Introduction

Intertrochanteric femur fractures account for 42% of all proximal femur fractures [1]. The majority of younger trauma patients sustain a proximal femur fracture due to a high-energy trauma [2]. This trauma mechanism is usually associated with additional injuries and leads to more complex fracture morphology that includes an increased risk of osteonecrosis and impaired fracture healing [3, 4]. Young male adults and elderly patients are at increased risk for intertrochanteric femur fractures [5]. These populations represent two different entities of patients, with their own risk factors and comorbidities, and are therefore worth looking at separately. Intertrochanteric proximal femur fractures are known for high mortality rates, increased complications and are usually accompanied by comorbidities such as osteoporosis and diabetes mellitus (DM) [6, 7]. Impaired healing is a feared complication with devastating outcomes for each patient [8, 9]. Two major treatment options exist: extramedullary and intramedullary implants [10, 11]. A substantial increase in the use of intramedullary nails (from 3% up to 67%) during the past two decades has been observed in the United States and Europe [12]. Previous studies have shown that most treatment failures of fractures of the proximal femur occur in unstable fracture patterns, after mal-reduction of the posteromedial cortex or in patients with reverse oblique fractures [13,14,15,16,17]. Most studies focus on geriatric fracture fixation and assess well known risk factors, such as infections or low-grade infections and advanced age for delayed fracture healing [13, 18, 19]. However, risk factors, others than infections and advanced age, for impaired healing of proximal femur fractures in non-geriatric adults are marginally assessed. This study aims to identify non-infectious related risk factors for impaired fracture healing of proximal femur fractures in non-geriatric trauma patients.

Methods

Ethical consideration

This retrospective cohort study was approved by the Cantonal Ethics Committee, Canton of Zurich review board (Basec No.:2020–00703). A written consensus of data collection was obtained from all patients during hospitalization.

Study population

Inclusion criteria

This study included non-geriatric patients (aged 69 years and younger) who were treated between 2013 and 2020 at one academic Level 1 trauma center due to a proximal femur fracture (PFF). The arbitrary cut-off of 70 years was chosen by our in-hospital protocol for geriatric co management. The geriatric co-management is the interdisciplinary collaboration of geriatricians and trauma surgeons. The interdisciplinary consensus has decided to set the cut-off age to 70 years [20]. Only patients that had a complete follow up of at least one year after surgery where included.

Exclusion criteria

Patients with genetic disorders affecting the musculoskeletal system, patients with oncologic diseases regarding the proximal femur, such a pathological fractures, fractures not classifiable as intertrochanteric fractures with the AO-classification and patients with postoperative infections were excluded from this study. Postoperative infections represent one of the major risk factors for impaired fracture healing. Numerous studies have investigated this association [21,22,23]. The goal of this study was to analyze other risk factor, that are not related with postoperative local infection, as defined as non-infection related. Therefore, these patients were excluded from the present study. Patients with additional ipsilateral lower limb fractures, open injuries, associated nerve and vascular injuries, and pelvic injuries were excluded. Patients requiring a staged procedure or primary plate fixation were also excluded from this study.

Treatment protocol

All patients were treated surgically with a femoral nail, using the institutional standard implant a Gamma 3 nail (Stryker). The standard approach in our institute is a closed reduction under radiological control and internal fixation (CRIF) with the patient in a supine position on a traction table. If the reduction is inadequate in either the anterior–posterior or the lateral view, the treating surgeon decides for open reduction and internal fixation (ORIF). In both cases, the entry point for the intramedullary implant is the tip of the trochanter. In cases where sufficient reduction cannot be maintained, cerclage wires (cable cerclage, 1.6 mm in diameter) may be used at the discretion of the treating surgeon. Postoperative rehabilitation includes full weight bearing, daily physiotherapeutic training, and optimized medical treatment. The follow-up appointments were performed after 6 weeks, 12 weeks, 6 months and 12 months. Radiographic imaging to assess the state of fracture healing was obtained at each appointment, focusing on persistent fracture lines, insufficient bone bridging, progressive deformity and the presence/absence of broken implants.

Outcomes and definitions

The primary outcome of this study was impaired fracture healing. Impaired fracture healing was subsequently specified as delayed union and nonunion. Per in-house protocol, the present study defined delayed union as callus formation on 3 out of 4 cortices after 3 to 6 months, in the absence of secondary fracture or material loosening. Nonunion was defined in cases of lack of callus-formation after 6 months, or the requirement of revision surgery. Equivalent definitions have been described in the literature [24]. The PFF was classified using the AO-Classification. Further, only fractures that where classifiable as intertrochanteric fractures with the AO-System for proximal femur fractures, 31 A1-3, were included [25]. Polytrauma was defined as an injury severity score (ISS) of 16 points or higher.

Statistical analysis

Continuous variables are summarized as a mean with standard deviation (± SD). Categorical variables are displayed as count and percentages. Two groups of continuous variables were compared using the students t-test. For groups of binary variables, the chi-square test was used. ANOVA was used when comparing more than two groups. A p-value below 0.05 was considered statistically significant. Regression analysis was performed with the primary outcome being impaired fracture healing. Multivariate regression analysis included variables of clinical significance, or risk factors that are associated with increased complication rates. All analyses were performed using R (R Core Team (2019). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria.)

Results

This study included 150 non-geriatric trauma patients at a mean age of 54.15 ± 12.52 years. 45 (30.0%) patients were female. The leading comorbidities were arrhythmia that required oral anticoagulation drugs (OAK) (N = 38, 25.3%, most of these patients aged 50 years and above N = 33, 87.1%), followed by osteoporosis (N = 22, 14.7%, most of these patients aged 50 years and above N = 21, 95.5%). The overall length of stay (LOS) was 11.27 ± 9.16 days (Table 1). The most common fracture type, was AO 31A1 (N = 61, 40.7%), followed by AO type 31A2 (N = 48, 33.0%). Delayed union was observed in 32 (21.3%) patients and nonunion with subsequent revision surgery occurred in 14 (9.3%) (Table 2).

Table 1 Patients demographics and fracture classification
Full size table
Table 2 Fracture classification and impaired fracture healing
Full size table

ORIF and Cerclage usage, were significantly more often used in AO 31 A3 type fractures (N = 27, 65.9%) and (N = 16, 39.0%) when compared to AO 31 A1 (N = 4, 6.6%) / (N = 2, 3.3%) and AO 31 A2 (N = 18, 37.5%) / (N = 15, 31.2%). With an increasing fracture classification (31 A1 up to 31 A3 type fractures), there was a significantly higher rate of delayed union: AO 31 A1 (N = 6, 9.8%), AO 31 A2 (N = 12, 25.0%) and AO 31 A3 (N = 14, 34.1%), (p ≤ 0.01, Table 3). Patients with delayed union had a significantly higher rate of ORIF (n = 18, 56.2%) when compared with patients without delayed union (n = 31, 26.3%, p = 0.003). Further, patients with diabetes had higher rates of delayed union (n = 7, 21.9%) when compared with non-diabetic patients. Both AO type 31A2 and 31A3 fractures were associated with an increased risk for the development of delayed union (95% CI 1.05 to 8.87, p = 0.04; 95% CI 1.64 to 13.73, p = 0.004) when compared with AO type 31A1 fractures. Additionally, ORIF (OR 6.17, (95% CI 1.54 to 24.70, p ≤ 0.01)) and DM (OR 5.74, (95% CI 1.39 to 23.72, p = 0.016)), were independent risk factors for delayed union (Table 4). The rate of nonunion was independent from fracture morphology, patient characteristics or comorbidities (Table 5).

Table 3 Distribution according to the AO-classification
Full size table
Table 4 Delayed union univariate and multivariate regression analysis
Full size table
Table 5 Nonunion univariate and multivariate regression analysis
Full size table

Discussion

Intertrochanteric proximal femur fractures are associated with increased mortality, increased complications and often comorbidities such as osteoporosis and DM. This study aimed to assess non-infectious related risk factors for delayed union and nonunion of PFF in non-geriatric trauma patients and found the following points:

  1. 1.

    AO 31 A3 type fractures have an increased rate of impaired healing

  2. 2.

    Additional injuries are not associated with delayed union or nonunion

  3. 3.

    ORIF and DM are independent risk factors for delayed union, but not for nonunion.

Certain risk factors for impaired bone healing include local infections, advanced age, multiple comorbidities, smoking, non-steroidal anti-inflammatory usage, sex, metabolic disease and nutritional deficiencies [26,27,28]. Further, fracture pattern and location, degree of displacement, severity of soft tissue injury, quality of surgical treatment, degree of bone loss and presence/absence of infection have also been reported as risk factors for impaired bone healing [29]. Intertrochanteric fractures are known to have an excellent blood supply and good cancellous bone in the intertrochanteric region; therefore, nonunion is uncommon [30]. Generally, intertrochanteric fractures treated with internal fixation heal [9, 18]. It is well known, that failure of fracture fixation and nonunion can occur due to poor bone quality, unfavorable fracture patterns, suboptimal internal fixation and delayed treatment of patients [13, 14, 31,32,33]. The most common fracture types for nonunion of the proximal femur are 31 A 1.1, 2.3 and 3.3 [33, 34]. These findings were comparable with the findings in this study. The reported incidence of intertrochanteric fractures resulting in nonunion ranges between 1–2% [35]. The investigated proximal fractures caused by high energy trauma in non-geriatric populations are known to be more difficult to reduce and more prone to complications [36, 37]. These finding could be an explanation for the deviation from our results.

The present results show that delayed union rate but not non-union was associated with fracture morphology. In younger trauma patients, it may therefore be appropriate to wait longer before initiating revision surgeries, even in cases of clinical or radiological suspicion of delayed fracture healing. As definitions of delayed union and nonunion are vague, there is no agreement on the exact timing when the diagnosis should be made [38]. Various studies come to different timeframes regarding the diagnosis of nonunion, with a variety between 6–8 weeks, 3 months and up to 6 months [39,40,41]. Since the clinical evaluation of fracture healing is a combination of both radiographic and clinical findings [42], patients with unremarkable clinical and radiological findings at 6 – 12 weeks postoperatively, and similar results for the rest of the consultations, were considered cured/healed.

This study showed that additional injuries seemed not to be associated with impaired bone healing. In femoral shaft fracture, ISS has been reported to not be associated with nonunion [43]. Controversially, other studies found nonunion in long bone fractures as a common problem in multiply injured patients [44,45,46,47,48]. Early final stabilization of long bone fractures during the first 24 h after the initial trauma has been shown to be the best treatment option for major fractures [49,50,51]. In this study, ORIF and DM serve as independent risk factors for impaired bone healing in non-geriatric trauma patients. Comparable findings are known for subtrochanteric femur fractures with risk factors including chronic diseases such as diabetes mellitus [52]. Nonunion is reported to occur in 95% of cases with unstable fractures with loss of medial calcar continuity following ORIF [31]. Our data showed that nonunion had no association to the reduction technique used and was therefore independent of ORIF or CRIF.

A systematic review further concluded that the reduction technique is not associated with nonunion in femoral neck fractures [53]. Conversely, it was been well studied that minimally invasive surgeries of the femur had shorter healing times and lower re-operation rates [54,55,56]. Our results may indicate that the 31A3 fracture types require special surgical attention for a beneficial outcome. The focus should be on an optimal surgical reduction, which can be achieved by an ORIF and the support of cerclage wires, if necessary. The femoral subtrochanteric region has a critical blood supply and is mainly composed of cortical bone [57]. Lack of medial cortical support, varus malreduction and residual displacement after reduction have been described as potential risk factors for nonunion of subtrochanteric femur fractures [58,59,60].

Strength & limitations

The retrospective design provides certain well-known limitations. One might argue that this study has an increased risk for type 2 error. We believe, that based on the standardized treatment protocol and the comparability of the study groups, the presented results provide some evidence for the identification of risk factors for nonunion. There is a lack of standardized quantification methods for the assessment of the quality of fracture reduction in the clinical setting. Therefore, documentation of the quality of fracture reduction is missing. This important factor is incorporated in our in-hospital protocol by utilizing the best-possible reduction prior nail fixation on the surgeons’ best judgement.

Occult infection is a potential cause of implant failure of a proximal femur fracture and should be considered [61]. In this cohort, there were no infections diagnosed in patients with nonunion or who required revision surgery, since patients with postoperative infections where excluded from this study. In addition, very limited information on tobacco consumption was available at the time of data analysis. The known negative effect of smoking on fracture healing could therefore not be taken into account in this study.

Conclusion

In this study, increasing fracture complexity, ORIF, and diabetes were found to be associated with delayed union of intertrochanteric femur fractures in non-geriatric patients. These variables, however, did not show a significant association with the development of non-union.