Injury, Int. J. Care Injured 46 (2015) 484–491
Contents lists available at ScienceDirect
Injury journal homepage: www.elsevier.com/locate/injury
Complications following young femoral neck fractures G.P. Slobogean a,b,*, S.A. Sprague b,c, T. Scott c, M. Bhandari b,c a
Department of Orthopaedic Surgery, University of British Columbia, Canada Division of Orthopaedic Surgery, Department of Surgery, McMaster University, Canada c Department of Clinical Epidemiology and Biostatics, McMaster University, Canada b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 23 July 2014 Accepted 6 October 2014
Background: Femoral neck fractures in patients 60 years of age or younger are challenging injuries to treat because of the high-energy trauma mechanisms and the displaced fracture patterns typically found in this patient population. Understanding the burden of disease is an important first step in addressing treatment controversies in this population. The purpose of the current study is to quantitatively pool the incidence of patient important complications following internal fixation of young femoral neck fractures. Methods: A comprehensive search of the Medline, Embase, CINAHL, Cochrane Database of Systematic Reviews, and Central databases was completed under the direction of a biomedical librarian. Multiple outcomes of interest (complications) were collected and included: reoperation, femoral head avascular necrosis, fracture non-union, infection, implant failure, and malunion. Results: 1558 fractures from 41 studies were included in the meta-analysis. An18.0% pooled reoperation incidence was observed for isolated femoral neck fractures. The total pooled incidence of avascular necrosis (AVN) was 14.3%, and the total incidence of nonunion was 9.3%. When stratified for fracture displacement displaced fractures were more likely to undergo reoperation and to result in AVN or nonunion. The total incidence of malunion was 7.1%, implant failure was 9.7%, and surgical site infection was 5.1%. Complications associated with a femoral neck fracture treated in conjunction with an ipsilateral femoral shaft fracture were lower overall than the pooled estimates for isolated neck fractures. Conclusions: The results of our analysis demonstrate that the incidence of complications experienced by young femoral neck fracture patients is relatively high. Reoperation following internal fixation of isolated femoral neck fractures occurred in nearly 20% of cases, and AVN and nonunion were the most common complications that likely contributed to repeat surgeries. These results highlight the importance of further efforts to improve the clinical outcomes in this population. ß 2014 Elsevier Ltd. All rights reserved.
Keywords: Non-geriatiric hip fracture Femoral neck fracture
Introduction Femoral neck fractures in patients 60 years of age or younger represent challenging injuries to treat because of the high-energy trauma mechanisms and the displaced fracture patterns typically found in this patient population [1]. The younger patient age and increased functional demands for work and recreational activities mandate a surgical treatment that preserves the native hip [2]. Although controversy exists surrounding methods of reduction and internal fixation, an understanding of the burden of disease is required.
* Corresponding author at: Division of Orthopaedic Trauma, Department of Orthopaedics, University of British Columbia, Centre for Hip Health & Mobility, 771 – 2635 Laurel Street, Vancouver, BC V5Z 1M9, Canada. Tel.: +1 604 875 5809. E-mail address: [email protected] (G.P. Slobogean). http://dx.doi.org/10.1016/j.injury.2014.10.010 0020–1383/ß 2014 Elsevier Ltd. All rights reserved.
The purpose of the current study is to quantitatively pool the incidence of patient important complications following internal fixation of young femoral neck fractures. This study aims to update the existing meta-analysis literature, expand previous reviews by including ipsilateral femoral neck and shaft fractures, and focus on multiple complications that impact quality of life and functional outcome. Methods Study eligibility Only studies that met the following criteria were considered eligible: [1] the population was comprised of non-geriatric adult patients (ages 16–60 years) with a femoral neck fracture [2], the patients were treated with any type of internal fixation [3], the authors reported original research, and [4] the study reported at
G.P. Slobogean et al. / Injury, Int. J. Care Injured 46 (2015) 484–491
least one complication outcome measure following fracture fixation. Only English language articles were included. Potentially eligible studies were also excluded if they focused on [1] stress fractures; [2] treatment of delayed fractures (greater than 14 days from injury) [3], management of femoral neck non-unions, or [4] management of osteonecrosis following femoral neck fractures. Studies involving combined femoral neck and femoral shaft fractures were also included based on a priori planned subgroup analyses. Identification of studies In November 2012, a comprehensive literature search was performed to identify studies involving the management of femoral neck fractures in patients age 60 years. Using the OVID interface, electronic searches of the following databases were performed: Medline, Embase, CINAHL, Cochrane Database of Systematic Reviews, and Central. Keywords and MeSH headings related to femoral neck fractures and surgical fixation were utilized under the direction of a biomedical librarian. A full description of the search strategy is found in Appendix A. Following the literature searches, two investigators reviewed the references lists of all key articles for further eligible articles. Frequently cited articles were identified and a separate Science Citation Index search (SciSearch) was performed to locate potentially relevant studies. We also conducted a ‘‘related articles’’ search on PubMed. Title review Two reviewers independently screened the titles identified in the literature searches. Titles that clearly did not meet the eligibility criteria were excluded; in all situations, the reviewers erred on the side of inclusivity and selected the abstract to be reviewed. Since the search strategy attempts to filter out elderly fractures, any paper that mentions fixation of femoral neck fractures was selected for further review. The abstract and full-text review was performed in a similar independent and duplicate fashion with two reviewers. When consensus could not be reached on study eligibility, a third reviewer was consulted. Data extraction Two reviewers independently performed the data extraction. Study characteristics, patient demographics, fracture patterns, and surgical details were recorded for each included study. Multiple outcomes of interest (complications) were collected and included: femoral head avascular necrosis, fracture non-union, reoperation, infection, implant failure, and malunion. The malunion outcome included any case described as malaligned, malreduced, malrotated, or malunited. Data analysis A random-effects model of DerSimonian and Laird was used to provide pooled estimates of the incidence of complications within the young femoral neck fracture literature [3]. This model assumes that the studies included in this review represent a random sample of all of the potentially available studies. While we are confident that our search strategy identified all relevant studies in this population, it remains plausible that not every study ever conducted was identified. The random-effects model accounts for this fact and assumes that we have a representative sample of all existing studies (published, non-published, and in progress). For each complication of interest we calculated the pooled incidence and 95% confidence interval (CI). The I2 statistic was reported for
485
each pooled estimate as a measure of study heterogeneity; values greater than 50% represent substantial study heterogeneity [4]. Subgroup analyses were performed based on fracture displacement as well as the presence of an ipsilateral femoral shaft fracture. Results Fig. 1 outlines the search results and selection of eligible studies. 41 articles were included in our analysis: 27 studies involved patients with femoral neck fractures only and 14 publications included femoral neck fractures associated with ipsilateral femoral shaft fractures [5–46]. Table 1 summarizes the outcomes reported by the included studies. Briefly, 1558 fractures were included. The mean sample size of included studies was 39.43 33.8 patients. The mean of the average age and duration of follow-up reported in each study was 39.4 5.6 years and 35.2 16.6 months, respectively. All research was published between 1964 and 2012. 60% of included studies were retrospective case series; only two studies were prospective randomised control trials. Table 2 presents the pooled results of isolated femoral neck fractures, combined neck–shaft fractures, and the overall incidence of complications when all results are combined. Reoperation 28 studies reported reoperation events within their study population. A total of 181 event in 1061 included patients. There was an overall 18.0% reoperation incidence for isolated femoral neck fractures (95% CI 13.1–24.2%, I2 = 19.8%; Fig. 2). When the pooled results were stratified for fracture displacement, similar reoperation estimates were obtained for displaced fractures (17.8%, 95% CI 12.4–24.9) and much lower estimates were observed for undisplaced fractures (6.9%, 95% CI 2.6–17.1%). Although a large difference in the point estimates between the subgroups is reported, the confidence intervals overlap and fail to demonstrate statistical significance. Femoral head avascular necrosis 39 studies reported femoral head avascular necrosis occurring in their study population. A total of 184 events in 1552 patients were included. The total pooled incidence of avascular necrosis for isolated femoral neck fractures was 14.3% (95% CI, 12.5–24.2%, I2 = 0%; Fig. 3). Similar to reoperation events, displaced fractures were associated with a statistically higher incidence of avascular necrosis than undisplaced fractures (14.7%, 95% CI 12.3–17.5% versus 6.4%, 95% CI 3.4–11.8%). Nonunion Fracture nonunion events were pooled from 35 studies. A total of 109 events in 1328 patients were included. The pooled incidence of nonunion after internal fixation for isolated fractures was 9.3% (95% CI 6.6–13.0%; Fig. 4). The point estimate for displaced fractures (10.0%) was nearly double the incidence of undisplaced fractures (5.2%), however, the confidence intervals were too wide to demonstrate significance (displaced 95% CI 6.9–14.3%; undisplaced 95% CI 2.0–13.1%).
Malunion, implant failure, and infection Malunion, implant failure, and infection outcomes were less commonly reported by the included studies (Table 1). The pooled incidence of reported femoral neck malunion was 7.1% (95% CI 2.7– 17.5%). The incidence of implant failure was 9.7% (95% CI 5.4–16.7%), and the incidence of surgical site infection was 5.1% (95% CI 3.2– 8.0%). Pooled analyses comparing these outcomes for displaced and
486
G.P. Slobogean et al. / Injury, Int. J. Care Injured 46 (2015) 484–491
Fig. 1. Study flow and selection of included studies.
G.P. Slobogean et al. / Injury, Int. J. Care Injured 46 (2015) 484–491
487
Table 1 Summary of included studies. Name
Shaft fractures included
Brown 1964 Swiontkowski 1984 Morwessel 1985 Tooke 1985 Srivastava 1989 Masetti 1990 Chaturvedi 1993 Parfenchuck 1993 Driesen 1984 He 1995 Robinson 1995 Koldenhoven 1997 Broos 1998 Gautam 1998 Chang 1999 Randelli 1999 Jain 2002 Verettas 2002 Okcu 2003 Upadhyay 2004 Jain 2004 Farooq 2005 Khallaf 2005 Oh 2006 Kao 2006 Liporace 2008 Singh 2008 Butt 2008 Vidyadhara 2009 Yildirim 2009 Stearns 2009 Tzachev 2009 Huang 2010 Wang 2010 Duckworth 2011 Huang 2011 Henari 2011 Pollack 2012 Razik 2012 Gardner 2012 Rawall 2012
N Y N N N N Y Y N N N Y N N N Y N N Y N Y N Y Y Y N Y N Y N N Y N Y N N N N N N N
Total
27
Sample size
Complications reported Reoperation
14 13 13 27 30 21 16 11 26 25 45 11 30 25 26 27 38 12 15 27 19 23 17 17 12 62 27 52 43 55 59 18 122 21 122 146 12 91 92 69 27
U U U U
U U U U U U U U U U U
U U U U U U U
U U U U
1558
AVN
Non-union
U U U U U U U U U U U U U U U U U U U U U U U U
U U U U U U U U U U
U U
U
U
U U U U U U U U U U U U U U
U U U U U U U U U U U U U
U
U
39
35
Table 2 Pooled complications of isolated femoral neck fractures. Outcome
Incidence (%)
95% CI
I2 statistic (%)
Reoperation AVN Nonunion Malunion Implant failure Infection
18.0 14.3 9.3 7.1 9.7 5.1
13.1–24.2 12.5–16.4 6.6–13.0 2.7–17.5 5.4–16.7 3.2–8.0
19.8 0 0 0 12.3 0
U U U
U U
U U U U U
U
U
U
U U
U U U U
U U
U U
U U U
U U U U U U U U U
Implant failure
U
U
28
Femoral neck outcomes associated with ipsilateral shaft fractures Complications associated with a femoral neck fracture treated in conjunction with an ipsilateral femoral shaft fracture were overall lower than the pooled estimates for isolated neck fractures (Table 2). Specifically, the incidence of AVN and nonunion following a combined neck–shaft fracture pattern most closely mirrored the isolated, undisplaced femoral neck fracture subgroup. Both complications were relatively uncommon, with point estimates 2–3 times lower than the isolated femoral neck fractures. The remaining malunion, implant failure, and infection
Infection
U
U U
undisplaced fractures were not performed due to the small amount of data available.
Malunion
U U
U
U
U U
U
U U
15
10
U
16
complications also displayed lower point estimates than the isolated neck fracture group (Table 3).
Discussion The current meta-analysis has pooled the results of 42 studies involving internal fixation of femoral neck fractures. The results of our analysis demonstrate that the incidence of complications experienced by young femoral neck fracture patients is relatively high. Nearly 20% of patients with an isolated femoral neck fracture had a reoperation related to their hip fracture. The incidence of other patient important complications is also relatively common. To the best of our knowledge, there has been one previous meta-analysis that has pooled the outcomes of young femoral neck fracture fixation. Damany and colleagues reviewed the incidence of avascular necrosis and nonunion in patients ages 50 years [1]. Their analysis included 564 patients in 18 studies published between 1976 and 2003. The authors reported a 23.0% overall incidence of AVN and an 8.9% incidence of nonunion. Their analysis also examined associations of reduction method (open versus closed) and timing ( 12 h) on the incidence of complications. Similar to our results, a much higher incidence of complications following displaced fractures was observed.
488
G.P. Slobogean et al. / Injury, Int. J. Care Injured 46 (2015) 484–491
Fig. 2. Forest plot and pooled analysis of reoperation events.
Our study extends the previous work of Damany et al. in several aspects. Our review updates the literature by including nine years of more recent publications. The updated search years, the inclusion of ipsilateral femoral neck and shaft fractures, and our other methods for ensuring a sensitive search strategy allowed us to include an additional 1000 fractures and 24 studies in our pooled analysis. Furthermore, we used a random-effects model to pool the results, rather than the simple weighted-averages method employed by Damany; this allowed us to provide a more conservative estimate of each outcome that accounts for sampling error of the true incidence of complications. In addition to the benefits of our larger sample size and improved statistical pooling methods, our meta-analysis included a subgroup of patients with ipsilateral femoral neck and shaft fractures. Clinically, treating this population of femoral neck fractures can be technically challenging due to significant displacement in the setting of an unstable shaft fracture, using multiple implants for adjacent fractures, and the potential of initially missing occult femoral neck fractures. When considering the incidence of complications in this subgroup, the point estimates were uniformly lower than the isolated neck fracture population and many of the pooled estimates were similar to the undisplaced subgroup. This suggests the outcomes of femoral neck fractures in a combined neck–shaft fracture pattern might be less severe than the displaced femoral neck group; this mirrors clinical observations in which 10% of these fractures are initially occult or minimally displaced [47]. When considering isolated femoral neck fractures only, displaced fractures were associated with a substantially higher incidence of reoperation, AVN, and nonunion. Although not
surprising, this observation underscores the challenging nature of treating the young femoral neck fracture population; the majority of young femoral neck fractures occur from high-energy trauma and are significantly displaced. The frequency of displaced fractures within the overall distribution of fracture patterns also explains why the estimates for the displaced subgroup closely mirrors the complication estimates for the entire isolated femoral neck fracture population. In addition, the pooled estimate of an 18.0% incidence of femoral neck reoperation is a compelling reminder that these injuries often cause significant patient Table 3 Pooled complications of combined femoral neck and shaft fractures. Outcome
Incidence (%)
95% CI
I2 statistic (%)
Reoperation AVN Nonunion Malunion Implant failure Infection
12.5 4.6 5.1 5.6 6.3 3.6
8.0–19.1 2.5–8.3 2.8–9.3 3.1–9.9 3.0–12.7 1.4–9.3
0 0 0 0 0 0
Table 4 Pooled complications of all femoral neck fractures combined. Outcome
Incidence (%)
95% CI
I2 statistic (%)
Reoperation AVN Nonunion Malunion Implant failure Infection
15.9 13.3 8.3 6.4 9.2 4.8
12.2–20.6 11.6–15.5 6.2–11.1 3.8–10.5 6.4–13.2 3.1–7.2
6.7 6.4 0 0 0 0
G.P. Slobogean et al. / Injury, Int. J. Care Injured 46 (2015) 484–491
Fig. 3. Forest plot and pooled analysis of AVN events.
Fig. 4. Forest plot and pooled analysis of nonunion events.
489
G.P. Slobogean et al. / Injury, Int. J. Care Injured 46 (2015) 484–491
490
Table 5 Pooled complications stratified by fracture displacement. Outcome
Reoperation AVN Nonunion
Displaced fractures
Undisplaced fractures 2
Incidence (%)
95% CI
I statistic (%)
Incidence (%)
95% CI
I2 statistic (%)
17.8 14.7 10.0
12.4–24.9 12.3–17.5 6.9–14.3
15.4 0 0
6.9 6.4 5.2
2.6–17.1 3.4–11.8 2.0–13.1
0 0 0
morbidity. Additionally, although reoperation represents a serious complication, it is important to recognize that many of the other complications of interest also lead to profound impacts on patient quality of life (Tables 4 and 5). Deep infection, implant failure, and nonunion are all serious events that will often lead to a repeat surgical procedure; however, the other complications of interest such as significant malunion or late onset AVN may not lead to an early reoperation during the follow-up periods of the included studies. This observation underscores the importance of measuring health related quality of life and other patient reported functional outcomes to ensure the true impact of these injuries; only 45% of the included studies reported any patient reported functional measures. Finally, it is important to recognize that our study estimates are limited by the available data; longer-term follow-up and increased availability of patient-reported outcomes would improve future meta-analyses. In conclusion, this meta-analysis provides pooled estimates for several patient important complications following femoral neck fracture fixation in patients ages
Contents lists available at ScienceDirect
Injury journal homepage: www.elsevier.com/locate/injury
Complications following young femoral neck fractures G.P. Slobogean a,b,*, S.A. Sprague b,c, T. Scott c, M. Bhandari b,c a
Department of Orthopaedic Surgery, University of British Columbia, Canada Division of Orthopaedic Surgery, Department of Surgery, McMaster University, Canada c Department of Clinical Epidemiology and Biostatics, McMaster University, Canada b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 23 July 2014 Accepted 6 October 2014
Background: Femoral neck fractures in patients 60 years of age or younger are challenging injuries to treat because of the high-energy trauma mechanisms and the displaced fracture patterns typically found in this patient population. Understanding the burden of disease is an important first step in addressing treatment controversies in this population. The purpose of the current study is to quantitatively pool the incidence of patient important complications following internal fixation of young femoral neck fractures. Methods: A comprehensive search of the Medline, Embase, CINAHL, Cochrane Database of Systematic Reviews, and Central databases was completed under the direction of a biomedical librarian. Multiple outcomes of interest (complications) were collected and included: reoperation, femoral head avascular necrosis, fracture non-union, infection, implant failure, and malunion. Results: 1558 fractures from 41 studies were included in the meta-analysis. An18.0% pooled reoperation incidence was observed for isolated femoral neck fractures. The total pooled incidence of avascular necrosis (AVN) was 14.3%, and the total incidence of nonunion was 9.3%. When stratified for fracture displacement displaced fractures were more likely to undergo reoperation and to result in AVN or nonunion. The total incidence of malunion was 7.1%, implant failure was 9.7%, and surgical site infection was 5.1%. Complications associated with a femoral neck fracture treated in conjunction with an ipsilateral femoral shaft fracture were lower overall than the pooled estimates for isolated neck fractures. Conclusions: The results of our analysis demonstrate that the incidence of complications experienced by young femoral neck fracture patients is relatively high. Reoperation following internal fixation of isolated femoral neck fractures occurred in nearly 20% of cases, and AVN and nonunion were the most common complications that likely contributed to repeat surgeries. These results highlight the importance of further efforts to improve the clinical outcomes in this population. ß 2014 Elsevier Ltd. All rights reserved.
Keywords: Non-geriatiric hip fracture Femoral neck fracture
Introduction Femoral neck fractures in patients 60 years of age or younger represent challenging injuries to treat because of the high-energy trauma mechanisms and the displaced fracture patterns typically found in this patient population [1]. The younger patient age and increased functional demands for work and recreational activities mandate a surgical treatment that preserves the native hip [2]. Although controversy exists surrounding methods of reduction and internal fixation, an understanding of the burden of disease is required.
* Corresponding author at: Division of Orthopaedic Trauma, Department of Orthopaedics, University of British Columbia, Centre for Hip Health & Mobility, 771 – 2635 Laurel Street, Vancouver, BC V5Z 1M9, Canada. Tel.: +1 604 875 5809. E-mail address: [email protected] (G.P. Slobogean). http://dx.doi.org/10.1016/j.injury.2014.10.010 0020–1383/ß 2014 Elsevier Ltd. All rights reserved.
The purpose of the current study is to quantitatively pool the incidence of patient important complications following internal fixation of young femoral neck fractures. This study aims to update the existing meta-analysis literature, expand previous reviews by including ipsilateral femoral neck and shaft fractures, and focus on multiple complications that impact quality of life and functional outcome. Methods Study eligibility Only studies that met the following criteria were considered eligible: [1] the population was comprised of non-geriatric adult patients (ages 16–60 years) with a femoral neck fracture [2], the patients were treated with any type of internal fixation [3], the authors reported original research, and [4] the study reported at
G.P. Slobogean et al. / Injury, Int. J. Care Injured 46 (2015) 484–491
least one complication outcome measure following fracture fixation. Only English language articles were included. Potentially eligible studies were also excluded if they focused on [1] stress fractures; [2] treatment of delayed fractures (greater than 14 days from injury) [3], management of femoral neck non-unions, or [4] management of osteonecrosis following femoral neck fractures. Studies involving combined femoral neck and femoral shaft fractures were also included based on a priori planned subgroup analyses. Identification of studies In November 2012, a comprehensive literature search was performed to identify studies involving the management of femoral neck fractures in patients age 60 years. Using the OVID interface, electronic searches of the following databases were performed: Medline, Embase, CINAHL, Cochrane Database of Systematic Reviews, and Central. Keywords and MeSH headings related to femoral neck fractures and surgical fixation were utilized under the direction of a biomedical librarian. A full description of the search strategy is found in Appendix A. Following the literature searches, two investigators reviewed the references lists of all key articles for further eligible articles. Frequently cited articles were identified and a separate Science Citation Index search (SciSearch) was performed to locate potentially relevant studies. We also conducted a ‘‘related articles’’ search on PubMed. Title review Two reviewers independently screened the titles identified in the literature searches. Titles that clearly did not meet the eligibility criteria were excluded; in all situations, the reviewers erred on the side of inclusivity and selected the abstract to be reviewed. Since the search strategy attempts to filter out elderly fractures, any paper that mentions fixation of femoral neck fractures was selected for further review. The abstract and full-text review was performed in a similar independent and duplicate fashion with two reviewers. When consensus could not be reached on study eligibility, a third reviewer was consulted. Data extraction Two reviewers independently performed the data extraction. Study characteristics, patient demographics, fracture patterns, and surgical details were recorded for each included study. Multiple outcomes of interest (complications) were collected and included: femoral head avascular necrosis, fracture non-union, reoperation, infection, implant failure, and malunion. The malunion outcome included any case described as malaligned, malreduced, malrotated, or malunited. Data analysis A random-effects model of DerSimonian and Laird was used to provide pooled estimates of the incidence of complications within the young femoral neck fracture literature [3]. This model assumes that the studies included in this review represent a random sample of all of the potentially available studies. While we are confident that our search strategy identified all relevant studies in this population, it remains plausible that not every study ever conducted was identified. The random-effects model accounts for this fact and assumes that we have a representative sample of all existing studies (published, non-published, and in progress). For each complication of interest we calculated the pooled incidence and 95% confidence interval (CI). The I2 statistic was reported for
485
each pooled estimate as a measure of study heterogeneity; values greater than 50% represent substantial study heterogeneity [4]. Subgroup analyses were performed based on fracture displacement as well as the presence of an ipsilateral femoral shaft fracture. Results Fig. 1 outlines the search results and selection of eligible studies. 41 articles were included in our analysis: 27 studies involved patients with femoral neck fractures only and 14 publications included femoral neck fractures associated with ipsilateral femoral shaft fractures [5–46]. Table 1 summarizes the outcomes reported by the included studies. Briefly, 1558 fractures were included. The mean sample size of included studies was 39.43 33.8 patients. The mean of the average age and duration of follow-up reported in each study was 39.4 5.6 years and 35.2 16.6 months, respectively. All research was published between 1964 and 2012. 60% of included studies were retrospective case series; only two studies were prospective randomised control trials. Table 2 presents the pooled results of isolated femoral neck fractures, combined neck–shaft fractures, and the overall incidence of complications when all results are combined. Reoperation 28 studies reported reoperation events within their study population. A total of 181 event in 1061 included patients. There was an overall 18.0% reoperation incidence for isolated femoral neck fractures (95% CI 13.1–24.2%, I2 = 19.8%; Fig. 2). When the pooled results were stratified for fracture displacement, similar reoperation estimates were obtained for displaced fractures (17.8%, 95% CI 12.4–24.9) and much lower estimates were observed for undisplaced fractures (6.9%, 95% CI 2.6–17.1%). Although a large difference in the point estimates between the subgroups is reported, the confidence intervals overlap and fail to demonstrate statistical significance. Femoral head avascular necrosis 39 studies reported femoral head avascular necrosis occurring in their study population. A total of 184 events in 1552 patients were included. The total pooled incidence of avascular necrosis for isolated femoral neck fractures was 14.3% (95% CI, 12.5–24.2%, I2 = 0%; Fig. 3). Similar to reoperation events, displaced fractures were associated with a statistically higher incidence of avascular necrosis than undisplaced fractures (14.7%, 95% CI 12.3–17.5% versus 6.4%, 95% CI 3.4–11.8%). Nonunion Fracture nonunion events were pooled from 35 studies. A total of 109 events in 1328 patients were included. The pooled incidence of nonunion after internal fixation for isolated fractures was 9.3% (95% CI 6.6–13.0%; Fig. 4). The point estimate for displaced fractures (10.0%) was nearly double the incidence of undisplaced fractures (5.2%), however, the confidence intervals were too wide to demonstrate significance (displaced 95% CI 6.9–14.3%; undisplaced 95% CI 2.0–13.1%).
Malunion, implant failure, and infection Malunion, implant failure, and infection outcomes were less commonly reported by the included studies (Table 1). The pooled incidence of reported femoral neck malunion was 7.1% (95% CI 2.7– 17.5%). The incidence of implant failure was 9.7% (95% CI 5.4–16.7%), and the incidence of surgical site infection was 5.1% (95% CI 3.2– 8.0%). Pooled analyses comparing these outcomes for displaced and
486
G.P. Slobogean et al. / Injury, Int. J. Care Injured 46 (2015) 484–491
Fig. 1. Study flow and selection of included studies.
G.P. Slobogean et al. / Injury, Int. J. Care Injured 46 (2015) 484–491
487
Table 1 Summary of included studies. Name
Shaft fractures included
Brown 1964 Swiontkowski 1984 Morwessel 1985 Tooke 1985 Srivastava 1989 Masetti 1990 Chaturvedi 1993 Parfenchuck 1993 Driesen 1984 He 1995 Robinson 1995 Koldenhoven 1997 Broos 1998 Gautam 1998 Chang 1999 Randelli 1999 Jain 2002 Verettas 2002 Okcu 2003 Upadhyay 2004 Jain 2004 Farooq 2005 Khallaf 2005 Oh 2006 Kao 2006 Liporace 2008 Singh 2008 Butt 2008 Vidyadhara 2009 Yildirim 2009 Stearns 2009 Tzachev 2009 Huang 2010 Wang 2010 Duckworth 2011 Huang 2011 Henari 2011 Pollack 2012 Razik 2012 Gardner 2012 Rawall 2012
N Y N N N N Y Y N N N Y N N N Y N N Y N Y N Y Y Y N Y N Y N N Y N Y N N N N N N N
Total
27
Sample size
Complications reported Reoperation
14 13 13 27 30 21 16 11 26 25 45 11 30 25 26 27 38 12 15 27 19 23 17 17 12 62 27 52 43 55 59 18 122 21 122 146 12 91 92 69 27
U U U U
U U U U U U U U U U U
U U U U U U U
U U U U
1558
AVN
Non-union
U U U U U U U U U U U U U U U U U U U U U U U U
U U U U U U U U U U
U U
U
U
U U U U U U U U U U U U U U
U U U U U U U U U U U U U
U
U
39
35
Table 2 Pooled complications of isolated femoral neck fractures. Outcome
Incidence (%)
95% CI
I2 statistic (%)
Reoperation AVN Nonunion Malunion Implant failure Infection
18.0 14.3 9.3 7.1 9.7 5.1
13.1–24.2 12.5–16.4 6.6–13.0 2.7–17.5 5.4–16.7 3.2–8.0
19.8 0 0 0 12.3 0
U U U
U U
U U U U U
U
U
U
U U
U U U U
U U
U U
U U U
U U U U U U U U U
Implant failure
U
U
28
Femoral neck outcomes associated with ipsilateral shaft fractures Complications associated with a femoral neck fracture treated in conjunction with an ipsilateral femoral shaft fracture were overall lower than the pooled estimates for isolated neck fractures (Table 2). Specifically, the incidence of AVN and nonunion following a combined neck–shaft fracture pattern most closely mirrored the isolated, undisplaced femoral neck fracture subgroup. Both complications were relatively uncommon, with point estimates 2–3 times lower than the isolated femoral neck fractures. The remaining malunion, implant failure, and infection
Infection
U
U U
undisplaced fractures were not performed due to the small amount of data available.
Malunion
U U
U
U
U U
U
U U
15
10
U
16
complications also displayed lower point estimates than the isolated neck fracture group (Table 3).
Discussion The current meta-analysis has pooled the results of 42 studies involving internal fixation of femoral neck fractures. The results of our analysis demonstrate that the incidence of complications experienced by young femoral neck fracture patients is relatively high. Nearly 20% of patients with an isolated femoral neck fracture had a reoperation related to their hip fracture. The incidence of other patient important complications is also relatively common. To the best of our knowledge, there has been one previous meta-analysis that has pooled the outcomes of young femoral neck fracture fixation. Damany and colleagues reviewed the incidence of avascular necrosis and nonunion in patients ages 50 years [1]. Their analysis included 564 patients in 18 studies published between 1976 and 2003. The authors reported a 23.0% overall incidence of AVN and an 8.9% incidence of nonunion. Their analysis also examined associations of reduction method (open versus closed) and timing ( 12 h) on the incidence of complications. Similar to our results, a much higher incidence of complications following displaced fractures was observed.
488
G.P. Slobogean et al. / Injury, Int. J. Care Injured 46 (2015) 484–491
Fig. 2. Forest plot and pooled analysis of reoperation events.
Our study extends the previous work of Damany et al. in several aspects. Our review updates the literature by including nine years of more recent publications. The updated search years, the inclusion of ipsilateral femoral neck and shaft fractures, and our other methods for ensuring a sensitive search strategy allowed us to include an additional 1000 fractures and 24 studies in our pooled analysis. Furthermore, we used a random-effects model to pool the results, rather than the simple weighted-averages method employed by Damany; this allowed us to provide a more conservative estimate of each outcome that accounts for sampling error of the true incidence of complications. In addition to the benefits of our larger sample size and improved statistical pooling methods, our meta-analysis included a subgroup of patients with ipsilateral femoral neck and shaft fractures. Clinically, treating this population of femoral neck fractures can be technically challenging due to significant displacement in the setting of an unstable shaft fracture, using multiple implants for adjacent fractures, and the potential of initially missing occult femoral neck fractures. When considering the incidence of complications in this subgroup, the point estimates were uniformly lower than the isolated neck fracture population and many of the pooled estimates were similar to the undisplaced subgroup. This suggests the outcomes of femoral neck fractures in a combined neck–shaft fracture pattern might be less severe than the displaced femoral neck group; this mirrors clinical observations in which 10% of these fractures are initially occult or minimally displaced [47]. When considering isolated femoral neck fractures only, displaced fractures were associated with a substantially higher incidence of reoperation, AVN, and nonunion. Although not
surprising, this observation underscores the challenging nature of treating the young femoral neck fracture population; the majority of young femoral neck fractures occur from high-energy trauma and are significantly displaced. The frequency of displaced fractures within the overall distribution of fracture patterns also explains why the estimates for the displaced subgroup closely mirrors the complication estimates for the entire isolated femoral neck fracture population. In addition, the pooled estimate of an 18.0% incidence of femoral neck reoperation is a compelling reminder that these injuries often cause significant patient Table 3 Pooled complications of combined femoral neck and shaft fractures. Outcome
Incidence (%)
95% CI
I2 statistic (%)
Reoperation AVN Nonunion Malunion Implant failure Infection
12.5 4.6 5.1 5.6 6.3 3.6
8.0–19.1 2.5–8.3 2.8–9.3 3.1–9.9 3.0–12.7 1.4–9.3
0 0 0 0 0 0
Table 4 Pooled complications of all femoral neck fractures combined. Outcome
Incidence (%)
95% CI
I2 statistic (%)
Reoperation AVN Nonunion Malunion Implant failure Infection
15.9 13.3 8.3 6.4 9.2 4.8
12.2–20.6 11.6–15.5 6.2–11.1 3.8–10.5 6.4–13.2 3.1–7.2
6.7 6.4 0 0 0 0
G.P. Slobogean et al. / Injury, Int. J. Care Injured 46 (2015) 484–491
Fig. 3. Forest plot and pooled analysis of AVN events.
Fig. 4. Forest plot and pooled analysis of nonunion events.
489
G.P. Slobogean et al. / Injury, Int. J. Care Injured 46 (2015) 484–491
490
Table 5 Pooled complications stratified by fracture displacement. Outcome
Reoperation AVN Nonunion
Displaced fractures
Undisplaced fractures 2
Incidence (%)
95% CI
I statistic (%)
Incidence (%)
95% CI
I2 statistic (%)
17.8 14.7 10.0
12.4–24.9 12.3–17.5 6.9–14.3
15.4 0 0
6.9 6.4 5.2
2.6–17.1 3.4–11.8 2.0–13.1
0 0 0
morbidity. Additionally, although reoperation represents a serious complication, it is important to recognize that many of the other complications of interest also lead to profound impacts on patient quality of life (Tables 4 and 5). Deep infection, implant failure, and nonunion are all serious events that will often lead to a repeat surgical procedure; however, the other complications of interest such as significant malunion or late onset AVN may not lead to an early reoperation during the follow-up periods of the included studies. This observation underscores the importance of measuring health related quality of life and other patient reported functional outcomes to ensure the true impact of these injuries; only 45% of the included studies reported any patient reported functional measures. Finally, it is important to recognize that our study estimates are limited by the available data; longer-term follow-up and increased availability of patient-reported outcomes would improve future meta-analyses. In conclusion, this meta-analysis provides pooled estimates for several patient important complications following femoral neck fracture fixation in patients ages
