria, 138 patients remained for analysis in our study (Fig. 1). Forty-one of these patients had insufficient clinical follow-up (<12 weeks), and 7 additional patients had inadequate radiographic follow-up (<12 weeks) and were thus excluded from the final analysis. Of the 90 patients with adequate clinical and radiographic follow-up, 11 patients (12%) were allowed immediate WBAT or 50% partial weight-bearing, and the remaining 79 patients (89%) were made NWB or TDWB. There were no cases of acute fixation failures (ie, within the first 6 weeks postoperatively) in either group.
;)
FIGURE 1.: Flow chart demonstrating process used for inclusion in the analysis.
At final radiographic follow-up, 78 (87%) achieved clinical fracture union, resulting in an overall nonunion rate of 13% for the entire cohort (Table 1). Female patients comprised 77% of the study population, and the mean age of patients was 68.83 (SD: 13.88) years. The median time to surgery was 2 days (interquartile range [IQR] = 1,3), and the median length of follow-up was 9 months (5,13). No significant difference was detected in nonunion rates between patients with restricted postoperative weight-bearing status and those allowed to weight-bear as tolerated immediately after surgery (P = 0.77). The impact of other previously identified risk factors on nonunion rates in our cohort are also given in Table 1. We identified a statistically significant association between the categorical rigidity score and nonunion in our cohort (P = 0.003, Table 1). However, there was no statistically significant difference in mean rigidity scores between patients who developed nonunion (2.83; SD = 0.94) and those who went on to fracture union (2.64; SD = 1.27, P = 0.62). Similarly, there was no difference in median rigidity scores between patients who achieved fracture union versus nonunion (median [IQR] rigidity scores of 3 [1,4] for union vs. 3 [2,3.5] for nonunion, P = 0.83). In addition, we did not observe a difference in the incidence of nonunion based on other potential risk factors including obesity, fracture type (open vs. closed), presence/absence of infection, or plate type/material (see Table 1).
TABLE 1 -
Study Participant Demographics and Clinical Characteristics Overall and by Union Status.
| Demographics and Clinical Characteristics |
Overall (N = 90) |
Union (N = 78) |
Nonunion (N = 12) |
P
*
|
| Sex |
|
|
|
0.46 |
| Male |
21 (23%) |
17 (22%) |
4 (33%) |
|
| Female |
69 (77%) |
61 (78%) |
8 (67%) |
|
| Age at time of surgery (years), mean (SD) |
68.83 (13.88) |
69.52 (14.65) |
63.50 (8.34) |
0.05 |
| BMI, mean (SD) |
32.43 (9.30) |
32.15 (9.25) |
34.17 (9.88) |
0.49 |
| Time to surgery (days), median (IQR) |
2 [1,3] |
2 [1,3] |
1 [1,2.5] |
0.31 |
| Length of follow-up (months), median (IQR) |
9 [5,13] |
8 [5,12] |
21 [12.5, 30] |
<0.001 |
| Tobacco use |
|
|
|
0.41 |
| Yes |
11 (12%) |
11 (14%) |
0 |
|
| No |
52 (58%) |
45 (58%) |
7 (58%) |
|
| Former |
27 (30%) |
22 (28%) |
5 (42%) |
|
| Postoperative weight-bearing |
|
|
|
0.77 |
| WBAT |
7 (8%) |
7 (9%) |
0 |
|
| 50% |
4 (4%) |
4 (5%) |
0 |
|
| TDWB or NWB |
79 (88%) |
67 (85%) |
12 (100%) |
|
| Obese (n = 87) |
|
|
|
0.75 |
| Yes |
51 (59%) |
43 (57%) |
8 (67%) |
|
| No |
36 (41%) |
32 (43%) |
4 (33%) |
|
| Fracture type |
|
|
|
0.23 |
| Open |
7 (8%) |
5 (6%) |
2 (17%) |
|
| Closed |
83 (92%) |
73 (94%) |
10 (83%) |
|
| Infection |
|
|
|
0.99 |
| Yes |
1 (1%) |
1 (1%) |
0 |
|
| No |
89 (99%) |
77 (99%) |
12 (100%) |
|
| Plate type |
|
|
|
0.31 |
| Stainless steel |
74 (82%) |
64 (82%) |
10 (83%) |
|
| Titanium |
16 (18%) |
14 (18%) |
2 (17%) |
|
| Mean rigidity score (SD) |
2.67 (1.23) |
2.64 (1.27) |
2.83 (0.94) |
0.62 |
| Median rigidity score (IQR) |
3 [1,4] |
3 [1,4] |
3 [2, 3.5] |
0.83 |
| Rigidity score |
|
|
|
0.003 |
| 0 |
1 (1%) |
1 (1%) |
0 |
|
| 1 |
26 (29%) |
25 (32%) |
1 (8%) |
|
| 2 |
3 (3%) |
0 |
3 (25%) |
|
| 3 |
32 (36%) |
27 (35%) |
5 (42%) |
|
| 4 |
28 (31%) |
25 (32%) |
3 (25%) |
|
| 5 |
0 |
0 |
0 |
|
*Fisher exact test used for all categorical analyses because of small, expected cell counts.
Analysis of the ICCs for RUST and modified RUST scores revealed moderate-to-good interobserver reliability, with ICC scores of 0.74 (95% confidence interval [CI]: 0.65–0.81) and 0.73 (95% CI: 0.65–0.8), respectively. Patients who went on to fracture union had significantly higher mean RUST (10.67 vs. 6.53, P < 0.001) and modified RUST (13.47 vs. 6.94, P < 0.001) scores than those who developed nonunion in our cohort (Table 2). Classification analysis of RUST scores identified a threshold of 9 for determining radiographic union, maximizing sensitivity at 93.6% and specificity at 91.7%. Similarly, the modified RUST score analysis revealed a threshold of 8, with a sensitivity of 93.6% and 91.7% (Table 3).
TABLE 2 -
Mean RUST Scores Overall and by Healing Status for Each Rater and Averaged Across all Raters.
| Variable |
Rater |
Overall |
Union (N = 78) |
Nonunion (N = 12) |
P
|
| RUST original |
NS |
9.36 (2.07) |
9.85 (1.67) |
6.17 (1.53) |
<0.001 |
| RUST modified |
NS |
12.09 (3.22) |
13.0 (2.31) |
6.17 (1.53) |
<0.001 |
| RUST original |
RG |
10.49 (2.16) |
11.09 (1.50) |
6.58 (1.68) |
<0.001 |
| RUST modified |
RG |
13.09 (3.4) |
13.94 (2.58) |
7.58 (2.75) |
<0.001 |
| RUST original |
PW |
10.51 (2.29) |
11.08 (1.75) |
6.83 (2.04) |
<0.001 |
| RUST modified |
PW |
12.63 (3.64) |
13.49 (2.97) |
7.08 (2.54) |
<0.001 |
| RUST original |
ALL |
10.12 (1.98) |
10.67 (1.37) |
6.53 (1.48) |
<0.001 |
| RUST modified |
ALL |
12.60 (3.09) |
13.47 (2.20) |
6.94 (1.79) |
<0.001 |
TABLE 3 -
Sensitivity Analyses for Different Thresholds.
| Cut Point |
Overall % Correct Classification |
Sensitivity |
Specificity |
False Positive |
False Negative |
| Using average RUST scores |
|  7 |
92.2% |
94.9 |
75.0 |
3.9 |
30.8 |
|  8 |
92.2% |
93.6 |
83.3 |
2.7 |
33.3 |
|  9 |
93.3% |
93.6 |
91.7 |
1.4 |
31.3 |
|  10 |
87.8% |
87.2 |
91.7 |
1.4 |
47.6 |
| Using average modified RUST scores |
|  7 |
93.3% |
96.2 |
75.0 |
3.8 |
25.0 |
|  8 |
93.3% |
93.6 |
91.7 |
1.4 |
31.3 |
|  9 |
92.2% |
92.3 |
91.7 |
1.4 |
35.3 |
|  10 |
92.2% |
92.3 |
91.7 |
1.4 |
35.3 |
4. Discussion
LLP of distal femur fractures became a popular treatment strategy in the 1990s, with initial reported nonunion rates as low as 0%–6%.16,17 More recently, however, larger studies have reported much higher nonunion rates, stimulating research focused on identifying risk factors that may account for these unfavorable outcomes. Multiple studies have identified independent risk factors for nonunion after LLP of distal femur fractures.7,8 By contrast, very few studies have evaluated postoperative weight-bearing status as a potential risk factor for nonunion after LLP of distal femur fractures. As such, there is no current consensus regarding appropriate postoperative weight-bearing status in this patient population, although the benefits of early weight-bearing (EWB) after operative fixation of geriatric fractures has been shown in the literature.
In a recent retrospective review by Consigliere et al,11 EWB in the setting of distal femur fractures treated with LLP was considered to be safe, resulting in a 0% nonunion rate at 3 months as compared with 8% (2/26) in the NWB group.11 Similarly, a multicenter study by Smith et al12 retrospectively reviewed 105 patients treated for distal femur fractures at 4 separate level-1 trauma centers and found that the EWB group did not have increased perioperative complications compared with patients who were kept NWB or TDWB.12
Our study did not identify a significant difference in nonunion rates or implant/fixation failure based on immediate postoperative weight-bearing status. This result may be attributable to the very low proportion of patients in our cohort (12%) who were permitted at least 50% weight-bearing immediately postoperatively, none of whom developed a nonunion. In addition, we observed a relatively low nonunion rate of 13% in our cohort, which is notably lower than the 19% nonunion rate reported by Ricci et al7 in a cohort of 335 distal femur fractures.7 As detailed in a recent systematic review by Wittauer et al,18 there is considerable variability in the literature regarding the definition of nonunion. We chose to use the pragmatic definition of nonunion as outlined by Rodriguez et al,8,13 as this represents the most objective, clinically relevant definition of nonunion.
Consistent with the results from previous literature,8,19 we identified a significant association between mechanical construct rigidity score and nonunion when rigidity score was analyzed as a categorical variable (ie, whole numbers, 0–5). However, there was no statistically significant difference in median or mean rigidity scores between patients who went on to clinical union and those who developed nonunion. This is likely related to the fact that a rigidity score of 3 was the most common score for patients who went on to union and those that developed nonunion (representing 35% and 42% of those cohorts, respectively). Similarly, the difference in mean rigidity scores for the 2 cohorts (2.64 for union, 2.83 for nonunion) did not reach statistical significance, likely related to the relatively small size of the nonunion cohort. The analysis of rigidity scores as a categorical variable was undertaken to investigate this variable with greater specificity. Our analysis did not identify any significant differences in nonunion rates in association with other previously identified risk factors, including obesity, fracture type, infection, and plate material. This is likely due to the smaller size of our study cohort (N = 90) compared with the studies by Rodriguez et al8 (N = 283) and Ricci et al7 (N = 335), thus making subgroup analysis more difficult. In addition, our study did not include any intraarticular fracture patterns, further narrowing the subset of patients included in our analysis.
Our study is among the first to use the RUST and modified RUST scores to evaluate fracture union in distal femur fractures. These radiographic scoring systems, originally developed and validated to assess union in tibial shaft fractures stabilized with an intramedullary nail,14,20 have recently been used to evaluate fracture union in metadiaphyseal fracture patterns.9 Our results demonstrated significantly higher mean RUST and modified RUST scores in patients who went on to achieve clinical union compared with those who developed nonunion. These results lend further support to the utility of the RUST and modified RUST scores for metadiaphyseal fractures. Furthermore, we identified moderate-to-good interrater reliability between observers for both the RUST and modified RUST scores in our study cohort. Only one other published study has used modified RUST scores to assess radiographic healing in distal femur fractures treated with locked plating, noting moderate interobserver agreement with ICC = 0.59.21
Strengths of our study include the specificity of the study cohort, which was limited to patients with distal femur fractures treated with LLP. This represents a clinically challenging fracture population with a relatively high nonunion rate. Correlating a validated radiographic score with a clinically relevant definition of nonunion adds to the generalizability of our study. Another key strength of this study is the inclusion of a classification analysis, which determined thresholds of 9 and 8 for the RUST and modified RUST scores, respectively. These thresholds maximized sensitivity and specificity while minimizing false positives and false negatives (see Table 3). We felt it important to prioritize minimizing the false positive rate to make these thresholds as clinically useful as possible. One other strength of our study is the use of the previously published rigidity score, which was correlated with nonunion in our cohort.
Our study suffers from the limitations inherent to a retrospective study. Because we did not use pressure sensors or another objective measurement of actual weight-bearing, we cannot be certain about patient adherence to surgeon-dictated weight-bearing regimens apart from what was documented in the electronic medical record. In addition, the relatively low overall nonunion rate of our cohort (13%) left us with only 12 cases of nonunion, making subgroup analysis challenging. In addition, the low proportion of patients who were allowed to initiate immediate weight-bearing (12% of all patients) resulted in 2 relatively uneven groups, thus compromising our power to detect potential differences. Finally, the RUST and modified RUST scores are somewhat difficult to measure in fractures stabilized with a locking plate construct. Depending on the size and position of the plate, an accurate determination of callus can be difficult because the plate itself often partially obscures one or more cortices (particularly on the lateral view). Finally, patients treated by multiple surgeons were included in the study. While this improves the generalizability of the study, it also introduces some heterogeneity in reduction techniques used and implant selection.
In conclusion, immediate postoperative weight-bearing status did not seem to affect nonunion rates in our cohort of distal femur fractures treated with LLP. These findings support other published studies documenting the safety of immediate postoperative weight-bearing in this patient population. Our study also adds to the existing literature documenting an increased risk of nonunion with more rigid fixation constructs for distal femur fractures. Finally, our study further supports the utility of the RUST and modified RUST scores for determining clinically relevant union in patients with distal femur fractures treated with LLP. To determine the true impact of immediate postoperative weight-bearing on union rates in distal femur fractures, prospective studies with greater statistical power, more evenly matched groups, and objective measurements of actual weight-bearing are needed.
References
1. Arneson TJ, Melton LJ 3rd, Lewallen DG, et al. Epidemiology of diaphyseal and distal femoral fractures in Rochester, Minnesota, 1965-1984. Clinic Orthop Relat Res. 1988:234:188–194.
2. Court-Brown CM, Caesar B. Epidemiology of adult fractures: a review. Injury. 2006;37:691–697.
3. Elsoe R, Ceccotti AA, Larsen P. Population-based epidemiology and incidence of distal femur fractures. Int Orthop. 2018;42:191–196.
4. Martinet O, Cordey J, Harder Y, et al. The epidemiology of fractures of the distal femur. Injury
. 2000;31:C62–C63.
5. Henderson CE, Lujan TJ, Kuhl LL, et al. 2010 mid-America Orthopaedic Association Physician in Training Award: healing complications are common after locked plating for distal femur fractures. Clinic Orthop Relat Res. 2011;469:1757–1765.
6. Lujan TJ, Henderson CE, Madey SM, et al. Locked plating of distal femur fractures leads to inconsistent and asymmetric callus formation. J Orthop Trauma. 2010;24:156–162.
7. Ricci WM, Streubel PN, Morshed S, et al. Risk factors for failure of locked plate fixation of distal femur fractures: an analysis of 335 cases. J Orthop Trauma. 2014;28:83–89.
8. Rodriguez EK, Boulton C, Weaver MJ, et al. Predictive factors of distal femoral fracture nonunion after lateral locked plating: a retrospective multicenter case-control study of 283 fractures. Injury
. 2014;45:554–559.
9. Litrenta J, Tornetta P 3rd, Mehta S, et al. Determination of radiographic healing: an assessment of consistency using RUST and modified RUST in metadiaphyseal fractures. J Orthop Trauma. 2015;29:516–520.
10. Carlin L, Sibley K, Jenkinson R, et al. Exploring Canadian surgeons' decisions about postoperative weight bearing for their hip fracture patients. J Eval Clin Pract. 2018;24:42–47.
11. Consigliere P, Iliopoulos E, Ads T, et al. Early versus delayed weight bearing after surgical fixation of distal femur fractures: a non-randomized comparative study. Eur J Orthop Surg Traumatol. 2019;29:1789–1794.
12. Smith JR, Halliday R, Aquilina AL, et al. Distal femoral fractures: the need to review the standard of care. Injury. 2015;46:1084–1088.
13. Rodriguez EK, Zurakowski D, Herder L, et al. Mechanical construct characteristics predisposing to non-union after locked lateral plating of distal femur fractures. J Orthop Trauma
. 2016;30:403–408.
14. Whelan DB, Bhandari M, Stephen D, et al. Development of the radiographic union score for tibial fractures for the assessment of tibial fracture healing after intramedullary fixation. J Trauma
. 2010;68: 629–632.
15. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics
. 1977;33:159–174.
16. Kolb W, Guhlmann H, Windisch C, et al. Fixation of distal femoral fractures with the Less Invasive Stabilization System: a minimally invasive treatment with locked fixed-angle screws. J Trauma
. 2008;65:1425–1434.
17. Kregor PJ, Stannard JA, Zlowodzki M, et al. Treatment of distal femur fractures using the less invasive stabilization system: surgical experience and early clinical results in 103 fractures. J Orthop Trauma. 2004;18:509–520.
18. Wittauer M, Burch MA, McNally M, et al. Definition of long-bone nonunion: a scoping review of prospective clinical trials to evaluate current practice. Injury. 2021;52:3200–3205.
19. Bottlang M, Doornink J, Fitzpatrick DC, et al. Far cortical locking can reduce stiffness of locked plating constructs while retaining construct strength. J Bone Joint Surg Am. 2009; 91: 1985–1994.
20. Bhandari M, Guyatt GH, Swiontkowski MF, et al. A lack of consensus in the assessment of fracture healing among orthopaedic surgeons. J Orthop Trauma. 2002;16:562–566.
21. Plumarom Y, Wilkinson BG, Marsh JL, et al. Radiographic healing of far cortical locking constructs in distal femur fractures: a comparative study with standard locking plates. J Orthop Trauma. 2019;33:277–283.
