Orthopaedics & Traumatology: Surgery & Research 105 (2019) 307–316

Contents lists available at ScienceDirect

Orthopaedics & Traumatology: Surgery & Research

journal homepage: www.elsevier.com

Original article

Deltoid-split approach versus deltopectoral approach for proximal

humerus fractures: A systematic review and meta-analysis
Linzhen Xie a,1 , Yingying Zhang b,1 , Chunhui Chen a , Wenhao Zheng a , Hua Chen a,∗∗,2 ,
Leyi Cai a,∗,2
a
Department of Orthopaedics Surgery, The Second Affiliated Hospital and Yuying Children’s Hospital of Wenzhou Medical University, No. 109, XueYuan
West road, 325000 Wenzhou, Luheng District, Zhejiang Province, People’s Republic of China
b
Department of Radiology, The Second Affiliated Hospital and Yuying Children’s Hospital of Wenzhou Medical University, No. 109, XueYuan West road,
325000 Wenzhou, Luheng District, Zhejiang Province, People’s Republic of China

a r t i c l e i n f o a b s t r a c t

Article history: Purpose: To investigate the evidence of deltoid-split approach (DS) versus deltopectoral approach (DP)
Received 10 June 2018 in treatment of proximal humerus fractures from current RCT and prospective literatures.
Accepted 19 December 2018 Methods: The electronic literature database of Pubmed, Embase, and Cochrane library was searched
at December 2017. The data complications (including implant failure, humeral head necrosis, infection,
Keywords: radiological adverse events, nonunion rate, subacromial impingement, and damage of the axillary nerve),
Proximal humerus fractures functional outcomes (including Constant, NEER, DASH, ADL, VAS score), operation time, hospital stay and
Deltoid-split approach
intraoperative blood loss were extracted and analyzed by STATA 11.0 software.
Deltopectoral approach
Humeral head necrosis
Results: Three RCTs and three prospective comparative studies were included in this meta-analysis. The
Axillary nerve. meta-analysis showed that the DS group had a significantly low humeral head necrosis rate and short
operation time. No significant difference was found in total complication rate, functional outcome, and
other Perioperative parameters between DS and DP groups.
Conclusion: The prospective evidence suggested that DS approach for proximal humerus fractures had
less humeral head necrosis and short operation time than DP approach. Both DS and DP approach had
similar results in functional outcomes, total complication, VAS, and hospital stay.
© 2019 Elsevier Masson SAS. All rights reserved.

1. Abbreviations PCs prospective comparative study

RR relative risk
CI confidence interval
DS deltoid-split approach SMD Standardized Mean Difference
DP deltopectoral approach PRISMA Preferred Reporting Items for Systematic Reviews and
DASH Disabilities of the Arm, Shoulder and Hand score Meta-Analyses
ADL activities of daily living SD standard deviation
VAS Visual analog scale
NA not available
ORIF open reduction internal fixation 2. Introduction
MIPPO minimal invasive percutaneous plate osteosynthesis
RCT randomized controlled trials Fractures of the proximal humerus are usually attributable to
osteoporosis [1] and are mostly caused by low-energy trauma [2].
A 3 fold increase of proximal humeral fractures is expected by
∗ Corresponding author. 2030 [3]. The optimal treatment for proximal humeral fractures is
∗∗ Co-corresponding author. controversial [4]. It includes conservative treatment, open reduc-
E-mail addresses: chenhua fey@163.com (H. Chen), caileyi@wmu.edu.cn (L. Cai). tion internal fixation (ORIF), minimal invasive percutaneous plate
1
These authors contributed equally to this work and should be considered as
co-first authors.
osteosynthesis (MIPPO), intramedullary nailing and arthroplasty
2
These authors contributed equally to this work and should be considered as reported in literature [5,6]. But a basic conclusion supported by
co-corresponding author. comparative study [7,8], review [9,10] and even meta-analysis [11]

https://doi.org/10.1016/j.otsr.2018.12.004
1877-0568/© 2019 Elsevier Masson SAS. All rights reserved.
308 L. Xie et al. / Orthopaedics & Traumatology: Surgery & Research 105 (2019) 307–316

Fig. 1. Schematic diagram for deltopectoral approach (DS) and deltoid-split approaches (DP).

has been drawn that surgical treatment can better of function and 3.2. Eligibility criteria
health-related quality, compared with the nonsurgical treatment.
Conventional surgical methods of ORIF use the deltopectoral The study was included in this meta-analysis if it was:
approach (DP) as Fig. 1A. However, some authors have argued that
this approach involves extensive soft tissue dissection and muscle • prospective randomized controlled trial (RCT) or nonrandomized
retraction to gain adequate exposure to the lateral aspect of the prospective comparative study (PCs);
humerus [12]. As an alternative, a less invasive, deltoid-split (DS) • it compared the clinical outcomes deltoid-splitting approach ver-
approach (Fig. 1B) has been described with the goal of minimizing sus deltopectoral approach for the proximal humeral fracture;
local soft tissue trauma [13]. However, the deltoid-split approach • it was with a follow-up term of at least 12 months.
using the MIPPO technique was recently reported to associated
with a risk of damage to blood supply of the humerus head [14]
Exclusion criteria were as follows:
and axillary nerve [15].
Which surgical approach has more advantages? Some RCTs
[16–18], prospective comparative study (PCs) [14,19] and ret- • respective studies, case series, case report, and review articles;
rospective study [20,21] were developed trying to answer this • follow-up of less than 12 months;
question, but the conclusions were not completely consistent. • duplicated publications from the same hospital or research cen-
Based on the current evidence, we performed this meta-analysis ter.
to compare the DS approach with the DP approach for the manage-
ment of proximal humerus fracture and expected to draw a certain 3.3. Selection of literature
and meaningful conclusion for this question.
We used the PRISMA flow diagram to select the included stud-
ies; the results of literature search were imported into the software
3. Methods Endnote X7. Two authors (the third and fourth author) indepen-
dently assessed the potentially eligible studies. Firstly, the titles and
This systematic review and meta-analysis was conducted fol- abstracts were screened to exclude the duplicated and apparently
lowing the Preferred Reporting Items for Systematic Reviews and irrelevant ones or those that do not meet our inclusion criteria. After
Meta-Analyses (PRISMA) Statement [22]. No primary personal data then, the remaining potential studies were full-text downloaded
will be collected; therefore, no additional ethical approval needs to and reviewed. Any disagreement between two above authors was
be obtained. sent and discussed with the third independent author (the sixth
author).

3.1. Search strategy 3.4. Data extraction

Two authors (The first and second authors) independently Two reviewers (the first and fifth author) independently
searched the electronic literature database of Pubmed, Embase, extracted data, and another reviewer (the six author) checked
and Cochrane library, without language limitation at December the consistency between them. A standard form was used; the
2017. The key words using a combination of different terms and extracted items included the following:
synonyms were used as follows: “proximal humeral fracture”,
“deltoid-splitting”, “deltopectoral”, and “approach”. In addition, the • the general study information, for example, the authors, pub-
reference lists of previously published randomized trials, review lishing date, country, study design, sample size, age, gender,
articles, and meta-analyses were manually searched for additional follow-up term;
eligible studies. Related articles and reference lists were searched • perioperative parameters, including operative time, estimated
to avoid original miss. blood loss and length of hospital stay;
 L. Xie et al. / Orthopaedics & Traumatology: Surgery & Research 105 (2019) 307–316 309

• clinical outcomes, including visual analog scale (VAS) of shoulder 3.6. Statistical analysis
pain and wound pain (0 = no pain at all to 10 = intolerable pain);
Constant-Murley score [23] or Disabilities of the Arm, Shoulder The data was collected and input into the STATA software
and Hand (DASH) [24] for range of shoulder motion; Activities of (version 11.0; StataCorp, College Station, TX) for meta-analysis.
daily living (ADL) were recorded according to the ADL score in A random-effects model was applied when heterogeneity was
line with Lawton and Brody [25]; detected or the statistical heterogeneity was high (P < 0.05 or
• complications, including implant failure (screw loosening, screw I2 > 50%). Otherwise, a fixed-effects model was used (P ≥ 0.05 or
cutout and plates breakage), humeral head necrosis, infec- I2 ≤ 50%). For heterogeneity data, sensitivity analysis was involved
tion, inadequate reduction (varus malalignment and secondary to remove one study and evaluate whether the other results would
displacement of a greater tuberosity fragment), nonunion, sub- be markedly affected. Relative risk (RR) was calculated for dichoto-
acromial impingement and axillary nerve damage. mous outcomes such as complications, standard mean difference
(SMD) was calculated for continuous outcomes such as operative
time, estimated blood loss, length of hospital stay, and function
For continuous outcomes, we extracted the mean and SD (stan- outcomes.
dard deviation) and participant number will be extracted. For
dichotomous outcomes, we extracted the total numbers and the 4. Results
numbers of events of both groups. The data in other forms was
recalculated when possible to enable pooled analysis. 4.1. Included studies

A total of 2635 potential records were identified through Med-

line (n = 1176), Embase (n = 1435), and Cochrane library (n = 24).
3.5. Quality assessment of included studies After removal of duplicates, 1546 articles were screened for rel-
evance on the basis of the title and abstract. Of the 17 articles that
Because both RCTs and prospective comparative studies (PCs) were possibly eligible for inclusion, 11 were excluded for reasons of
were included, the risk of bias in included RCT studies was assessed “the papers were review or retrospective studies” and some other
using the Risk of Bias Tool recommended by the Cochrane Collabo- reasons (details were showed in Fig. 2). The remaining 6 studies
ration [26]. PCs were evaluated using the Newcastle-Ottawa Scale (3 RCT and 3 nonrandom prospective comparative studies) were
[27]. included in this meta-analysis.

Fig. 2. Flowchart of study selection.

310 L. Xie et al. / Orthopaedics & Traumatology: Surgery & Research 105 (2019) 307–316

Table 1
The characteristics of the included studies.

Study Year Location Study design Patients enrolled Age (years) Gender(M/F) Follow-up term(months) Operation time

Buecking et al. 2014 Germany RCT DS: 60 DS: 69 ± 1.53 DS: 12/48 12 DS: 62 ± 2.56
DP: 60 DP: 67 ± 2.04 DP: 16/44 DP: 67 ± 3.32
Zhao et al. 2017 China RCT DS: 17 DS: 64.3 ± 6.7 DS: 9/8 24 DS: 53.6 ± 7.3
DP: 19 DP: 63.6 ± 5.0 DP: 12/7 DP: 61.4 ± 7.0
Martetschl et al. 2012 Germany RCT DS: 37 DS: 59 ± 13.5 DS: 13/24 DS: 20 DS: 107
DP: 33 DP: 56 ± 9.95 DP: 21/12 DP: 48 DP: 106
Hepp et al. 2008 Germany PCT DS: 39 DS: 64 ± 17.9 DS: 12/27 12 DS: 66.5 ± 17.8
DP: 44 DP: 65.5 ± 15.1 DP: 7/37 DP: 85.9 ± 28.1
Bandalo et al. 2014 Croatia PCT DS: 25 DS: >65 NA 14.8 NA
DP: 42 DP: >65
Fischer et al. 2016 Germany PCT DS: 20 DS: 57.6 ± 13.5 DS: 10/20 DS: 22.8 ± 17.0 NA
DP: 30 DP: 60.6 ± 14.5 DP: 6/14 DP: 20.7 ± 15.2

DS: deltoid-splitting group; DP: deltopectoral group; M/F: male/female; NA: Not available.

Table 2
Risk of bias assessment of the RCTs.

Study Random Allocation Blinding of Blinding of Incomplete Selective Other bias

sequence concealment participants outcome outcome data reporting
generation and personnel assessment

Buecking et al. Low risk Low risk Low risk Low risk Low risk Low risk Low risk
Zhao et al. Low risk Low risk Low risk Low risk Low risk Low risk Low risk
Martetschl et al. High risk Unclear risk Low risk Unclear risk Low risk Low risk Low risk

Table 3
Risk of bias assessment of the PCs.

Study Selection Comparability Outcome Total score

Exposed cohort Nonexposed Ascertainment Outcome of Assessment of Length of Adequacy of

cohort of exposure interest outcome follow-up follow-up

Hepp et al. * * * * ** * * * 9
Bandalovi et al. * * * * – * * * 7
Fischer et al. * * * * ** * * * 9

Risk of bias was assessed with use of the Newcastle-Ottawa Scale. “*” means a score of 1; “**” means a score of 2; the total score of this scale is 9. A higher overall score
corresponds to a lower risk of bias; a total score of 5 or less indicates a high risk of bias.

4.2. Characteristics and qualifications of included studies between both DS and DP groups, with RR = 0.98 (95% CI: 0.50, 1.94),
with no heterogeneity (I2 = 0%, P = 0.45) (Fig. 3). Three studies
The characteristics of all five included studies were summarized [16,18,19] reported they did revision with a prosthesis or removal
and shown in Table 1. All the 6 included studies [14,16–19,28] were the implant. Bandaloviü, et al. [14] found intraarticular screw
RCT and prospective comparative studies without random. They placement was seen in 1 patient in DP group during follow-up, but
were from three different countries (1 from China, 3 from Germany they did not mention process of dealing with it.
and 1 from Croatia) and all of them were published between 2008
and 2017. Total of 198 participants in DS group and 228 in DP group 4.3.2. Humeral head necrosis
were included in this meta-analysis. Risk of bias assessment of RCTs Four studies [14,18,19,28] reported of humeral head necrosis
was presented in Table 2. When using the Newcastle-Ottawa Scale during the fellow-up. The meta-analysis showed that there was sig-
to assess the risk of bias of the PCs, the total scores were all higher nificant difference between both DS and DP groups, with RR = 0.28
than 5 indicating a low risk of bias (Table 3). (95% CI: 0.08, 0.94, P = 0.04), with no heterogeneity (I2 = 0%, P = 0.87)
(Fig. 3). Two of 198 (1%) patients in the DS group and 11 of 228 (4.8%)
4.3. Complications patients in the DP group developed the head necrosis.
The method for dealing with patients with a head necrosis
Complications including implant fail (screw loosening, screw differed among the studies. Martetschlager, et al. [18] used the
cutout and plates breakage), humeral head necrosis, infec- additional criteria of most relevant predictors of ischemia accord-
tion, inadequate reduction (varus malalignment and secondary ing to Cruess [29] and Hepp, et al. [19].These predictors were mostly
displacement of a greater tuberosity fragment), nonunion and sub- based on the determining factors which had been reported by Her-
acromial impingement were reported. The meta-analysis showed tel et al. [30]. Unfortunately, most of the studies did not mention
that there was no significant difference between both DS and DP any further treatment for this complication, just Fischer, et al. [28]
groups, with RR = 0.74 (95% CI: 0.50, 1.08), while, the heterogeneity report One woman in the DP group showed radiologic signs of
among studies was very low (I2 = 0%, P = 0.95) (Fig. 3). avascular head necrosis without suffering from any functional lim-
itation.
4.3.1. Implant failure
There were four papers [14,16,18,19] reporting the com- 4.3.3. Infection
plications. It included screw loosening, screw cutout and plates The rate of infection is very low in the two groups: 1 of 198 (0.5%)
breakage. Another two studies [17,28] had no similar complication. in DS group, 4 of 228 (1.8%) in DP group. Just three studies [16–18]
The meta-analysis showed that there was no significant difference reported this adverse event, other three studies [14,19,28] found
L. Xie et al. / Orthopaedics & Traumatology: Surgery & Research 105 (2019) 307–316 311

Fig. 3. Forest plot showing the meta-analysis of the complications.

312 L. Xie et al. / Orthopaedics & Traumatology: Surgery & Research 105 (2019) 307–316

Table 4
Summary of the outcomes of function scores.

Variable Number of study First author DS cases Mean SD DP cases Mean SD P

Constant score (6 weeks) 1 Buecking 48 46 2.04 42 47 7.14 0.62

Constant score (3 months) 1 Hepp 39 57.9 16.6 44 49.6 17.9 0.02*
Constant score (6 months) 2 Buecking 48 68 2.81 42 64 3.32 0.41
Hepp 39 69.4 13.8 44 71.4 20.9 0.4
Constant score (12 months) 3 Buecking 48 81 3.32 42 73 4.34 0.13
Zhao 17 88.8 1 19 86.9 2.1 0.001*
Hepp 39 73.1 12.1 44 81 18.7 0.004*
Constant score (24 months) 2 Martetschlager 37 75.1 14.8 33 72.8 15.7 0.7
Fischer 20 81.6 16.1 30 76.3 18.6 0.37
NEER score (12 months) 1 Zhao 17 87.4 1.2 19 85.7 2.6 0.019*
DASH (12 months) 1 Hepp 39 14.6 NA 44 14.6 NA 0.43
DASH (24 months) 2 Martetschlager 37 14.1 No 33 13.8 No 0.94
Fischer 20 21.1 8.6 30 20.1 20.3 0.68
ADL (6 weeks) 1 Buecking 48 13 0.77 42 12 1.02 0.49
ADL (3 months) 1 Hepp 39 14.5 3.51 44 12.5 3.2 0.01*
ADL (6 months) 2 Buecking 48 19 0.77 42 15 2.55 0.82
Hepp 39 16.9 2.5 44 16.2 3.5 0.6
ADL (12 Months) 2 Buecking 48 18 0.77 42 17 1.28 0.31
Hepp 39 17.7 2.7 44 17.5 2.9 0.9
ADL (24 Months) 1 Martetschlager 37 19 NA 33 19 NA 0.6
VAS (6 weeks) 1 Buecking 48 3.6 0.28 42 3.9 0.31 0.56
VAS (6 months) 2 Buecking 48 2.7 0.33 42 3.1 0.38 0.45
Zhao 17 3.1 1.3 19 3.2 1.3 0.73
VAS (12 Months) 1 Buecking 48 1.8 0.31 42 2.5 0.41 0.15

ADL: activities of daily living; VAS: visual analog scale; DASH: Disabilities of the Arm, Shoulder and Hand score; NA: not available; The asterisk and bold present mean there
was a significant difference between DP and DS group.

there was no infection during the fellow-up. The meta-analysis No damage of the axillary nerve was detected on clinical neuro-
showed that there was no significant difference between both DS logical examination in all the studies. One study [14] reported the
and DP groups, with RR = 0.41 (95% CI: 0.08, 2.01), with no hetero- deltoid muscles were weak initially, however returned to normal
geneity (I2 = 0%, P = 0.94) (Fig. 3). after rehabilitation, but the author did not mention which group
Different processing method dependent on different infection. happened.
Buecking et al. [16] made a second surgery to remove the plate, thus
the deep infection was under control. The infection in Martetschl 4.4. Functional outcomes
et al. [18] was treated successfully by revision surgery and antibiotic
therapy. While Zhao et al. [17] did not mention elective treatment As detailed in Table 4, at the postoperative 6 weeks [16], 6
operations for the infections. months [16,19] and 2 years [18,28], the functional outcomes of
Constant (Fig. 4A) and DASH (Fig. 4B) score in the DS group did not
better than the DP groups. However, one study [19] reported the
4.3.4. Radiological adverse events
Constant score in at the postoperative 3 months was significant dif-
We determine the radiology adverse events as inadequate
ferent (Fig. 4A). At the postoperative 12 months, most studies found
reduction, varus malalignment and secondary displacement. Four
that the DS group did better than the DP group significantly, regard-
studies [14,18,19,28] reported the radiology complication. There
less of which function scoring system was used (Constant [17,19]
was no significant difference between both DS and DP groups by the
and NEER [17] score) (Fig. 4A). However, when the meta-analysis
meta-analysis RR = 0.85 (95% CI: 0.42, 1.72), with no heterogeneity
was made, because of detecting the heterogeneity with P = 0.000,
(I2 = 0%, P = 0.61) (Fig. 3).
I2 = 96.5%, the random-effect model was performed and there was
All the 4 studies referred a varus malalignment during the
no significant difference between the two groups (SMD = 0.9, 95%
fellow-up, but none of them mentioned the way to deal with it.
CI −0.78 to 2.59; P = 0.293) (Fig. 5).
Hepp et al. [19] and Bandalo et al. [14] reported a secondary dis-
For the variable of ADL [16,18,19] (Fig. 4C) and VAS [16,17]
placement of a greater tuberosity fragment, a revision surgery in
(Fig. 4D), most studies found there was no significant difference
DS group was performed in Hepp et al.’s study.
between both DS and DP groups regardless of the postoperative
time (6 weeks, 3, 6, 12 and 24 months). Just one study [19] found
4.3.5. Other complications there was significant difference of ADL with postoperative 3months
Two studies [18,19] reported the data of nonunion rate, the between the two groups.
meta-analysis showed that there was no significant difference
between both DS and DP groups, with RR = 0.68 (95% CI: 0.12, 4.03), 4.5. Perioperative parameters
with no heterogeneity (I2 = 0%, P = 0.94) (Fig. 3). One research [18]
did not perform revision surgery due to without pain or clini- Operation time, hospital stay and intraoperative blood loss were
cal restrictions. The nonunion in another one [19] was treated by defined as the perioperative parameter. Four studies [16–19] could
reosteosynthesis combined with cancellous grafting after 6 and 8 provide data for the operation time, but one of them [18] could
months. provide standard format for calculating the SMD, but with standard
Subacromial impingement was reported in two papers [19,28] deviation was not available. Then, the meta-analysis was made,
as complication. There was no significant difference between both because of detecting the heterogeneity with P = 0.018, I2 = 75%, the
DS and DP groups by the meta-analysis RR = 1.13 (95% CI: 0.33, 3.89), random-effect model was performed and a significant difference
with no heterogeneity (I2 = 0%, P = 0.99) (Fig. 3), while the authors was found between the two groups (SMD = −1.25, 95% CI −1.53 to
did not mention how to solve this problem. −0.97; P = 0.018) (Fig. 6). To eliminate the heterogeneity and obtain
 L. Xie et al. / Orthopaedics & Traumatology: Surgery & Research 105 (2019) 307–316 313

Fig. 4. Histogram of the comparison of the two groups of functions. A. Constant score. B. Disability Arm, Shoulder Hand (DASH). C. Activities of daily living (ADL). D. Visual
Analogue Scale (VAS).

Fig. 5. Forest plot showing the meta-analysis of the Constant Score at the postoperative 12 months with a random-effect model.

a more objective result, we performed a sensitivity analysis [16]. (129.2 ± 17.8 vs. 145.3 ± 23.0 ml P = 0.026), the DS group show less
After excluding the study of Buecking et al. [16], the heterogeneity blood loss. Buecking et al. [16] reported the hospital stay(day), the
disappeared (P = 0.514, I2 = 0%), then we found operation time in the two groups did the similar hospital stay(P = 0.86).
DS group was significantly lower than that in the DP group (SMD
−0.89, 95% CI −1.27 to −0.52; P = 0.001) using the fixed effect model 5. Discussion
(Fig. 7).
About the intraoperative blood loss was reported by Zhao et al. Proximal humerus fractures are common [2], and there was
[17]. They found there was a significant difference between them no meta-analysis in the previous study to compare the two
314 L. Xie et al. / Orthopaedics & Traumatology: Surgery & Research 105 (2019) 307–316

Fig. 6. Forest plot showing the meta-analysis of the operation time.

Fig. 7. Forest plot showing the meta-analysis of the operation time excluding the study of Buecking et al.

approaches. According to our pooled data, the meta-analysis results [33]. What’s more, plating through a minimally invasive antero-
showed there was a better result for DS group about the humerus lateral acromial approach allows direct access to the appropriate
head necrosis and function outcome in 12 months. While, one RCT plating zone, a bare spot between the humeral head – penetrat-
[17] showed the intraoperative blood loss was more less in the ing vessels from the anterior and posterior circumflex system. The
DS group. What’s more, the result did not show any damage of meta-analysis showed the low rate in the DS group (2/198, 1%), it
the axillary nerve in all the studies. These findings indicate that was comparable with the Neviaser et al.’s study [34]. They used
through a DS approach, we not only can reduce the problems of the the anterolateral approach, the anterior deltoid raphe was split,
head necrosis and operation blood loss but also achieve the bet- while no patient suffered complete osteonecrosis (0%) and only one
ter functional outcomes with the similar operation time and other patient suffered partial necrosis (2.8%) of the humeral head.
complications compared with the DP approach. The risk of damaging the axillary nerve in less invasive surgery is
The humeral head articular surface has a tenuous blood sup- a frequently discussed and feared problem [35]. To protect the axil-
ply [31,32]. The arcuate branch of the anterior humeral circumflex lary nerve, Buecking et al. [16] indicated the nerve with the index
artery provides a significant proportion of the flow to the humeral finger in the subdeltoid bursa and its course was marked on the skin.
head articular surface in a retrograde fashion. This is akin to the Additionally, some used a five-hole plate that was inserted with
lateral epiphyseal branch of the medial femoral circumflex artery its tip contacting the bone, and Ruchholtz et al. [36] fixed screws
that supplies the femoral head. The traditional surgical approach in the three distal holes, far away from the axillary nerve. In our
(DP group) will separate more soft tissue, resulting in an increased meta-analysis, just one reported the deltoid muscles were weak
surface blood loss, thereby increasing the risk of humeral head initially, however returned to normal after rehabilitation. With the
necrosis. The DS approach provides direct access to the lateral progress of surgical techniques and internal fixation development,
humeral bald spot obviating the need for circumferential dissection we believe this complication will be less, and will not be a limitation
and potential vascular disruption either anteriorly or posteriorly of the DS approach for Proximal humerus fractures.
 L. Xie et al. / Orthopaedics & Traumatology: Surgery & Research 105 (2019) 307–316 315

Yves et al. [37] reported a prospective analysis of minimally References

invasive plating in proximal humerus fractures. They found this
minimally invasive procedure resulted in a shorter operation time [1] Viprey M, Caillet P, Canat G, Jaglal S, Haesebaert J, Chapurlat R, et al. Low
osteoporosis treatment initiation rate in women after distal forearm or prox-
and may have reduced the avascular necrosis rate. In our meta- imal humerus fracture: a Healthcare Database Nested Cohort Study. PloS one
analysis, the DS approach showed a better result in the operation 2015;10 [e0143842].
time compared with the DP approach. While, less avascular necro- [2] Court-Brown CM, Garg A, McQueen MM. The epidemiology of proximal
humeral fractures. Acta Orthop Scand 2001;72:365–71.
sis in the DS group. In contrast to the commonly used delta-pectoral
[3] Kannus P, Palvanen M, Niemi S, Parkkari J, Jarvinen M, Vuori I. Osteoporotic
standard approach, the trans-deltoid MIPO approach provides a fractures of the proximal humerus in elderly Finnish persons: sharp increase
direct visualisation of the plating zone under minimal soft tissue in 1970-1998 and alarming projections for the new millennium. Acta Orthop
dissection and involves simpler exposure of the tuberosity. Further- Scand 2000;71:465–70.
[4] Neuhaus V, Bot AG, Swellengrebel CH, Jain NB, Warner JJ, Ring DC. Treatment
more, because no retractors are placed dorsal to the humeral shaft, choice affects inpatient adverse events and mortality in older aged inpatients
no distracting malalignment results, and restoring the anatomic with an isolated fracture of the proximal humerus. J Shoulder Elbow Surg
head-shaft axis using the DS approach is easier. However, there 2014;23:800–6.
[5] Horak J, Nilsson BE. Epidemiology of fracture of the upper end of the humerus.
was just two studies for the meta-analysis after the sensitivity anal- Clin Orthop Relat Res 1975;112:250–3.
ysis. The evidence that DS approach is better for operation time still [6] Kralinger F, Schwaiger R, Wambacher M, Farrell E, Menth-Chiari W, Lajtai G,
needs further research. et al. Outcome after primary hemiarthroplasty for fracture of the head of the
humerus. A retrospective multicentre study of 167 patients. J Bone Joint Surg
Less soft tissue manipulation and reduced exposure, for exam- Br 2004;86:217–9.
ple, by short duration of operation and direct access to the [7] Ge W, Sun Q, Li G, Lu G, Cai M, Li S. Efficacy comparison of intramedullary nails,
fracture in the deltoid-splitting group, might be a reason for higher locking plates and conservative treatment for displaced proximal humeral frac-
tures in the elderly. Clin Interv Aging 2017;12:2047–54.
early functional values [19]. But our meta-analysis showed it was [8] Brouwer ME, Reininga IHF, El Moumni M, Wendt KW. Outcomes of opera-
not significant different in the two groups with random-effect tive and nonoperative treatment of 3- and 4-part proximal humeral fractures
model, even two RCTs [16,17] reported there was significant dif- in elderly: a 10-year retrospective cohort study. Eur J Trauma Emerg Surg
2017.
ference between. Whatever, for a long time fellow-up, it was
[9] Maier D, Jager M, Strohm PC, Sudkamp NP. Treatment of proximal humeral
no significant difference in the study. All studies included did fractures – A review of current concepts enlightened by basic principles. Acta
not mention any difference of osteosynthesis between the two Chir Orthop Traumatol Cech 2012;79:307–16.
approached. [10] Lowry V, Bureau NJ, Desmeules F, Roy JS, Rouleau DM. Acute proximal humeral
fractures in adults. J Hand Ther 2017;30:158–66.
This study had many strengths; all of the included studies in this [11] Mao F, Zhang DH, Peng XC, Liao Y. Comparison of surgical versus non-surgical
meta-analysis were RCT and prospective studies, which therefore treatment of displaced 3- and 4-part fractures of the proximal humerus: a meta-
overcomes the shortcomings of recall or selection bias in retrospec- analysis. J Invest Surg 2015;28:215–24.
[12] Gardner MJ, Griffith MH, Dines JS, Briggs SM, Weiland AJ, Lorich DG. The
tive studies. The methodological index for nonrandomized studies extended anterolateral acromial approach allows minimally invasive access
(Newcastle – Ottawa Scale) was used to assess the quality of the to the proximal humerus. Clin Orthop Relat Res 2005;434:123–9.
included studies, which had minimized scores of 7, the score range [13] Roderer G, Abouelsoud M, Gebhard F, Bockers TM, Kinzl L. Minimally invasive
application of the non-contact-bridging (NCB) plate to the proximal humerus:
of 7 to 9. However, there were still some limitations of present an anatomical study. J Orthop Trauma 2007;21:621–7.
study, not all of them was randomized control trials, the sample [14] Bandalovic A, Cukelj F, Knezevic J, Ostojic M, Pavic A, Parac Z, et al. The results
size was not very large, and the duration of follow-up was less than of internal fixation of proximal humeral osteoporotic fractures with PHILOS
locking plate. Psychiatr Danub 2014;26:376–81.
5 years. Therefore, we suggested further long-term, larger sample [15] Gavaskar AS, Chowdary N, Abraham S. Complex proximal humerus fractures
size, and randomized control trials to be conducted. treated with locked plating utilizing an extended deltoid split approach with a
shoulder strap incision. J Orthop Trauma 2013;27:73–6.
[16] Buecking B, Mohr J, Bockmann B, Zettl R, Ruchholtz S. Deltoid-split or deltopec-
6. Conclusion toral approaches for the treatment of displaced proximal humeral fractures?
Clinical Orthop Relat Res 2014;472:1576–85.
[17] Zhao LYP, Zhu L, Chen AM. Minimal invasive percutaneous plate osteosynthe-
The prospective evidence suggested that DS approach for prox- sis (MIPPO) through deltoidpectoralis approach for the treatment of elderly
imal humerus fractures had less humeral head necrosis and short proximal humeral fractures. BMC Musculoskeletal Disord 2017;18:187.
[18] Martetschlager F, Siebenlist S, Weier M, Sandmann G, Ahrens P, Braun K, et al.
operation time than DP approach. Both DS and DP approach had Plating of proximal humeral fractures. Orthopedics 2012;35 [e1606–e1612].
similar results in functional outcomes, total complication, VAS, and [19] Hepp P, Theopold J, Voigt C, Engel T, Josten C, Lill H. The surgical approach
hospital stay. for locking plate osteosynthesis of displaced proximal humeral fractures influ-
ences the functional outcome. J Shoulder Elbow Surg 2008;17:21–8.
[20] Wu CHMC, Yeh JJ, Yen CY, Yu SW, Tu YK. Locked plating for proximal
humeral fractures: differences between the deltopectoral and deltoid-splitting
Disclosure of interest approaches. J Trauma 2011;71:1367–70.
[21] Lin TXB, Ma X, Fu D, Yang S. Minimally invasive plate osteosynthesis with a
locking compression plate is superior to open reduction and internal fixation
The authors declare that they have no competing interest.
in the management of the proximal humerus fractures. BMC Musculoskelet
Disord 2014;15:206.
[22] Moher D, Liberati A, Tetzlaff J, Altman DG, Group P. Preferred reporting items
Funding for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med
2009;6 [e1000097].
[23] Constant CR, Murley AH. A clinical method of functional assessment of the
This work was supported by the National Natural Science Foun- shoulder. Clin Orthop Relat Res 1987;214:160–4.
dation of China (81701928). The funders had no role in the study [24] Russo R, D’Auria D, Ciccarelli M, Della Rotonda G, D’Elia G, Siciliano B. Triangu-
design, data collection and analysis, decision to publish, or prepa- lar block bridge method for surgical treatment of complex proximal humeral
fractures: theoretical concept, surgical technique and clinical results. Injury
ration of the manuscript. 2017;48:S12–9.
[25] Lawton MP, Brody EM. Assessment of older people: self-maintaining and
instrumental activities of daily living. Gerontologist 1969;9:179–86.
Contribution [26] Armijo-Olivo S, Stiles CR, Hagen NA, Biondo PD, Cummings GG. Assessment
of study quality for systematic reviews: a comparison of the Cochrane Col-
laboration Risk of Bias Tool and the Effective Public Health Practice Project
H. C. and LY. C. designed the study and LZ. X. wrote this Quality Assessment Tool: methodological research. J Eval Clin Pract 2012;18:
manuscript. LZ. X. and YY. Z. searched database and reviewed stud- 12–8.
ies. H. C., LZ. X. and LY. C. collected and analyzed data. X. L. All of [27] Tao E, Wan L, Wang W, Luo Y, Zeng J, Wu X. The prognosis of infective endocardi-
tis treated with biological valves versus mechanical valves: a meta-analysis.
the authors have read and approved the final manuscript. PloS one 2017;12 [e0174519].
316 L. Xie et al. / Orthopaedics & Traumatology: Surgery & Research 105 (2019) 307–316

[28] Fischer C, Frank M, Kunz P, Tanner M, Weber MA, Moghaddam A, et al. [33] Gardner MJ, Voos JE, Wanich T, Helfet DL, Lorich DG. Vascular implications of
Dynamic contrast-enhanced ultrasound (CEUS) after open and minimally minimally invasive plating of proximal humerus fractures. J Orthop Trauma
invasive locked plating of proximal humerus fractures. Injury 2016;47: 2006;20:602–7.
1725–31. [34] Neviaser AS, Hettrich CM, Dines JS, Lorich DG. Rate of avascular necrosis fol-
[29] Cruess RL. Osteonecrosis of bone. Current concepts as to etiology and patho- lowing proximal humerus fractures treated with a lateral locking plate and
genesis. Clin Orthop Relat Res 1986;208:30–9. endosteal implant. Arch Orthop Trauma Surg 2011;131:1617–22.
[30] Hertel R, Hempfing A, Stiehler M, Leunig M. Predictors of humeral head [35] Handoll HH, Ollivere BJ, Rollins KE. Interventions for treating proximal humeral
ischemia after intracapsular fracture of the proximal humerus. J Shoulder fractures in adults. Cochrane Database Syst Rev 2012;12 [CD000434].
Elbow Surg 2004;13:427–33. [36] Ruchholtz S, Hauk C, Lewan U, Franz D, Kuhne C, Zettl R. Minimally invasive
[31] Cadet ER, Ahmad CS. Hemiarthroplasty for three- and four-part proximal polyaxial locking plate fixation of proximal humeral fractures: a prospective
humerus fractures. J Am Acad Orthop Surg 2012;20:17–27. study. J Trauma 2011;71:1737–44.
[32] Schlegel TF, Hawkins RJ. Displaced proximal humeral fractures: evaluation and [37] Acklin YP, Stoffel K, Sommer C. A prospective analysis of the functional and radi-
treatment. J Am Acad Orthop Surg 1994;2:54–78. ological outcomes of minimally invasive plating in proximal humerus fractures.
Injury 2013;44:456–60.