Arch Orthop Trauma Surg

DOI 10.1007/s00402-013-1798-2

Trauma Surgery

Operative versus non‑operative treatment for two‑part surgical

neck fractures of the proximal humerus
O. Hauschild · G. Konrad · L. Audige · P. de Boer ·
S. M. Lambert · R. Hertel · N. P. Südkamp

Received: 31 December 2011

© Springer-Verlag Berlin Heidelberg 2013

Abstract range of motion and absolute and relative Constant scores

Introduction Aim of this study was to evaluate outcomes at 3, 6 and 12 months following injury and coronal plane
of operative as compared to conserveative treatment for alignment at 12 months.
two-part humerus fractures at the surgical neck. Results Operative (n = 133) and non-operative (n = 31)
Methods Data from a prospective multi-centre cohort groups were comparable with regard to all parameters
study on four treatment options (conservative treatment and assessed including mean age (62.9 vs. 65.6, P = 0.479),
three implants, i.e. LPHP, PHILOS and PHN) for proximal gender (27 vs. 29 % male, P = 0.826) and fracture distribu-
humerus fractures were evaluated in this post hoc analysis. tion (65 vs. 77 % A3 type, P = 0.207). 26 of the 31 con-
All patients with two-part fractures of the surgical neck servatively treated and 103 of the 133 operatively treated
(AO types A2, n = 54 and A3, n = 110) were identified and patients (84 and 77 %, respectively) were available for final
included for the analysis. All operatively treated patients follow-up. There was a continuous improvement for all
were gathered and compared to those receiving conserva- outcome parameters in both treatment groups (P < 0.001).
tive treatment. Primary outcome parameters were pain, Operative treatment resulted in a more effective reduction
of pain at 3 months (51 vs. 76 % reporting pain at frac-
ture site, P = 0.03) and a reduction of coronal plane mala-
O. Hauschild and G. Konrad state that they have contributed equally lignment. Both range of motion and Constant scores were,
to this manuscript and therefore apply for shared first authorship.
however, comparable in both groups at all follow-up vis-
O. Hauschild (*) · N. P. Südkamp its. Relative and absolute Constant scores were generally
Department of Orthopaedic Surgery and Traumatology, Freiburg excellent at final follow-up (74 vs. 74, P = 0.528 and 89 vs.
University Medical Centre, Hugstetter Str. 55, 79106 Freiburg, 91, P = 0.494, respectively).
Germany
e-mail: oliver.hauschild@uniklinik‑freiburg.de
Conclusions Both non-operative treatment and opera-
URL: www.uniklinik-freiburg.de tive treatment using modern implants (LPHP, PHILOS and
PHN) can be considered safe and effective treatment options
G. Konrad for two-part fractures of the proximal humerus. Operative
Kreiskrankenhaus Erding, Erding, Germany
treatment may result in better range of motion and reduced
L. Audige pain in the early postoperative course of treatment.
AO Clinical Investigation and Documentation, Dübendorf,
Switzerland Keywords Proximal humerus fracture · Two-part
P. de Boer
fracture · Non-operative treatment · Operative treatment
Honorary Consultant Surgeon, York Hospital, York, UK

S. M. Lambert Introduction

Royal National Orthopaedic Hospital, Stanmore, UK

R. Hertel Fractures of the proximal humerus can be regarded as

Lindenhofspital, Bern, Switzerland common injuries with a reported yearly incidence of

13
 Arch Orthop Trauma Surg

~66/10,000 [13]. Associated with osteoporosis, proximal i.e. all patients were treated conservatively in the “conserv-
humerus fractures rarely affect patients under 60 years of ative” centre and all patients were treated operatively using
age and reduced bone quality poses significant challenges the assigned implant in the various “operative” centres.
for treatment of these injuries [18, 20]. Operative therapy All patients with two-part fractures of the surgical neck
is prone to inherent and partially unsolved problems and (i.e. A2 and A3 type fractures according to AO classifica-
has been shown to be associated with markedly high com- tion) included in the underlying cohort study were iden-
plication rates [18]. Conservative treatment has proved a tified and taken for this post hoc analysis. Exclusion cri-
safe treatment option associated with reasonable functional teria included pseudarthrosis, pathological fractures and
outcome for fractures of the proximal humerus of differ- refractures, open fractures or concomitant fractures of the
ent complexity [5, 6, 8, 22, 23]. Owing to the good results ipsilateral elbow or distal radius. In addition, patients with
of conservative treatment, operative treatment for two-part existing disorders having a relevant effect on the healing
fractures of the proximal humerus has been questioned process and function such as multiple sclerosis, paraplegia
by some authors [6, 8]. However, the most relevant study or other relevant neurological disorders, polytraumatized
[6] comparing operative to conservative treatment of this patients with an Injury Severity Score (ISS) exceeding 16
entity had several flaws, namely a small number of opera- or patients with pre-existing plexus injury or nerve palsy
tively treated patients and high rates of incomplete reduc- were excluded.
tion. Moreover, flexible nailing and tension band wiring, Eight and 70 surgeons were involved in the non-opera-
techniques that have more and more been abandoned in the tive and operative groups, respectively. Treating surgeons
presence of modern fixation devices, were used for stabili- were fellowship-trained trauma surgeons and had to have
sation in all cases. performed at least 30 proximal humerus fracture stabilisa-
Fostered by advancements in implant design, namely the tion procedures and 5 with the respective Synthes implant
introduction of locked plates and intramedullary nails with used in the study. All operative procedures were performed
spiral blades (angular stable proximal fixation) promising using standard approaches (i.e. delta-split approach for
enhanced stability even in osteopenic cancellous bone and PHN and anterior deltoideo-pectoral approach for LPHP
in presence of severe comminution operative treatment has and PHILOS) and AO reduction techniques and implants
more recently been advocated by various authors, particu- were used according to manufacturer’s recommendations.
larly for more displaced fractures and in younger patients Postoperatively, the arm was immobilised in a sling and
[1, 9, 11, 19]. The aim of the present analysis was there- passive ROM exercises were started within 2 days after
fore to compare the outcomes of non-operatively treated surgery. Controlled active mobilisation with abduction and
patients with two-part fractures of the proximal humerus to flexion beyond 90° was started 1–3 weeks postoperatively,
those treated operatively using modern implants (intramed- depending on the stability of the osteosynthesis and bone
ullary nails and locking plates). quality.
Conservative treatment included immobilisation of the
shoulder in an arm sling for 1–3 weeks with passive ROM
Materials and methods exercises starting after 1 week. Controlled active mobi-
lisation was allowed after 4 weeks. Initial closed reduc-
Patient inclusion and treatment tion manoeuvres were allowed at discretion of the treating
physician.
Data used in this analysis have been collected prospec-
tively in the context of a large cohort study that was per- Data collection
formed as a series of four separate case series (i.e. treat-
ment arms, with similar protocols and case report forms). During hospitalisation, patient demographics (i.e. gender,
There was one non-operative group, which was exclusively age, dexterity, smoking, concomitant diseases, and medi-
enrolled in one centre [8], and three operative groups [2, cation) and baseline characteristics (i.e. date of accident,
3, 21] using three different standard AO implants (Synthes, accident type, energy level of trauma, concomitant inju-
Solothurn, Switzerland) in 27 other level I trauma centres, ries, fracture classification, date of surgery, operation time,
whereby each centre used only one, i.e. the locking proxi- c-arm counter time, additional implants and sutures, addi-
mal humerus plate (LPHP), the proximal humerus inter- tional medication, type and duration of immobilisation,
nal locking system (PHILOS) or the proximal humerus and beginning of active assisted and unrestricted mobili-
nail (PHN). The study was designed in a way that within sation) were recorded. Fractures were classified according
the study period all patients who met the inclusion crite- to the AO-Müeller classification [14] by the treating sur-
ria were treated with the treatment option assigned to the geon using plain radiographs and intraoperative fracture
respective centre, irrespective of displacement or stability, visualisation.

13
Arch Orthop Trauma Surg

Scheduled follow-up visits were performed 3, 6 and examination, occurrence of local complications within
12 months following injury. At each follow-up visit, the 1 year) were analysed by multivariable binomial regres-
patients were examined and interviewed concerning their sion, and adjusted risk ratios (RR) were used to quantify
pain, shoulder mobility and strength and the Constant treatment effect.
scores [4] of the injured and the contralateral shoulder were We made a post hoc power analysis using repeated
obtained. Isometric muscle strength was assessed using a measures ANOVA and the Constant and Murley score.
Nottingham Mecmesin Myometer (Mecmesin Co, Not- Group sizes of 26 and 103 patients, respectively (i.e. as
tingham, UK). The tests were performed with the shoul- examined at 1-year follow-up), provided >99 % power to
der at 90° abduction, or, if 90° could not be reached, in detect a minimum difference of 10 points in the Constant
maximum active abduction as described by Constant [4]. and Murley score with a known standard deviation of 13;
Patients were asked to maintain this resisted abduction the correlations between contralateral and injured sides,
for 5 s. The mean of three tests was recorded as strength and between follow-ups were set to 0.50 and 0.80, respec-
of the shoulder. Patients with a history of trauma or prior tively. For achieving a power of 90 %, 14 and 61 patients in
surgery involving the contralateral shoulder were not taken both groups would have been required.
into consideration for the contralateral Constant scores at
follow-up visits. At the 12-month follow-up visit we addi-
tionally obtained Neer scores [15] for each patient. Results
True AP and trans-scapular Y view radiographs were
obtained postoperatively (in case of operative treatment) This analysis included 31 patients (one clinic) with two-
and subsequently at each follow-up visit. The treating sur- part fracture in the non-operative and 133 patients (21 clin-
geon evaluated the radiographs primarily for fracture heal- ics) in the operative group. Final 1-year follow-up data
ing and possible occurrence of complications. All radio- were available from 26 to 103 patients (84 and 77 %) in
graphs and collected clinical data were jointly reviewed non-operative and operative groups, respectively. Both
by two of the authors to validate complication records and groups were comparable with regard to all demographic
define their most likely influencing factors, as well as docu- parameters assessed (Table 1); analyses, however, were
ment final deformity (valgus/varus deviation) at the fracture adjusted for age, presence of comorbidities, energy level of
site. injury, dominant arm injured and fracture type to control
for a potential confounding effect.
Data management and statistics There was a continuous and highly significant improve-
ment over time for all outcome parameters (pain, range of
Study monitoring, database management and statistics were motion, absolute and relative Constant scores) assessed in
carried out at a central monitoring organisation. Patients patients from both non-operative and operative treatment
who had undergone conservative treatment were compared groups (P < 0.001).
to those operatively treated regarding baseline demograph-
ics and injury parameters using standard descriptive sta-
tistics and non-parametric univariable tests. Observed dif-
Table 1  Patient demographics
ferences between the groups were considered carefully by
clinical judgement; variables with differences potentially Parameter Operative Non-operative P value
confounding the outcome comparisons were considered for
N 133 31
adjustment in multivariable analyses.
Mean age (SD) (years) 62.9 (17.2) 65.6 (13.3) 0.479a
Treatment groups were compared regarding shoulder
Gender (male/female) 36/97 9/22 0.826b
function and health status at 3, 6, and 12 months. For each
Dexterity (%) 92 87 0.413b
continuous outcome, any repeated measurements of each
Comorbidities (%) 47 65 0.110b
patient were pooled and analysed together in one overall
Energy (high/low) 27/106 2/29 0.113b
multivariable linear regression model. The likelihood ratio
Dominant arm injured (%) 39 29 0.409b
test was used to test the null hypothesis that there would
Injured side (right/left) 47/86 13/18 0.537b
be no difference in shoulder function and health status
between the two groups. The analyses of absolute Con- Worked before accident (%) 28 29 1.000b
stant score and range of motion values were adjusted for AO Classification [N (%)] 0.207b
respective contralateral healthy side values, as appropriate. A2 47 (35) 7 (23)
For each patient the mean of contralateral values recorded A3 86 (65) 24 (77)
a
across follow-up examinations was used as reference value. Wilcoxon rank-sum test
Four dichotomous parameters (pain at each follow-up b
Fisher’s exact test

13
 Arch Orthop Trauma Surg

Table 2  Pain at fracture site (A) Continuous Neer score

b a
Follow-up Operative Non-operative RR (95 % CI) P value 100
Excellent
(months)
Satisfactory
3 56/115 (51 %) 22/29 (76 %) 0.73 (0.55; 0.97) 0.030* 80
Unsatisfactory
6 49/101 (49 %) 14/28 (50 %) 1.0 (0.64; 1.56) 1.000

Neer score
12 27/103 (26 %) 6/26 (23 %) 1.28 (0.59; 2.76) 0.984 60 Failure

Portion of patients expressing pain at fracture site during follow-up

interview
40
* Significant at 0.05 level
a
Wald test
b Non-operative Operative (PHILOS/LPHP/PHN)
 RR adjusted risk ratio for operative over non-operative treatment 20
with 95 % confidence interval 1 year Scale

(B) Caterogized Neer score

Operatively treated patients were less likely to suf-
60
fer from pain at fracture site at the 3-month follow-up

Percentage of patients (%)

visit as compared to conservatively treated patients (51
vs. 76 %, P = 0.03). This difference was, however, not 40
observed at 6- and 12-month follow-up (Table 2). Shoul-
der function was generally excellent at 12-month follow-
up and average relative Constant scores approximating 20

90 % of the uninjured side were observed in both treat-

ment groups (Table 3). Mean Neer scores at 12-month
0
follow-up exceeded 80 in both groups and were slightly failure unsatisfactory satisfactory excellent
better in the non-operative group (88.6 vs. 84.7, P = 0.02; Neer score categories
Fig. 1). According to Neer’s outcome criteria, 85 and Non-operative Operative (PHILOS/LPHP/PHN)

73 % of patients in the non-operative and operative group

were found to have satisfactory or excellent results at final Fig. 1  Absolute and categorised Neer scores at 1-year follow-up
follow-up, respectively (P = 0.28, n.s.). With the exemp- examination. Linear regression analysis revealed significant inferior-
ity of operative as compared to non-operative treatment. The respec-
tion of passive external rotation at 3 (in favour of con-
tive group difference was −6.6 (95 % CI −12.2; −1.0), P = 0.02
servative treatment) and 12 months (in favour of operative using Wald test)
treatment), operative treatment did not result in superior
or inferior ranges of motion at any of the follow-up vis-
its even though there was a tendency towards improved In particular, there was a significant reduction of varus
abduction and flexion at 3 months in the operative treat- malalignment exceeding 15° (OR 0.38, 95 % CI 0.15–0.95,
ment group (Table 4). P = 0.038). Interestingly, there was a significant impair-
Anatomical reduction defined as varus malalignment of ment of absolute Constant scoring in the operative group
no more than 15° was achieved in 93 of 133 operatively when varus malalignment exceeded 15° (92.2 vs. 83.0,
treated patients (70 %). Radiographic analyses at 1-year P = 0.0006). On the contrary, varus malalignment did not
follow-up revealed that operative treatment resulted in a have a significant effect on Constant scoring in the conserv-
marked improvement of coronal plane alignment (Fig. 2). ative group (Fig. 3).

Table 3  Absolute and relative Parameter Follow-up Operative Non-operative Group difference (95 % P valueb
Constant scores (months) CI)a
Absolute and relative (% of
healthy side) Constant scores at Healthy sidec – 82.8 (7.5) 82.1 (5.6)
follow-up visits [mean (SD)] Absolute injured side 3 57.5 (14.1) 51.2 (15.2) 3.0 (−2.2; 8.1) 0.257
a
95 % confidence interval 6 65.8 (14.7) 67.3 (13.2) −3.1 (−8.2; 2.0) 0.237
b
Wald test 12 74.2 (13.0) 74.3 (9.9) −1.7 (−6.9; 3.5) 0.528
c
For each patient, the mean of % of healthy side 3 69.1 (16.6) 62.2 (18.1) 4.0 (−2.2; 10.2) 0.202
contralateral values recorded 6 79.0 (15.9) 81.7 (14.6) −3.1 (−8.2; 2.0) 0.529
across follow-up examinations 12 88.9 (12.8) 90.8 (9.6) −1.7 (−6.9; 3.5) 0.494
was used as reference value

13
Arch Orthop Trauma Surg

Table 4  Range of motion Direction Follow-up Operative Non-operative Group difference P valueb
(months) (95 % CI)a

Flexion
Healthy side (°)c – 165.7 (17.1) 169.9 (13.5)
Absolute injured 3 111.9 (34.3) 112.9 (31.9) −3.4 (−15.5; 8.8) 0.585
side (°) 6 132.4 (31.6) 133.2 (33.8) 0.1 (−12.3; 12.5) 0.989
12 145.0 (32.7) 136.2 (37.8) 10.2 (−2.4; 22.8) 0.114
% of healthy side 3 67.2 (20.0) 66.8 (18.2) −1.9 (−9.2; 5.5) 0.621
6 80.1 (17.8) 78.7 (18.7) 0.6 (−6.9; 8.1) 0.872
12 87.8 (17.9) 80.4 (19.4) 7.1 (−0.5; 14.8) 0.069
Abduction
Healthy side (°)c – 156.2 (27.7) 166.6 (17.7)
Absolute injured 3 96.9 (37.0) 101.2 (37.6) −3.6 (−17.0; 9.8) 0.601
side (°) 6 118.7 (37.7) 125.2 (39.4) −0.6 (−14.2; 13.1) 0.934
Absolute and relative (% of 12 131.9 (38.0) 131.2 (35.2) 7.4 (−6.4; 21.3) 0.294
healthy side) ranges of motion % of healthy side 3 62.9 (22.4) 60.8 (21.5) −1.1 (−9.4; 7.3) 0.798
for flexion, abduction and 6 77.3 (22.1) 75.3 (20.7) 1.9 (−6.6; 10.4) 0.656
passive external rotation at
follow-up examination 12 85.3 (19.5) 79.1 (16.8) 6.7 (−1.9; 15.4) 0.126
Group difference of operative Passive external rotation
over non-operative treatment Healthy side (°)c – 66.8 (16.9) 58.5 (14.0)
* Significant at 0.05 level Absolute injured 3 42.3 (23.7) 47.8 (23.1) −12.8 (−20.1; −5.4) 0.001*
a
95 % confidence interval side (°) 6 51.0 (22.2) 47.0 (23.5) −3.2 (−10.7; 4.4) 0.411
b
Wald test 12 58.9 (21.2) 42.3 (16.6) 8.6 (0.8; 16.3) 0.031*
c
For each patient the mean of % of healthy side 3 61.6 (30.1) 80.5 (35.0) −20.3 (−31.8; −8.9) 0.001*
contralateral values recorded 6 75.0 (24.9) 79.3 (32.5) −4.3 (−16.0; 7.4) 0.470
across follow-up examinations 12 86.4 (25.9) 74.2 (26.0) 13.1 (1.1; 25.2) 0.033*
was used as reference value

60 CI 0.66–36.0, P = 0.1 for operative over non-operative

Percentage of patients (%)

treatment). Moreover, implant-associated complications

40
(such as plate impingement, screw cut out, etc.) correlated
with excessive varus malalignment (RR = 2.5; 95 % CI
1.7–3.6; P = 0.0002).
20

0
° ° ° ° 15°
Discussion
>45 -4 5 -30 / -15 s>
rus > 30° > 15° al + lgu
Va ru s ru s mic Va
Va Va ato
An
The aim of this analysis was to compare the outcomes of
Varus/Valgus deviation categories operative to non-operative treatment strategies in patients
Non-operative Operative with isolated two-part surgical neck fractures of the proxi-
mal humerus. The most important finding of the present
Fig. 2  Coronal plane alignment as assessed on plain radiographs at analysis was that both operatively and non-operatively
1-year follow-up. Binomial regression analysis revealed a significant treated patients can expect reasonable shoulder function
reduction of varus malalignment exceeding 15° in the operative group
and, if any, superiority of operative over non-operative
(RR = 0.70; 95 % CI 0.47; 1.05, P = 0.083)
treatment can only be expected in the first 3 months follow-
ing injury.
Complications are summarised in Table 5. All but one Appreciation for the strengths and limitations of the
operated fractures showed bony union at 12-month fol- present study is warranted. First and foremost patients
low-up, there was one case of avascular necrosis and one were not randomly assigned to either one of the treatment
deep infection in the operative group. Generally, there groups; the choice of therapy was left at discretion of the
was a non-significant tendency towards a higher compli- treating physician and lastly of the patient. Therefore,
cation rate in surgically treated patients (RR = 4.9; 95 % some selection bias cannot be excluded. Randomisation

13
 Arch Orthop Trauma Surg

versa, throughout the study period, all patients who met

the inclusion criteria and agreed to participate in the study
were treated operatively in the hospitals allocated to one of
the three operative arms, limiting the possible bias that less
displaced fractures were treated conservatively. It, however,
remains unclear, whether and if, how many patients opted
for conservative treatment and refused to participate in the
study. However, both groups were comparable with respect
to all demographic parameters assessed. It is a particular
strength of the analysis that only two-part fractures of the
surgical neck were included resulting in a rather homog-
enous patient sample despite the fact that initial displace-
ment was not assessed. We, moreover, statistically adjusted
for available factors that could have had an influence on the
results.
Follow-up rates were reasonably high at all visits. Still,
more unfavourable results in patients lost to follow-up
cannot be fully excluded—a phenomenon that has been
reported for studies on the outcome of patients undergo-
ing total knee replacement [10]. It does, however, not seem
unreasonable to presume that if there was an effect of drop-
outs, it would comparably have affected both groups. The
maximum follow-up of 1 year may seem relatively short.
It has, however, been shown that long-term outcome can be
predicted fairly well at this time and little improvement nor
deterioration of function can be expected after that point
[16].
Fig. 3  Absolute and relative Constant scores at 1-year follow-up over Expectedly and in accordance with most studies on
coronal plane alignment. Relative Constant scores were significantly proximal humerus fractures, both subjective and objective
lower in the operative group when varus malalignment exceeded 15° outcome parameters improved continuously over the obser-
as compared to anatomical or valgus alignment (92.2 vs. 83.0 %, vation period [5, 6, 8, 11]. Constant scores at 1-year fol-
P = 0.0006 using a two-sample t test)
low-up were generally excellent in both groups and in the
range of those reported for locking plate fixation of two-
of patients would have been desirable to increase internal part fractures [7, 9, 17]. Neer scores at final follow-up were
validity. However, the original study behind this post hoc slightly better than those in the series of Court-Brown et al.
analysis was designed as four separate prospective cohort [6]. Given the well-known reciprocal correlation of age and
series (PHN, LPHP, PHILOS and conservative) in which outcome of proximal humerus fractures [6], this may most
one centre recruited conservatively treated patients and likely be attributed to the younger patient sample in the
each of the remaining participating centres was to exclu- present analysis as compared to the cohort reported on by
sively enrol patients treated with either one of the respec- Court-Brown et al. [6].
tive implant to yield the highest possible quality for the Overall Neer scores at 1-year follow-up were slightly
individual surgical procedures. Since all conservatively better in the non-operative as compared to the operative
treated patients were recruited at a single institution, the treatment group while Constant scores were comparable.
results may not necessarily reflect general results of con- Differences were, however, small and most likely reflect
servative treatment and be influenced by both patient char- that pain at fracture site contributes more points to Neer
acteristics and treatment experience. In this context it is, as compared to Constant scoring. Given the multiplicity of
however, important to state that during the study period outcomes and statistical tests implemented, such minor dif-
not a single two-part fracture was treated operatively in ference and marginal significance may also have occurred
the institution that enrolled conservatively treated patients, by chance alone, and thus should be interpreted with
i.e. there was no selection bias in that more displaced frac- caution.
tures were assigned to operative treatment. Yet of course, Operative treatment resulted in reduced pain and a
it remains unrevealed whether patients had elected not to tendency towards better range of motion in both flex-
show up again and move to another (operative) centre. Vice ion, abduction and passive rotation as compared to

13
Arch Orthop Trauma Surg

Table 5  Complications Complications Non-operative Operative

n % 95 % CI n % (95 % CI)

Total number of patients 31 133

Primary complications 0 0 0–11.2 7 5.3 2.1–10.5
Screw perforation (prim.) 0 0 0–11.2 6 4.5 1.7–9.6
Plate impingement 0 0 0–11.2 1 0.8 0.02–4.1
Secondary complications 1 3. 2 0.08–16.7 18 13.5 8.2–20.5
Implant complications 0 0 0–11.2 7 5.3 2.1–10.5
  Screw perforation (sec.) 0 0 0–11.2 2 1.5 0.18–5.3
  Implant loosening 0 0 0–11.2 0 0 0–2.7
  Screw backing out 0 0 0–11.2 5 3.8 1.2–8.6
  Plate and/or screw pull-out 0 0 0–11.2 1 0.8 0.02–4.1
  Implant breakage 0 0 0–11.2 0 0 0–2.7
  Other implant/surgery 0 0 0–11.2 0 0 0–2.7
Bone/fracture complications 1 3.2 0.08–16.7 12 9 4.7–15.2
  Loss of reduction 1 3.2 0.08–16.7 8 6 2.6–11.5
  Dislocation fragment (sec.) 0 0 0–11.2 2 1.5 0.18–5.3
  Impaction 0 0 0–11.2 3 2.3 0.47–6.5
  Delayed union 0 0 0–11.2 2 1.5 0.18–5.3
  Non-union 0 0 0–11.2 1 0.8 0.02–4.1
  Head necrosis 0 0 0–11.2 1 0.8 0.02–4.1
  Impingement 0 0 0–11.2 0 0 0–2.7
  Other bone/fracture 0 0 0–11.2 0 0 0–2.7
Soft tissue/wound 0 0 0–11.2 3 2.3 0.47–6.5
complications
  Superficial infection 0 0 0–11.2 0 0 0–2.7
  Deep infection 0 0 0–11.2 2 1.5 0.18–5.3
Overall risk ratio for any   Nerve complication 0 0 0–11.2 0 0 0–2.7
complication was 4.9 (95 %   Haematoma 0 0 0–11.2 1 0.8 0.02–4.1
CI 0.66; 36.0), P = 0.1 for   Other soft tissue 0 0 0–11.2 0 0 0–2.7
operative over non-operative Any local complication 1 3.2 0.08–16.7 21 15.8 10–23.1
treatment

non-operative treatment at the 3 month follow-up. It is rea- of operatively treated patients. One possible explanation
sonable to conceive that a higher initial stability translating might be a higher rate of implant-associated complications
to pain reduction and allowing for more aggressive physi- (such as plate impingement or implant failure) in cases
otherapy in the early course of treatment may account for with excessive varus malalignment. In fact, implant related
this observation. complications were significantly more frequent in cases
Expectedly, radiographic analyses revealed a significant of excessive varus malalignment. The causality, however,
reduction of coronal plane malalignment in the operative remains unclear. One could argue that such complications
group. The improvement of alignment observed in the oper- lead to a loss of reduction and therefore resulted in mala-
ative group did, however (at least for two-part fractures), lignment. Secondary dislocation can certainly account for
not translate to better overall outcomes. Interestingly, some of the malalignments. Dislocation rates were, how-
varus malalignment exceeding 15° was associated with an ever, not high enough to explain for all the malalignments
impairment of Constant scores in the operative group. On observed. They may more likely be the result of insufficient
the contrary, varus malalignment did not have a significant reduction at time of surgery. We therefore hypothesise that
effect on Constant scoring in the non-operative group– increased implant-associated complications may be a con-
–an observation that is in accordance with that of Court- sequence of insufficient restoration of alignment (i.e. insuf-
Brown et al. [6] who found no correlation of outcome and ficient reduction).
alignment. We can only speculate why varus malalign- Non-union is rarely considered a problem in two-part
ment seems to have a detrimental effect on the outcome fractures of the proximal humerus and rates of <5 % have

13
 Arch Orthop Trauma Surg

been reported [6]. Accordingly, only one case of non-union (C. Sinopidis); Johannes Gutenberg Universitätsklinikum, Klinik
was observed in the present series. In accordance with für Unfallchirurgie, Mainz, Germany (J. Blum); Università di
Padova, Ortopedia e Traumatologia, Padova, Italy (A. Olmeda);
previous reports on operative stabilisation of proximal Vogtlandklinikum Plauen, Plauen, Germany (W. Merbold); Kardi-
humerus fractures, most of the complications were implant nal Schwarzenberg’schen Krankenhaus, Abteilung für Unfallchirur-
failures (particularly primary and secondary screw perfora- gie, Schwarzach, Austria (F. Genelin); Unfallkrankenhaus Meidling,
tion and loss of reduction) and could partially be attributed Wien, Austria (H. Matuschka); Klinikum Worms gGmbH, Worms,
Germany (J. Blum) PHILOS: Rätisches Kantons- und Regionalspi-
to an insufficient operative procedure. tal, Chur, Switzerland (C. Sommer); Hôpital Cantonal de Fribourg,
Infections are uncommon affecting ~1 % of all opera- Fribourg, Switzerland (G. Kohut); Westpfalz-Klinikum GmbH,
tively treated patients. Accordingly, there was only one Unfallchirurgie Klinik, Kaiserslautern, Germany (H. Winkler); Kan-
case of deep infection in the present series. The overall tonsspital Luzern, Chirurgie/Traumatologie, Luzern, Switzerland (R.
Babst); Klinikum Rosenheim, Unfall- und Wiederherstellungschi-
complication rates were low as compared to other studies rurgie, Rosenheim, Germany (G. Regel); BG Unfall- und Univer-
on operative treatment of proximal humerus fractures [12, sitätsklinik, Tübingen, Germany (D. Höntzsch) LPHP: Charité Uni-
18–20]. This can, however, almost certainly be attributed to versitätsmedizin Berlin, Zentrum für Muskuloskeletale Chirurgie,
the fact that only two-part fractures were included in the Berlin, Germany (N. Haas); Allgemeines Krankenhaus Celle, Celle,
Germany (H-J. Oestern); Albert-Ludwigs-Universität, Orthopädie und
present analysis which are notoriously less frequently asso- Traumatologie, Freiburg, Germany (N. Südkamp); Universitätsklinik
ciated with complications when compared with three- and für Unfallchirurgie Graz, Graz, Austria (M. Plecko); Evangelisches
four-part fractures or fracture dislocations of the proximal Diakoniewerk Friederikenstift, Unfallklinik, Hannover, Germany (H.
humerus. In summary, with regard to the present analyses Lill); Universität Leipzig, Klinik für Unfall- und Wiederherstellung-
schirurgie, Leipzig, Germany (C. Josten). Non-operative study arm:
both non-operative and operative treatment can be consid- York District Hospital, York, United Kingdom (P. De Boer)
ered viable options for two-part humerus fractures. Fur-
ther sufficiently powered randomised controlled trials are
warranted to identify patients who will benefit most from References
operative treatment.
1. Agudelo J et al (2007) Analysis of efficacy and failure in proxi-
mal humerus fractures treated with locking plates. J Orthop
Conclusion Trauma 21(10):676–681
2. Blum J et al (2009) Angle-stable intramedullary nailing of proxi-
mal humerus fractures with the PHN (proximal humeral nail).
Both non-operative treatment and operative treatment using Oper Orthop Traumatol 21(3):296–311
modern implants (LPHP, PHILOS and PHN) can be con- 3. Brunner F et al (2009) Open reduction and internal fixation of
sidered safe and effective treatment options for two-part proximal humerus fractures using a proximal humeral locked
plate: a prospective multicenter analysis. J Orthop Trauma
fractures of the proximal humerus. Operative treatment 23(3):163–172
may result in better range of motion and reduced pain in 4. Constant CR, Murley AH (1987) A clinical method of functional
the early postoperative course of treatment. However, these assessment of the shoulder. Clin Orthop Relat Res 214:160–164
effects diminish over time and outcome will be similarly 5. Court-Brown CM et al (2002) Impacted valgus fractures (B1.1)
of the proximal humerus. The results of non-operative treatment.
good 1 year upon injury regardless of choice of treatment. J B Jt Surg Br 84(4):504–508
Varus malalignment, despite more frequently observed in 6. Court-Brown CM et al (2001) The translated two-part fracture of
the conservative group, did not affect the outcome of con- the proximal humerus. Epidemiology and outcome in the older
servatively treated patients. On the contrary, higher compli- patient. J B Jt Surg Br 83(6):799–804
7. Handschin AE et al (2008) Functional results of angular-stable
cation rates and impaired shoulder function were associated plate fixation in displaced proximal humeral fractures. Injury
with excessive varus malalignment in the operative group. 39(3):306–313
With regard to the good results achieved by conservative 8. Hanson B et al (2009) Functional outcomes after nonoperative
treatment, operative treatment for two-part fractures of the management of fractures of the proximal humerus. J Shoul Elb
Surg 18(4):612–621
proximal humerus should be indicated with caution, pos- 9. Helwig P et al (2009) Does fixed-angle plate osteosynthesis solve
sibly reserved to cases in which early pain reduction and the problems of a fractured proximal humerus? A prospective
regain of shoulder function is of importance, e.g. younger series of 87 patients. Acta Orthop 80(1):92–96
patients or patients with functional impairment of the con- 10. Kim J et al (2004) Response bias: effect on outcomes evalua-
tion by mail surveys after total knee arthroplasty. J B Jt Surg Am
tralateral arm. Meticulous restoration of alignment should 1(86-A):15–21
be aspired whenever opting for operative treatment. 11. Konrad G et al (2010) Open reduction and internal fixation
of proximal humeral fractures with use of the locking proxi-
Acknowledgments The authors wish to thank all the following mal humerus plate. Surgical technique. J B Jt Surg Am 92(1 Pt
investigators and clinics for their participation in this study: Opera- 1):85–95
tive study arm: PHN: Centro Traumatologico Ortopedico, Firenze, 12. Kostler W et al (2003) Proximal humerus fracture in advanced
Italy (R.Angeloni); UZ Gasthuisberg, Leuven, Belgium (S. Nijs); age. Treatment with fixed angle plate osteosynthesis. Chirurg
Royal Liverpool University Hospital, Liverpool, United Kingdom 74(11):985–989

13
Arch Orthop Trauma Surg

13. Lanting B et al (2008) Proximal humeral fractures: a systematic 19. Strohm PC et al (2007) Locking plates in proximal humerus frac-
review of treatment modalities. J Shoul Elb Surg 17(1):42–54 tures. Acta Chir Orthop Traumatol Cech 74(6):410–415
14. Mueller M, Narzarian S (1990) The comprehensive classifica- 20. Strohm PC et al (2008) Proximal humerus fracture––what to do?
tion for fractures of long bones. Heidelberg, Springer, New York, Z Orthop Unfall 146(3):312–317
Berlin 21. Sudkamp N et al (2009) Open reduction and internal fixation
15. Neer CS 2nd (1970) Displaced proximal humeral fractures I. of proximal humeral fractures with use of the locking proximal
Classification and evaluation. J B Jt Surg Am 52(6):1077–1089 humerus plate. Results of a prospective, multicenter, observa-
16. Olsson C et al (2005) Long-term outcome of a proximal humerus tional study. J B Jt Surg Am 91(6):1320–1328
fracture predicted after 1 year: a 13-year prospective population- 22. Urgelli S et al (2005) Conservative treatment versus prosthetic
based follow-up study of 47 patients. Acta Orthop 76(3):397–402 replacement surgery to treat 3- and 4-fragment fractures of the
17. Siwach R et al (2008) Internal fixation of proximal humeral frac- proximal epiphysis of humerus in the elderly patient. Chir Organi
tures with locking proximal humeral plate (LPHP) in elderly Mov 90(4):345–351
patients with osteoporosis. J Orthop Traumatol 9(3):149–153 23. Zyto K (1998) Non-operative treatment of communited fractures
18. Sproul RC et al. (2011) A systematic review of locking plate fixa- of the proximal humerus in elderly patients. Injury 29(5):349–352
tion of proximal humerus fractures. Injury 42(4):408–413

13