Eur J Anaesthesiol 2019; 36:418–426

ORIGINAL ARTICLE

Functional recovery after knee arthroplasty with

regional analgesia
A systematic review and meta-analysis of randomised controlled
trials
Downloaded from https://journals.lww.com/ejanaesthesiology by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWnYQp/IlQrHD3tjcLwhL8g9bZhhig7WeJvqoknH7nWA1nmF2ZQbxpqlAEIDoJqsn42A== on 05/11/2019

Thomas Osinski, Samir Bekka, Jean-Philippe Regnaux, Dominique Fletcher and Valeria Martinez

BACKGROUND Regional analgesia (RA) has been widely
evaluated for pain relief after total knee arthroplasty (TKA). Its
impact on functional recovery is less well known.
OBJECTIVES To evaluate the functional benefits of RA after
TKA.
DESIGN Systematic review with a random-effects meta-
analysis of randomised controlled trials comparing LRA with
systemic analgesia on function in adults undergoing TKA for
osteoarthritis.
DATABASE SOURCES MEDLINE, EMBASE, LILAC,
Cochrane, CTRD databases.
OUTCOMES Length of stay (LOS) in hospital and early knee
flexion range of motion (ROM), early and long-term knee
function, serious adverse effects.
RESULTS Twenty-three studies (1246 patients) were
included. LOS was significantly shorter for RA than for
systemic analgesia (0.90 days, 95% confidence interval
0.3 to 1.4). Subgroup analyses found that only infiltration
analgesia decreased the LOS. ROM during the first week
was significantly higher for all techniques of RA than for
systemic analgesia (9.238, 95% confidence interval 4.6 to
13.9). No impact of regional analgesia techniques on global
function in the longer term was demonstrated. No difference
in serious adverse effects was found between RA and
systemic analgesia.
CONCLUSION RA techniques compared with systemic
analgesia have a beneficial impact on the LOS and the
ROM achieved in the early postoperative period. Global
function in the longer term after surgery seems unaffected
by peri-operative RA.
TRIAL REGISTRATION CRD42014013995.
Published online 3 April 2019

Introduction
Osteoarthritis, the most common type of arthritis, is
estimated to affect more than 40 million people across
Europe1 and has a lifetime risk of 45% for knee osteoar-
thritis.2 Symptomatic knee osteoarthritis has a prevalence
of 16.7% in subjects over the age of 45 years, and more
than 500 000 knee replacements are performed annually
in the USA.3,4 This number was projected to grow by
673% from 2005 to 2030.4 Total knee arthroplasty (TKA)
is widely performed to improve mobility and quality of
life. TKA is a very painful surgical procedure.5 Effective
analgesia in the immediate postoperative phase is
essential to allow the patient to exercise and regain
mobility, thereby facilitating recovery and decreasing
the length of hospital stay. Several techniques of regional
analgesia (RA) have been developed for postoperative
pain relief. RA has been widely evaluated and several
systematic reviews have shown its benefit for pain relief
after TKA.6–13 However, the impacts on functional
recovery and adverse effects are less well known. We
undertook a systematic review of randomised controlled
trials that compared RA with systemic analgesia in adults
undergoing major knee surgery for osteoarthritis. We

From the Service d’anesthesie, Hôpital Raymond Poincare, Garches, Assistance Publique Hôpitaux de Paris (SB, DF, VM), INSERM, U-987, Hôpital Ambroise Par e, Centre
d’Evaluation et de Traitement de la Douleur (TO, DF, VM), Universit
e Versailles Saint-Quentin, Paris (DF, VM) and D
epartement sciences infirmières et param
edicales Ecole
des Hautes Etudes en sant e publique, Rennes, France (J-PR)
Correspondence to Valeria Martinez, MD, PhD, Service d’anesth
esie, Hôpital Raymond Poincar
e, Garches, Assistance Publique Hôpitaux de Paris, Paris, France
E-mail: valeria.martinez@aphp.fr

0265-0215 Copyright ß 2019 European Society of Anaesthesiology. All rights reserved. DOI:10.1097/EJA.0000000000000983

Copyright © European Society of Anaesthesiology. Unauthorized reproduction of this article is prohibited.


assessed length of hospital stay, range of motion (ROM),
global function and severe adverse effects.
Method
The systematic review of randomised controlled trials
(RCTs) was reported in accordance with the criteria of
the Preferred Reporting Items for Systematic Reviews
and Meta-Analyses (PRISMA) statement and the current
recommendations of the Cochrane Collaboration.14,15 The
study was registered at PROSPERO (CRD42014013995).
We searched the Cochrane Central Register of Con-
trolled Trials, MEDLINE, EMBASE and LILACS, from
the inception of each database to January 2017. The
search equation is available in Supplementary data 1,
http://links.lww.com/EJA/A194. The articles identified
had to be published in English. We also searched the
Cochrane Database of Systematic Reviews and the Data-
base of Abstracts of Reviews of Effects for previous
relevant systematic reviews. We hand-searched the
annual conference proceedings of the American Society
of Anesthesiologists and the European Society of Anaes-
thesiology from June 2013 to June 2017 and searched for
completed trials in ClinicalTrials.gov and the WHO
International Clinical Trials Registry Platform. We iden-
tified randomised trials with the highly sensitive search
strategy of the Cochrane Collaboration.16
We included all RCTs on adults undergoing TKA for
osteoarthritis. For trials on mixed populations, we con-
sidered only those in which more than 50% of the patients
were suffering from osteoarthritis. The interventions of
interest were: epidural analgesia, peripheral nerve block,
local infiltration analgesia (LIA), regardless of the anaes-
thetic drug, volume administered, type of block or type of
local infiltration (peri-articular tissue and/or intra-articu-
lar space). The control group of interests was either a
group with a sham technique with saline administration
or a group with no RA. In either case, systemic analgesia
had to be clearly defined. We excluded trials in which
functional outcomes were not provided. Two authors
(TO and SB) independently screened titles, abstracts
and full texts for the inclusion criteria. Any disagreement
between these two authors was settled by discussion with
a third author (VM) until a consensus was reached.
The primary outcomes assessed were length of stay
(LOS) in hospital in days, and knee flexion ROM in
degrees on day 4 after surgery (if this value was not
provided, we used the ROM recorded for the week
closest to this time point). The secondary outcomes were:
time to first ambulation (in days), knee flexion ROM at
least 1 month after surgery, number of patients achieving
active straight leg raising (SLR), patient satisfaction and
functional score [Western Ontario and McMaster Uni-
versities Osteoarthritis Index (WOMAC), Knee Oxford
Score, knee society score] in the first week and at least 1
month after surgery. We assessed the incidence of severe
Eur J Anaesthesiol 2019; 36:418–426
Copyright © European Society of Anaesthesiology. Unauthorized reproduction of this article is prohibited.
Functional recovery after knee arthroplasty 419
adverse events (SAEs) corresponding to the definition of
Food and Drug Administration (FDA).17 We expected
only small total numbers of SAEs to be reported. We
therefore analysed specific adverse effects corresponding
to the FDA definition and reported in the trials included
in the meta-analysis. Three experts independently pre-
determined which adverse effects frequently reported in
previous studies corresponded to SAEs. Following
discussion to resolve any conflicts/disagreements, we
classified the following adverse effects as SAEs: haemo-
dynamic instability, respiratory depression, venous
thrombosis, wound infection or necrosis (whatever the
site), arrhythmia or bradycardia, syncope, knee infection,
falls, pneumonia, cerebral stroke, myocardial infarction
or death.
Pairs of authors independently reviewed and extracted
data from each study. Disagreements were resolved by
consensus with a third author. We extracted information
about the general characteristics of the study (first author,
number of arms in the study, country, sponsorship), parti-
cipants [age, BMI, American Society of Anesthesiologists’
(ASA) physical status, characteristics of the population,
population randomised and analysed], experimental inter-
vention (local anaesthetic used, administration route, tim-
ing of administration and doses), the use of discharge
criteria, the existence of an enhanced recovery programme
and all functional outcomes evaluated. Dichotomous out-
comes were extracted as the presence or absence of an
effect. For continuous data, we calculated mean and SD. If
necessary, means and measures of dispersion were approx-
imated from figures generated with dedicated software
(http://www.datathief.org/). If not reported, the SDs were
obtained from the confidence intervals (CIs) or P values for
the differences between the means of two groups.16,18 If
medians with ranges were reported, we obtained the mean
and SD as described by Hozo.19 If only means were
reported, we contacted the authors. We followed the
recommendations of the Cochrane group to manage mul-
tiple groups.20 First, we selected only interventions of
interest in our meta-analysis. Second when possible, we
combined groups to create a single pair-wise comparison.
Third, when we needed to include each pair-wise compar-
ison separately, we split the ‘shared’ group into two groups
with smaller sample size, and include two comparisons.
For dichotomous outcomes, both the number of events
and the total number of patients were recorded.
Two of the authors (TO, SB) independently assessed the
methodological quality of the trials with the Cochrane
Risk of Bias tool, with any discrepancies resolved by
consensus.21 We documented the methods used for gen-
erating allocation sequences, allocation concealment, the
blinding of investigators and participants, the blinding of
outcome assessors, and for dealing with incomplete out-
come data. Each item was classified as having a low,
unclear or high risk of bias. The overall risk of bias was
defined as the highest risk of bias documented.
 420 Osinski et al.

Data synthesis and analysis
We calculated risk ratios with 95% CIs for dichotomous
data and mean differences with 95% CI for continuous
data. We expected there to be heterogeneity (because of
the diverse populations included), and we therefore used
the Dersimonian and Lairs random effects meta-analysis
modules. We used the Wells calculator to obtain the
number needed to treat for an additional beneficial
outcome for continuous measures (available at the
Cochrane Musculoskeletal Group editorial office
https://musculoskeletal.cochrane.org/). We assumed a
minimal clinically importance difference (MCID) of 1
day for LOS and 108 for ROM to interpret the clinical
importance of our results.22 We assessed statistical hetero-
geneity by a visual inspection of graphs and by using the I2
statistic, which describes the proportion of variability in
effect estimates due to heterogeneity rather than sampling
error (I2 > 50% indicates substantial heterogeneity).23 The
software RevMan 5.30 (Review Manager (RevMan) [Com-
puter program]. Version 5.3. Copenhagen: The Nordic
Cochrane Centre, The Cochrane Collaboration, 2014)
was used for the meta-analysis. We rated the quality of
evidence for each outcome following the Grades of Rec-
ommendation, Assessment, Development, and Evaluation
(GRADE) Working Group system.24 sufficient evidence
had been accrued. We rated the quality of evidence for
each outcome following the GRADE Working Group
system.24
Results
There were 463 potentially eligible reports. We exam-
ined 65 full-text articles, and we selected 23 studies on a
total of 1246 patients (Fig. 1). All the studies involved
single sites. The median target sample size was 60 (range
16 to 210) patients. The median publication date was
2009 (range 1990 to 2015). Participants were adults with
ASA physical status classes 1 or 2. The characteristics of
the selected trials are reported in Supplemental data
Table 1, http://links.lww.com/EJA/A194.
Seven trials (30%) were classified as being at low risk of
bias, eight (33%) at an unclear risk of bias and eight (33%)

Fig. 1

Studies from other sources:

Studies found in electronic databases
Identification

- Conferences (n = 40)
(n = 636)
- Register of ongoing studies (n = 24)
- References in other studies (n = 63)

Studies found
(n = 463)
Selection

Studies excluded:
- Study design (n = 397)

Selection on full text

(n = 65)
Studies excluded on full text (n = 42):
Eligibility

- Study design (n = 35)

- Inadequate data (n = 5)
- Language (n = 2)
Studies included in
qualitative review
(n = 23)

Studies included in quantitative analysis

Inclusion

- Length of stay (n = 13)

- Range of flexion the first week (n = 16)
- SAE (n = 17)
- First deambulation (n = 3)
- Range of flexion after 1 month (n = 4)
- Function at distance (n = 5)

PRISMA flow diagram showing literature search results.

Eur J Anaesthesiol 2019; 36:418–426

Copyright © European Society of Anaesthesiology. Unauthorized reproduction of this article is prohibited.
 Functional recovery after knee arthroplasty 421

at high risk of bias. The randomisation procedure was Fig. 2

adequately described in 16 trials (66%) and the con-

Blinding of participants and personnel (performance bias)

cealment of treatment allocation was described in
12 trials (50%). Eleven (48%) were double-blind. Four
studies (17%) had an unclear or high risk of incomplete

Blinding of outcome assessment (detection bias)

data outcomes (Fig. 2 details the risk of bias for each

Random sequence generation (selection bias)

trial).

Incomplete outcome data (attrition bias)

Allocation concealment (selection bias)
Primary outcomes
Thirteen studies,27–39 including 872 patients, reported
data for LOS, which was found to be shorter for RA than
for systemic analgesia by 0.9 days (range 0.36 to 1.45,
I2 ¼ 82%). Subgroup analysis by type of RA showed that

Overall risk of biais

this small but significant difference concerned only the
LIA group [1.05 days (range 0.46 to 1.64, I2 ¼ 60%)].
Sixteen studies,29–32,34–36,39–47 including 958 patients,
reported data for range of flexion on day 4 after surgery.
ROM was significantly higher for patients given RA than
for those receiving systemic analgesia, regardless of the
Baranovic 2011 + ? – – – –
type of RA [9.28 (range 4.7 to 13.7, I2 ¼ 95%)]. The effect
estimates for both outcomes met our criteria for MCID Busch 2006 + ? ? ? ? ?
(i.e. 1 day for LOS and 108 for ROM) and were statisti-
Chan 2014 + + – + + –
cally significant (P < 0.05). The number needed to treat
for a beneficial outcome was four (range 2.5 to 10) for Chen 2012 + + + + + +
LOS and two (range 1.5 to 3.7) for ROM. The level of
Essving 2010 + + + + + +
quality of evidence was downgraded twice for both out-
comes due to inconsistency (I2 > 50%) and insufficient Fu 2009 + + + + + +
quality of data, so the quality of evidence is low (Fig. 3).
Fu 2010 + + + + + +
Secondary outcomes Gomez 2010 + + + + + +
Three studies,33,34,48 including 155 patients, reported
data for time to first ambulation. There was no significant Good 2007 ? + + + + ?
global difference in this time between LIA and systemic Kadic 2009 + + ? ? + ?
analgesia [0.29 days (range 0.82 to 0.2), I2 ¼ 73%].
Only one study reported a significant difference in the Kardash 2007 + ? + + + ?
time to ambulation between epidural analgesia and sys- Mahoney 1990 ? ? ? ? + ?
temic analgesia [0.80 (range 0.08 to 1.52), P ¼ 0.03].34
Three studies45,46,49 comparing LIA and systemic anal- McDonald 2016 + + – – + –
gesia, including 261 patients, evaluated the time until the Ng 2001 ? ? + + ? ?
patient could perform SLR after surgery. The time to first
SLR was significantly shorter for LIA than for systemic Niemlainen 2014 + + + + + +
analgesia [20.6 h (range 17.7 to 23.5), I2 ¼ 91%]. Four Ong 2010 – – + –
? ?
studies35,45,46,50 comparing LIA and systemic analgesia,
including 310 patients, reported data for ROM in the Seet 2006 ? ? – – + –
longer term. No difference in median ROM was found Shum 2009 ? ? – + + –
between LIA and systemic analgesia 3 months after
surgery [1.58 (range 1.1 to 4.2)], with a median ROM Singelyn 1998 + ? – – + –
for the control group of 1088 at this time point. Two
Vaishya 2015 + + + + + +
studies41,51 indicated no difference in long-term
WOMAC score between femoral nerve block and sys- Venditolli 2006 + + – – + –
temic analgesia. Three studies41,48,52 showed no differ-
Wang 2002 + ? + + + ?
ence in long-term Knee Society Score (KSS) between
femoral nerve block and systemic analgesia. One study Zhang 2011 ? ? ? + ? ?
showed no difference in long-term KSS and WOMAC
scores between LIA and systemic analgesia.27
Three33,41,48 of five33,41,48,53,54 studies reported higher Risk of bias.
satisfaction in the peripheral nerve block group than in

Eur J Anaesthesiol 2019; 36:418–426

Copyright © European Society of Anaesthesiology. Unauthorized reproduction of this article is prohibited.
 422 Osinski et al.

Fig. 3

(a) RA SA Mean difference Mean difference

Study or subgroup Mean SD Total Mean SD Total Weight IV, Random, 95% CI IV, Random, 95% CI
1.1.1 Epidural analgesia
Mahoney 1990 9.6 1.34 56 11.2 1.94 42 9.2% –1.60 (–2.28, –0.92)
Singelyn 1998 16 4 15 21 3 7 2.5% –5.00 (–8.01, –1.99)
Subtotal (95% Cl) 71 49 11.7% –2.97 (–6.24, 0.30)
Heterogeneity: Tau2 = 4.54; Chi2 = 4.67, df = 1 (P = 0.03); I2 = 79%
Test for overall effect: Z = 1.78 (P = 0.07)

1.1.1 Femoral block

Chan 2014 5.71 1.77 134 5.4 1.4 66 10.0% 0.31 (–0.14, 0.76)
Kardash 2007 6.46 1.33 39 6.1 1.34 20 9.1% 0.36 (–0.36, 1.08)
Ng 2001 9.2 3.19 36 8.8 2.9 12 4.6% 0.40 (–1.54, 2.34)
Seet 2006 6.51 1.94 35 7 2.22 20 7.2% –0.49 (–1.66, 0.68)
Singelyn 1998 17 3 15 21 3 8 3.2% –4.00 (–6.57, –1.43)
Wang 2002 3.5 1.15 15 4.25 1 15 8.9% –0.75 (–1.52, 0.02)
Subtotal (95% Cl) 274 141 42.9% –0.29 (–1.00, 0.43)
Heterogeneity: Tau2 = 0.47; Chi2 = 16.52, df = 5 (P = 0.006); I2 = 70%
Test for overall effect: Z = 0.78 (P = 0.43)

1.1.3 Infiltration analgesia

Busch 2006 5.2 1.34 32 5.1 1.94 32 8.7% 0.10 (–0.72, 0.92)
Essving 2010 4.5 2 24 6.25 2.08 23 7.2% –1.75 (–2.92, –0.58)
Gomez 201 0 5.72 1.34 25 7.32 1.94 25 8.2% –1.60 (–2.52, –0.68)
Ong 2010 5.39 1.51 37 7.25 4.04 17 4.5% –1.86 (–3.84, 0.12)
Vaishya 2015 4.5 0.67 40 5.7 0.64 40 10.5% –1.20 (–1.49, –0.91)
Venditolli 2006 4.8 2.1 22 5.2 2.5 20 6.3% –0.40 (–1.80, 1.00)
Subtotal (95% Cl) 180 157 45.3% –1.051–1.64, –0.46)
Heterogeneity: Tau2 = 0.28; Chi2 = 12.57, df = 5 (P = 0.03); I2 = 60%
Test for overall effect: Z = 3.48 (P = 0.0005)

Total (95% Cl) 525 347 100.0% –0.90 (–1.45, –0.36)

Heterogeneity: Tau2 = 0.72; Chi2 = 72.54, df = 13 (P < 0.00001); I2 = 82%
Test for overall effect: Z = 3.24 (P = 0.001) –10 –5 0 5 10
Test for subgroup differences: Chi2 = 4.37, df = 2 (P = 0.11), I2 = 54.2% Favour (RA) Favour (SA)

Forest plots. Comparison between regional analgesia and systemic analgesia. (a) Length of stay, (b) range of motion, (c) severe adverse effects.

the systemic analgesia group. Two studies compared
patient satisfaction between LIA and systemic analgesia
and reported conflicting results.27,29 Fifteen studies27,30–
34,36,39–41,45,46,49,50,53,55
including 1221 patients, com-
pared the occurrence of SAEs between systemic analge-
sia and RA groups. The incidence of SAEs tended to
be smaller in the RA group but failed to achieve signifi-
cance with a risk ratios of 0.75 (range 0.52 to 1.08,
I2 ¼ 11%) (Fig. 3). The level of quality of evidence
was downgraded for imprecision because the optimal
information size was not reached. The quality of evi-
dence was moderate (Fig. 3). Falls were not reported in
the included studies.
Discussion
The systematic review summarises the available evi-
dence concerning the impact of regional analgesia on
functional recovery after TKA. It shows that all RA
techniques provide a similar transient improvement in
knee ROM in the first week over that observed in
patients receiving systemic analgesia. However, none
of the regional analgesia techniques influenced global
postoperative function or the recovery process in the long
term after surgery.
Overall, the randomised clinical trials reported that all RA
techniques slightly increased ROM in the early postoper-
ative period but had no impact on long-term function. It
has been suggested that peripheral nerve blocks provide
better pain control, particularly for pain on movement,
than systemic analgesia alone.7,10 Better pain management
probably improves mobility, as indicated by the greater
ROM recorded in the early postoperative period. How-
ever, this benefit does not seem to be systematically
sustained beyond the period covered by the peripheral
nerve block, with conflicting results reported for functional
outcomes after discharge35,45,46,50 and no impact on long-
term functional recovery.56 Although, the decrease in LOS
was very small, it reached the minimal clinical importance
difference and was similar to that reported in other sys-
tematic reviews.11,12 However, this decrease in LOS is
much lower than the 2 to 3 days reported in cohort studies
including an enhanced recovery programme.55,57,58 It
could be explained by the lack of standardised discharge
criteria in more than two-third of trials included.

Eur J Anaesthesiol 2019; 36:418–426

Copyright © European Society of Anaesthesiology. Unauthorized reproduction of this article is prohibited.
 Functional recovery after knee arthroplasty 423

Fig. 3

(b) RA SA Mean difference Mean difference

Study or subgroup Mean SD Total Mean SD Total Weight IV, Random, 95% CI IV, Random, 95% CI
1.1.1 Epidural analgesia
Mahoney 1990 84.3 14.3 56 66.2 6.8 42 6.7% 18.10 (13.83, 22.37)
Singelyn 1998 79 12 15 62 18 7 4.1 % 17.00 (2.35, 31.65)
Subtotal (95% Cl) 71 49 10.8% 18.01 (13.91, 22.12)
Heterogeneity: Tau2 = 0.00; Chi2 = 0.02, df = 1 (P = 0.89); I2 = 0%
Test for overall effect: Z = 8.61 (P < 0.0001)

1.1.2 Femoral block

Chan 2014 85.59 1 0.92 41 85.24 9.88 41 6.7% 0.35 (–4.16, 4.86)
Good 2007 67 16 19 62 10 19 5.7% 5.00 (–3.48, 13.48)
Kadic 2009 87.5 7.41 16 75 11.12 16 6.2% 12.50 (5.95, 19.05)
Kardash 2007 61.03 19.88 39 55 17.89 20 5.3% 6.03 (–3.99, 16.05)
Singelyn 1998 81 7 15 62 18 8 4.5% 19.00 (6.03, 31.97)
Wang 2002 76.7 9.5 15 71 10.2 15 6.1 % 5.70 (–1.35, 12.75)
Subtotal (95% Cl) 145 119 34.5% 7.04 (11.92, 12.16)
Heterogeneity: Tau2 = 24.26; Chi2 = 13.59, df = 5 (P = 0.02); I2 = 63%
Test for overall effect: Z = 2.70 (P = 0.007)

1.1.3 Infiltration
Essving 2010 80 17.56 22 64.75 14.11 22 5.5% 15.25 (5.84, 24.66)
Fu 2009 63.9 8.9 40 57.2 10 40 6.8% 6.70 (2.55, 10.85)
Fu 2010 64.4 8.9 50 56.9 10 50 6.8% 7.50 (3.79, 11.21)
Gomez 201 0 102.8 8.9 25 93.5 10 25 6.5% 9.30 (4.05, 14.55)
Niemlainen 2014 75.8 22 22 76 18 22 4.8% –0.20 (–12.08, 11.68)
Ong 2010 76.42 18.52 37 75 28.43 17 4.0% 1.42 (–13.35, 16.19)
Vaishya 2015 90.125 10 40 59.25 10 40 6.7% 30.88 (26.49, 35.26)
Venditolli 2006 90 8.9 22 94 10 20 6.4% –4.00 (–9.75, 1.75)
Zhang 2011 93.15 4.76 54 87.5 5.7 26 7.0% 5.65 (3.12, 8.18)
Subtotal (95% Cl) 312 262 54.6% 8.48 (1.73, 15.23)
Heterogeneity: Tau2 = 92.75; Chi2 = 128.89, df = 8 (P = 0.00001); I2 = 94%
Test for overall effect: Z = 2.46 (P = 0.01)

Total (95% Cl) 528 430 100.0% 9.20 (4.70, 13.69)

Heterogeneity: Tau2 = 73.25; Chi2 = 167.01, df = 16 (P < 0.00001); I2 = 90%
–20 –10 0 10 20
Test for overall effect: Z = 4.01 (P = 0.0001)
Favour (SA) Favour (RA)
Test for subgroup differences: Chi2 = 12.65, df = 2 (P = 0.002), I2 = 84.2%

(continued)

Our systematic review was based on numerous small
single-site trials. There was therefore a risk of overesti-
mation of treatment effects and underreporting of rele-
vant severe adverse effects. We observed a significant
imbalance in terms of the amount of evidence available
for individual types of intervention, with an underrepre-
sentation of epidural analgesia. In addition, the included
trials reported many different nonstandardised end-
points, precluding the comparison of trial results and
the pooling of data from independent studies. The
retrieved trials dealt with highly diverse drug regimens
(volume, dose, timing and molecule administered), and
the limited functional outcome data reported made it
impossible to carry out more detailed subgroup analyses.
Our study has several strengths. First, we conducted a
rigorous and extensive literature search, including registry
searches, contact with the authors of the studies considered
and searches of the abstract proceedings for the two main
congresses in the field for up to 5 years. Second, our meta-
analysis provides a complete overview of current scientific
knowledge concerning the contribution of regional anal-
gesic techniques to functional recovery after knee
arthroplasty compared with systemic analgesia. This anal-
ysis could therefore be used to develop a rational basis for
future research. Based both on the frequency of outcomes
reported in this systematic review and on the unmet needs
for which future clinical research is required, we suggest
the following evaluation of function after TKA. Global
function should be evaluated before surgery and in the
longer term after surgery, by calculating the WOMAC
score. Maximal flexion of the knee and SLR tests could
be used as intermediate functional outcomes in the imme-
diate postoperative period. The lack of standardised crite-
ria of discharge is an issue and precludes the evaluation of
the LOS based on a combination of criteria used in the
studies included in this review and cohort studies regard-
ing the addition of an enhanced recovery program for
TKA55,57 could be proposed. We suggest that the following
criteria should be considered; to standardise LOS endpoint
in trials independently mobile with sticks or elbow
crutches, ability to climb or descend a single flight of stairs

Eur J Anaesthesiol 2019; 36:418–426

Copyright © European Society of Anaesthesiology. Unauthorized reproduction of this article is prohibited.
 424 Osinski et al.

Fig. 3

(c) RA SA Risk ratio Risk ratio

Study or subgroup Events Total Events Total Weight M-H, Random, 95% CI M-H, Random, 95% CI
1.3.1 Epidural
Mahoney 1990 5 56 1 42 2.8% 3.75 (0.45, 30.91)
Singelyn 1998 1 15 0 8 1.3% 1.69 (0.08, 37.26)
Subtotal (95% Cl) 71 50 4.2% 2.91 (0.51, 16.63)
Total events 6 1
Heterogeneity: Tau2 = 0.00; Chi2 = 0.18, df = 1 (P = 0.67); I2 = 0%
Test for overall effect: Z = 1.20 (P = 0.23)

1.3.2 Femoral block

Baranovic 2011 6 35 23 36 16.6% 0.27 (0.12, 0.58)
Good 2007 2 19 3 19 4.4% 0.67 (0.13, 3.55)
Kadic 2009 0 27 0 26 Not estimable
Kardash 2007 13 39 4 20 11.3% 1.67 (0.62, 4.45)
Seet 2006 4 35 3 20 6.2% 0.76 (0.19, 3.07)
Singelyn 1998 0 15 1 7 1.4% 0.17 (0.01, 3.65)
Subtotal (95% Cl) 170 128 39.8% 0.60 (0.25, 1.46)
Total events 25 34
Heterogeneity: Tau2 = 0.52; Chi2 = 9.08, df = 4 (P = 0.06); I2 = 56%
Test for overall effect: Z = 1.12 (P = 0.26)

1.3.3 Infiltration analgesia

Busch 2006 5 32 5 32 8.8% 1.00 (0.32, 3.12)
Chen 2012 1 40 2 40 2.3% 0.50 (0 .05, 5.30)
Fu 2009 9 40 10 40 16.1 % 0.90 (0.41,1.98)
Fu 2010 11 50 12 50 18.4% 0.92 (0.45, 1.88)
McDonald 2016 0 113 0 109 Not estimable
Ong 2010 0 37 0 17 Not estimable
Vaishya 2015 0 40 1 40 1.3% 0.33 (0.01,7.95)
Venditolli 2006 1 22 3 20 2.7% 0.30 (0.03, 2.68)
Zhang 2011 5 54 3 26 6.5% 0.80 (0.21, 3.10)
Subtotal (95% Cl) 428 374 56.0% 0.84 (0.54,1.29)
Total events 32 36
Heterogeneity: Tau2 = 0.00; Chi2 = 1.55, df = 6 (P = 0.96); I2 = 0%
Test for overall effect: Z = 0.80 (P = 0.42)

Total (95% Cl) 669 552 100.0% 0.75 (0.52, 1.08)

Total events 63 71
Heterogeneity: Tau2 = 0.05; Chi2 = 14,63, df = 13 (P < 0.33); I2 = 11%
0.01 0.1 1 10 100
Test for overall effect: Z = 1.54 (P = 0.12)
Favour (RA) Favour (SA)
Test for subgroup differences: Chi2 = 2.49, df = 2 (P = 0.29), I2 = 19.7%

(continued)

safely, 908 knee flexion, satisfactory analgesia (maximum regional analgesia techniques on global function in the
30 mm on a visual analogue scale) and good quadriceps longer term has been demonstrated.
strength (able to stand up from a sitting position and to
maintain knee extension while weight-bearing).55 The Acknowledgements relating to this article
achievement of discharge criteria should be the gold Assistance with the study: none.
standard in all studies evaluating analgesia approaches
Financial support and sponsorship: none.
in postoperative care.
Conflicts of interest: none.
Presentation: none.
Conclusion
There is an urgent need to standardise functional evalu- References
ation after knee arthroplasty. Nevertheless, we can con- 1 World Health Organization. The burden of musculoskeletal conditions at
firm with confidence that all RA techniques are superior the start of the new millennium. Geneva: World Health Organization. World
Health Organ Tech Rep Ser 2003; 919:i–x; 1–218.
to systemic analgesia in terms of the ROM achieved in 2 Murphy L, Schwartz TA, Helmick CG, et al. Lifetime risk of symptomatic
the early postoperative period. However, no impact of knee osteoarthritis. Arthritis Rheum 2008; 59:1207–1213.

Eur J Anaesthesiol 2019; 36:418–426

Copyright © European Society of Anaesthesiology. Unauthorized reproduction of this article is prohibited.

3 Kurtz S, Ong K, Lau E, et al. Projections of primary and revision hip and knee
arthroplasty in the United States from 2005 to 2030. J Bone Joint Surg Am
2007; 89:780–785.
4 Lawrence RC, Felson DT, Helmick CG, et al. Estimates of the prevalence of
arthritis and other rheumatic conditions in the United States. Part II. Arthritis
Rheum 2008; 58:26–35.
5 Gerbershagen HJ, Aduckathil S, van Wijck AJ, et al. Pain intensity on the
first day after surgery: a prospective cohort study comparing 179 surgical
procedures. Anesthesiology 2013; 118:934–944.
6 Albrecht E, Guyen O, Jacot-Guillarmod A, et al. The analgesic efficacy of
local infiltration analgesia vs femoral nerve block after total knee
arthroplasty: a systematic review and meta-analysis. Br J Anaesth 2016;
116:597–609.
7 Andersen LO, Kehlet H. Analgesic efficacy of local infiltration analgesia in
hip and knee arthroplasty: a systematic review. Br J Anaesth 2014;
113:360–374.
8 Johnson RL, Kopp SL, Hebl JR, et al. Falls and major orthopaedic surgery
with peripheral nerve blockade: a systematic review and meta-analysis. Br J
Anaesth 2013; 110:518–528.
9 Li D, Yang Z, Xie X, et al. Adductor canal block provides better performance
after total knee arthroplasty compared with femoral nerve block: a
systematic review and meta-analysis. Int Orthop 2016; 40:925–933.
10 Chan EY, Fransen M, Parker DA, et al. Femoral nerve blocks for acute
postoperative pain after knee replacement surgery. Cochrane Database
Syst Rev 2014; 5:CD009941.
11 Paul JE, Arya A, Hurlburt L, et al. Femoral nerve block improves analgesia
outcomes after total knee arthroplasty: a meta-analysis of randomized
controlled trials. Anesthesiology 2010; 113:1144–1162.
12 Marques EM, Jones HE, Elvers KT, et al. Local anaesthetic infiltration for
peri-operative pain control in total hip and knee replacement: systematic
review and meta-analyses of short- and long-term effectiveness. BMC
Musculoskelet Disord 2014; 15:220.
13 Terkawi AS, Mavridis D, Sessler DI, et al. Pain management modalities after
total knee arthroplasty: a network meta-analysis of 170 randomized
controlled trials. Anesthesiology 2017; 126:923–937.
14 Higgins J, Green S. Chapter 4: Guide to the contents of a Cochrane
protocol and review.Cochrane handbook for systematic reviews of
interventions version 5.1.0. The Cochrane Collaboration; 2011; Available
from: www.cochrane-handbook.org. [Updated March 2011].
15 Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting
systematic reviews and meta-analyses of studies that evaluate healthcare
interventions: explanation and elaboration. BMJ 2009; 339:b2700.
16 Lefebvre C, Manheimer E, Glanville J. Chapter 6.4: searching for
studies.Cochrane handbook for systematic reviews of interventions
version 5.1.0. The Cochrane Collaboration; 2011; Available from:
www.cochrane-handbook.org. [Updated March 2011]; 2008.
17 FDA. The FDA safety information and adverse event reporting program:
reporting serious problems to FDA: what is a serious adverse event? 2010.
https://www.fda.gov/Safety/MedWatch/HowToReport/ucm053087.htm
[Updated 2 January 2016]
18 Higgins JP, White IR, Wood AM. Imputation methods for missing outcome
data in meta-analysis of clinical trials. Clin Trials 2008; 5:225–239.
19 Hozo SP, Djulbegovic B, Hozo I. Estimating the mean and variance from the
median, range, and the size of a sample. BMC Med Res Methodol 2005; 5:13.
20 Chapter 16: special topics in statistics.Higgins JPT, Spiegelhalter DJ,
Altman DG, editors. et al. Cochrane handbook for systematic reviews of
interventions version 5.1.0. The Cochrane Collaboration; 2011; Available
from: www.handbook.cochrane.org. [Updated March 2011].
21 Higgins JP, Altman DG, Gotzsche PC, et al. The Cochrane Collaboration’s
tool for assessing risk of bias in randomised trials. BMJ 2011; 343:d5928.
22 Man-Son-Hing M, Laupacis A, O’Rourke K, et al. Determination of the
clinical importance of study results. J Gen Intern Med 2002; 17:469–476.
23 Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis.
Stat Med 2002; 21:1539–1558.
24 Balshem H, Helfand M, Schunemann HJ, et al. GRADE guidelines: 3.
Rating the quality of evidence. J Clin Epidemiol 2011; 64:401–406.
25 Wetterslev J, Thorlund K, Brok J, et al. Estimating required information size
by quantifying diversity in random-effects model meta-analyses. BMC Med
Res Methodol 2009; 9:86.
26 Pogue JM, Yusuf S. Cumulating evidence from randomized trials: utilizing
sequential monitoring boundaries for cumulative meta-analysis. Control
Clin Trials 1997; 18:580–593.
27 Busch CA, Shore BJ, Bhandari R, et al. Efficacy of periarticular multimodal
drug injection in total knee arthroplasty. A randomized trial. J Bone Joint
Surg Am 2006; 88:959–963.
28 Chan MH, Chen WH, Tung YW, et al. Single-injection femoral nerve block
lacks preemptive effect on postoperative pain and morphine consumption
in total knee arthroplasty. Acta Anaesthesiol Taiwan 2012; 50:54–58.
Eur J Anaesthesiol 2019; 36:418–426
Copyright © European Society of Anaesthesiology. Unauthorized reproduction of this article is prohibited.
Functional recovery after knee arthroplasty 425
29 Essving P, Axelsson K, Kjellberg J, et al. Reduced morphine consumption
and pain intensity with local infiltration analgesia (LIA) following total knee
arthroplasty. Acta Orthopaedica 2010; 81:354–360.
30 Kardash K, Hickey D, Tessler MJ, et al. Obturator versus femoral nerve
block for analgesia after total knee arthroplasty. Anesth Analges 2007;
105:853–858.
31 Mahoney OM, Noble PC, Davidson J, et al. The effect of continuous
epidural analgesia on postoperative pain, rehabilitation, and duration of
hospitalization in total knee arthroplasty. Clin Orthop Rel Res 1990;
260:30–37.
32 Ong JC, Chin PL, Lin CP, et al. Continuous infiltration of local anaesthetic
following total knee arthroplasty. J Orthop Surg 2010; 18:203–207.
33 Seet E, Leong WL, Yeo AS, et al. Effectiveness of 3-in-1 continuous
femoral block of differing concentrations compared to patient controlled
intravenous morphine for post total knee arthroplasty analgesia and knee
rehabilitation. Anaesth Intensive Care 2006; 34:25–30.
34 Singelyn FJ, Deyaert M, Joris D, et al. Effects of intravenous patient-
controlled analgesia with morphine, continuous epidural analgesia, and
continuous three-in-one block on postoperative pain and knee rehabilitation
after unilateral total knee arthroplasty. Anesth Analg 1998; 87:88–92.
35 Gomez-Cardero P, Rodriguez-Merchan EC. Postoperative analgesia in
TKA: ropivacaine continuous intraarticular infusion. Clin Orthop Relat Res
2010; 468:1242–1247.
36 Vendittoli PA, Makinen P, Drolet P, et al. A multimodal analgesia protocol for
total knee arthroplasty. A randomized, controlled study. J Bone Joint Surg
Am 2006; 88:282–289.
37 Ng F-Y, Chiu K-Y, Yan CH, et al. Continuous femoral nerve block versus
patient-controlled analgesia following total knee arthroplasty. J Orthop
Surg 2012; 20:23–26.
38 Wang F, Zhou Y, Sun J, et al. Influences of continuous femoral nerve block
on knee function and quality of life in patients following total knee
arthroplasty. Int J Clin Exp Med 2015; 8:19120–19125.
39 Vaishya R, Wani AM, Vijay V. Local infiltration analgesia reduces pain and
hospital stay after primary TKA: randomized controlled double blind trial.
Acta Orthop Belg 2015; 81:720–729.
40 Good RP, Snedden MH, Schieber FC, et al. Effects of a preoperative
femoral nerve block on pain management and rehabilitation after total knee
arthroplasty. Am J Orthop 2007; 36:554–557.
41 Kadic L, Boonstra MC, De Waal Malefijt MC, et al. Continuous femoral
nerve block after total knee arthroplasty? Acta Anaesthesiol Scand 2009;
53:914–920.
42 Wang HJ, Zhang DZ, Li SZ. Comparing the analgesic efficacy of
continuous femoral nerve blockade and continuous intravenous analgesia
after total knee arthroplasty. Zhonghua Yi Xue Za Zhi 2010; 90:2360–
2362.
43 Zaric D, Boysen K, Christiansen C, et al. A comparison of epidural
analgesia with combined continuous femoral-sciatic nerve blocks after total
knee replacement. Anesth Analg 2006; 102:1240–1246.
44 Niemelainen M, Kalliovalkama J, Aho AJ, et al. Single periarticular local
infiltration analgesia reduces opiate consumption until 48 h after total knee
arthroplasty. A randomized placebo-controlled trial involving 56 patients.
Acta Orthop 2014; 85:614–619.
45 Fu P, Wu Y, Wu H, et al. Efficacy of intra-articular cocktail analgesic
injection in total knee arthroplasty – a randomized controlled trial. Knee
2009; 16:280–284.
46 Fu PL, Xiao J, Zhu YL, et al. Efficacy of a multimodal analgesia protocol in
total knee arthroplasty: a randomized, controlled trial. J Int Med Res 2010;
38:1404–1412.
47 Wang H, Boctor B, Verner J. The effect of single-injection femoral nerve
block on rehabilitation and length of hospital stay after total knee
replacement. Reg Anesth Pain Med 2002; 27:139–144.
48 Shum CF, Lo NN, Yeo SJ, et al. Continuous femoral nerve block in total
knee arthroplasty: immediate and two-year outcomes. J Arthroplasty 2009;
24:204–209.
49 Chen Y, Zhang Y, Zhu YL, et al. Efficacy and safety of an intra-operative
intra-articular magnesium/ropivacaine injection for pain control following
total knee arthroplasty. J Int Med Res 2012; 40:2032–2040.
50 Zhang S, Wang F, Lu ZD, et al. Effect of single-injection versus continuous
local infiltration analgesia after total knee arthroplasty: a randomized,
double-blind, placebo-controlled study. J Int Med Res 2011; 39:1369–
1380.
51 Chan E-Y, Teo Y-H, Assam PN, et al. Functional discharge readiness
and mobility following total knee arthroplasty for osteoarthritis: a
comparison of analgesic techniques. Arthritis Care Res 2014; 66:
1688–1694.
52 Tugay N, Saricaoglu F, Satilmis T, et al. Effects on the independence level in
functional activities in the early postoperative period in patients with total
knee arthroplasty. Neurosciences 2006; 11:175–179.
 426 Osinski et al.
53 Baranovi S, Maldini B, Milosevi M, et al. Peripheral regional analgesia with
femoral catheter versus intravenous patient controlled analgesia after total
knee arthroplasty: a prospective randomized study. Coll Antropol 2011;
35:1209–1214.
54 Parvataneni HK, Shah VP, Howard H, et al. Controlling pain after total hip and
knee arthroplasty using a multimodal protocol with local periarticular
injections: a prospective randomized study. J Arthroplasty 2007; 22:33–38.
55 McDonald DA, Siegmeth R, Deakin AH, et al. An enhanced recovery
programme for primary total knee arthroplasty in the United Kingdom –
follow up at one year. Knee 2012; 19:525–529.
Eur J Anaesthesiol 2019; 36:418–426
Copyright © European Society of Anaesthesiology. Unauthorized reproduction of this article is prohibited.
56 Atchabahian A, Schwartz G, Hall CB, et al. Regional analgesia for
improvement of long-term functional outcome after elective large joint
replacement. Cochrane Database Syst Rev 2015; 8:CD010278.
57 Christelis N, Wallace S, Sage CE, et al. An enhanced recovery after
surgery program for hip and knee arthroplasty. Med J Aust 2015;
202:363–368.
58 Khan SK, Malviya A, Muller SD, et al. Reduced short-term complications
and mortality following enhanced recovery primary hip and knee
arthroplasty: results from 6,000 consecutive procedures. Acta Orthop
2014; 85:26–31.