Nonunion of Distal Femoral Fractures: a Systematic Review

Abstract

Nonunions of the distal femur are inherently difficult to treat and associated with significant patient disability. An extensive search of published reports was conducted and identified articles reviewed to determine the prevalence of various fracture patterns, initial treatment methods, reported risk factors, and definitive treatment methods. The definitive treatment modalities for distal femoral nonunions were then analyzed according to union rate, time to union and complications. The purpose of this review was to determine the risk factors that predispose to nonunion of the distal femur and the most effective treatment methods to implement after nonunion has occurred. The current review assesses 19 studies published over the last 12 years in which there were patients with nonunion of the distal femur and treatment specific to those patients was described. It was found that the most common fracture pattern involved metaphyseal comminution, most common initial treatment was open reduction and internal fixation with plating, most common reported risk factor was an open fracture, and most common definitive treatment was fixed angle plating combined with cancellous autografting. Taken together, the reviewed reports suggest that this form of treatment has a successful union rate of 97.4% and average time to heal of 7.8 months.

Introduction

Distal femur fractures occur relatively commonly. The vast majority of these fractures are caused by either high‐energy trauma in young patients or low‐energy falls in elderly patients with concomitant osteoporosis. Precise reduction and fixation of these fractures, with enough stability to enable prompt movement and rehabilitation, is crucial. In current practice, distal femur fractures are commonly treated with open reduction and internal fixation (ORIF) using locking plates, condylar screws, blade plates or intramedullary nails (IMN)1, 2. Nonunions of the distal femur are rare; however, those that do occur have a significant influence on patient quality of life and represent a considerable treatment challenge because of the associated bone loss and soft tissue scarring3. Poor bone quality or unfavorable fracture patterns are commonly associated with nonunions in this part of the femur4.

Pain‐free weight‐bearing and formation of bridging callus are the usual clinical and radiographic evidence, respectively, of fracture union. Consequently, typical criteria for diagnosis of nonunion are pain over the fracture site on weight‐bearing and no visible progressive radiologic evidence of healing for three months when a minimum of six months has elapsed since fracture2.

There are numerous published reports regarding management of nonunions of the distal femur; however, there is no recent review on this topic. For the purposes of this study, data were collected from published reports specifically about patients with nonunions of the distal femur to analyze the prevalence of various fracture patterns, initial treatments, reported risk factors, definitive treatment methods, and outcomes of those treatment methods. The purpose of this review was to identify treatment factors that produce the best outcomes, based on recent published reports.

Methods

The “PubMed” and “Google Scholar” electronic databases were searched for articles containing the terms “distal femur nonunion” and “nonunion of supracondylar femur fracture” published from September 2000 to June 2012. All studies identified by these searches were then reviewed. Exclusion criteria were defined to help focus the search and identify the articles that would assist attainment of the stated objectives. First, articles were excluded if the definitive treatment method used was not surgical and if it was not possible to determine the specific treatment method. Next, studies were excluded if they included periprosthetic fractures or fractures other than those of the distal femur because inclusion of such fracture types made analysis difficult. Studies in languages other than English and those with animal or pediatric subjects were also excluded.

In all, 19 peer reviewed articles passed all the stated exclusion criteria and together provided data on 169 nonunions of the distal femur for the present study3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21. These articles consisted of eight retrospective case series, one prospective case series, six one‐case reports and four two‐case reports.

Data collected included sex, age, mechanism of injury (MOI), whether open or closed, fracture patterns, time to diagnosis of nonunion, time to union, initial treatments, definitive treatments, outcomes, revisions, and complications.

The 19 articles provided 169 nonunions in 77 men and 92 women with an average age of 56.7 years. There were 133 closed and 36 open fractures. The MOI was high‐energy trauma in 65, falls in 49, gun‐shot wounds in two, a car versus pedestrian accident in one, an insufficiency fracture in one, an osteotomy in one, a pathologic fracture in one, and unknown in 49. The initial treatment was ORIF with plating in 108, ORIF with IMN in 24, external fixation in 13, multiple Steinmann pins in one, allograft in one, plate and cast in one, non‐operative in one, and unknown in 20.

Reported risk factors in patients with nonunion were open fractures in 37, bone loss in 25, broken implants and/or screws in 19, infections in 16, failed fixations in 13, loss of reduction in two, a stiff construct in one, and excessive shearing in one. The definitive treatments were ORIF with plating in 83, ORIF with IMN in 36, external fixation in 28, insertion of megaprostheses in 10, IMN and plate in seven, low intensity pulsed ultrasound (LIPUS) in two, “dynamization” of a plate in one, and osteotomy with plating in one. Bone grafts were used in 104 patients. Table 1 summarizes the reviewed articles. Data were charted for individual patients.

Table 1.

Data from reviewed reports

References	Year	Cases (n)	Male	Female	Mean age (years)	Level of evidence	Type of report
Ali and Saleh5	2002	15	9	6	35.4	IV	Retrospective case series
Alt et al.6	2007	1	1	0	49	V	Case report
Bellabarba et al.7	2002	20	12	8	48	IV	Prospective case series
Chen et al.8	2011	1	1	0	23	V	Case report
Davila et al.9	2001	2	0	2	68	V	Case report
Gardner et al.3	2008	31	9	22	57.6	IV	Retrospective case series
Graves et al.10	2005	1	0	1	55	V	Case report
Haidukewych et al.11	2003	22	8	14	55	IV	Retrospective case series
Hailer and Hoffmann12	2006	1	0	1	82	V	Case report
Oh et al.13	2011	1	0	1	58	V	Case report
Pao and Jiang14	2005	2	2	0	72	V	Case report
Saridis et al.15	2006	13	10	3	34.6	IV	Retrospective case series
Tis et al.16	2005	1	0	1	68	V	Case report
Vaishya et al.4	2011	8	5	3	74	IV	Retrospective case series
Wang and Weng17	2003	13	2	11	60	IV	Retrospective case series
Waseem et al.18	2010	2	0	1	77	V	Case report
Wu19	2011	13	11	2	36	IV	Retrospective case series
Wu20	2011	20	5	15	77	IV	Retrospective case series
Yoshida et al.21	2009	2	2	0	48	V	Case report
Total		169	77	92

Results

Fracture Patterns

Of the 169 nonunions reviewed, 133 (78.7%) were in patients with closed fractures and 36 (21.3%) in patients with open fractures. Of the 161 for whom the type of fracture was reported, 130 (80.7%) were Orthopaedic Trauma Association (OTA) 33A1, 18 (11.2%) metaphyseal comminution fractures OTA 33A3, and nine (5.6%) OTA 33C1 (not metaphyseal comminution fractures), whereas four (2.5%) articular comminution fractures were OTA 33C3. Of the 21 open fractures classified according to the Gustillo‐Anderson (GA) fracture classification, 11 (52.4%) were GA IIIB, four (19.0%) GA IIIC, three (14.3%) GA IIIA and three (14.3%) GA II.

Initial Treatment

Of the 169 cases reviewed, the initial treatment was ORIF with plating in 108 (63.9%), IMN in 24 (14.2%), external fixation in 13 (7.7%) and unknown in 20 (11.8%). The remaining four initial treatments included one instance of multiple Steinmann pins, one allograft, one plate and cast and one non‐operative. Of the 169 nonunions, the specific initial treatments were dynamic condylar screws (DCS) in 21 (12.4%), anterograde IMN (AIMN) in 18 (10.7%), angled blade plates (ABP) in 16 (9.47%), retrograde IMN (RIMN) in six (3.55%), and locking compression plates (LCP) in five (2.96%). In addition, 30 (17.8%) received bone grafts during the initial treatment.

Reported Risk Factors in Patients with Nonunion

Of the 113 reported risk factors for nonunion in the cases reviewed, 36 (31.9%) were open fractures, 25 (22.1%) bone loss, 19 (16.8%) broken implants and/or screws, 16 (14.2%) infection, 13 (11.5%) failed fixation caused by union failure that did not result from hardware malfunction, two (1.8%) loss of reduction caused by fractures that had been initially surgically reduced but at follow‐up had lost their reduction, one (0.9%) a stiff construct, and one (0.9%) excessive shearing upon weight‐bearing.

Definitive Treatment

Of the 169 recorded cases of distal femoral nonunion reviewed for whom the type of definitive treatment was provided, 104 (61.5%) had been augmented with bone grafting, 78 (45.2%) had fixed angle plating, 35 (20.7%) IMN, 28 (16.6%) external fixation, 10 (5.9%) megaprostheses, seven (4.1%) IMN combined with plating, six (3.6%) condylar buttress plating, two (1.2%) LIPUS, one (0.6%) “dynamization” of a plate, one (0.6%) osteotomy with plating, and one (0.6%) IMN with recombinant human bone morphogenetic protein 2. “Dynamization of a plate‐bone construct” entails removal of the lag screw to decrease the stability and rigidity of the fixation.

Of the 169 recorded cases of distal femoral nonunion reviewed, 164 (97%) healed in an average time of 9.86 months. In all, 86 complications occurred (Table 2), of which 26 were pin site infections, 23 malunions, 13 persistent pain at the nonunion site, 11 significantly decreased range of motion, five persistent nonunion, two wound necrosis, two re‐fracture, one significant soft‐tissue swelling, one deep venous thrombosis, one symptomatic hardware requiring removal, and one avascular necrosis.

Table 2.

Definitive treatment

References	Cases (n)	Methods	Bone graft	Union	Union time (months)	Pain at nonunion site (n)	Number of prior operations (mean)	Number of complications	Additional surgery (n)
Ali and Saleh5	15	EF	0	14	10.4	No	3.3	15	Yes (5)
Alt et al.6	1	IMN+rhBMP‐2	1	1	7	No	2	1	No
Bellabarba et al.7	20	P	9	20	3.5	Yes (2)	1	4	Yes (1)
Chen et al.8	1	P	1	1	15	No	2	0	No
Davila et al.9	2	Megaprosthesis	0	2	—	No	4.5	0	No
Gardner et al.3	31	P	31	30	16.1	Yes (1)	1	21	Yes (1)
Graves et al.10	1	P	1	1	6	No	2	0	No
Haidukewych et al.11	22	P‐18/IMN‐4	21	21	—	Yes (1)	—	3	No
Hailer and Hoffmann12	1	P	1	1	48	No	1	1	No
Oh et al.13	1	Dynamization	0	1	6	No	1	0	No
Pao and Jiang 14	2	IMN	1	2	5	Yes (1)	1	2	No
Saridis et al.15	13	EF	2	13	—	No	3	10	Yes (10)
Tis et al.16	1	Osteotomy+P	1	1	9	No	2	1	No
Vaishya et al.4	8	Megaprosthesis	0	8	—	No	2	3	No
Wang and Weng17	13	P‐10/IMN‐2/B‐1	13	13	5	No	1.77	4	Yes (1)
Waseem et al.18	2	LIPUS	0	2	2	No	4	0	No
Wu19	13	IMN‐7/P + IMN‐6	0	13	4.5	Yes (2)	1.73	0	No
Wu20	20	IMN	20	18	4.9	Yes (6)	1.5	6	No
Yoshida et al.21	2	P	2	2	5.5	No	2	0	No
Total	169		104	164

EF, external fixation; IMN, intramedullary nail; LIPUS, low intensity pulsed ultrasound; P, plate.

Of the 169 recorded cases of nonunion reviewed for whom the type of specific definitive treatment was provided, 33 (19.5%) underwent RIMN, three (1.8%) AIMN, 29 (17.2%) ABP, 10 (5.9%) DCS, and three (1.8%) LCP. Bone grafts were used 128 times in the 104 patients who underwent bone grafting. Of these, 98 (76.6%) were autografts, 20 (15.6%) were allografts, nine (7.0%) were demineralized bone matrix and one (0.8%) was bone morphogenetic protein.

Discussion

Nonunions represent a relatively rare subset of complications following distal femur fractures. However, they are often associated with substantial disability, including pain and decreased knee function. In addition, nonunions are often accompanied by bone loss, poor bone quality, extensive soft tissue scarring and unfavorable fracture patterns. Presently, published reports describe use of a variety of implants with or without bone grafting, substantiating the absence of consensus about the optimal treatment for nonunions of the distal femur.

Gardner et al. reported that nonunions of the distal femur occur most often after open and extensively comminuted fractures3. According to our review, the most common fracture pattern associated with nonunion is metaphyseal comminution. Our data substantiate the contributions of instability and disruption of blood flow to development of nonunion; both of these factors are associated with highly comminuted fractures. Panagiotis reported that the blood supply is vital to sustaining a favorable biological environment, whereas stability provides favorable mechanics22.

Traditionally, RIMN or locking plate osteosynthesis have been the two major treatment techniques used for distal femur fractures. Accurate preoperative planning, combined with a thorough understanding of fracture patterns and the advantages and disadvantages of each type of implant, is essential for both methods. Hierholzer et al. found no difference in outcomes between these two types of implants in terms of fracture healing, rate of nonunion and subsequent infection23. The findings of our review are consistent with this data; both RIMN and locking plates had a low incidence of subsequent development of nonunion.

Kayali et al. note that more recently locking plates have emerged as the most frequently used method for treating fractures of the distal femur, particularly when the involved bone is of poor quality, namely osteoporotic or comminuted24. One of the reported benefits of locking plates in the distal femur is that they confer better stability than do traditional plates. Our review is consistent with this notion; dynamic condylar screws and angled blade plates were associated with a higher incidence of nonunion than were locking plates. Scolaro and Ahn found that locking plates are ideal for highly comminuted fractures, for which IMNs cannot provide sufficient reduction and stabilization25. Data from Bellabarba et al. support this finding; in their study fractures in 18 patients with metaphyseal comminution and three with articular comminution all healed after treatment with plating7. Lastly, locking plates are amenable to minimally invasive insertion, which helps preserve the vascular supply to the fracture site.

Despite the aforementioned benefits, Kubiak et al. reported that locking plates pose a risk of becoming too rigid, thus inhibiting fracture healing26. Henderson et al. also noted that in published reports distal femoral nonunion rates vary between 0% and 10% after treatment with locking plates1. According to our review, the most common initial treatment reported in patients with nonunion is ORIF with plate fixation. This could simply be due to the high incidence of use of plates to treat the original distal femur fractures in the reports reviewed. In addition, there was significant diversity in the types of plate used. Henderson et al. reported that overly rigid fixation with locking plates can reduce motion at the fracture site1, which may suppress callus formation1, 2. However, in our review, only five patients with nonunion were initially treated with locking plates.

Inadequate stabilization, insufficient blood supply, poor bone quality, troublesome fracture patterns and infection have all been linked to increased risk of nonunion. Lynch et al. noted that poor vascularity can occur in association with open fractures, periosteal destruction or gap defects in the affected bone27. According to our review, the most frequently reported risk factor in patients with nonunion is open fractures. Open fractures are associated with relatively high rates of infection and disruption of the vascular supply, both of which may contribute to nonunion. In the past, ORIF was associated with nonunion because it involved substantial stripping of the periosteum and damage to the bone and surrounding soft tissue. However, with the advent of minimally invasive plate osteosynthesis, there has been a decrease in complication rates28, 29, 30, 31.

When treating nonunions of the distal femur, the ultimate goal is to achieve stable reduction while simultaneously preserving sufficient blood flow. As noted by Chan et al., various treatment options exist and they are somewhat dependent on the initial treatment of the distal femur fracture32. To illustrate, Nadkarni et al. reported successful treatment of nonunions following IMN with exchange nailing33. Other procedures used in this particular situation include “dynamization” of an IMN, augmentation with locking compression plates and bone grafting.

Despite the variety of therapeutic options, Hakeos et al. mention that there are currently no standard guidelines for the treatment of distal femoral nonunions34. According to our review, the most common definitive treatment employed is fixed angle plating combined with cancellous autografting. Using this method, bony union was achieved in 75 of 77 patients (97.4%). Of those articles reporting more than two cases, fixed angle plating with bone grafting healed in the shortest average time (7.8 months). In a review of nonunions around the knee joint, Chan et al. found that most nonunions of the distal femur can be reliably treated with a fixed angle device, lag screws and bone grafting32. There are few published reports regarding functional outcomes after treatment of distal femoral nonunions. However, an article by Gardner et al. demonstrated predictable functional outcomes in 31 patients treated with fixed angle implants, lag screws and bone graft supplementation3. They reported an improvement from 43.0 to 78.3 in Knee Society Rating System (KSRS) Knee Assessment subsection scores. Likewise, KSRS Knee Function subsection scores had improved from 11.1 to 61.2 by final follow‐up. The overall rate of successful bony healing for all methods was 97.0%.

The primary goal of treatment of nonunion is to provide an environment in which new bone can be created by establishing and maintaining mechanical stability and biological conditions that are conducive to new bone formation. In distal femur fractures, this can be accomplished through a large variety of types of fixation with or without the addition of bone grafts. Selection of a particular technique is based on the type of nonunion and whether or not alignment of the fragments is adequate. According to an article by Panagiotis, instability is the primary cause of hypertrophic nonunions, which are therefore best treated with stable fixation with minimal vascular disturbance22. On the other hand, insufficient bone forming cells or blood supply, being the main cause of atrophic nonunions, often necessitate resection of the nonviable bone and fibrous tissue combined with bone grafting.

This review is limited by a lack of published randomized control studies. Most of the data collected was from level V case reports. Data reporting varied among the authors. Where some authors reported nonunion results for individual patients, others reported results for the studied group as a whole. Methods of evaluation also varied among the authors. In particular, the functional outcome measures that were used differed significantly in each article.

Nonunions of the distal femur are unusual complications for which multiple treatment methods are available. The data collected for this report indicate that treatment with ORIF using a fixed angle plate and supplemental autograft produces the best results. Depending on the situation, IMN with or without bone grafting, and external fixation remain reasonable alternatives for orthopaedic surgeons. As noted by Chan et al., IMN is less effective in cases with small distal fragments32. Also, arthroplasty is usually reserved for elderly patients with osteoporosis and severe comminution. Future topics of research directed at improving results include the benefit of bone growth factors, such as bone morphogenetic protein, and the utility of dynamization of plate fixation to reduce construct stiffness.