Intramedullary nailing versus plating for extra-articular distal tibial metaphyseal fracture: A systematic review and meta-analysis

Accepted Manuscript Title: Intramedullary Nailing versus Plating for Extra-Articular Distal Tibial Metaphyseal Fracture: A Systematic Review and Meta-Analysis Authors: Xing-He Xue Shi-Gui Yan Xun-Zi Cai Ming-Min Shi Tiao Lin PII: DOI: Reference:

S0020-1383(13)00488-9 http://dx.doi.org/doi:10.1016/j.injury.2013.10.024 JINJ 5540

To appear in:

Injury, Int. J. Care Injured

Received date: Revised date: Accepted date:

15-7-2013 21-9-2013 11-10-2013

Please cite this article as: Xue X-H, Yan S-G, Cai X-Z, Shi M-M, Lin T, Intramedullary Nailing versus Plating for Extra-Articular Distal Tibial Metaphyseal Fracture: a Systematic Review and Meta-Analysis, Injury (2013), http://dx.doi.org/10.1016/j.injury.2013.10.024 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ip t cr

us

Intramedullary Nailing versus Plating for Extra-Articular Distal Tibial Metaphyseal Fracture: a Systematic

M an

Review and Meta-Analysis

Running Title: Intramedullary Nailing versus Plating Xing-He Xue, M.S., Shi-Gui Yan, M.D., Xun-Zi Cai, M.D*, Ming-Min Shi, M.D, Tiao Lin, Ph.D.

ed

First author: Xing-He Xue, M.S.

Department of Orthopaedic Surgery, Second Affiliated Hospital, School of Medicine, Zhejiang University; No.88 Jiefang Road, Hangzhou,

ce pt

310009, China P.R. Phone: (86) 571-15088681606, Fax: (86) 571- 86806079. E-mail: [email protected]

*Corresponding co-authors: Xun-Zi Cai, M.D.

Ac

Department of Orthopaedic Surgery, Second Affiliated Hospital, School of Medicine, Zhejiang University; No.88 Jiefang Road, Hangzhou, 310009, China P.R. Phone: (86) 571-13750838226, Fax: (86) 571- 86806079. E-mail: [email protected]

Page 1 of 41

ip t cr

The Corresponding Author has the right to grant on behalf of all authors and does grant on behalf of all authors, a worldwide licence to the

us

Publishers and its licensees in perpetuity, in all forms, formats and media (whether known now or created in the future), to i) publish, reproduce, distribute, display and store the Contribution, ii) translate the Contribution into other languages, create adaptations, reprints, include within

M an

collections and create summaries, extracts and/or, abstracts of the Contribution, iii) create any other derivative work(s) based on the Contribution, iv) to exploit all subsidiary rights in the Contribution, v) the inclusion of electronic links from the Contribution to third party material where-ever

Ac

ce pt

ed

it may be located; and, vi) licence any third party to do any or all of the above.

Word count: 3560 Abstract: 298 Reference count: 34 Figure count: 2 Table count: 6 Supplementary file count: 1

Page 2 of 41

ip t cr

us

Intramedullary Nailing versus Plating for Extra-Articular Distal Tibial Metaphyseal Fracture: a Systematic

M an

Review and Meta-Analysis

Running Title: Intramedullary Nailing versus Plating

Abstract

ed

Introduction With development in the techniques of reduction and fixation, there has been a controversy in comparison between intramedullary nailing (IMN) and plating for the treatment of distal tibial metaphyseal fracture (DTF). We asked (1) which fixation, IMN or plating, was better

between two modalities.

ce pt

in the clinical outcomes and the complications for the treatment of DTF, and (2) which modifying variables affected the comparative results

Methods PubMed, EMBASE, OVID, Scopus, ISI Web of Science, the Cochrane Library, Google Scholar and specific orthopaedic journals were

Ac

searched from inception to July 2013, using the search strategy of “("Fracture Fixation, Intramedullary"[Mesh]) AND ("Tibial Fractures"[Mesh]) AND (plate OR plating)”. All prospective and retrospective controlled trials comparing function, pain, bone union and complications between IMN and plating for DTF were identified. Our analysis had no limitation of the language or the publication year. The primary outcome measurements were complication rate, union time, operation time and hospital stays, while the secondary outcome measurements were

Page 3 of 41

ip t cr us

functional score and pain score.

Result Fourteen of 6620 literatures with 842 patients were included. IMN was probably preferential to plating for DTF given its higher function

M an

score (P=0.01), lower risk of infection (p=0.02), and comparable pain score (P=0.33), total complication rate (P=0.53) and time to union (p=0.86). However, plating had the lower malunion rate than IMN (p<0.0001). All the results were based on the GRADE evidence of moderate quality.

ed

Conclusions With a satisfying alignment obtained may IMN be preferential to plating for fixation of DTF with better function and lower risk of infection. However, IMN showed higher malunion rate for fixation of DTF. With the biases in our meta-analysis it will ultimately require a

ce pt

rigorous and adequately powered RCT to prove.

Level of Evidence Level III, therapeutic study (systematic review). See Guidelines for Authors for a complete description of levels of evidence

Ac

Keyword intramedullary nailing; plating; distal tibial fracture; internal fixation

Page 4 of 41

ip t cr us

Introduction

Tibial metaphyseal fractures contains both distal tibial metaphyseal fractures (DTF, AO/OTA type 43 or distal 42) and proximal metaphyseal

M an

tibial fractures (PTF, AO/OTA type 41 or proximal 42) 1, which account for 3%-10% and 5%-11% of total tibial fractures, respectively 2, 3. With the severe damage of soft tissue and the extreme instability, TMF have a high risk of unsatisfactory function, severe pain, delayed union, malunion, and infection 2, 4. The established treatments include intramedullary nailing (IMN) and plating.

ed

Plating has been accepted as the first choice for DTF 5, which ensures accurate reduction and rigid fixation. Unfortunately, the extensive dissection of the host bone and the soft tissue is mandatory. It inevitably raises the risk of infection and nonunion. Furthermore, the complaint

ce pt

about hardware irritation makes it prone to be removed 5. IMN is the gold standard for tibial diaphyseal fractures. It has a small influence on the blood supply of the host tissue, which would contribute to a low rate of nonunion and infection 4, 5. Initially the extreme high malunion rate and the poor function frustrated orthopaedic surgeons from using IMN for DTF 5. Biomechanical experiments showed that even the reamed IMN

Ac

could not fit with the lenient medullary canal of the tibia metaphysis 6. Of the two fracture ends, the short part lacked the cortical friction with implants and the adequate purchase of locking screws so that the tibial alignment could be neither obtained nor maintained 6. Given these inherent defects, IMN were limited or even relatively contradicted for DTF. With the emerging shortened and multidirectional interlocking IMN, e.g. the distal locking nail 2, and the evolving reduction techniques, e.g.

Page 5 of 41

ip t cr

us

the blocking screw and other percutaneous reduction techniques 2, the interest of applying IMN to tibial metaphyseal fractures has been renewed. The claimed improvement in the alignment and the stability has been confirmed both in the laboratory and the clinical follow-up

2, 3, 6

.

M an

Meanwhile the novel technique of minimally invasive plating osteosynthesis (MIPO) has been developed to further alleviate the local damage of plating 3.

At present there has been a great controversy on the ideal surgical option for DTF. A large amount of studies compared IMN with plating 7-20.

ed

Limited by the sample size they failed to show a clear superiority of one modality over the other. To address this, the present systematic review and meta-analysis is to cover all the comparative evidence with the purpose of determining (1) which fixation, IMN or plating, was better in the

two modalities.

Ac

Materials and Methods

ce pt

clinical outcomes and the complications for the treatment of DTF, and (2) which modifying variables affected the comparative results between

Three reviewers (XHX, XZC and MMS) searched PubMed (1966 to July 2013), EMBASE (1974 to July 2013), Ovid (1966 to July 2013), Scopus (1966 to July 2013), ISI Web of Science (1945 to July 2013), Cochrane Library, Clinical Trial Grade Center, Google Scholar (1966 to July 2013), Chinese VIP Database (1986 to July 2013) and Chinese Wan-Fang Database (1992 to July 2013), using the search strategy of

Page 6 of 41

ip t cr

us

“("Fracture Fixation, Intramedullary"[Mesh]) AND ("Tibial Fractures"[Mesh]) AND (plate OR plating)”, plus “clinical trial” and “comparative study”, with no limitation of publication year or language. The reference lists of all the selected articles and the related orthopaedic journals were

M an

hand-searched for any additional trials. In addition, we searched the Clinical Trial Registry, the Current Controlled Trials, the Trials Central and the Centre Watch for grey literatures. We defined the criteria of inclusion and exclusion before searching. We only included studies where: (1) DTF (AO/OTA type 43 or distal 42) was involved, (2) the age was no less than 18, (3) both IMN and plating were adopted, (4) functional score,

ed

pain score or complication rate was assessed, and (5) the design was comparative either prospectively or retrospectively. Exclusion criteria included studies where: (1) tibial isthmal fractures or AO type 43-C with serious intra-articular damages were involved, (2) Neither of the

ce pt

outcomes was available, (3) the follow-up was less than one year, and (4) no control data was provided. All the redundant publications were excluded. The abstracts of the rest publications were reviewed for relevance. Without the redundant publications and the unsatisfactory

criteria.

Ac

publications, the full texts of the rest publications were acquired and read in detail. We included the publications which satisfy our inclusion

We contacted the corresponding authors of the eligible trials if necessary to verify the accuracy of the data abstraction as well as the methodological assessment. We also tried to get any further data or unpublished data which were useful for our data analysis. Methodological quality

Page 7 of 41

ip t cr

us

Three reviewers (XHX, SGY and MMS) assessed the methodological quality of literatures according to the 12-items scale 21. The 12-items scale contained randomized adequately, allocation concealed, similar baseline, patient blinded, care provider blinded, outcome assessor blinded,

M an

avoided selective reporting, similar or avoided cofactor, patient compliance, acceptable drop-out rate, similar timing and ITT analysis. The inconsistent opinions were judged by another author (XZC). The disagreements were evaluated by the means of kappa (κ) test and resolved by discussion. According to the 12-items standard (Table 1), five literatures 8-12 explicitly described the randomization and the concealment of the 8-9, 11, 14-15, 17

described the proper blinding, and only one study [9] described ITT analysis. The weighted kappa

ed

allocation assignment, six studies

Analysis of the data

ce pt

for the agreement on the trial quality between the reviewers was 0.85 (95% CI, 0.77–0.93).

Three reviewers (XHX, TL and MMS) extracted the relevant data and checked the accuracy (Table 2). Study design, sample size, age, gender, loss to follow-up, AO and Gustilo classification of DTF, reduction technique, implants, fibular fixation, protocol of weight bearing, and

Ac

outcomes were abstracted. We used the ITT data from trials whenever it is possible. If the data were not reported in the original article, we extrapolated them from the accompanying graphs. Most of the studies were small-scaled with the sample size ranging from 14-160. The total sample size was 443 for IMN and 399 for plating. As for IMN, one study 15 adopted the distal locking nail and the blocking screw, and nine studies attempted the percutaneous reduction techniques

8-9, 10, 12, 14, 16, 18-20

. Eight trials

7, 10-11, 13, 15, 17, 19-20

chose the locking plate and MIPO was

Page 8 of 41

ip t cr

7-10, 15-16, 19-20

. The fixation rate of the associated fibular fracture were >50% in seven studies

four studies 8-9, 13, 16 and unclear in three studies

7, 11, 14, 15, 17-18, 20

, <50% in

us

applied in eight studies

10, 12, 19

. According to the Gustilo Classification 22, the fractures in six studies were exclusively

M an

closed or Gustilo I 7, 10-14 and the remaining eight also included Gustilo II or Gustilo III open fractures 8, 9, 15-20. The primary outcome measurements (Table 3) included the complication rate, the union time, the operation time and the hospital stays. The secondary outcome measurements included the functional score and the pain score. All the literatures mentioned the malunion and the nonunion

ed

(13/13). The literatures showed a higher malunion rate in IMN group and three literatures showed a higher nonunion rate in IMN group. All but one literature mentioned the infection rate (12/13) and only three literatures showed the inferiority of IMN group. More than half of the included

ce pt

literatures described the secondary surgery rate (7/13) and the union time (8/13). Six literatures mentioned the implant removal rate and the operation time (6/13). Few of literature mentioned the hospital stays (3/13). Functional scores were assessed with three different criteria: Olerud and Molander functional Ankle score (OMAS) 11-12, 15, 17-18 in five literatures, American Orthopaedic Foot and Ankle surgery scores (AOFAS) 10

Ac

in one and Musculoskeletal Function Assessment (MFA) 8 in one (7/13 in total). All seven literatures showed a higher functional score in IMN group and Mauffery et al. 12 also showed statistical difference (p<0.05). In addition, four literatures assessed pain score, of which one did knee pain

14

while the other three did ankle pain 7, 8, 10. As a result we used the standard mean difference (SMD) as the outcome measure of the

functional scores and the pain scores.

Page 9 of 41

ip t cr us

Subgroup analysis

Subgroup analysis was done in our meta-analysis which mainly focuses on the types of internal fixations such as the blocking screw, the

M an

locking plate, the distal locking nail and the reamed IMN which would affect the mechanical stability. The degree of injury of different reduction techniques, MIPO or non-MIPO and fracture type made a difference. The time of weight bearing also influence the recovery of the fracture. In addition, the quality, the study design and the ITT analysis were included in the subgroup analysis.

ed

Statistical analysis

We converted all outcome measurements using Review Manager 5.1.3 software and Stata 11.0. We used relative risk (RR) for the

ce pt

dichotomous data and weighted mean difference (WMD) or standardized mean difference (SMD) for the continuous data. A Chi-squared test on N-1 degrees of freedom was used to calculate the statistical heterogeneity, with significance at 0.05. I² (I² = [(Q-df)/Q] x 100%) was used to calculate the percentage of the variability in effect estimates according to the heterogeneity. Q is the χ² statistic and df is the degree of freedom.

Ac

We considered I² values of 25%, 50%, and 75% as low, medium, and high heterogeneity, respectively. A fixed effects model was used if I²< 50%; otherwise, we used the random effects model. If substantial heterogeneities across studies (I2 > 50%) were detected in the index five main meta-analysis, we performed post hoc sensitivity analysis by omitting the outlier studies to determine the sources of heterogeneity. The outliers were detected as the studies of which the confidence interval of the estimated effect size did not well overlap with the pooled overall effect size

Page 10 of 41

ip t cr

. For skewed distribution, if the sample size was less than 60, SD = (upper limit –lower limit)/4. We also calculated SD = SE × N1/2 if we could

us

23

got SE related to Z score. In order to keep the consistent trend in the forest plot, we used the computational method that adjusted mean score =

M an

total score – mean score, and the SD kept unchanged for several outcomes in the functional score and the pain score. The funnel plot24 were used to assess publication bias. If there were asymmetrical plots, we used the trim and fill analysis to assess the stability

25

. When allowed, the

subgroup analyses were isolatedly performed for DTF. We also used the Grading of Recommendations Assessment, Development and

ce pt

Results

ed

Evaluation (GRADE) system to evaluate the quality of evidence by each outcome 26.

The literature search initially yielded 6620 relevant studies, of which 2391 redundant publications were excluded. According to our criteria of inclusion and exclusion, 4213 studies were excluded. With the rest full texts, three literatures had data duplication and only the newest one was

Ac

included 7. One study was another’s further study 8-9 and we included both. Finally 14 literatures including five prospective trials 8-12 and 12 retrospective trials 7, 13-20 with 842 participants were included (Fig. 1). The weighted kappa for the agreement on eligibility between reviewers was 0.84 (95% CI: 0.71-0.93). No publication bias was found in either the funnel plot (Fig. 2). Complication rate

Page 11 of 41

ip t cr

us

IMN decreased the infection rate by 48% (N=695, RR: 0.52 [0.30, 0.89]; p=0.02), but increased the malunion rate by 147% (N=842, RR: 2.47 [1.58, 3.85]; p<0.0001) compared with plating. There was no significant difference in the total complication rate (N=842, RR: 1.14 [0.75, 1.72];

M an

p=0.53), the nonunion rate (N=842, RR: 1.16 [0.51, 2.67]; p=0.72), the secondary surgery rate (N=457, RR=0.87 [0.52, 1.43]; p=0.57), the implant removal rate (N=411, RR=0.82 [0.51, 1.31] (Table 4). All the outcomes were not changed if the studies with moderate to low quality were omitted.

ed

We found medium heterogeneity (I2=57%) in the total complication rate. We conducted the subgroup analysis, and found the origin of the heterogeneity. After the separation in the subgroup analysis of locking plate, the heterogeneity was significantly reduced (p=0.04). We also found

23

ce pt

medium heterogeneity in the secondary surgery rate (I2=67%) and the implant removal rate (I2=52%). When excluding the data of Seyhan et al. which included two different IMNs, the heterogeneity reduced to 37% (p=0.007) and 16% (p=0.04), respectively.

Function score and pain score

Ac

IMN increased the functional score (N=383, SMD: -0.26 [-0.47--0.06]; P=0.01) compared with plating. However, no difference was found in the pain score (N=245, SMD: 0.23 [-0.23-0.70]; P=0.33). All the outcomes were not changed if the studies with moderate to low quality were omitted. We found medium heterogeneity (I²=67%) in the pain score. We noted three studies 7-8, 10 assessed the ankle pain except for Janssen et al. 14

assessing the knee pain. The heterogeneity reduced to 22% after excluding Janssen et al (p=0.01). All the outcomes were not changed if the

Page 12 of 41

ip t cr us

studies with moderate to low quality were omitted. Other data

M an

IMN decreased the operation time (N=320, WMD=-10.66 [-16.64, -4.68]; p=0.0005), but had no influence on the time to union (N=408, WMD: -0.21 [-2.46-2.05]; p=0.86) or the hospital stays (N=97, WMD=-1.08 [-3.33, 1.17]; p=0.35). All the outcomes were not changed if the studies with moderate to low quality were omitted. The high heterogeneity in the time to union (I²=88%) reduced to 42% (p<0.00001) after

ed

excluding Feng et al. 7 Subgroup analysis

ce pt

IMN conferred more than twice the risk of malunion than plating. However, if the technique of either the blocking screw or the distal locking nail was adopted, the inferiority of IMN would disappeared (RR of BS: 3.00 [0.34-26.76], RR of non-BS: 2.45 [1.55-3.85]; RR of DLN: 3.00 [0.34-26.76], RR of non-DLN: 2.33 [1.46, 3.71]) (Table 5). For DTF (Table 5), There was a reduced trend of PRT in the malunion rate for IMN

Ac

(p=0.13). MIPO also showed no statistical significance in the infection rate compared with IMN (RR: 0.61 [0.30-1.23]). However, no benefit of MIPO was found in the nonunion rate (p=0.70). GRADE analysis

Our GRADE analysis showed the comprehensively moderate quality in all the outcomes (Table 6). The most important reasons for the

Page 13 of 41

ip t cr

us

reduced level of evidence were inadequate blinding and lack of concealed allocation. Little sample size also decreased the evidence grade of the pain score, the union time, and the functional score. Furthermore, the heterogeneity in the pain score and the union time had a negative effect on

M an

the quality.

Discussion

ed

Plating is always accepted as the first choice for DTF until the late twentieth century when IMN has gained satisfactory results and wide popularity with the improvement in techniques

2-5

. In this systematic review and meta-analysis we asked (1) which fixation is better regarding

ce pt

the clinical function and the complications, and (2) which modifying factors affect the comparative effect between both techniques. To our knowledge, the present meta-analysis is the first to comprise all the available comparative controlled evidence and comprehensively investigate the difference in function, pain and complications between IMN and plating for DTF. As the previous systematic reviews only

Ac

included retrospective observational studies, their validity was limited by imbalances between groups, lack of independent assessment, failure for blinding of outcome measurements, and inadequate follow-up 5. Recently, a systematic review 27 focusing on complication rate, was done by including 2 RCTs

10-11

and 4 retrospective comparative studies

14, 16-17, 28

. All the studies were also included in our analysis except one study,

which didn’t meet our inclusion criteria, involved serious intra-articular fracture and lacked concrete data of complication rate or functional

Page 14 of 41

ip t cr

us

score in group IMN 28. This systematic review turned out extra same results in complication rate, but the lacking of enough sample size limited the level of evidence. The present sample size in our analysis is larger with the elevated quality of GRADE evidence compared with the previous

M an

reviews. Furthermore, our analysis has no language restriction so that the publication bias is reduced as far as possible. We acknowledge limitations: (1) some might argue against the inclusion of the retrospective studies because of their inherent risk of bias. However, they formed most of the patients and, despite methodological limitations, ignoring this source of data might underpower the analysis,

ed

raise the risk of false negative error, and influence the accuracy of our findings. Most of them balanced the demographic parameters between two groups, which limited the opportunity of selection bias. Furthermore, the results of the subgroup analysis remained unchanged after excluding

ce pt

these retrospective studies. (2) The unavailable raw data of the early literatures, such as the weight bearing and the fibular fixation, made part of our subgroup analyses impossible. In addition, the different scoring criteria for function and pain across the studies might lead to the heterogeneity. These drawbacks necessitated the uniform and standardized format of follow-up in future. (3) The heterogeneity was significant in

Ac

the outcomes of the total complication rate, the pain score, and the time to union. The sensitive analysis and the subgroup analysis were done to find the origins. (4) The small sample size in the subgroup analysis reduced the precision of the pooled estimates and the ability to detect the statistical significance of some variables, i.e. the blocking screw, the distal locking nail, and the percutaneous reduction techniques in the evaluation of the malunion rate, and MIPO in the evaluation of the infection rate. More RCTs would be warranted to clarify them.

Page 15 of 41

ip t cr

us

One of the most significant results of our analysis is IMN had higher function score and comparable pain score of DTF compared with plating. However, we were unable to explain the clinical implications of the statistical difference because SMD was merely an absolute value without

M an

unit or cut-off reference. Of the included four studies in the pain comparison, three assessed the ankle 7-8, 10 while one assessed the knee 14. The present data, indicating the similar rates of hardware removal and the ankle pain of both modalities, might confirm the possible relationship between the ankle pain and the hardware irritation7-8, 10. In addition, the present analysis failed to show the correlation between the knee pain and 8, 14

. Vallier et al. 8 found a trend of more knee pain with IMN, whereas 95% of the patients returned to work without activity

ed

the leg function

limitation. Also, the more severe knee pain with IMN reported by Janssen et al. 14 mainly occurred during kneeling or squatting, which did not

ce pt

influence the excellent knee function.

Our data showed the malunion rate of IMN was considerably higher than plating, which were in agreement with the biomechanical data 29 and previous systematic review 27. Interestingly, when the studies with combination of IMN and either the blocking screw or the distal locking nail

Ac

were analyzed separately in the subgroup analysis, the inferiority of IMN disappeared. In addition, there was a trend of reduced malunion rate of IMN when the percutaneous reduction techniques were adopted. In a multicenter case series of tibial metaphyseal fractures treated with the distal locking nail, Attal et al. 30 found the malalignment rate was only 5.4% for DTF, respectively. Nork et al. 31-32 treated tibial metaphyseal fractures using reamed IMN combined with the percutaneous locking techniques, which showed acceptable alignment in 92% of the patients. All the

Page 16 of 41

ip t cr

us

clinical and biomechanical evidence supported the potential advantages of the blocking screw, the distal locking nail, and the percutaneous reduction techniques. Interestingly, the subgroup analysis revealed benefit towards IMN rather than plating when the weight-bearing was

M an

allowed before callus forming. Considering the nature of load-sharing device, IMN biomechanically tolerated more axial load than plating. Whether this biomechanical advantage enables IMN of early weight-bearing needs further evidence. With regard to DTF, our results demonstrated excellence of IMN over plating in the infection rate, which one previous systematic review

ed

failed 5 but the other one turned out the same result 27. It turned out the previous result was potentially biased by the higher percentage of the open fractures in IMN group. Besides, MIPO had the similar infection rate with IMN in the subgroup analysis, which was consistent with the

ce pt

cadaveric data 33. An enough-powered RCT with long-term follow-up is necessary to verify it. The present analysis and the previous systematic reviews 5, 27 both suggested similar nonunion rate between IMN and plating, which might be attributed to the minimal exposure, the precise reduction and the rigid fixation. Despite its theoretical advantage in local vascularity 33, we did

Conclusion

Ac

not detect the statistical significance of MIPO in lowering the nonunion risk. More comparative data would be needed to address this hypothesis.

IMN and plating appear to share the similar nonunion rate, secondary surgery rate and implant removal rate for DTF. Specifically for DTF, IMN may be preferential given its higher function score, lower risk of infection and comparable pain score and time to union. Plating could be

Page 17 of 41

ip t cr

Ac

ce pt

ed

M an

require a rigorous and adequately powered RCT to prove.

us

an alternative when BS, DLN, and PRT are unavailable. However, with the biases in our meta-analysis all these viewpoints will ultimately

Page 18 of 41

ip t cr

us

REFERENCES

1. Marsh JL, Slongo TF, Agel J, et al. Fracture and Dislocation Classification Compendium - 2007: Orthopaedic Trauma Association

M an

Classification, Database and Outcomes Committee. J Orthop Trauma. 2007;21:S1-S133.

2. Hiesterman TG, Shafiq BX, Cole PA. Intramedullary nailing of extra-articular proximal tibia fractures. J Am Acad Orthop Surg. 2011; 19: 690-700.

ed

3. Bedi A, Le TT, Karunakar MA. Surgical treatment of nonarticular distal tibia fractures. J Am Acad Orthop Surg. 2006; 14: 406-416. 4. Bhandari M, Audige L, Ellis T, et al. Operative treatment of extra-articular proximal tibial fractures. J Orthop Trauma. 2003;17:591-595.

ce pt

5. Zelle BA, Bhandari M, Espiritu M, et al. Treatment of distal tibia fractures without articular involvement: a systematic review of 1125 fractures. J Orthop Trauma. 2006;20:76-79.

2007;15:97-106.

Ac

6. Bong MR, Kummer FJ, Koval KJ, et al. Intramedullary nailing of the lower extremity: biomechanics and biology. J Am Acad Orthop Surg.

7. Feng YZ, Hong JJ, Peng L, et al. Comparison of two minimally invasiveinternal fixed methods for the treatmentof distal tibio-fibula. Chin J Surg. 2011;2:113-118.

8. Vallier HA, Cureton BA, Patterson BM. Factors influencing functional outcomes after distal tibia shaft fractures. J Orthop Trauma.

Page 19 of 41

ip t cr us

2012;26:178-183.

shaft fractures. J Orthop Trauma. 2011;25:736-741.

M an

9. Vallier HA, Cureton BA, Patterson BM. Randomized, prospective comparison of plating versus intramedullary nail fixation for distal tibia

10. Guo JJ, Tang N, Yang HL, et al. A prospective, randomised trial comparing closed intramedullary nailing with percutaneous plating in the treatment of distal metaphyseal fractures of the tibia. J Bone Joint Surg Br. 2010;92:984-988.

ed

11. Im GI, Tae SK. Distal metaphyseal fractures of tibia: a prospective randomized trial of closed reduction and intramedullary nail versus open reduction and plating and screws fixation. J Trauma. 2005;59:1219-1223.

ce pt

12. Mauffrey C, McGuinness K, Parsons N, et al. A randomised pilot trial of "locking plating" fixation versus intramedullary nailing for extra-articular fractures of the distal tibia. J Bone Joint Surg Br. 2012;94:704-708. 13. Chen N, He QQ. Clinical analysis for two fixation methods in distal tibiofibular fracture. Chin Prac Med. 2008;3:116-117.

Ac

14. Janssen KW, Biert J, van Kampen A. Treatment of distal tibial fractures: plating versus nail: a retrospective outcome analysis of matched pairs of patients. Int Orthop. 2007;31:709-714. 15. Li Y, Liu L, Tang X, et al. Comparison of low, multidirectional locked nailing and plating in the treatment of distal tibial metadiaphyseal fractures. Int Orthop. 2012;36:1457-1462.

Page 20 of 41

ip t cr

versus intramedullary nailing. J Orthop Trauma. 2008;22:307-311.

us

16. Vallier HA, Le TT, Bedi A. Radiographic and clinical comparisons of distal tibia shaft fractures (4 to 11 cm proximal to the plafond): plating

Injury. 2006;37:531-535.

M an

17. Yang SW, Tzeng HM, Chou YJ, et al. Treatment of distal tibial metaphyseal fractures: Plating versus shortened intramedullary nailing.

18. Zhang C, Jiang Y, An ZQ. Interlocking intramedullary nailing versus percutaneous plating in osteosynthesis of metaphyseal fractures of

ed

distal tibia. Chin J Orthop Trauma. 2007;9:131-134.

19. Zhu HW. Interlocking itramedullary nail for distal tibial fracture. Chin Prac Med. 2009;4:33.

ce pt

20. Seyhan M, Unay K, Sener N. Intramedullary nailing versus percutaneous locked plating of distal extra-articular tibial fractures: a retrospective study. Eur J Orthop Surg Traumatol 2012; s00590-012-1016-x. 21. Furlan AD, Pennick V, Bombardier C, et al. 2009 updated method guidelines for systematic reviews in the Cochrane Back Review Group.

Ac

Spine (Phila Pa 1976). 2009;34:1929-1941.

22. Fernandez-valencia JA. How to quote Gustilo open fracture classification? J Shoulder Elbow Surg. 2009;18:e32. 23. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med. 2002; 21: 1539-1558. 24. Higgins JPT, Green S, Cochrane Collaboration. Cochrane handbook for systematic reviews of interventions. Chichester, England Hoboken

Page 21 of 41

ip t cr us

NJ: Wiley-Blackwell; 2008.

25. Hayashino Y, Noguchi Y, Fukui T. Systematic evaluation and comparison of statistical tests for publication bias. J Epidemiol. 2005; 15:

M an

235-243.

26. Phi L, Ajaj R, Ramchandani MH, et al. Expanding the Grading of Recommendations Assessment, Development, and Evaluation (Ex-GRADE) for Evidence-Based Clinical Recommendations: Validation Study. Open Dent J. 2012;6:31-40.

ed

27. Iqbal HJ, Pidikiti P. Treatment of distal tibia metaphyseal fractures; plating versus intramedullary nailing: A systematic review of recent evidence. Foot Ankle Surg. 2013 Sep; 19(3):143-7.

2010; 34(4): 583–8.

ce pt

28. Joveniaux P, Ohl X, Harisboure A, et al. Distal tibia fractures: management and complications of 101 cases. International Orthopaedics.

29. Zlowodzki M, Williamson S, Cole PA, et al. Biomechanical evaluation of the less invasive stabilization system, angled blade plate, and

Ac

retrograde intramedullary nail for the internal fixation of distal femur fractures. J Orthop Trauma. 2004; 18: 494-502. 30. Attal R, Hansen M, Kirjavainen M, et al. A multicentre case series of tibia fractures treated with the Expert Tibia Nail (ETN). Arch Orthop Trauma Surg. 2012; 132: 975-984.

31. Nork SE, Schwartz AK, Agel J, et al. Intramedullary nailing of distal metaphyseal tibial fractures. J Bone Joint Surg Am. 2005; 87:

Page 22 of 41

ip t cr us

1213-1221.

32. Nork SE, Barei DP, Schildhauer TA, et al. Intramedullary nailing of proximal quarter tibial fractures. J Orthop Trauma. 2006; 20: 523-528.

Ac

ce pt

ed

cadaveric study. J Orthop Trauma. 2002;16:691-695.

M an

33. Borrelli J, Jr., Prickett W, Song E, et al. Extraosseous blood supply of the tibia and the effects of different plating techniques: a human

Page 23 of 41

Acknowledge

7

We thank the corresponding authors of included studies, especially Vallier HA, for their

8

sincere assistance in obtaining and verifying data. The project was funded by Zhejiang

9

Provincial Natural Science Foundation of China (Y2110239), National Natural Science

10

Foundation of China (81101345), and Zhejiang Key Program Science and Technology

11

(2011C13033).

us

cr

ip t

6

12

Ac ce p

te

d

M

an

13

37

Page 24 of 41

13

Conflicts of Interest and Source of Funding: Xun-Zi Cai is currently receiving grants from Zhejiang Provincial Natural Science

15

Foundation of China (Y2110239), National Natural Science Foundation of China

16

(81101345) and Zhejiang Key Program Science and Technology (2011C13033). For the

17

remaining authors none were declared.

ip t

14

18

cr

19

Ac ce p

te

d

M

an

us

20

38

Page 25 of 41

Figure legends

21

Fig. 1. A QUOROM flowchart illustrated the selection of studies included in our

22

meta-analysis

23

Fig. 2. Funnel plot for total complication rate between IMN and plating showed no

24

publication bias in visual. IMN = intramedullary nailing; RR = risk ratio.

cr

ip t

20

Ac ce p

te

d

M

an

us

25

39

Page 26 of 41

cr

ip t

Table

Table 1. Methodological quality of the included studies based on the 12-items scoring system Outcome

Avoided

Similar or

Patient

Acceptable

provider

assessor

selective

avoided

complianc

drop-out

blinded

blinded

reporting

cofactor

e#

rate$

Unclear

Unclear

Unclear

Yes

Yes

Yes

Yes

No

Unclear

Yes

Yes

Unclear

No

Yes

Unclear

Unclear

No

No

Yes

Unclear

Unclear

Vallier HA [9]

Yes

yes

Yes

Yes

Vallier HA [8]

Yes

yes

Yes

Yes

Mauffery C [12]

Yes

yes

Yes

Unclear

Li Y [15]

No

No

Yes

Seyhan M [20]

No

No

Zhang C [18]

No

No

Feng YZ [7]

No

No

Chen N [13]

No

No

Zhu HW [19]

No

No

Similar

Patient

adequately*

concealed

baseline

blinded

Guo JJ [10]

Yes

Yes

Yes

Im GI [11]

Yes

Yes

Janssen K [14]

No

No

Yang SW [17]

No

Vallier HA [16]

Similar

ITT

timing

analysis&

Yes

Yes

No

quality+

High

Yes

Yes

Yes

Yes

Yes

No

High

Yes

Yes

Yes

No

Yes

No

Moderate

Yes

Yes

Yes

Yes

Unclear

Yes

No

Moderate

Unclear

Yes

Yes

Yes

Unclear

Yes

No

Moderate

Unclear

Yes

Yes

Yes

Yes

Yes

Yes

Yes

High

Unclear

Yes

Yes

Yes

Yes

Yes

Yes

No

High

Unclear

Unclear

Yes

Yes

Yes

Yes

Yes

No

High

Unclear

Unclear

Yes

Yes

Yes

Yes

Unclear

Yes

No

Moderate

Yes

Unclear

Unclear

Unclear

Yes

Yes

Yes

Unclear

Yes

No

Moderate

Yes

Unclear

Unclear

Unclear

Yes

No

Yes

Unclear

Yes

No

Moderate

Yes

Unclear

Unclear

Unclear

Yes

Unclear

Yes

Unclear

Yes

No

Moderate

unclear

Unclear

Unclear

Unclear

Yes

No

Yes

Unclear

Yes

No

Low

Yes

Unclear

Unclear

Unclear

Yes

Yes

Yes

Yes

unclear

No

Moderate

d

M

Yes

Unclear

ep te

Study

an

Allocation

us

Care

Randomized

Ac c

* Only if the method of sequence generated was explicitly described could get a ‘‘Yes’’; sequence generated by ‘‘Dates of Admission’’ or ‘‘Patients Number’’ received a ‘‘No’’; # intermittent treatment or therapy duration less than 6 months means ‘‘Yes’’, otherwise ‘‘No’’; $ drop-out rate > 20% means ‘‘No’’, otherwise ‘‘Yes’’; & ITT = intention-to-treat, only if all randomized patients are analyzed in the group they were allocated to could receive a ‘‘Yes’’; + ‘‘Yes’’ items greater than 7 means ‘‘High’’; greater than 4 but no more than 7 means ‘‘Moderate’’; no more than 4 means ‘‘Low’’.

Page 27 of 41

cr

ip t 44.3 (23-70)

Im GI [11]

Prospective, randomized

39/39

41.1 (17-65)

Janssen K [14]

Retrospective, nonrandomized

12/12

42.1 (25-84)

Yang SW [17]

Retrospective, nonrandomized

Vallier HA [16]

Multicenter, retrospective, nonrandomized

Vallier HA [9]

Prospective, randomized

13/14

50/35

Loss to follow up

12

13/13

51.5 (20-86)

76/37

39.1 (16-77)

56/48

38.3 (18-95)

AO classific ation

46/18

24

12/12

12/15

77/34

85/19

63

33

24

20

Internal fixation

Reduction (IMN/plate)

Fracture type

fixed fibula (percentage+)

Inclusion criteria

R/MIPO,L CP

Percutaneou s/ Percutaneou s

Closed, Gustilo I

Unclear

Extra-articular (≥ 3cm of distal fragment)

R/LCP

Manual/Ope n

Closed, Gustilo I

38 (59%)

Extra-articular or minimally displaced extension into the ankle joint

R/Non-LC P

Percutaneou s/ Open

Closed, Gustilo I

22 (92%)

Extra-articular, ≥ 18 years, closed or Gustilo I fracture

Non-R/LC P

Manual/ Open

Closed, Gustilo I/II/III

27(100%)

Extra-articular (3-5 cm of the distance from the fracture line to ankle articular surface)

Closed, Gustilo I/II/III

42 (36%)

Extra-articular (4-11cm proximal from the tibial plafond)

Closed, Gustilo I/II/III

28 (27%)

Extra-articular (4-11cm proximal from the tibial plafond)

an

57/54

Follow -up(mo nth)

5/9

ep te

Prospective, randomized

Ac c

Guo JJ [10]

Gender (male/fe male)

43-A

M

Design

Age (years)

d

Study

Sample size (IMN/pl ate)

us

Table 2. Study characteristics and details of interventions of the included studies

NA/NA

NA/NA

NA/NA

NA/NA

43-A/C1 *

42-A/B2 #

43-A

42-A

#

42-A

#

R/MIPO,N on-LCP

R/ MIPO,Non -LCP

Percutaneou s/ Percutaneou s Percutaneou s/ Percutaneou s

Page 28 of 41

Retrospective, nonrandomized

Zhang C [18]

Retrospective, nonrandomized

Feng YZ [7]

Retrospective, nonrandomized

Chen N [13]

Retrospective, nonrandomized

Zhu HW [19]

Retrospective, nonrandomized

25/36

39.9 (19-81)

27/24

41.2 (18-70)

36/10

34/27

12

24.7

21.34

31/20

21

cr

ip t Seyhan M [20]

23/23

38.0 (21-59)

8/16

43-A

Non-R/Non -LCP

Percutaneou s/ Percutaneou s

Closed, Gustilo I

43-A

ETN/MIPO ,LCP

Manual/Mn ual

28 (27%)

Extra-articular (4-11cm proximal from the tibial plafond)

Unclear

Extra-articular (within two Müller squares of the ankle), >18 years, closed or Gustilo I

Closed, Gustilo I/II/III

42 (91%)

Closed or Gustilo I/ II fractures, skeletally mature, early failure of conservative treatment

ETN/MIPO ,LCP

Percutaneou s/ Percutaneou s

Closed, Gustilo I/II/III

31(51%)

Extra-articular (4-10cm proximal from the plafond)

43-A

R/Non-LC P

Percutaneou s/ Open

Closed, Gustilo I/II/III

45 (88%)

Extra-articular

Manual/ Percutaneou s

Closed, Gustilo I

44 (88%)

Extra-articular, Gustilo I/II, Tscherne 0~2

42-A

us

11/7

Closed, Gustilo I/II/III

#

an

Li Y [15]

Retrospective, nonrandomized

12/12

41.5 (24-60)

22

NA/NA

M

Prospective, randomized

73/13

NA/NA

d

Mauffery C [12]

56/48

38.3 (18-95)

NA/NA

ep te

Prospective, randomized

Ac c

Vallier HA [8]

Percutaneou s/ Percutaneou s

R/ MIPO,Non -LCP

NA/NA

R， 43-A

22/28

44 (19-74)

30/20

12

NA/NA

43-A

Non-R/MI PO,LCP

25/21

31 (21-50)

38/8

24

NA/NA

43-A

R/LCP

Manual/ Open

Closed, Gustilo I

5 (11%)

Extraaticular, Middle and lower fracture, closed fracture

80/80

38.8 (20-59)

43-A

R/MIPO,L CP

Percutaneou s/ Open

Closed, Gustilo I/II/III

Unclear

Extraarticular

85/62

22.4

6/7

IMN = intramedullary nailing; AO = Arbeitsgemeinschaft für Osteosynthesefragen; *43-C1 in this study had minimally displaced extension into articular surface which was al

Page 29 of 41

ip t cr

Ac c

ep te

d

M

an

us

tibial plafond.R=reamed; LCP = locking compression plating; ETN = Expert Tibia Nail; N.A.= not applicable; + percentage = number of fixed fibular fracture / total fracture

Page 30 of 41

ip t cr

Table 3. the details of outcomes of included studies Implant removal

surgery rate(%)

rate(%)

Nonunion(%)

IMN

Plate

IMN

Plate

IMN

Plate

IMN

Plate

IMN

Plate

Guo JJ [10]

0

0

0

0

7

15

52

59

52

59

Im GI [11]

12

0

9

7

3

23

N.A.

N.A.

N.A.

Janssen K [14]

50

17

0

0

8

8

100

75

Yang SW [17]

23

7

0

0

0

0

N.A.

N.A.

29

5

7

3

5

3

30

16

8

25

19

7

4

5

6

18

8

0

0

8

25

8

8

13

4

0

0

4

13

16

3

4

6

0

Zhang C [18]

19

0

0

0

Feng YZ [7]

23

0

0

Chen N [13]

0

0

Zhu HW [19]

1

8

Vallier HA [8-9] Mauffery C [12] Li Y [15] Seyhan M [20]

Functional score

Pain score

[M(SD)]

[M(SD)]

[M(SD)]*

[M(SD)]#

[M(SD)]

IMN

Plate

IMN

Plate

IMN

Plate

IMN

Plate

17.7

17.6

81.23

97.9

N.A.

N.A.

86.1

83.9

7.5

8.5

(2.9)

(2.2)

(11.43)

(9.61)

(8.16)

(6.93)

(4.05)

(3.91)

18

20

72

89

N.A.

N.A.

88.5

88.3

(13)

(15)

(15)

(27.5)

(1.76)

(1.76)

N.A.

N.A.

21.4

19.3

123

107

9.8

9.5

N.A.

N.A.

43.17

6.92

(4.35)

(4.97)

(30)

(33.75)

(4)

(2.75)

(37.91)

(14.76)

22.6

27.8

N.A.

N.A.

6.4

6.5

86.2

83.9

(4.3)

(7.6)

(2)

(1.6)

(3.2)

(7.1)

N.A.

N.A.

8

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

23

9

10

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

73

72

0.35

0.31

(17.25)

(16.25)

(0.22)

(0.24)

42

8

33

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

82.3

66.7

(13.51)

(13.07)

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

21.3

23.1

76

90

5.8

8.9

89

87.6

(3.5)

(3.6)

(16.6)

(20.3)

(2.1)

(3.1)

(7.1)

(8.4)

N.A.

N.A.

17

12

56

8

36

15.7

17.24

(3.92)

(3.05)

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

0

17

N.A.

N.A.

N.A.

N.A.

18.59

20

82.22

89.16

79

(3.96)

(12.51)

(14.12)

N.A.

82

(3.75)

N.A.

(7)

(7)

N.A.

N.A.

0

14

11

N.A.

N.A.

N.A.

N.A.

21.1

15.4

94

100

(12)

(14)

N.A.

N.A.

6.1

(2.9)

N.A.

6.7

(3)

N.A.

(2.5)

(2.6)

0

0

0

0

4

0

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

0

1

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

an

Plate

M

N.A.

92

75

N.A.

N.A.

ep te

[16]

Hospital stays

IMN

Ac c

Vallier HA

Operation time

d

Study

Time to union

us

Infection(%)

Secondary

Malunion(%)

Page 31 of 41

ip t cr

IMN = intramedullary nailing; MIPO = minimally invasive plating osteosynthesis; N.A.= not applicable; OMAS = olerud and molander functional ankle score; AOFAS = American Orthopaedic Foot and Ankle surgery scores ;

us

MFA = Musculoskeletal Function Assessment; *of the functional score, Guo JJ adopted AOFAS; Im GI, Yang SW, Mauffery C, Li Y and Zhang C adopted OMAS; Vallier adopted MFA.#of the pain score,Janssen K adopted

Ac c

ep te

d

M

an

anterior knee pain score; Feng YZ adopted VAS pain score;Guo JJ adopted the pain score of AOFAS; Vallier HA adopted the pain scoe of MFA.

Page 32 of 41

ip t cr

Ac c

ep te

d

M

an

us

Table 4. The statistical comparison in outcomes between IMN and plating Sample size Outcomes Fracture Type Event (IMN/plating) (IMN/plating) Mean [CI] I2 P 1.RR total complication rate DTF 118/89 443/399 1.14 [0.75, 1.72] 57% 0.53 Malunion rate DTF 68/22 443/399 2.47 [1.58, 3.85] 29% <0.0001 Nonunion rate DTF 13/9 443/399 1.16 [0.51, 2.67] 0% 0.72 Infection rate DTF 19/35 369/326 0.52 [0.30, 0.89] 15% 0.02 Secondary surgery rate DTF 73/75 250/207 0.87 [0.52, 1.43] 67% 0.57 Implant removal rate DTF 48/58 225/186 0.82 [0.51, 1.31] 53% 0.42 2.WMD Time to union DTF N.A. 200/208 -0.21 [-2.46, 2.05] 88% 0.86 Operation time DTF N.A. 162/158 -10.66 [-16.64, -4.68] 65% 0.0005 Hospital stays DTF N.A. 48/49 -1.08 [-3.33, 1.17] 79% 0.35 3.SMD Functional score DTF N.A. 198/185 -0.26 [-0.47, -0.06] 0% 0.01 Pain score DTF N.A. 123/122 0.23 [-0.23, 0.70] 67% 0.33 DTF = distal tibial fracture; IMN = intramedullary nailing; RR = relative risk; WMD = weighted mean difference; SMD = standardized mean difference; N.A. = not applicable; RR > 1 or WMD > 0 or SMD > 0 means the results favoring plating, vice versa.

Page 33 of 41

Nonunion rate

Infection rate

Factors RR(95%CI)

Subgroup

RR(95%CI)

Subgroup

RR(95%CI)

Pro (4)

1.00

Pro (4)

1.74

Pro (4)

1.10

Quality*

Yes (4)

1.30

3.01

1.44

Pain score

RR(95%CI)

Subgroup

RR(95%CI)

Subgroup

SMD(95%CI)

Subgroup

SMD(95%CI)

Pro (4)

0.54

Pro (3)

0.81

Pro (4)

-0.23

Pro (2)

-0.04

(0.30, 4.06)

Ret (9)

Functional score

Ret (8)

(0.25, 1.13)

0.50

(0.54, 1.22)

Ret (4)

0.97

(-0.48, 0.01)

Ret (3)

-0.33

(-0.45, 0.38)

Ret (2)

0.66

(0.69, 2.45)

(1.68, 5.39)

(0.61, 3.40)

(0.23, 1.10)

(0.36, 2.60)

(-0.68, 0.03)

(-0.30, 1.62)

p=0.49

p=0.23

p=0.91

p=0.90

p=0.75

p=0.68

p=0.19

1.00

(0.68, 1.47)

No (9)

Ret (9)

1.37

Yes (4)

1.74

Ac c

Ret (9)

(0.87, 3.47)

d

(0.68, 1.47)

Secondary operative rate

Subgroup

M

Subgroup

ep te

Design*

cr

ip t Malunion rate

an

General complication

us

Table 5. Subgroup analysis of the included studies between IMN and plating based on influential factors

Pro (4)

(0.87, 3.47)

No (9)

3.01

1.10

Pro (4)

(0.30, 4.06)

Ret (9)

1.44

0.54

Pro (3)

(0.25, 1.13)

Ret (8)

0.50

0.81

Yes (4)

(0.54, 1.22)

Ret (4)

0.97

-0.23

Yes (2)

(-0.48, 0.01)

No (3)

-0.33

-0.04

(-0.45, 0.38)

No (2)

0.66

(0.87, 2.14)

(1.68, 5.39)

(0.61, 3.40)

(0.23, 1.10)

(0.36, 2.60)

(-0.68, 0.03)

(-0.30, 1.62)

p=0.49

p=0.23

p=0.91

p=0.90

p=0.75

p=0.68

p=0.19

Page 34 of 41

ETN

Yes (1)

1.12

(0.63, 2.80)

No (12)

Yes (1)

Screw

p=0.71

p=0.07

p=0.59

1.27

Yes (1)

3.00

(0.34,26.76)

No (11)

2.33

(0.81, 1.99)

(1.46, 3.71)

p=0.73

p=0.82

1.00

(0.28, 3.52)

No (12)

1.02

(0.39, 2.68)

1.15

(0.74, 1.79)

Ac c

Blocking

No (12)

(1.81, 5.56)

1.00

Yes (1)

3.00

Yes (1)

(0.33, 8.95)

(0.69, 1.80)

(0.28, 3.52)

No (11)

3.18

1.71

Yes (1)

N.A.

No (11)

N.A.

No (11)

Yes (1)

N.A.

No (10)

2.45

(1.55, 3.85)

0.48

0.78

Yes (1)

(0.36, 1.67)

No (6)

0.87

No (6)

N.A.

No (11)

-0.32

0.17

(-0.25, 0.60)

No (3)

0.31

(0.48, 1.57)

(-0.55, -0.09)

(-0.42,1.04)

p=0.50

p=0.83

p=0.28

p=0.75

0.33

N.A

N.A.

Yes (1)

0.62

0.33

0.53

(0.31, 0.94)

-0.18

Yes (0)

N.A.

No (4)

N.A.

(-0.76, 0.40)

N.A

N.A.

No (6)

-0.27

(-0.49, -0.06)

N.A.

N.A

N.A.

p=0.76

Yes (1)

(0.04, 2.97)

No (12)

Yes (1)

(0.27, 0.87)

p=0.59

Yes (1)

-0.66

(-0.48, 0.36)

(0.35, 1.10)

(0.34,26.76)

No (12)

Yes (1)

(0.04, 2.97)

N.A.

Yes (1)

0.86

(0.18, 4.05)

M

No (12)

Yes (1)

cr

ip t (0.74, 2.24)

1.33

us

Yes (1)

an

1.29

d

Yes (1)

ep te

ITT

-0.18

N.A.

Yes (0)

N.A.

No (4)

N.A.

(-0.76, 0.40)

N.A

N.A.

No (6)

-0.27

(-0.49, -0.06)

Page 35 of 41

2.60

0.85

(0.25, 2.92)

No (5)

1.49

Yes (7)

0.36

Yes (3)

(0.17, 0.75)

No (5)

0.87

N.A.

0.60

p=0.76

Yes (4)

(0.17, 2.16)

No (4)

1.10

-0.23

N.A.

Yes (2)

(-0.49, 0.04)

No (3)

-0.31

-0.04

(-0.51, 0.42)

No (2)

0.62

(1.05, 2.72)

(1.48, 4.57)

(0.48, 4.67)

(0.38, 2.00)

(0.63, 1.92)

(-0.63, 0.00)

(-0.39, 1.63)

p=0.04

p=0.77

p=0.51

p=0.12

p=0.24

p=0.68

p=0.24

1.08

Yes (7)

(0.53, 2.21)

1.17

(0.74, 1.85)

2.04

Yes (7)

(1.23, 3.39)

No (6)

1.30

4.43

Yes (6)

(0.47, 3.60)

No (6)

Ac c

No (6)

Yes (8)

(1.10, 4.65)

No (5)

cr

ip t Yes (7)

1.69

2.26

us

(0.46, 1.36)

No (5)

MIPO

Yes (8)

p=0.68

M

plate

0.79

d

Yes (8)

N.A.

ep te

Locking

p=0.86

an

p=0.84

(1.68,

0.91

0.61

Yes (4)

(0.31, 1.18)

No (6)

0.38

0.80

Yes (3)

(0.41, 1.53)

No (3)

0.84

-0.17

Yes (3)

(-0.44, 0.09)

No (4)

-0.38

0.03

(-0.27, 0.33)

No (1)

1.22

(0.22, 3.86)

(0.15, 0.98)

(0.16, 4.33)

(-0.69, -0.07)

(0.33, 2.10)

p=0.70

p=0.43

p=0.95

p=0.31

p=0.01

11.65)

p=0.86

Fibular fixation#

Yes (7)

1.27

(0.73, 2.23)

p=0.16

Yes (7)

5.15

(2.31,11.51)

Yes (7)

1.06

(0.27, 4.19)

Yes (7)

0.30

(0.14, 0.68)

N.A.

N.A.

Yes(1)

-0.31

(-0.73, 0.12)

Yes (2)

0.66

(-0.30, 1.62)

Page 36 of 41

(1.15, 4.24)

(0.54, 7.32)

p=0.80

p=0.11

p=0.51

1.11

Yes (8)

1.15

2.05

Yes (8)

(1.26, 3.34)

No (5)

1.18

p=0.07

Yes (7)

(0.44, 2.89)

No (5)

1.32

(1.71,

N.A.

N.A.

No (6)

(0.34, 4.07)

No (5)

0.58

N.A.

Yes (6)

(0.30, 1.13)

0.41

0.84

Yes (4)

(0.50, 1.41)

No (1)

2.54

-0.03

No (1)

(-0.28, 0.21)

(-0.25, 0.60)

p=0.27

p=0.36

-0.30

Yes (3)

(-0.56, -0.05)

No (3)

0.17

-0.19

0.27

(-0.38, 0.92)

No (1)

0.23

(0.24, 7.40)

(0.16, 1.05)

(0.11, 59.23)

(-0.53, 0.15)

(-0.33, 0.79)

p=0.87

p=0.54

p=0.50

p=0.59

p=0.93

ep te

(0.59, 2.23)

5.61

1.12

No (3)

us

1.99

(0.55, 3.98)

(0.64, 1.93)

No (5)

No (3)

cr

ip t for IMN

2.21

an

Yes (8)

No (3)

M

PRT

1.48

d

No (3)

18.46)

Reamed

Yes (9)

IMN

1.01

(0.60, 1.69)

No (4)

1.50

p=0.13

Ac c

p=0.94

Yes (9)

2.17

Yes (9)

(1.34, 3.52)

No (4)

4.73

1.31

Yes (8)

(0.54, 3.14)

No (4)

0.33

0.38

Yes (6)

(0.20, 0.75)

No (4)

1.11

0.94

Yes (4)

(0.58, 1.53)

No (1)

0.20

-0.20

Yes (3)

(-0.44, 0.03)

No (3)

-0.45

0.27

(-0.38, 0.92)

No (1)

0.23

(0.77, 2.95)

(1.43,15.65)

(0.01, 7.45)

(0.41, 2.98)

(0.03, 1.47)

(-0.86, -0.04)

(-0.33, 0.79)

p=0.36

p=0.24

p=0.41

p=0.08

p=0.14

p=0.30

p=0.93

Page 37 of 41

ip t cr

us

IMN = intramedullary nailing; ITT = intention-to-treat; ETN = expert tibia nail; MIPO = minimally invasive plating osteosynthesis; PRT = percutaneous reduction technique; Pro = prospective, Ret = retrospective; * all the prospective studies were with high quality according to 12-items, while retrospective studies were all with moderate or low quality, yes =

an

high quality according to 12-items scoring system; no = moderate and low quality; # yes = cases with intact fibula or fibular fixation / total cases >50%, no = cases with intact fibula or

M

fibular fixation / total cases <50%; $ yes = exclusively included closed fracture and Gustilo type I fractures, no = also included Gustilo type II and Gustilo type III fractures; & yes =

Ac c

ep te

d

weight bearing before callus formed, no = weight bearing after callus formed.

Page 38 of 41

ip t cr

Ac c

ep te

d

M

an

us

Table 6. GRADE evidence of comparison between IMN and plating in efficacy and safety for treatment of TMF Summary of findings Quality assessment Sample size Outcome RR/WMD/SMD [95%CI] Limitations* Inconsistency# Indirectness Imprecision$ Others& Quality (IMN/plating) Functional score Serious No serious Serious Serious None Moderate 198/185 -0.26 [-0.47, -0.06] Pain score 123/122 0.23 [-0.23, 0.70] Serious Serious No serious Serious None Moderate Time to union 200/208 -0.21 [-2.46, 2.05] Serious Serious No serious No serious None Moderate Malunion rate Serious No serious No serious No serious None Moderate 443/399 2.47 [1.58, 3.85] Infection rate Serious No serious No serious No serious None Moderate 369/326 0.52 [0.30, 0.89] Nonunion rate Serious No serious No serious No serious None Moderate 443/399 1.16 [0.51, 2.67] Secondary surgery rate 250/207 0.87 [0.52, 1.43] Serious No serious No serious No serious None Moderate Implant removal rate Serious No serious No serious No serious None Moderate 225/186 0.82 [0.51, 1.31] Total complication rate Serious Serious No serious No serious None Moderate 443/399 1.14 [0.75, 1.72] Operation time Serious Serious No serious Serious None Moderate 162/158 -10.66 [-16.64, -4.68] Hospital stay 48/49 -1.08 [-3.33, 1.17] Serious Serious No serious Serious None Moderate GRADE = Grading of Recommendations Assessment, Development and Evaluation; RR = risk ratio; WMD = weighted mean difference;* inadequate blinding, lack of allocation concealed may cause limitations; # inconsistent report of outcomes and significant heterogeneity, but we used subgroup analysis to explain them; $ a study with wide confidence interval around the estimate of the effect, or included sample less than 400, it would cause imprecision; &‘‘Other’’ included publication bias and upgraded quality of evidence (large effect, plausible residual confounding and dose-response gradient).

Page 39 of 41

Ac

ce

pt

ed

M

an

us

cr

i

Figure 1

Page 40 of 41

Ac

ce

pt

ed

M

an

us

cr

i

Figure 2

Page 41 of 41