Clinical outcomes following long versus short cephalomedullary devices for fixation of extracapsular hip fractures: a systematic review and meta-analysis

Although both long and short cephalomedullary devices (CMDs) are used in the treatment of extracapsular hip fractures, the advantages of either option are subject to debate. This study aims to evaluate the differences in clinical outcomes with long versus short CMDs for extracapsular hip fractures. Studies included must have included subjects with at least 1 year of follow-up and reported on at least one of the following outcomes: rate of reoperation; rate of peri-implant fracture; operating time; blood loss; complication rate; length of hospital stay; 1-year mortality. Only articles written in the English language were included in this study. A search was conducted across the databases of Medline, Embase, CENTRAL (Cochrane Central Register of Controlled Trials), CINAHL and Scopus for articles published from the inception of the database to 1 November 2020. Included studies were assessed for their risk of bias using the Risk of Bias Tool (RoB2) and the risk-of-bias in non-randomized studies – of interventions (ROBINS-I) tool. A total of 8460 fractures from 16 studies were included in the analysis, with 3690 fixed with short, and 4770 fixed with long CMDs. A meta-analysis of the results revealed that short CMDs offer peri-operative advantages, while long CMDs could offer longer-term advantages. Limitations of this study include a lack of randomized control trials included in the analysis. In conclusion, when planning for the treatment of extracapsular hip fractures, a patient specific approach may be necessary to make a decision according to the individual risk profile of the patient.

Hip fractures are known to be increasing in frequency globally due to an ageing and active global population, with annual incidences set to rise to 6.26 million by 2050 1 . Furthermore, with a 1-year mortality rate of about 20%, being able to choose the optimal treatment of a hip fracture is of increasing importance and relevance 2,3 . Of the different types of hip fractures, extracapsular fractures such as basicervical neck of femur fractures and intertrochanteric (IT) fractures are thought to be the most associated with falls from a standing height in elderly patients who have osteoporosis 4 .
Currently there is an increasing trend towards the using a cephalomedullary device (CMD) to manage both stable and unstable extracapsular hip fractures (EHFs) as these nails act as an intramedullary buttress to prevent excessive shaft medialization [5][6][7] . When compared with previous surgical options for management, CMDs have shown more favourable long term outcomes and a lower rate of complications 8 . CMDs which are < 250 mm in length are generally considered to be short, while those longer than 250 mm are classified as long CMDs, the difference being short CMDs do not cross the isthmus of the femur 9 .
Previous studies have compared the biomechanical properties of long and short CMDs and their effectiveness in fixation of EHFs. While it has been found that axial stiffness is greater in the use of short CMDs, the overall results show no significant differences between short or long CMDs and suggest that either option can be employed for fixation of unstable EHFs 10,11 . www.nature.com/scientificreports/ Fracture pattern. Twelve studies reported on the fracture classification, with ten further subclassifying the fractures under the length of nail used. Many (n = 6) focused on 31A1/2 fractures, but some studies (n = 4) also elected to include patients from all 31A fractures. Of note, Okcu et al. only studied patients with 31A3 fractures 28 .
Blood loss. Eight studies reported on the estimated blood loss from either procedure, with 7 studies reporting statistically significantly higher mean estimated blood loss in long CMD operations as opposed to short CMDs 20,[24][25][26]30,31,33 . Five studies included information on the number of patients requiring blood transfusion, however only Boone et al. reported that there was a statistically significant difference in transfusion rates 18,[25][26][27]31 .  Table 2.

Length of stay (LOS
Post-operative results and complications. Seven studies reported on the 1-year mortality rates of patients, with none of them finding any significant difference between the two groups 18,21,22,[27][28][29]31 .
Peri-implant infection rates 1-year post-operatively were also reported to be similar in the 6 studies which reported on them 18,20,24,26,32,33 .
Only three studies noted patient reported outcomes under the Harris Hip Score 1-year post-operatively. Of the studies, only Shannon et al. reported a significant difference between the groups, with the short CMD group reporting higher results 28,30,33 .
Detailed information on post-operative results and complications can be found in Table 3.

Meta-analysis.
We performed a meta-analysis to compare several outcomes of interest between long and short CMD groups. The outcomes analysed were mean operating time, mean estimated blood loss, mean length of stay, peri-implant fracture rates, reoperation rates and 1-year mortality rates. All 16 included studies were included in the meta-analysis.
Mean operating time. A total of 8 studies were used to analyse the difference in mean operating time between the two groups. Results from the analysis favoured the group which used short CMDs, reporting a statistically significant lower mean operating time of 13.99 min (95% CI − 15.15 to − 12.84; p value < 0.00001) (Fig. 3).
Mean estimated blood loss. A total of 6 studies were used to analyse the difference in mean estimated blood loss between the two groups. Results from the analysis favoured the group which used short CMDs, reporting   (Fig. 4).
Mean length of stay. A total of 6 studies were used to analyse the difference in mean length of stay between the two groups. Results from the analysis favoured the group which used short CMDs, however this difference was found to not be statistically significant (95% CI − 0.35 to 0.09; p value = 0.23) (Fig. 5).
Peri-implant fractures. A total of 10 studies were used to analyse the overall risk ratio for peri-implant fractures between the two groups, with results showing a statistically significant difference, favouring the group with long CMDs. The risk ratio of peri-implant fractures among patients with short CMDs was 1.85 (95% CI 1.14-2.98; p value = 0.01) times as likely as the risk among patients who had long CMDs (Fig. 6).
Reoperation rate. A total of 6 studies were used to analyse the risk ratio of reoperation rate among patients, with results showing no statistically significant differences between the groups. The risk ratio of peri-implant fractures among patients with short CMDs was 1.08 (95% CI 0.78-1.49; p value = 0.63) times as likely as the risk among patients who had long CMDs (Fig. 7).

1-Year Mortality Rate.
A total of 7 studies were used to analyse the risk ratio for 1-year mortality rate between the two groups, with results showing no statistically significant differences between the groups. The risk ratio of 1-year mortality rate among patients with long CMDs was 1.03 (95% CI 0.83-1.27; p value = 0.78) times as likely as the risk among patients who had short CMDs (Fig. 8).

Discussion
The aim of the study was to evaluate differences in clinical outcomes between using short and long CMDs in the treatment of extracapsular hip fractures. With the current lack of pooled analysis and comparison between these CMDs, it is difficult for surgeons to decide with conviction the optimal nail length for their patients. Through this systematic review and meta-analyses, we have found that long and short CMDs have different short-term and long-term outcomes. Short CMDs offered significant advantages in some of the peri-operative outcomes over the long CMDs. These were shorter operative times and less blood loss. However, long CMDs have shown advantages over short CMDs in their long-term benefits. With a significant reduction in risk of peri-implant fracture rate, the longer survivorship of the prosthesis could be a reflection of the theoretical increased stability provided by having a long CMD (Table 4). Mean operating time between the two groups showed a significantly shorter operating time when using a short CMD as compared to a long CMD. This finding is in agreement with what has been reported in the existing literature where the number of steps and their complexity is lower when using a short CMD. This has been attributed to the additional time needed for reaming, as well the freehand placement of distal interlocking screws when inserting long CMDs 36,37 .
Estimated blood loss between the groups favoured the use of short CMDs over long CMDs, with the pooled data showing a statistically significant difference between the two groups. This has also been reported in the existing literature, with short CMDs showing a consistently lower blood loss over long CMDs 36,38 . The lower Table 1. Eligible studies, study characteristics, population demographics. RCT randomized controlled trial, NI no information, NC not possible to calculate. *Indicates statistically significant difference reported. All included studies were subject to an assessment of bias, with the revised Cochrane risk-of-bias tool (RoB 2) for the randomized controlled trials, and Risk-of-bias in Non-randomized Studies-of Interventions (ROBINS-I) tool for the retrospective cohort studies 34,35 . The detailed information on the assessment for bias is reported in Tables 2 and 3 respectively. None of the studies received external funding. www.nature.com/scientificreports/ blood loss may also confer more advantages towards the use of a short nail, such as a reduction in the transfusion requirements, which had also been reported in 2 of the studies included 25,26 . The short CMD group displayed a shorter LOS post-operatively as compared to the long CMD group. While this difference was not statistically significant, the difference in the length of stay could be due to a multitude of factors involving the condition and care of the patient. This may include the availability of community healthcare resources for the patient to be discharged to, rehabilitation services such as physiotherapy or other patient specific factors which may have altered their LOS. Another postulation is that patients who required long CMDs in the studies could have had a more complex or unstable fracture pattern which could have delayed their postoperative rehabilitation.

S/N Study Country
Rates of peri-implant fractures were significantly higher in the short CMD group, with a risk ratio of 1.85 times as compared to the long CMD group. The differences in these results may be due to the nail spanning the full length of the femur, and therefore providing additional stability and strength to the bone 16,17 . Periimplant fractures were reported to be at the distal tip of the implant in several cases of long nails 18,23,33 . In some reports, the peri-implant fractures were due to identifiable incidents of trauma, such as falls from standing height 12,23,25,32,33 . Kleweno et al. reported no significant difference in the time to peri-implant fracture between the short and long nail cohorts 23 . In the studies included, the range of incidence for post-operative peri-implant fracture rates was reported to be from as early as 21 days post-operatively, up to 563 days post-operatively. It has been proposed that the timing of peri-implant fracture rates is not predictable as there is a large range of timings of reported incidence of this complication 24 .
Reoperation rates were similar in both groups, with no significant differences calculated. While the rate of peri-implant fractures 1-year post-operatively was higher in the short CMD group, rates of other complications was not well reported and thus a pooled analysis for the overall complication rate was not possible.
Finally, the 1-year mortality rate between the groups was found to be similar, with no significant difference between them. While the 1-year mortality rate has been reported to be as high as up to 58%, the current study Most studies included in this study have been deemed to have low levels of bias. However, those papers which have been identified to have potential issues with bias are commonly due to confounding variables, which most commonly are due to a significant difference between the ages of the short and long CMD groups. The studies which had these differences postulated that the difference could be attributed to shorter operative time which would be favourable in older age groups 12,25,26 .
In light of our analysis, we postulate that the use of short CMDs may be better suited for high-risk patients who may not be able to tolerate longer operating times. This includes those who have multiple medical comorbidities, and the elderly. The decrease in the amount of time under anaesthesia could be advantageous for these patients as it may decrease the risk of perioperative complications 13,40 . However, this should also be weighed against benefit in long-term outcomes that have been shown with the use of a long CMD.
While the study has presented several findings that could help guide the decision between the use of a short or long CMD, the strength of this study is in the volume of patients included in the analysis. This is the largest meta-analysis on the topic thus far, with the inclusion of large multi-centre data in the pooled analysis. Limitations of the paper would include the lack of RCTs included. While there were 3 included in the analysis, a larger pool of RCTs would be helpful in ascertaining the differences between short and long CMDs. Additionally, Table 3. Post-operative results and complications. NI no information. *Indicates statistically significant difference reported.      www.nature.com/scientificreports/ there was limited data available for some of the outcomes of interest, including 1-year post-operative HHS and transfusion rate, preventing a pooled analysis on these outcomes. Furthermore, the mean follow-up period of the included studies was varied (10.2-40.6 months), making a cross sectional study of the outcomes of interest difficult. While a key point of interest would be to investigate incidence and timeline of peri-implant fracture rates, few studies reported the time to the incidence of peri-implant fracture. This precluded further analysis on this subject, and the current study is only able to draw a conclusion to the overall rate of complications and peri-implant fractures. Other potentially interesting areas of further research on this topic would include a comparison between the peri-implant fracture rate in different periods post-operatively, including short-and long-term studies on the topic.

Conclusion
Short CMDs offer advantages of shorter operative time and lesser blood loss. However long CMDs could offer longer-term protection against peri-implant fractures. Therefore, in planning for the operation, a patient specific approach may be necessary to make a decision according to the individual risk profile of the patient.  *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI) CI: confidence interval; RR: risk ratio GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect