The Journal of Foot & Ankle Surgery 52 (2013) 622–628
Contents lists available at ScienceDirect
The Journal of Foot & Ankle Surgery
journal homepage: www.jfas.org
Rerupture Rate after Early Weightbearing in Operative Versus Conservative
Treatment of Achilles Tendon Ruptures: A Meta-analysis
Dorien M. van der Eng, MD 1, Tim Schepers, MD, PhD 2, J. Carel Goslings, MD, PhD 2,
Niels W.L. Schep, MD, PhD, MSc 2
1
2
Department of Surgery, Academic Medical Centre, Amsterdam, The Netherlands
Trauma Unit, Department of Surgery, Academic Medical Centre, Amsterdam, The Netherlands
a r t i c l e i n f o
a b s t r a c t
Level of Clinical Evidence: 1
Keywords:
Achilles tendon
braces
rupture
surgical procedure
weightbearing
Whether Achilles tendon rupture benefits from surgery or conservative treatment remains controversial.
Moreover, the outcome can be influenced by the rehabilitation protocol. The goal of the present meta-analysis
was to compare the rerupture rate after surgical repair of the Achilles tendon followed by weightbearing
within 4 weeks versus conservative treatment with weightbearing within 4 weeks. In addition, a secondary
analysis was performed to compare the rerupture rates in patients who started weightbearing after 4 weeks.
Seven randomized controlled trials published from 2001 to 2012, with 576 adult patients, were included. The
primary outcome measure was the rerupture rate. The secondary outcomes were minor and major complications other than rerupture. In the early weightbearing group, 7 of 182 operatively treated patients (4%)
experienced rerupture versus 21 of 176 of the conservatively treated patients (12%). A secondary analysis of the
patients treated with late weightbearing showed a rerupture rate of 6% (7 of 108) for operatively treated
patients versus 10% (11 of 110) for conservatively treated patients. The differences concerning the rerupture
rate in both groups were not statistically significant. No differences were found in the occurrence of minor or
major complications after early weightbearing in both patient groups. In conclusion, we found no difference in
the rerupture rate between the surgically and nonsurgically treated patients followed by early weightbearing.
Weightbearing after 4 weeks also resulted in no differences in the rupture rate in the surgical versus
conservatively treated patients. However, surgical treatment was associated with a twofold greater complication rate than conservative treatment.
Ó 2013 by the American College of Foot and Ankle Surgeons. All rights reserved.
Rupture of the Achilles tendon is a common injury, mostly
affecting adult males who participate in sports (1). Whether Achilles
tendon rupture requires surgery or nonoperative treatment remains
controversial. The outcome can also be influenced by the methods of
rehabilitation. Specifically, weightbearing or non-weightbearing
might be of importance.
Research has shown lower rerupture rates after surgery compared
with nonsurgical treatment but not always with a significant difference. In previously conducted meta-analyses, the rerupture rates for
surgically treated patients varied from 3.1% to 5.0% versus 8.8% to 13%
for nonsurgically treated patients (1–4).
One drawback of surgery has been the greater rate of complications. The pooled results of the Cochrane collaboration showed a 29%
Financial Disclosure: None reported.
Conflict of Interest: None reported.
Address correspondence to: Dorien M. van der Eng, MD, Department of Surgery,
Academic Medical Centre, Nachtwachtlaan 120, Amsterdam 1058 ED, The Netherlands.
E-mail address: d.m.vandereng@hotmail.com (D.M. van der Eng).
complication rate in the surgical group versus 8% in the nonsurgical
group (2).
Animal studies have shown that immobilization compromises the
healing process of tendons and could indicate that a functional
rehabilitation protocol is warranted (5,6). Krapf et al investigated
operatively and nonoperatively treated Achilles tendon ruptures in
rat models with immediate weightbearing (7). They found no
difference in the maximum failure load between the operative and
nonoperative groups, nor did they find a difference in tendon stiffness during the 8-week research period. However, a nonoperatively
treated Achilles tendon rupture showed a nearly normal histologic
pattern, although a sutured tendon showed scar tissue and less
organized collagen fibers (7). Earlier research also showed that
loading of healing tendons leads to essential changes in the biologic
process of tendon healing (8).
Moreover, early weightbearing could theoretically prevent muscle
atrophy, stiffness, adhesions, and deep venous thrombosis and has
been associated with faster healing and stronger tendons because of
improved vascularization and an improved immunologic response
1067-2516/$ - see front matter Ó 2013 by the American College of Foot and Ankle Surgeons. All rights reserved.
http://dx.doi.org/10.1053/j.jfas.2013.03.027
D.M. van der Eng et al. / The Journal of Foot & Ankle Surgery 52 (2013) 622–628
(5–7). It is also considered to be more patient friendly. Therefore, early
mobilization after operative and nonoperative treatment of Achilles
tendon ruptures has been advocated (9–11).
Previous reviews and trials that compared surgical and nonsurgical treatment did not specifically focus on the rehabilitation
protocol. Therefore, the goal of the present meta-analysis was to
compare the rerupture rate after surgical repair of the Achilles tendon
followed by early weightbearing versus conservative treatment with
early weightbearing. An additional analysis was performed of surgical
versus conservative management with weightbearing after 4 weeks.
The secondary outcomes were minor and major complications.
Patients and Methods
The present study was reported according to the Preferred Reporting Items for
Systematic Reviews and Meta-Analyses guidelines (12).
Eligibility Criteria
Randomized controlled trials of adults patients with acute rupture of the Achilles
tendon, published from 2001 to 2012, were included. The intervention was defined as
surgical repair of the Achilles tendon followed by early weightbearing. The control
group consisted of those who had undergone nonoperative treatment with early
weightbearing. The surgical treatment could include open or minimally invasive repair.
Nonoperative treatment could consist of casting, bracing, orthosis, and taping. Early
weightbearing was defined in the present study as weightbearing initiated within the
first 4 weeks after treatment. The follow-up period had to be at least 1 year.
In addition, a subgroup analysis was performed to compare the rerupture rates in
surgically versus nonsurgically treated patients for whom weightbearing was initiated
4 weeks after the beginning of treatment.
Primary and Secondary Outcomes
The primary outcome measure was the rerupture rate. The secondary outcomes
were minor and major complications. Minor complications included superficial wound
infections, painful or hypertrophic scars, and transient pain. Major complications
included deep vein thrombosis, pulmonary embolism, deep infection, and sural nerve
injury.
623
Results
Data Search
The search resulted in 32 potentially relevant studies. Of these, 10
met the inclusion criteria. However, 3 studies were excluded because
no comparison had been done between operative and nonoperative
treatment. Thus, 7 randomized controlled trials, with 624 patients,
were included in the present meta-analysis (Fig. 1).
After 1 year of follow-up, a total of 576 patients were eligible for
evaluation. Of these 576 patients, 290 were treated operatively and
286 were treated nonoperatively.
Of the 7 studies, 6 had included 496 patients and compared open
operative treatment and nonoperative treatment, followed by functional rehabilitation after 1 year of follow-up (13–18). The seventh
study compared minimally invasive surgical treatment and conservative treatment and included 80 patients (9).
Description of Included Studies
The characteristics and rehabilitation protocols of the studies are
summarized in Tables 1 and 2.
The study by Costa et al (13) consisted of 2 independent trials. The
patients chose either operative or nonoperative treatment. Subsequently, the patients from the 2 treatment arms were randomized by
computer for the rehabilitation method.
In the first trial, all the patients underwent surgical repair of the
Achilles tendon. Rehabilitation consisted of functional bracing with
early weightbearing or a plaster cast without weightbearing. The
choice of surgical technique was left to the operating surgeon.
In the second trial, every patient was treated nonoperatively.
Similar to their first trial, the patients were randomized for functional
rehabilitation after treatment with immediate weightbearing or
a plaster cast without weightbearing. After 6 months, an independent
Information Sources and Search Methods
The Ovid MEDLINE electronic database (January 2001 to December 2012) and the
Cochrane Library were used for the data search. A systematic search was performed
with the following terms: Achilles tendon (MeSH), rupture (MeSH) AND surgery (all
fields) or surgical procedures, operative (MeSH). The studies were limited to
randomized controlled trials.
The following search strategy was used: ((“achilles tendon”[MeSH Terms] OR
(“achilles”[All Fields] AND “tendon”[All Fields]) OR “achilles tendon”[All Fields]) AND
(“rupture”[MeSH Terms] OR “rupture”[All Fields])) AND (“surgery”[Subheading] OR
“surgery”[All Fields] OR “surgical procedures, operative”[MeSH Terms] AND Randomized Controlled Trial[ptyp] AND (“2001/01/01”[PDAT] : “2012/12/31”[PDAT])).
Study Selection
Two physicians assessed the abstracts of the studies during the initial search. Fulltext copies of potentially relevant studies were assessed by the same physicians and
subsequently verified by a trauma surgeon.
Quality Assessment
For evaluation of the methodologic quality of the included studies, the guidelines of
the Cochrane Bone, Joint and Muscle Trauma group were used. The quality assessment
of the included studies was performed in duplicate by 2 physicians.
32 articles identified through
search strategy
3 articles removed (duplicates)
29 articles screened
19 articles excluded,
because not meeting
eligibility criteria
10 full-text articles
assessed for eligibility
3 articles excluded,
because no comparison
between operative and
non-operative treatment
7 studies included in
qualitative synthesis
Data Collection and Analysis
The Cochrane Review Manager software, version 5.1 (RevMan, Cochrane Collaboration, Oxford, UK) was used for data analysis. Two physicians extracted the data from
all included studies. All outcomes were dichotomous variables. The risk ratios (RRs) and
95% confidence intervals (CIs) were calculated. A treatment effect was defined as
significant if p < .05.
The random effects model was used. Heterogeneity was explored using the chisquare test, with significance set at p < .100. For quantification, I2 was used, with
values less than 30% indicating low heterogeneity.
7 of studies included in
quantitative synthesis
(meta-analysis)
Fig. 1. Study flow diagram of the systematic review.
624
D.M. van der Eng et al. / The Journal of Foot & Ankle Surgery 52 (2013) 622–628
Table 1
Characteristics of included studies, including number of patients, follow-up, and study type
Investigator
Costa et al (13), 2006
Keating et al (14), 2011
Metz et al (9), 2008
€ller et al (15), 2001
Mo
Nilsson-Helander
et al (16), 2010
Twaddle et al (17) 2007
Willits et al (21), 2010
Total
Initially Included Patients (n)
After 1 y of Follow-Up
(n ¼ 576)
Total
Operative
Operative
Conservative
Patients Lost to
Follow-Up (%)
Conservative
Mean Age (y)
Operative Technique
Study Type
Operative
Conservative
55
80
83
112
100
23
39
42
59
50
22
41
41
53
50
20
37
41
59
49
19
39
39
53
48
10.9
5
3.6
0
3
42.0
41.2
40.0
39.6
40.9
53.0
39.5
41.0
38.5
41.2
Unknown
Open, Kessler stitch
Minimally invasive, Bunnell
Open, Kessler stitch
Open, Kessler stitch
Multicenter
Single center
Multicenter
Multicenter
Single center
50
144
624
25
72
310
25
72
304
22
62
290
23
65
286
10
11.8
41.8
39.7
40.3
41.1
Open, Krackow
Open, Krackow
Single center
Multicenter
physiotherapist performed muscle function tests and a clinical
assessment to measure the ankle range of motion. For the present
meta-analysis, the immediate weightbearing group of the surgical
trial was compared with the immediate weightbearing group of the
conservative trial (13).
In the study by Keating and Will (14), the patients were randomly
allocated to operative or nonoperative treatment. The surgically
treated patients had undergone surgery within 7 days of presentation.
The operation was performed by 1 of the 4 involved consultant
orthopedic surgeons or by a senior trainee under supervision. The
complete postoperative treatment was a plaster cast in both groups.
The patients were allowed to start weightbearing after 6 and 8 weeks
in the operative and nonoperative groups, respectively. Range of
motion and isokinetic tests were performed by a research physiotherapist (14).
The study by Metz et al (9) was unique in that minimal invasive
surgery was used (i.e., an incision of <5 cm). In the randomization
process, block stratification was applied to balance the groups among
the 4 participating hospitals. Treatment in the surgical and conservative group was initiated within 3 days after rupture. All the physicians were familiar with the minimally invasive technique. The
patients in the surgical group received 1 week of plaster casting and
were not allowed to load their leg during that week. The rehabilitation
phase for the nonsurgically treated patients consisted of a brace
(VacopedÒ, OPED, Valley, Germany). The patients were not allowed
to remove it by themselves. Both groups were allowed to start
weightbearing after 1 week (9).
€ ller et al (15) randomized their patients between surgical repair
Mo
with bracing and nonsurgical treatment with a plaster cast. The
treatment started within 7 days after rupture. Surgically treated
patients were allowed to remove the brace and were encouraged to
exercise. Weightbearing was allowed in the third week for the
surgically treated patients and in the fourth week for the nonsurgically treated patients. Diagnostic tests were performed by
independent physiotherapists (15).
The patients from the surgical and nonsurgical groups in the study
by Nilsson et al (16) were treated within 3 days. Surgery was performed by 1 of 28 orthopedic surgeons. The patients from both groups
wore a brace and were not allowed to remove it (DonJoy ROM Wal€ , Sweden). Weightbearing was allowed after
kerÒ, DJO Nordic, Malmo
6 weeks. The evaluation of leg function and physical activity was
performed by 2 independent physiotherapists (16).
In the study by Twaddle and Poon (17), the patients had to have
presented within 10 days after their injury. Allocation was performed
by tossing a coin 50 times. The sequence of “heads” (operative) and
“tails” (nonoperative) was used to allocate the sequence of blinded
envelopes with the treatment option. Within 2 days after presentation, the patients in the surgical group underwent surgery by a senior
Table 2
Description of postoperative protocol of surgical and nonsurgical groups
Investigator
Costa et al (13), 2006
Keating et al (14), 2011
Metz et al (9), 2008
€ller et al (15), 2001
Mo
Nilsson-Helander
et al (16), 2009
Twaddle et al (17), 2007
Willits et al (18), 2010
Protocol Surgical Group
Brace: 4.5-, 3.0-, 1.5-, 0-cm heel
raise for 2, 2, 2, 2 wk
Cast: equinus, semi-equinus for
4, 2 wk
Cast: 30 plantar flexion for 1 wk,
followed by functional taping
þ heel raise: 2 cm, 1 cm, 0 cm
for 2, 2, 2 wk
Cast: equinus for 12 days,
followed by brace: maximal
30 plantar flexion, 10 plantar
flexion, maximal 10
dorsiflexion for 4, 2, 2 wk
Cast: equinus for 2 wk, followed
by brace: 30 plantar flexion,
10 plantar flexion, dynamic,
maximal 10 dorsiflexion for
2, 2, 2 wk
Cast: equinus for 10 days,
followed by brace: 20 plantar
flexion, neutral for 2, 4 wk
Cast: 20 plantar flexion for 2 wk,
followed by brace: 20 plantar
flexion, neutral for 4, 2 wk
Protocol Conservative Group
Brace: 4.5-, 3.0-, 1.5-, 0-cm heel
raise for 2, 2, 2, 2 wk
Cast: equinus, semi-equinus,
neutral for 4, 4, 2 wk
Cast: 30 plantar flexion for 1 wk,
followed by brace: 30 plantar
flexion, 15 plantar flexion,
dynamic neutral for 2, 2, 2 wk
Cast equinus, neutral for 4, 4 wk
Cast: equinus for 2 wk, followed
by brace: 30 plantar flexion,
10 plantar flexion, dynamic,
maximal 10 dorsiflexion for
2, 2, 2 wk
Cast: equinus for 10 days,
followed by brace: 20 plantar
flexion, neutral for 2, 4 wk
Cast: 20 plantar flexion for 2 wk,
followed by brace: 20 plantar
flexion, neutral for 4, 2 wk
Start Weightbearing
Follow-Up Visits
Operative
Nonoperative
Immediate
Immediate
2, 3, 6, 12 mo
6 wk
8 wk
3, 4, 6, 12 mo
1 wk
1 wk
1, 3, 5, 7 wk, 3, 6, 12 mo
3 wk
4 wk
2, 3, 6, 12, 24 mo
6–8 wk
6–8 wk
2 wk, 2, 3, 6, 12 mo
6 wk
6 wk
2 wk, 2, 3, 6, 12 mo
2 wk
2 wk
3, 6, 12, 24 mo
D.M. van der Eng et al. / The Journal of Foot & Ankle Surgery 52 (2013) 622–628
surgeon. After 2 weeks of plaster casting, the patients in both groups
received a removable, functional orthosis for 6 weeks. The patients
were instructed to remove the brace for 5 minutes every hour to
practice active dorsiflexion (not beyond neutral) and passive plantar
flexion. At 6 weeks after the beginning of treatment, the patients were
allowed to remove the brace during the night and were allowed to
start weightbearing (17).
In the study by Willits et al (18), the patients had to have presented
within 14 days after the injury. No mention was made of the number
or experience level of the surgeons. The rehabilitation protocol was
identical in the 2 groups. After 2 weeks of plaster casting, the patients
were placed in a brace (Aircast, Summit, NJ). After 2 weeks, partial
weightbearing was permitted, and weightbearing as tolerated was
allowed in the fourth week. The patients were required to wear the
brace during the night but were allowed to remove it for bathing (18).
Methodologic Quality
The methodologic scoring items and individual scores for each
study are listed in Table 3. Low scores indicate poor methodology. Two
items from the list (blinding of patients and care providers) were
excluded because they proved negative for all 7 studies.
Outcome Parameters
Rerupture Rate
All 7 studies (n ¼ 576 patients) reported information on the
rerupture rate. The 4 studies in which weightbearing was initiated
within 4 weeks after treatment demonstrated no significant difference in the rerupture rate (9,13,15,18). Of the 182 patients in the
operative group, 7 (4%) experienced a rerupture compared with 21 of
176 patients (12%) in the nonoperative group. The RR was 0.4 (95% CI
0.12 to 1.32, p ¼ .13, I2 ¼ 38%).
In the subgroup of 3 studies in which weightbearing was allowed
after 4 weeks, no difference between the groups could be demonstrated
(11,16,17). Of the 108 patients in the operative group, 7 (6%) experienced
rerupture compared with 11 of 110 (10%) patients in the nonoperative
group. The RR was 0.64 (95% CI 0.25 to 1.63, p ¼ .35, I2 ¼ 0%).
An analysis of all the included studies combined showed no
significant difference in the rerupture rate between the operatively
and nonoperatively treated patients (9,13–18). The rerupture rate was
5% (14 of 290) in the operatively treated group and 11% (32 in 286) in
the nonoperatively treated group (RR 0.50, 95% CI 0.25 to 1.00, p ¼ .05,
I2 ¼ 8%; Fig. 2).
Complications
Data on the complications other than reruptures were reported in
all 7 studies.
625
Minor Complications. In the early weightbearing group, the incidence
of minor complications in the surgical group was 15% (28 of 182
patients) compared with 9% (15 of 176 patients) in the nonsurgical
group. The RR was 3.54 (95% CI 0.4 to 31.61, p ¼ .26, I2 ¼ 82%). In the
group that started weightbearing after 4 weeks, 5 complications
developed in the 108 surgically treated patients (5%) versus
0 complications in the 110 nonsurgically treated patients. The RR was
5.53 (95% CI 0.65 to 46.88, p ¼ .12, I2 ¼ 0%). The pooled results of the 2
groups showed a greater incidence of complications in the surgically
treated group; however, no significant difference was found. Of the
290 patients in the surgical group, 33 (11%) developed minor
complications compared with 15 of the 286 patients (5%) in the
nonsurgical group (RR 3.85, 95% CI 0.72 to 20.56, p ¼ .12, I2 ¼ 74%;
Fig. 3A).
Major Complications. The incidence of major complications was lower
than that of the minor complications. In the immediate weightbearing
patients, 6% of the surgically treated group versus 3% of the nonsurgically treated group developed a major complication (RR 1.79, 95%
CI 0.64 to 5.01, p ¼ .27, I2 ¼ 0%).
In the group that started weightbearing after 4 weeks, complications occurred in 6 of 108 surgically treated patients (6%) and 4 of 110
nonsurgically treated patients (4%; RR 1.24, 95% CI 0.19 to 7.93, p ¼ .82,
I2 ¼ 38%). For the total group, the incidence of major complications
was greater in the surgically treated patients (16 of 290 [6%] versus 7
of 286 [2%]). Again, no significant difference was found (RR 1.56, 95%
CI 0.68 to 3.59, p ¼ .29, I2 ¼ 0%; Fig. 3B).
Fig. 4 shows a greater total complication rate for the surgically
versus conservatively treated patients. In the surgical group, 49 of
290 (17%) developed a complication versus 22 of 286 (8%) in the
nonsurgical group (RR 3.00, 95% CI 0.88 to 10.21, p ¼ .08, I2 ¼ 74%).
Strength and Range of Motion
Four studies provided information on isokinetic ankle strength,
tested with dynamometry (13–15,18). All tests were performed using
dynamometers with different settings and patient positioning.
Furthermore, the results were not uniformly reported. Because of
various measuring methods used and nonuniform reporting of the
results, no comparison could be made in the present study. Costa et al
(13) found no significant difference in eccentric or concentric peak
torque strength between the surgical and nonsurgical groups. Keating and Will (14) found no difference in dorsiflexion or plantar
flexion peak strength after 1 year in the operative and nonoperative
€ ller et al (15) found that
groups. Both Keating and Will (14) and Mo
both groups showed increased range of dorsiflexion. In neither group
was a significant difference provable (14,15). Willits et al (18) could
not demonstrate clinically important differences between the
Table 3
Quality assessment of included studies using methodologic quality evaluation guidelines of Cochrane Bone, Joint and Muscle Trauma group
Investigators
Factors Affecting Quality
Randomization or Allocation
Concealment
Costa et al (13), 2006
Keating et al (14), 2011
Metz et al (9), 2008
€ ller et al (15), 2001
Mo
Nilsson-Helander
et al (16), 2009
Twaddle et al (17), 2007
Willits et al (18), 2010
*
1,
2,
2,
2,
2,
Patient choice þ computer
Blinded envelope
Computer
Blinded envelope
Computer
1, Tossing a coin
2, Computer
Total
(maximum
18 points)
Intention
to Treat
Blinding
Treatment
to Outcome
Assessors
Comparable
Treatment
Groups
Care
Programs
Identical
Clear Inclusion
and Exclusion
Criteria
Interventions
Clearly
Defined
Clear
Outcome
Measures
Adequate
Active
Follow-Up
Length
0
1
1
1
1
1
0
0
1
1
0
2
2
2
2
0
0
0
0
2
0
2
2
2
2
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
7
13
13
14
16
1
1
0
0
2
2
2
2
2
2
2
2
1*
2
2
2
13
15
No confidence intervals or standard deviations mentioned.
626
D.M. van der Eng et al. / The Journal of Foot & Ankle Surgery 52 (2013) 622–628
Operative
Conservative
Risk Ratio
Study or Subgroup Events Total Events Total Weight M-H, Random, 95% CI
1.1.1 Start weight before 4 weeks
1.90 [0.19, 19.27]
8.5%
19
1
20
2
Costa 2006
0.32 [0.07, 1.48]
39 18.1%
6
41
2
Metz 2008
0.08 [0.01, 0.61]
53 11.1%
11
59
1
Möller 2001
0.70 [0.12, 4.04]
65 14.3%
3
62
2
Willits 2010
0.40 [0.12, 1.32]
176 51.9%
182
Subtotal (95% CI)
Risk Ratio
M-H, Random, 95% CI
21
7
Total events
Heterogeneity: Tau² = 0.55; Chi² = 4.81, df = 3 (P = 0.19); I² = 38%
Test for overall effect: Z = 1.50 (P = 0.13)
1.1.2 Start weight bearing after 4 weeks
Keating 2011
Nilsson 2009
Twaddle 2007
Subtotal (95% CI)
2
3
2
37
49
22
108
4
6
1
39
48
23
110
16.2%
23.5%
8.4%
48.1%
0.53 [0.10, 2.71]
0.49 [0.13, 1.85]
2.09 [0.20, 21.45]
0.64 [0.25, 1.63]
11
7
Total events
Heterogeneity: Tau² = 0.00; Chi² = 1.20, df = 2 (P = 0.55); I² = 0%
Test for overall effect: Z = 0.94 (P = 0.35)
Total (95% CI)
290
286 100.0%
32
14
Total events
Heterogeneity: Tau² = 0.07; Chi² = 6.52, df = 6 (P = 0.37); I² = 8%
Test for overall effect: Z = 1.96 (P = 0.05)
Test for subgroup differences: Chi² = 0.35, df = 1 (P = 0.56), I² = 0%
0.50 [0.25, 1.00]
0.01
0.1
1
10
100
Favours operative Favours conservative
Fig. 2. Table and Forest plot illustrating the rate of reruptures, divided into 2 groups. The first group started weightbearing before 4 weeks after initiation of treatment. The second group
started weightbearing after 4 weeks. Chi2, chi-square; CI, confidence interval; df, degrees of freedom; M-H, Mantel-Haenszel.
operative and nonoperative patients concerning the range of dorsiflexion or plantar flexion. Also, no difference in strength could be
measured (18).
Discussion
The present meta-analysis showed no significant differences in
the rate of rerupture in surgically treated patients and conservatively
treated patients after early weightbearing. An analysis of patients
with delayed weightbearing after 4 or more weeks also did not
demonstrate a difference in the rerupture rate. Also, no differences
were seen in the incidence of minor or major complications after
early weightbearing in the surgically and conservatively treated
patients. However, for the total group of patients, surgery resulted in
significantly more complications compared with conservative treatment. No less than 1 of 6 surgically treated patients experienced
a complication compared with 1 of 13 conservatively treated
patients.
Three previously conducted reviews concluded that surgical repair
of Achilles tendon rupture was associated with the lowest rerupture
rate (1–3). Nevertheless, the results of these previous reviews cannot
simply be transferred to our findings, because the present analysis
focused especially on the effect of weightbearing and used only recent
randomized controlled trials.
In the review by Soroceanu et al (19), 10 studies, including
826 patients, were analyzed. The pooled results showed that
surgical repair significantly reduced the rerupture rate (RR 0.4,
p ¼ .001) compared with nonsurgical treatment. However, the
results of the stratified analysis showed that when early range of
motion was practiced, surgical and nonsurgical treatment were
equivalent regarding the rerupture rate (p ¼ .45). This suggests that
surgical intervention did not contribute to a risk reduction if
a rehabilitation protocol with early weightbearing was used. The
focus in the review by Soroceanu et al (19) was on early range of
motion, which is not the same as weightbearing. However, the
threshold to discriminate between early and late range of motion
was not reported (19).
Wallace et al (20) followed up more than 900 patients from 1996
to 2008 and treated all patients with functional bracing and
weightbearing within the fourth week. They found a surprisingly
low rerupture rate of 2.8% (20). The rerupture rate in the early
versus late weightbearing group was more or less the same, stratified
by the treatment modality. Consequently, we found no evidence
that early weightbearing is unsafe.
One limitation of the present meta-analysis was the heterogeneity
present when analyzing minor complications, mainly caused by the
inclusion of 1 study (9). In that study, the complication rate in the
nonoperatively treated group was greater than in the surgically
treated group, in contrast to all the other studies. In that study, skinrelated problems occurring in the patients who had worn a brace
were responsible for the high number of complications in the
conservatively treated group. The patients were not allowed to
remove the brace. If that study were excluded from our subgroup
analysis, the RR would change from 3.54 to 8.96. Moreover, the
difference in the minor complication rate between the operative and
nonoperative patients in the early weightbearing group would reach
significance (p ¼ .0008, I2 ¼ 0%).
Another limitation was that the incidence of deep venous thrombosis (DVT) could not be accurately reported. In 1 study (21), all the
patients underwent ultrasonography of the leg at 8 weeks postoperatively or after the beginning of bracing. A high number of
(asymptomatic) DVTs was found in 32 of 95 screened patients; 14 of
49 surgically treated patients developed a DVT compared with 18 of
46 nonsurgically treated patients. The study reported that 10 of 32
patients had complaints. However, it was not stated how many DVTs
were considered clinically relevant. Therefore, it was impracticable to
combine these DVTs with those from the other studies. Thus, the
number of DVTs from that study (21) was not included in the present
analysis.
Another point of concern for our study was the difference in the
rehabilitation protocols of the included studies. Some studies had
D.M. van der Eng et al. / The Journal of Foot & Ankle Surgery 52 (2013) 622–628
A
Operative
Conservative
Study or Subgroup Events Total Events Total Weight
1.2.1 Start weight bearing before 4 weeks
Costa 2006
Metz 2008
Möller 2001
Willits 2010
Subtotal (95% CI)
6
5
9
8
20
41
59
62
182
0
13
1
1
19
39
53
65
176
Risk Ratio
M-H, Random, 95% CI
627
Risk Ratio
M-H, Random, 95% CI
12.38 [0.75, 205.75]
0.37 [0.14, 0.93]
8.08 [1.06, 61.71]
8.39 [1.08, 65.12]
3.54 [0.40, 31.61]
14.5%
22.6%
17.9%
17.9%
72.9%
15
28
Total events
Heterogeneity: Tau² = 3.96; Chi² = 17.12, df = 3 (P = 0.0007); I² = 82%
Test for overall effect: Z = 1.13 (P = 0.26)
1.2.2 Start weight bearing after 4 weeks
Keating 2011
Nilsson 2009
Twaddle 2007
Subtotal (95% CI)
1
4
0
37
49
22
108
0
0
0
39
48
23
110
13.0%
14.1%
27.1%
3.16 [0.13, 75.16]
8.82 [0.49, 159.51]
Not estimable
5.53 [0.65, 46.88]
0
Total events
5
Heterogeneity: Tau² = 0.00; Chi² = 0.23, df = 1 (P = 0.63); I² = 0%
Test for overall effect: Z = 1.57 (P = 0.12)
286 100.0%
290
Total (95% CI)
3.85 [0.72, 20.56]
15
Total events
33
Heterogeneity: Tau² = 3.00; Chi² = 19.15, df = 5 (P = 0.002); I² = 74%
Test for overall effect: Z = 1.58 (P = 0.12)
Test for subgroup differences: Chi² = 0.08, df = 1 (P = 0.78), I² = 0%
B
Operative
Conservative
Study or Subgroup Events Total Events Total Weight
1.3.1 Start weight bearing before 4 weeks
Costa 2006
Metz 2008
Möller 2001
Willits 2010
Subtotal (95% CI)
2
4
1
3
20
41
59
62
182
0
2
1
2
19
39
53
65
176
7.8%
25.7%
9.1%
22.4%
65.0%
0.01
0.1
1
10
100
Favours operative Favours conservative
Risk Ratio
M-H, Random, 95% CI
Risk Ratio
M-H, Random, 95% CI
4.76 [0.24, 93.19]
1.90 [0.37, 9.81]
0.90 [0.06, 14.01]
1.57 [0.27, 9.10]
1.79 [0.64, 5.01]
Total events
10
5
Heterogeneity: Tau² = 0.00; Chi² = 0.69, df = 3 (P = 0.88); I² = 0%
Test for overall effect: Z = 1.11 (P = 0.27)
1.3.2 Start weight bearing after 4 weeks
Keating 2011
Nilsson 2009
Twaddle 2007
Subtotal (95% CI)
2
4
0
37
49
22
108
2
0
2
39
48
23
110
19.0%
8.2%
7.8%
35.0%
1.05 [0.16, 7.10]
8.82 [0.49, 159.51]
0.21 [0.01, 4.12]
1.24 [0.19, 7.93]
Total events
6
4
Heterogeneity: Tau² = 1.04; Chi² = 3.22, df = 2 (P = 0.20); I² = 38%
Test for overall effect: Z = 0.22 (P = 0.82)
Total (95% CI)
290
286 100.0%
Total events
16
9
Heterogeneity: Tau² = 0.00; Chi² = 4.06, df = 6 (P = 0.67); I² = 0%
Test for overall effect: Z = 1.05 (P = 0.29)
Test for subgroup differences: Chi² = 0.12, df = 1 (P = 0.73), I² = 0%
1.56 [0.68, 3.59]
1
0.01
0.1
10
100
Favours operative Favours conservative
Fig. 3. (A) Table and forest plot illustrating the number of minor complications. A subdivision in weightbearing before and after 4 weeks was made. (B) Table and Forest plot illustrating
the number of major complications. A subdivision in weightbearing before and after 4 weeks was made. Chi2, chi-square; CI, confidence interval; df, degrees of freedom; M-H, MantelHaenszel.
different protocols for the surgical and nonsurgical groups. No
previously completed meta-analysis could answer the question of
which protocol is the most effective and safe. This was mainly because
of the small number of patients.
Future randomized controlled trials should be of high methodologic rigor (concealment and blinding of outcome assessors and large
patient numbers) and should contain clear operative protocols and
similar rehabilitation protocols for the surgical and nonsurgical
groups. The use of the same, validated instruments to measure the
functional outcome should be encouraged.
In conclusion, we found no difference in the rerupture rate between
the surgically and nonsurgically treated patients followed by early
weightbearing. The analysis of patients who started weightbearing
after 4 weeks also showed no difference in the rerupture rate in the
surgically versus conservatively treated patients.
Concerning the similar rerupture rate in the surgical and nonsurgical groups, no rehabilitation protocol seemed superior for patients
with acute Achilles tendon rupture. However, surgical treatment was
associated with a twofold greater complication rate compared with
conservative treatment.
628
D.M. van der Eng et al. / The Journal of Foot & Ankle Surgery 52 (2013) 622–628
Study or Subgroup
Costa 2006
Keating 2011
Metz 2008
Möller 2001
Nilsson 2009
Twaddle 2007
Willits 2010
Total (95% CI)
Operative
Conservative
Events Total Events Total Weight
19 10.9%
0
20
8
39 16.7%
2
37
3
39 23.2%
15
41
9
53 18.4%
2
59
10
48 10.7%
0
49
8
23
0
22
0
65 20.0%
3
62
11
290
286 100.0%
22
49
Total events
Heterogeneity: Tau² = 1.56; Chi² = 19.52, df = 5 (P = 0.002); I² = 74%
Test for overall effect: Z = 1.76 (P = 0.08)
Risk Ratio
M-H, Random, 95% CI
Risk Ratio
M-H, Random, 95% CI
16.19 [1.00, 262.45]
1.58 [0.28, 8.93]
0.57 [0.28, 1.15]
4.49 [1.03, 19.58]
16.66 [0.99, 280.85]
Not estimable
3.84 [1.13, 13.13]
3.00 [0.88, 10.21]
0.01
1
0.1
10
100
Favours operative Favours conservative
Fig. 4. Table and Forest plot illustrating the minor and major complications and the subgroups together. Chi2, chi-square; CI, confidence interval; df, degrees of freedom;
M-H, Mantel-Haenszel.
Acknowledgment
The authors would like to thank N. Keulen, MD, for her contribution during the initial search, the assessment of the abstracts, studies,
and quality assessment, and data extraction.
References
1. Khan RJ, Fick D, Keogh A, Crawford J, Brammar T, Parker M. Treatment of acute
Achilles tendon ruptures. A meta-analysis of randomized, controlled trials. J Bone
Joint Surg Am 87:2202–2210, 2005.
2. Khan RJ, Carey Smith RL. Surgical interventions for treating acute Achilles tendon
ruptures. Cochrane Database Syst Rev 9:CD003674, 2010.
3. Bhandari M, Guyatt GH, Siddiqui F, Morrow F, Busse J, Leighton RK, Sprague S,
Schemitsch EH. Treatment of acute Achilles tendon ruptures: a systematic overview and meta-analysis. Clin Orthop Relat Res 400:190–200, 2002.
4. Wilkins R, Bisson LJ. Operative versus nonoperative management of acute Achilles
tendon ruptures: a quantitative systematic review of randomized controlled trials.
Am J Sports Med 40:2154, 2012.
5. Andersson T, Eliasson P, Aspenberg P. Tissue memory in healing tendons: short
loading episodes stimulate healing. J Appl Physiol 107:417–421, 2009.
6. Enwemeka CS, Spielholz NI, Nelson AJ. The effect of early functional activities on
experimentally tenotomized Achilles tendons in rats. Am J Phys Med Rehabil
67:264–269, 1988.
7. Krapf D, Kaipel M, Majewski M. Structural and biomechanical characteristics after
early mobilization in an Achilles tendon rupture model: operative versus nonoperative treatment. Orthopaedics 35:1383–1388, 2012.
8. Palmes D, Spiegel HU, Schneider TO, Langer M, Stratmann U, Budny T, Probst A.
Achilles tendon healing: long-term biomechanical effects of postoperative mobilization and immobilization in a new mouse model. J Orthop Res 20:939–946,
2002.
9. Metz R, Verleisdonk EJ, van der Heijden GJ, Clevers GJ, Hammacher ER,
Verhofstad MH, van der Werken C. Acute Achilles tendon rupture: minimally
invasive surgery versus nonoperative treatment with immediate full weight
bearingda randomized controlled trial. Am J Sports Med 36:1688–1694,
2008.
€m a
€ la
€inen M, Leppilahti J. Early functional treatment
10. Kangas J, Pajala A, Siira P, Ha
versus early immobilization in tension of the musculotendinous unit after Achilles
rupture repair: a prospective, randomized, clinical study. J Trauma 54:1171–1180,
2003.
LA, Durand DC, Jomha NM. The influence of early
11. Suchak AA, Bostick GP, Beaupre
weight-bearing compared with non-weight-bearing after surgical repair of the
Achilles tendon. J Bone Joint Surg Am 90:1876–1883, 2008.
12. Moher D, Liberati A, Tetzlaff J, Altman DG; PRISMA Group. Preferred reporting
items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med
6:e1000097, 2009.
13. Costa ML, MacMillan K, Halliday D, Chester R, Shepstone L, Robinson AH, Donell ST.
Randomised controlled trials of immediate weight-bearing mobilisation for
rupture of the tendo Achillis. J Bone Joint Surg Br 88:69–77, 2006.
14. Keating JF, Will EM. Operative versus non-operative treatment of acute rupture of
tendo Achillis: a prospective randomised evaluation of functional outcome. J Bone
Joint Surg Br 93:1071–1078, 2011.
€ ller M, Movin T, Granhed H, Lind K, Faxe
n E, Karlsson J. Acute rupture of tendon
15. Mo
Achillis: a prospective randomised study of comparison between surgical and nonsurgical treatment. J Bone Joint Surg Br 83:843–848, 2001.
R, Faxe
n E, Olsson N, Eriksson BI,
16. Nilsson-Helander K, Silbernagel KG, Thomee
Karlsson J. Acute Achilles tendon rupture: a randomized, controlled study
comparing surgical and nonsurgical treatments using validated outcome
measures. Am J Sports Med 38:2186–2193, 2010.
17. Twaddle BC, Poon P. Early motion for Achilles tendon ruptures: is surgery
important? A randomized, prospective study. Am J Sports Med 35:2033–2038,
2007.
18. Willits K, Amendola A, Bryant D, Mohtadi NG, Giffin JR, Fowler P, Kean CO,
Kirkley A. Operative versus nonoperative treatment of acute Achilles tendon
ruptures: a multicenter randomized trial using accelerated functional rehabilitation. J Bone Joint Surg Am 92:2767–2775, 2010.
19. Soroceanu A, Sidhwa F, Aarabi S, Kaufman A, Glazebrook M. Surgical versus
nonsurgical treatment of acute Achilles tendon rupture: a meta-analysis of
randomized trials. J Bone Joint Surg Am 94:2136–2143, 2012.
20. Wallace RG, Heyes GJ, Michael ALR. The non-operative functional management of
patients with a rupture of the tendo Achillis leads to low rates of rerupture. J Bone
Joint Surg Br 93:1362–1366, 2011.
21. Nilsson-Helander K, Thurin A, Karlsson J, Eriksson BI. High incidence of deep
venous thrombosis after Achilles tendon rupture: a prospective study. Knee Surg
Sports Traumatol Arthrosc 17:1234–1238, 2009.