Arch Orthop Trauma Surg (2010) 130:241–249

DOI 10.1007/s00402-009-0866-0

TRAUMA SURGERY

Simple elbow dislocations: a systematic review of the literature

J. de Haan · N. W. L. Schep · W. E. Tuinebreijer ·
P. Patka · D. den Hartog

Received: 27 January 2009 / Published online: 2 April 2009

© The Author(s) 2009. This article is published with open access at Springerlink.com

Abstract Introduction
Objective To identify if functional treatment is the best
available treatment for simple elbow dislocations. The elbow joint is the second most commonly dislocated
Search strategy Electronic databases MEDLINE, EMBASE, joint in adults. The annual incidence of simple and complex
LILACS, and the Cochrane Central Register of Controlled elbow dislocations in children and adults is 6.1 per 100,000
Trials. [1]. Elbow dislocations are classiWed as simple or complex
Selection criteria Studies were eligible for inclusion if types [2]. The simple dislocation is characterised by the
they were trials comparing diVerent techniques for the absence of fractures, while the complex dislocation is asso-
treatment of simple elbow dislocations. ciated with fractures. The terrible triad is an example of a
Data analysis Results were expressed as relative risk for complex posterior dislocation with intra-articular fractures
dichotomous outcomes and weighted mean diVerence for of the radial head and coronoid process. The annual inci-
continuous outcomes with 95% conWdence intervals. dence of complex elbow dislocations in children and adults
Main results This review has included data from two tri- is 1.6 per 100,000, or 26% percent of all elbow dislocations
als and three observational comparative studies. Important [1]. Conn et al. [3] found 414 injuries of the elbow in their
data were missing from three observational comparative fracture service, including 58 elbow dislocations in children
studies and the results from these studies were extracted for and adults. Elbow injuries accounted for 6.8% of all treated
this review. No diVerence was found between surgical fractures. Seventy-six percent of the patients with elbow
treatment of the collateral ligaments and plaster immobili- dislocations were older than 20 years. In 51% of these
sation of the elbow joint. Better range of movement, less adults, the dislocations were simple, a lower percentage
pain, better functional scores, shorter disability and shorter than the 74% found in Josefsson’s study [1]. Elbow dislo-
treatment time were seen after functional treatment versus cations can also be classiWed by the direction of their dis-
plaster immobilisation. placement. Nearly all the dislocations are of the posterior or
posterolateral types. In Conn’s study, 96% of the disloca-
Keywords Elbow · Elbow joint · Dislocation · Review · tions were posterior or lateral [3] and JoseVson reported no
Therapy anterior dislocations in his study of 52 patients [4]. In 58%
of patients, the simple elbow dislocations were on the non-
dominant side [4]. Following reposition and treatment in
plaster of simple dislocations, recurrent dislocations and
J. de Haan
chronic instability are not or only rarely seen [2]. For
Department of Surgery and Traumatology, Westfriesgasthuis,
Maelsonstraat 3, 1624 NP Hoorn, The Netherlands instance in JoseVson’s study an obviously unstable joint
was seen in his study of 52 patients after a mean follow-up
N. W. L. Schep · W. E. Tuinebreijer · P. Patka · D. den Hartog (&) of 24 years [4]. After reposition of the simple dislocation,
Department of Surgery-Traumatology, Erasmus MC, University
treatment options include immobilisation in a static plaster
Medical Center Rotterdam, Gravendijkwal 230, OYce H-960,
3015 CE Rotterdam, The Netherlands for diVerent periods, surgical treatment of the ruptured
e-mail: [email protected] medial and lateral collateral ligaments or so-called functional

123
242 Arch Orthop Trauma Surg (2010) 130:241–249

treatment, which is characterised by early active move- Dichotomous outcomes (e.g., presence/absence of nor-
ments within the limits of pain with or without the use of a mal extension) were reported as proportions and were
sling, hinged brace or functional plaster. In theory, after directly compared (diVerence in proportions). We used
repositioning of a simple dislocated elbow, the joint retains these proportions to calculate risk ratios (RRs) and absolute
an inherent stability caused by the contour of the intact joint risk reductions (risk diVerences) with 95% conWdence
surfaces. This stability may allow the patient to exercise the intervals (CIs). For continuous data (e.g., range of motions,
joint shortly after the reposition. This functional treatment function scores) results are presented as weighted mean
should prevent stiVness or restricted range of motion with- diVerences (WMD). We used Review Manager 5.0 soft-
out risking increased joint instability. ware (RevMan 5.0, Cochrane software) for generating the
The primary objective of this systematic review of the Wgures and statistical analyses. We explored heterogeneity
literature was to identify if functional treatment is the best using the chi-squared test with signiWcance set at a P value
available treatment for simple elbow dislocations after less than 0.10. The quantity of heterogeneity was estimated
closed reduction. by the I-squared statistic.
Because prior statistical evidence existed for homogene-
ity of eVect sizes, the planned analysis used a Wxed eVect
Materials and methods model.

We conducted an electronic search including MEDLINE,

EMBASE, LILACS and the Cochrane Central Register of Results
Controlled Trials (CENTRAL). We did not limit the search
by language or publication date. We used the following A total of two randomised controlled trials (RCTs) and six
search terms in diVerent combinations as MeSH (Medical observational comparative studies comparing diVerent treat-
Subject Heading) terms and as text words: elbow joint, dis- ments for elbow dislocations were included with a total
location, treatment outcome, surgery, controlled clinical enrolment of 342 patients with available follow-up (see
trial, comparative study. Manual searches including refer- Tables 1, 2, 3, 4, 5, 6, 7 and 8 for characteristics of the
ence lists of all included studies were used to identify trials included studies). The full text of every study was retrieved.
that the electronic search may have failed to identify. Because only two RCTs were retrieved we expanded the
Two reviewers independently assessed the titles and review with observational comparative studies. Non-com-
abstracts of all reports identiWed by electronic and manual parative observational studies were excluded. All studies
searches. Each report was labelled as (a) deWnitely exclude, included simple elbow dislocations. One study consisted of
(b) unsure or (c) deWnitely include. Full text articles of patients with simple and complex elbow dislocations [5].
abstracts labelled as “unsure” were reassessed according to No RCTs or comparative studies of complex elbow disloca-
the inclusion criteria for this review. Any diVerences were tions were retrieved.
resolved through discussion. Studies labelled as “deWnitely One RCT comparing surgical and non-surgical treatment
exclude” were excluded from the review, while studies of simple elbow dislocation was included [6]. The other
labelled as “deWnitely include” were further assessed for RCT compared functional treatment with immobilisation in
methodological quality. plaster during 3 weeks [7]. The observational comparative
Two reviewers independently extracted the data for the studies compared functional treatment with immobilisation
primary and secondary outcomes and entered the data into in plaster [5, 8, 9] or compared diVerent periods of immobi-
data collection forms developed for this purpose. Discrep- lisation [10–12].
ancies were resolved by discussion. All data were entered Observational studies that did not compare diVerent
into Review Manager [RevMan, (Computer program. Ver- treatments were excluded because they provide a low level
sion 5.0. Copenhagen: The Nordic Cochrane Centre, The of evidence (level IV evidence, no control group).
Cochrane Collaboration, 2008)]. In Josefsson’s study [6] random selection was by the use
Two reviewers independently assessed the included of sealed envelopes, but in Rafai’s study [7] no information
studies for sources of systematic bias in trials. The studies on randomisation was published. In the observational stud-
were evaluated with the following criteria: allocation con- ies from Schippinger [12] and Maripuri [8] the period of
cealment (selection bias), rates of follow-up and intention immobilisation, and thus the treatment allocation, was
to treat analysis (attrition bias). Allocation concealment dependent on the preference of the treating doctor.
was graded as (a) adequate, (b) inadequate or (c) unsure. Since blinding of treatment is diYcult or impossible in
DiVerences between the two reviewers were resolved by surgical treatments, the RCTs did not blind doctors or
discussion. Masking of outcome assessors in the included patients to treatment. No information is provided about
studies was assessed. blinding of the evaluators of the outcomes.

123
Arch Orthop Trauma Surg (2010) 130:241–249 243

Table 1 Characteristics of the study of Josefsson et al. [6]

Methods Randomised controlled trial

Participants 30 consecutive patients included, acute dislocation of the elbow, age ¸16 years, mean
age 34.5 years, free from elbow symptoms before injury. Dislocation with fracture excluded
except small avulsed fragments <2 £ 3 mm, 10 males, 20 females, 18 dislocations left,
12 dislocations right side, 28 posterior or posterolateral and 2 lateral dislocations. Reduction
in emergency room. Examination under general anaesthesia after mean of 4 days for examination
stability: all elbows medial instability and 16 lateral instability. N = 11 re-dislocated easily, most
often in 45° of Xexion
Interventions Surgical treatment: N = 15, exploration medial and lateral side joint through separate incisions. Medial
and lateral collateral ligaments found to be totally ruptured, although only 8 showed lateral instability.
Suturing and re-Wxation of ligaments. 6 of the 11 easily re-dislocated elbows treated surgically.
Immobilisation in plaster, 90°, 19 days (SD = 3). 1 patient in this group lost to follow-up
Non-surgical treatment: N = 15, 5 of the 11 easily re-dislocated elbows treated non-surgically.
Immobilisation 17 days (SD = 2). 1 patient in this group lost to follow-up
Outcomes Follow-up surgical group 31 months (SD = 15), non surgical 24 months (SD = 11). Range of motions
at 5, 10 weeks and Wnal examination >1 year: no diVerence in motion, grip strength, pain, instability
Loss of extension >1 year: surgical group 18° (SD = 15) and non-surgical group 10° (SD = 14)
Loss of Xexion >1 year: surgical group 1° (SD = 2) and non-surgical group 1° (SD = 2)
For unstable elbows (N = 11 of which 6 were treated surgically) the loss of extension >1 year
was 20° (SD = 19), and loss of Xexion was 2° (SD = 3)
No recurrent dislocations or episodes of instability in both groups
Allocation concealment Random selection by 30 sealed envelopes, 15 envelopes for surgical treatment and 15 for non-surgical
treatment

In the observational studies of Protzman [11] and MehlhoV

Table 2 Characteristics of the study of Rafai et al. [7]
[10] no standard deviations of the outcome measures are
Methods Randomised controlled trial given and in the study of MehlhoV [10] the sample sizes of the
Participants 50 pure posterior luxations, adults, treatment groups are also not provided. In Schippinger’s study
normal psychological proWle, stable [12] the sample sizes and outcome scores of the three groups
after reposition and tested under
with diVerent immobilisation periods are not provided.
general anaesthesia, no previous elbow
injury. Mean age 25 years In the observational studies of Schippinger [12] and Mari-
(range 16–67 years), 43 males, puri [8] the period of immobilisation was dependent on the
7 females, 30 right arm, 20 left arm treating doctor and was most likely biased by the severity of
Interventions Group I: N = 26, reduction in general the trauma so that the patients with the most severe trauma
anaesthesia and testing stability.
received the longest period of immobilisation.
Immobilisation for 3 weeks
The results were expressed as relative risk (RR) for
Group II: N = 24, reduction in general
anaesthesia and testing stability. dichotomous outcomes and weighted mean diVerence
Mobilisation after 3 days. Functional (WMD) for continuous outcomes with 95% conWdence
treatment intervals (CI).
Outcomes Normal extension: group I 81% and Only data from two observational studies comparing
group II 96% (statistically signiWcant functional treatment with plaster immobilisation could be
diVerence concluded by authors)
pooled [5, 8]. The percentages of excellent or good results
StiVness (=loss of Xexion): group I = 19%
and group II = 4% (statistically were pooled with the Mantel-Haenszel statistical method.
signiWcant diVerence concluded For this pooling, the Wxed eVects model was used since we
by authors) assumed that all variation between the two studies was
No diVerence in pain and ossiWcations caused by chance and that the studies measured the same
No recurrent dislocations or episodes overall eVect. Even if a random-eVects model was used, our
of instability in both groups conclusions remained the same. Data from the other studies
Notes No P values are given, but only that compared diVerent types of treatment and used diVer-
remarks declaring signiWcant results
ent outcome measures could not be pooled due to clinical
Allocation No details about randomisation
and methodological heterogeneity, and thus are described
concealment
individually.

123
244 Arch Orthop Trauma Surg (2010) 130:241–249

Table 3 Characteristics of the study of Royle [5]

Methods Retrospective, observational study with 2 comparative groups with mean follow-up of 31 months
Participants N = 38, follow-up of N = 32, period 1982–1987, mean age 35.8 years, 17 males (53%), 15 females (47%),
N = 23 (72%) posterolateral dislocation, N = 9 (28%) posterior, N = 20 (62%) associated fractures:
N = 12 radial head, N = 6 coronoid, N = 4 olecranon avulsion fracture, N = 4 medial epicondyle, N = 1
lateral condyle, N = 1 capitellum, average time for reduction 3.8 h, general anaesthesia N = 27 (84%),
N = 1 internal Wxation radial head fracture, instability after reduction N = 8 (tested in extension with val-
gus stress)
Interventions Group I: N = 9, closed reduction and plaster, mean duration 24.7 days
Group II: N = 23, reduction and sling, mean 17.5 days
Outcomes Group I excellent (no pain and full extension) or good (minimal pain and extension loss <15°) in 33.3 ver-
sus 83% in group II. Results were graded according to Lindscheid and Wheeler
No recurrent dislocations
Notes Age range 11–75 years; thus included children, also associated fractures N = 20 (62%)
Posterior dislocation 100% good/excellent result versus N = 18 (56%) posterolateral dislocation
Better outcome if reduction <3 h, 87 versus 53% good/excellent result
Associated fractures N = 8 (40%) fair (exertional pain and 15–30° extension loss) or poor (constant pain
and >30° extension loss) versus N = 2 (17%) without fractures
The results of group I versus group II could be confounded by associated fractures, time of reduction and
direction of dislocation
Bias Heterogeneity of groups, children included, confounded by associated fractures, time of reduction and
direction of dislocation

Table 4 Characteristics of the study of Maripuri et al. [8]

Methods Observational retrospective comparative study

Participants 47 simple elbow dislocations in period 2000–2004, mean age 42.5 years, follow-up >2 years, N = 42 avail-
able for review. Inclusion criteria: age ¸16 years, simple dislocation, closed reduction, concentric relo-
cation conWrmed by radiography, follow-up >2 years, no associated fractures, no neurovascular deWcit.
Posterolateral dislocation 60%, direct posterior 30%, posteromedial 10%
Interventions Group I: N = 20, plaster immobilisation, mean 14 days followed by physiotherapy until range of motions
(ROM) 100°
Group II: N = 22, sling application and early mobilisation within pain limits
Outcomes Group I: mean score Mayo Elbow Performance Index (MEPI) 83.8 (SEM = 4.2, SD = 18.8). Group II:
mean score MEPI 96.5 (SEM = 8.9, SD = 8.9), P < 0.05. MEPI score components are pain, ROM, sta-
bility, daily function, which are graded as excellent 90–100, good 75–89, fair 60–74, poor <60
Group I: mean score Quick Disabilities of the Arm, Shoulder and Hand (DASH) questionnaire 12.8
(SEM = 3.5, SD = 15.7). Group II: mean score DASH 2.7 (SEM = 1.5, SD = 7.0), P < 0.05. Of the
DASH the disability and symptom section was used
Weeks oV work: group I mean 6.6 weeks (SEM = 0.64, SD = 2.86); group II 3.2 weeks (SEM = 0.29,
SD = 1.36), P < 0.001
20 patients (of 22) with excellent or good result in group II (depends on MEPI score). 12 patients (of 20)
with excellent or good result in group I (depends on MEPI score)
One recurrent dislocation in group I, treated surgically
Notes Period of immobilisation depended on preference of the treating doctor
Allocation concealment Retrospective study
Bias Selection bias for therapy, attending physician decides, instability, time period, for co-interventions: only
50% of group 1 received physiotherapy at 2 weeks versus 100% of group II

Surgical versus non-surgical treatment of simple elbow ¡2.75 to 18.75; P = 0.14) and loss of Xexion (Comparison
dislocations 1.2: WMD 0.00, 95% CI ¡1.48 to 1.48; P = 1.00) were not
statistically diVerent between the two groups. Furthermore,
Only one RCT was found that compared surgical with non- at 10 weeks the loss of extension (Comparison 1.3: WMD
surgical treatment [6] (Table 9). At more than 1 year the 11.00, 95% CI ¡4.19 to 26.19; P = 0.16) and loss of Xexion
loss of extension (Comparison 1.1: WMD 8.00, 95% CI (Comparison 1.4: WMD 6.00, 95% CI ¡0.11 to 12.11;

123
Arch Orthop Trauma Surg (2010) 130:241–249 245

Table 5 Characteristics of the study of Riel et al. [9]

Methods Observational retrospective comparative study with a historical control group. Mean follow-up 8.2
(SD = 4.5) years
Participants In period 1976–1992 50 simple elbow dislocations, N = 6 treated surgically, N = 44 conservatively, last
group re-examined. Reduction without anaesthesia (N = 31) or in local anaesthesia
Interventions Group I: period 1976–1985, N = 20, reduction and immobilisation in plaster for 3–4 weeks, N = 17 patients
re-examined, N = 1 telephone enquiry, follow-up 11 (SD = 2.6) years, mean plaster period 24 (SD = 3)
days plus data from medical records, last examination after a mean of 6 months
Group II: period 1985–1992, N = 24, reduction and functional treatment day after reposition, N = 18
patients re-examined, N = 3 telephone enquiry, follow-up 4 (SD = 1.8) years, mean plaster period 2
(SD = 1) days plus data from medical records, last examination after a mean of 4 months
Outcomes Range of motions, stability and power not diVerent between groups
After-treatment period group I 12 (SD = 3) weeks, group II 8 (SD = 3) weeks, disability period group I 16
(SD = 8) weeks, group II 8 (SD = 3) weeks, physical rehabilitation period group I 6 (SD = 3) months,
group II 4 (SD = 3) months
Notes Sex had no inXuence on result. No recurrent dislocations
Allocation concealment No RCT, observational comparative study with a historical control group

Table 6 Characteristics of the study of Protzmann [11]

Methods Retrospective observational study with 3 comparative groups, mean follow-up 24.5 months
Participants 49 consecutive patients, military service, 1971–1976, from N = 47 follow-up, range age 17–44 years,
N = 15 associated fractures of which 1 radial head fracture and one coronoid process (=only one which
could inXuence stability). From N = 25 X-ray: 19 posterolateral, 5 posterior, 1 posteromedial, no anterior
dislocation
Interventions Closed reduction without anaesthesia and group I immobilisation <5 days, N = 27; group II immobilisation
10–15 days, N = 13; group III immobilisation >20 days, N = 7
Outcomes Mean extension loss group I = 3°, group II = 11°, group III = 21°. Mean duration disability group
I = 6 weeks, group II = 19 weeks, group III = 24 weeks. No SD given
Notes No standard deviations given for outcome measures. No recurrent dislocations and no subjective com-
plaints of instability. 28 patients of the 47 with follow-up had periarticular or ligamentous calciWcations
Allocation concealment No RCT, observational study, probably retrospective, comparative study, comparison = post-hoc, immobi-
lisation period was decision of orthopaedic surgeon
Bias Selection bias for therapy, treating doctor decided

P = 0.05) were not statistically diVerent. Moreover, at normal extension (Comparison 2.2: RR 1.78, 95% CI 1.23–
5 weeks the loss of extension (Comparison 1.5: WMD 2.57; P = 0.002) at 3 months was statistically higher in the
11.00, 95% CI ¡4.93 to 26.93; P = 0.18) and loss of Xexion functional treatment group. A post hoc power calculation
(Comparison 1.6: WMD 9.00, 95% CI ¡0.88 to 18.88; on the percentages of normal extension and Xexion at
P = 0.07) were not statistically diVerent. A post hoc power 1 year with G*Power software (version 3.03, Kiel, Ger-
calculation on the mean loss of extension after 1 year with many) showed a power of 19%.
G*Power software (version 3.03, Kiel, Germany) showed a Data from two studies could be pooled to analyse the
power of 29%. percentage of excellent and good outcomes (Fig. 1) [5, 8].
At a follow-up time greater than 2 years, there was a sig-
Functional treatment versus plaster treatment niWcant diVerence between functional and plaster treatment
for the outcome excellent and good results (Comparison
One RCT and three observational studies are described 2.6: RR 1.76, 95% CI 1.19–2.60; P = 0.004). The other out-
(Table 10). The results of the RCT are described individu- come measures of Maripuri [8] study are described individ-
ally [7]. The percentages of normal extension (Comparison ually. Several other measures were statistically diVerent:
2.1: RR 1.19, 95% CI 0.97–1.46; P = 0.10) and Xexion the mean diVerences of the Mayo Elbow Performance
(Comparison 2.3: RR 1.19, 95% CI 0.97–1.46; P = 0.10) Index (MEPI) (Comparison 2.7: WMD 12.70, 95% CI
and pronation and supination (Comparison 2.5: RR 1.25, 3.66–21.74; P = 0.006), short Quick Disabilities of the
95% CI 0.99–1.56; P = 0.06) at 1 year and normal Xexion Arm, Shoulder and Hand (Comparison 2.8: WMD ¡10.10,
(Comparison 2.4: RR 1.25, 95% CI 0.99–1.56; P = 0.06) at 95% CI ¡17.58 to ¡2.62; P = 0.008) and weeks oV work
3 months were not statistically diVerent. The percentage of (Comparison 2.9: WMD ¡3.40, 95% CI ¡4.78 to ¡2.02;

123
246 Arch Orthop Trauma Surg (2010) 130:241–249

Table 7 Characteristics of the study of MehlhoV et al. [10]

Methods Observational retrospective comparative study with 3 comparative groups, mean follow-up 34.3 months
Participants 90 consecutive patients, adults, simple dislocations, follow-up >12 months, age >18 years, no associated
fractures. Stable after reduction. Period 1978–1985, follow-up from N = 52 (56% follow-up), N = 34
males, N = 18 females. Dislocations: 90% posterolateral + posterior, 10% posteromedial + medial
Interventions Closed reduction, after reduction stability and ROM were tested and gravity stress photos were taken.
Group I immobilisation 0–13 days; group II immobilisation 14–24 days; group III immobilisation
¸25 days
Outcomes Ratings extension loss: <5° excellent, <15° good, <30° fair, ¸30° poor
Groups divided according to immobilisation period: Group I 0–13 days. Group II 14–24 days, Group III
>24 days
Mean Xexion contracture = loss of extension: group I: 5.1°; group III 30.1°; loss of Xexion: group I 2.7°,
group II 5.6°, group III 18.6°. Pain (McGill Pain Questionnaire): group I 80% no pain, group II 45% no
pain, group III 10% no pain
Instability non signiWcant. No sample sizes of the groups and no SDs for the outcome measures are pre-
sented
No gross instability of the elbow or recurrent dislocation
Notes No correlation between age, sex or length of follow-up and Xexion contracture, pain or instability
(Chi-square test, multiple testing)
Heterotopic ossiWcation was seen in 55% of the radiographs, but there was no correlation with impairment
of motion
Allocation concealment No RCT, observational study, probably not prospective, comparative study, groups were formed post-hoc
Bias Selection bias, 31 of 84 patients did not participate, selection bias for therapy, treating doctor decided

Table 8 Characteristics of the study of Schippinger et al. [12]

Methods Retrospective observational study with comparative groups (post hoc). Mean follow-up 61.5 (SD = 22.2)
months
Participants 45 simple elbow dislocations, no or minor fractures (<2 £ 3 mm), 2 trauma centres, period 1989–1995,
N = 27 posterior, N = 12 posterolateral, N = 2 bilateral posterior, N = 1 medial, N = 1 anterior, N = 1
divergent, N = 1 anterolateral dislocation, age 44.5 years (SD = 15.9)
Interventions Closed reduction without general anaesthesia. Check for re-dislocation in various Xexion positions. Group
I immobilisation <2 weeks; group II immobilisation 2–3 weeks; group III immobilisation >3 weeks
Outcomes Morrey scores and pain group I and II better than group III, but nonsigniWcant. Number of groups and
scores of groups not given
N = 28 periarticular ossiWcations and N = 11 heterotopic calciWcations, but no correlation of ossiWcations
with impairment of motion
No recurrent dislocations
Notes Period of immobilisation was dependent on preference of the orthopaedic surgeon
Allocation concealment No RCT, observational study, retrospective, comparative study, groups were formed post hoc, immobilisa-
tion period was decision of orthopaedic surgeon

P < 0.0001) all suggested better results following func- P < 0.0001), disability period in weeks (Comparison 2.11:
tional treatment. MEPI is one of the most commonly used WMD ¡8.00, 95% CI ¡11.71 to ¡4.29; P < 0.0001) and
physician-based elbow rating systems. This index consists after-treatment time in months (Comparison 2.12: WMD
of four parts: pain (with a maximum score of 45 points), ¡2.00, 95% CI ¡3.78 to ¡0.22; P = 0.03) were statistically
ulnohumeral motion (20 points), stability (10 points) and signiWcant shorter in the functional group.
the ability to perform Wve functional tasks (25 points). The
DASH disability/symptom score is a summation of the DiVerent periods of plaster immobilisation
responses to 11 questions on a scale of 1–5, with 0 (no dis-
ability) to 100 (severe disability). The results of the observational studies [10–12] comparing
The results of the observational study of Riel [9] are diVerent periods of plaster immobilisation could not be
described individually. The physiotherapy time in weeks expressed as RR or WMD because data (sample sizes of the
(Comparison 2.10: WMD ¡4.00, 95% CI ¡5.78 to ¡2.22; groups or scores and/or standard deviations) were missing.

123
Arch Orthop Trauma Surg (2010) 130:241–249 247

Table 9 Surgical versus non-surgical treatment of simple elbow dislocation

Outcome Studies Participants Statistical method EVect estimate

1.1 Loss of extension at more than 1 year 1 28 Mean diVerence (IV, Wxed, 95% CI) 8.00 [¡2.75, 18.75]
1.2 Loss of Xexion at more than 1 year 1 28 Mean diVerence (IV, Wxed, 95% CI) 0.00 [¡1.48, 1.48]
1.3 Loss of extension at 10 weeks 1 28 Mean diVerence (IV, Wxed, 95% CI) 11.00 [¡4.19, 26.19]
1.4 Loss of Xexion at 10 weeks 1 28 Mean diVerence (IV, Wxed, 95% CI) 6.00 [¡0.11, 12.11]
1.5 Loss of extension at 5 weeks 1 28 Mean diVerence (IV, Wxed, 95% CI) 11.00 [¡4.93, 26.93]
1.6 Loss of Xexion at 5 weeks 1 28 Mean diVerence (IV, Wxed, 95% CI) 9.00 [¡0.88, 18.88]
IV inverse variance; CI conWdence interval

Table 10 Functional treatment versus plaster immobilisation

Outcome Studies Participants Statistical method EVect estimate

2.1 Percentage of patients with normal 1 50 Risk ratio (M-H, Wxed, 95% CI) 1.19 [0.97, 1.46]
extension at 1 year
2.2 Percentage of patients with normal 1 50 Risk ratio (M-H, Wxed, 95% CI) 1.78 [1.23, 2.57]
extension at 3 months
2.3 Percentage of patients with normal 1 50 Risk ratio (M-H, Wxed, 95% CI) 1.19 [0.97, 1.46]
Xexion at 1 year
2.4 Percentage of patients with normal 1 50 Risk ratio (M-H, Wxed, 95% CI) 1.25 [0.99, 1.56]
Xexion at 3 months
2.5 Percentage of patients with normal 1 50 Risk ratio (M-H, Wxed, 95% CI) 1.25 [0.99, 1.56]
pro- and supination at 1 year
2.6 Percentage patients with excellent 2 74 Risk ratio (M-H, Wxed, 95% CI) 1.76 [1.19, 2.60]
or good results at >2 years
2.7 Mayo Elbow Performance Index (MEPI) 1 42 Mean diVerence (IV, Wxed, 95% CI) 12.70 [3.66, 21.74]
2.8 Quick Disabilities of the Arm, 1 42 Mean diVerence (IV, Wxed, 95% CI) ¡10.10 [¡17.58, ¡2.62]
Shoulder and Hand (DASH)
2.9 Weeks oV work 1 42 Mean diVerence (IV, Wxed, 95% CI) ¡3.40 [¡4.78, ¡2.02]
2.10 Physiotherapy time (weeks) 1 44 Mean diVerence (IV, Wxed, 95% CI) ¡4.00 [¡5.78, ¡2.22]
2.11 Period disability (weeks) 1 44 Mean diVerence (IV, Wxed, 95% CI) ¡8.00 [¡11.71, ¡4.29]
2.12 After-treatment time (months) 1 44 Mean diVerence (IV, Wxed, 95% CI) ¡2.00 [¡3.78, ¡0.22]
M-H Mantel-Haenszel statistical method; CI conWdence interval; IV inverse variance

Fig. 1 Forest plot comparing

functional treatment (sling) and
plaster immobilisation for the
percentage of excellent or good
results

Thus, we could not judge the following conclusions made did not analyse this data statistically. The number of
by the authors of the studies. Without making statistical patients with symptoms of instability of the elbow joints
inferences, Protzman [11] describes less extension loss and increased from the shorter immobilisation group to the
shorter mean disability in weeks for the shorter immobilisa- longer immobilisation groups without reaching statistical
tion group. MehlhoV [10] describes less extension loss for signiWcance at the 5% level. Schippinger [12] saw better
the two shorter immobilisation groups (group I 0–13 days, Morrey scores, which are composed of the items pain,
group II 14–24 days), with a signiWcant correlation between movement, strength, instability and function (activities of
extension loss and duration of follow-up (P = 0.001). He daily living), and better separate pain scores in the shorter
also reported less Xexion loss and less prevalence and immobilisation groups, though without statistical signiW-
severity of pain for the shorter immobilisation groups but cance.

123
248 Arch Orthop Trauma Surg (2010) 130:241–249

Stability testing of the elbow joint after reposition tional treatment or plaster immobilisation, two observa-
tional comparative studies were pooled [5, 8]. This
Do the above results diVer for stable or instable elbow classiWcation of excellent or good depends on the amount
joints after reduction? Nearly all cited studies only included of pain and range of movement. The results favoured the
stable joints after reduction. An exception is the study of functional group. This functional treatment after the reposi-
Josefsson et al. [6]. In this study, the elbows were tested for tion consisted of early mobilisation in a sling without a
instability after reduction in general anaesthesia and com- plaster or brace.
pared with the other elbow in full but unforced extension. For the outcome measures MEPI score, quick DASH
All the elbows showed medial instability and 16 of 30 score and weeks of work we used an individual observa-
elbows showed lateral instability. Eleven elbows re-dislo- tional study. Functional treatment resulted in signiWcantly
cated easily. Royle’s [5] study also included unstable elbow better outcomes. In addition, an individual observational
joints. The elbows were tested mainly in general anaesthe- study showed that patients in the functional group needed
sia in extension with valgus stress and eight of the 38 less time for physiotherapy and after-treatment and had a
elbows showed instability. MehlhoV et al. [10] and Schipp- shorter disability period. Importantly, since treatment allo-
inger et al. [12] tested for instability and did not include cation was determined by the attending physician in these
unstable elbows. Maripuri et al. [8], Riel et al. [9] and observational studies, it is likely that severe cases were pre-
Protzman et al. [11] did not test the elbows for instability. scribed longer immobilisation. In one study, outcome was
We carefully conclude that the majority of the patients, in fact correlated with the presence of fractures, delay to
included in these studies, had simple dislocations, which reduction, and direction of dislocation [5]. Any of these
remained stable after reposition. variables could be a confounding factor in analysing the
eVect of treatment in study, as the heterogeneity could be
introduced by combining patients with simple and complex
Discussion dislocations.
Data from the studies comparing diVerent periods of
This review has included data from two trials and three plaster immobilisation could not be extracted, while the
observational comparative studies. Important data were authors of all three observational studies observed less
missing from three observational comparative studies and movement loss after shorter immobilisations, but this Wnd-
the results from these studies were extracted for this ing was only statistically signiWcant in one study. These
review. studies could also be confounded by the severance of the
Only one RCT assessed suture repair of the collateral injury, as worse cases probably underwent longer immobi-
ligaments of the elbow joint versus conservative treatment lisation periods.
with plaster [6]. No statistically signiWcant diVerences were In the eight included studies only one recurrent dislocation
found either for loss of extension at 5 weeks, 10 weeks or after treatment was mentioned [8] i.e., one recurrence on 342
after more than 1 year, or for loss of Xexion after more than patients (0.3%). No subjective or gross objective signs of
1 year. A trend was found for less loss of Xexion at 5 and instability were found after treatment, indicating that recur-
10 weeks for the conservative group. This study lacked the rent dislocations and instability are not a problem after
power to Wnd a signiWcant diVerence because of its small simple posterior dislocations. The majority of the patients
sample size. (323 out of 342 patients) probably had a stable elbow joint
Only one RCT compared functional treatment and plas- after reduction of the dislocation, although it was not clear in
ter [7]. The percentages of patients with normal extension three studies if the patients were tested for instability.
and Xexion at 1 year were not statistically diVerent. A sig-
niWcantly higher percentage of patients with normal exten-
sion at 3 months was found for the group with functional Summary of main results
treatment. A trend was found for a higher percentage of
patients with normal Xexion at 3 months and normal pro- No diVerence was found between surgical treatment of the
and supination at 1 year for the functional treatment group. collateral ligaments and plaster immobilisation of the elbow
This study also lacked the power to Wnd a signiWcant diVer- joint. Better range of movement, less pain, better functional
ence because of its small sample size. An important short- scores, shorter disability and shorter treatment time were
coming of this study is that it did not describe the seen after functional treatment versus plaster immobilisation.
randomisation process, so allocation bias cannot be Since we did not Wnd any RCTs or comparative studies that
excluded. studied complex elbow dislocations, our conclusions can
To analyse the percentage of patients with excellent or only address simple elbow dislocations. Our conclusions
good results at more than 2 years following either func- only apply to stable elbow joints after reduction.

123
Arch Orthop Trauma Surg (2010) 130:241–249 249

The quality of the evidence is very low because of the 3. Conn J Jr, Wade PA (1961) Injuries of the elbow: a ten year
lack of high-quality RCTs. Moreover, the available RCTs review. J Trauma 1:248–268
4. Josefsson PO, Johnell O, Gentz CF (1984) Long-term sequelae of
lack power due to their small sample sizes. The observa- simple dislocation of the elbow. J Bone Joint Surg Am 66(6):927–
tional studies could be biased by confounding due to the 930
use of a historical control group or treatment allocation by 5. Royle SG (1991) Posterior dislocation of the elbow. Clin Orthop
the treating physician rather than by randomisation. In addi- Relat Res 269:201–204
6. Josefsson PO, Gentz CF, Johnell O et al (1987) Surgical versus
tion, the treatment groups were not balanced for important non-surgical treatment of ligamentous injuries following disloca-
potential confounders, and some observational studies did tion of the elbow joint. A prospective randomized study. J Bone
not provide important data as sample size and/or standard Joint Surg Am 69(4):605–608
deviations. 7. Rafai M, Largab A, Cohen D et al (1999) Pure posterior luxation
of the elbow in adults: immobilization or early mobilization. A
Since we did not Wnd any RCTs or comparative studies randomized prospective study of 50 cases. Chir Main 18(4):272–
that studied complex elbow dislocations, our conclusions 278
can only address simple elbow dislocations, which are con- 8. Maripuri SN, Debnath UK, Rao P et al (2007) Simple elbow dis-
sidered stable after reposition. location among adults: a comparative study of two diVerent meth-
ods of treatment. Injury 38(11):1254–1258
We advise to test the elbow after reposition for instabil- 9. Riel KA, Bernett P (1993) Simple elbow dislocation. Comparison
ity by valgus and varus testing and by the lateral pivot-shift of long-term results after immobilization and functional treatment.
test [13]. When the elbow is considered stable one may Unfallchirurg 96(10):529–533
consider functional after treatment with a pressure bandage. 10. MehlhoV TL, Noble PC, Bennett JB et al (1988) Simple disloca-
tion of the elbow in the adult. Results after closed treatment.
When plaster immobilisation is preferred to treat simple J Bone Joint Surg 70(2):244–249
elbow dislocations one has to realise that immobilisation of 11. Protzman RR (1978) Dislocation of the elbow joint. J Bone Joint
more than 14 days may be associated with stiVness. Surg Am 60 (4):539–541
12. Schippinger G, Seibert FJ, Steinbock J et al (1999) Management of
simple elbow dislocations. Does the period of immobilization aVect
Open Access This article is distributed under the terms of the Cre-
the eventual results? Langenbecks Arch Surg 384(3):294–297
ative Commons Attribution Noncommercial License which permits
13. O’Driscoll SW, Jupiter JB, King GJ et al (2001) The unstable
any noncommercial use, distribution, and reproduction in any medium,
elbow. Instr Course Lect 50:89–102
provided the original author(s) and source are credited.

References

1. Josefsson PO, Nilsson BE (1986) Incidence of elbow dislocation.

Acta Orthop Scand 57(6):537–538
2. Hildebrand KA, Patterson SD, King GJ (1999) Acute elbow dislo-
cations: simple and complex. Orthop Clin North Am 30(1):63–79

123