ARTICLE IN PRESS

doi:10.1510/icvts.2009.221127

Interactive CardioVascular and Thoracic Surgery 10 (2010) 92–96

www.icvts.org

Best evidence topic - Thoracic non-oncologic

In patients undergoing thoracic surgery is paravertebral block
as effective as epidural analgesia for pain management?
Marco Scarci*, Abhishek Joshi, Rizwan Attia
Department of Thoracic Surgery, Guy’s and St Thomas Hospital, Great Maze Pond, London, SE1 9RT, UK

Received 5 September 2009; received in revised form 5 October 2009; accepted 5 October 2009

Abstract

A best evidence topic in cardiac surgery was written according to a structured protocol. The question addressed was: in patients
undergoing thoracic surgery is paravertebral block (PVB) as effective as epidural analgesia for pain management? Altogether )184 papers
were found using the reported search, seven of which represented the best evidence to answer the clinical question. All studies agreed
that PVB is at least as effective as epidural analgesia for pain control post-thoracotomy. In one paper, the visual analogue pain score (VAS)
at rest and on cough was significantly lower in the paravertebral group (Ps0.02 and 0.0001, respectively). Pulmonary function, as assessed
by peak expiratory flow rate (PEFR), was significantly better preserved in the paravertebral group. The lowest PEFR as a fraction of
preoperative control was 0.73 in the paravertebral group in contrast with 0.54 in the epidural group (P-0.004). Oximetric recordings were
better in the paravertebral group (96%) compared to the epidural group (95%) (Ps0.0001). Another article reported that statistically
significant differences (forced vital capacity 46.8% for PVB and 39.3% for epidural group P-0.05; forced expiratory volume in 1 s (FEV1)
48.4% in PVB group and 35.9% in epidural group, P-0.05) were reached in day 2 and continued until day 3. Plasma concentrations of
cortisol, as marker of postoperative stress, increased markedly in both groups, but the increment was statistically different in favour of
the paravertebral group (Ps0.003). Epidural block was associated with frequent side-effects wurinary retention (42%), nausea (22%), itching
(22%) and hypotension (3%) and, rarely, respiratory depression (0.07%)x. Additionally, it prolonged operative time and was associated with
technical failure or displacement (8%). Epidurals were also related to a higher complication rate (atelectasisypneumonia) compared to the
PVB (2 vs. 0). PVB was found to be of equal efficacy to epidural anaesthesia, but with a favourable side effect profile, and lower
complication rate. The reduced rate of complication was most marked for pulmonary complications and is accompanied by quicker return
to normal pulmonary function. We conclude intercostal analgesia, in the form of PVB, can be at least as effective as epidural analgesia.
䊚 2010 Published by European Association for Cardio-Thoracic Surgery. All rights reserved.

Keywords: Intercostal analgesia; Epidural analgesia

1. Introduction 4. Search strategy

A best evidence topic was constructed according to a Medline 1950 to August 2009 using OVID interface
structured protocol. This is fully described in the ICVTS w1x. wpostthoracotomy.mp OR thoracotomy.mpx AND wparaver-
tebral.mpx AND wepidural.mpx.
2. Three-part question
5. Search outcome
In wpatients undergoing thoracic surgeryx is wparavertebral
blockx as effective as wepidural analgesiax for wpain One hundred and eighty-four papers were found using the
managementx? reported search. From these seven papers were identified,
that provided the best evidence to answer the question.
3. Clinical scenario These are presented in Table 1.

You are in theatre about to perform a procedure requiring 6. Results

a thoracotomy. The anaesthetist wants to insert an epidural
catheter for analgesia but you are worried about possible Joshi et al. w2x conducted a systematic review of data
complications and would prefer a paravertebral block between 1966 and May 2004. Seven articles specifically
(PVB). He contends that epidural remains the gold standard dealt with paravertebral analgesia techniques vs. epidural
to compare other pain relief modalities. You resolve to analgesia. Scatter-plot analysis of pain scores among stud-
check the literature yourself. ies demonstrated that, overall, continuous PVB was as
effective as thoracic epidural analgesia with local anaes-
*Corresponding author. Tel.: q44-751-5542899. thetic (both with or without opioid) at day 1, whereas
E-mail address: [email protected] (M. Scarci). when opiod was added to local anaesthetic in either or
䊚 2010 Published by European Association for Cardio-Thoracic Surgery
 ARTICLE IN PRESS
M. Scarci et al. / Interactive CardioVascular and Thoracic Surgery 10 (2010) 92–96 93

Table 1
Best evidence papers

Author, date and country Patient group Outcomes Key results Comments
Study type
(level of evidence)

Joshi et al., (2008), 64 randomised trials Primary outcome Continuous PVB was as Thoracic epidural analgesia
Anaesthe Analg, were included in the measures were effective as thoracic epidural with LA plus opioid or
USA, w2x review postoperative analgesia with local anaesthetic continuous PVB
pain scores, analgesic but was associated with a with LA can be recommended
Systematic review 171 patients in 7 trials use and complication reduced incidence of
(level 1a) specifically looked to rate hypotension
evaluate paravertebral
analgesia compared to PVB reduced the
epidural analgesia incidence of pulmonary
complications

Davies et al., (2006), 520 patients from 10 Pain scores There was no significant Paravertebral and epidural
Br J Anaesth, trials difference between paravertebral analgesia provide comparable
Australia, w3x and epidural groups for pain pain relief after thoracic
scores at 4–8, 24 or 48 h, surgery, but paravertebral
Meta-analysis of weighted mean difference analgesia has a better side-effect
randomised trials 0.37 (95% CI: –0.5, 121), 0.05 profile and is associated with
(level 1a) (–0.6, 0.7), –0.04 (–0.4, 0.3), a reduction in pulmonary
respectively complications

Pulmonary Pulmonary complications

complications occurred less often with
paravertebral analgesia, OR
0.36 (0.14, 0.92)

Complications Urinary retention, OR 0.23

(0.10, 0.51), nausea and
vomiting, OR 0.47 (0.24, 0.53),
and hypotension, OR 0.23
(0.11, 0.48), were less common
with paravertebral analgesia.
Rates of failed block were
lower in paravertebral analgesia
group, OR 0.28 (0.2, 0.6)

Detterbeck, (2005), 619 patients from 17 Pain control It was as good if not better in PVB is as good as epidural
Ann Thorac Surg, trials the paravertebral group analgesia if not better, with
USA, w4x fewer side effects
Respiratory function All studies except one found a
Systematic review preservation better preservation of the FEV1
(level 1a)
Complications Epidural block was associated
with frequent side-effects
wurinary retention (42%), nausea

Best Evidence
(22%), itching (22%) and

Topic
hypotension (3%) and rarely
respiratory depression (0.07%)x.
Additionally, it prolonged
operative time and was
associated with technical failure
or displacement (8%)

Richardson et al., 100 adult patients Pain control The VAS at rest and on cough PVB is as effective as epidural
(1999), Br J was significantly lower in the and better in terms of
Anaesth, UK, w5x paravertebral group (Ps0.02 and pulmonary function,
0.0001, respectively) neuroendocrine stress response
Prospective and side effects
randomized trial Pulmonary function Pulmonary function, as assessed
(level 1b) by PEFR, was significantly
better preserved in the
paravertebral group. The lowest
PEFR as a fraction of
preoperative control was 0.73
in the paravertebral group in
contrast with 0.54 in the
epidural group (P-0.004)

(Continued on next page)

 ARTICLE IN PRESS
94 M. Scarci et al. / Interactive CardioVascular and Thoracic Surgery 10 (2010) 92–96

Table 1 (Continued)
Author, date and country Patient group Outcomes Key results Comments
Study type
(level of evidence)

Oximetry Oximetric recording were better

in the paravertebral group (96%)
compared to the epidural group
(95%) (Ps0.0001)

Plasma cortisol Plasma concentrations of

concentration cortisol increased markedly in
both groups, but the increment
was statistically different in
favour of the paravertebral
group (Ps0.003)

Hypotension Seven patients in the epidural group

became hypotensive compared
to 0 in the paravertebral group

Length of stay Mean hospital stay was 6.7

days (range 4–11) for the
paravertebral group and 6.7
(range 3–16) (PsNS) for the
epidural group

Kaiser et al., (1998), 30 patients were Pain control PVB was superior to epidural Extrapleural intercostal
Ann Thorac Surg, evaluated for extra- in the first 24 h postoperatively analgesia is a valuable alternate
Switzerland, w6x pleural vs. epidural to epidural analgesia
analgesia post- Recovery in Statistically significant
Prospective thoracotomy ventilatory function differences (FVC 46.8% for
randomized trial PVB and 39.3% for epidural
(level 1b) group P-0.05; FEV1 48.4% in
PVB group and 35.9% in
epidural group, P-0.05) were
reached in day 2 and continued
until day 3

Pulmonary Epidurals were related to a

complications higher complication rate
(atelectasisypneumonia)
compared to the PVB (2 vs. 0)

Casati et al., (2006), 42 patients Area under the curve The AUCVAS over time was PVB is as effective as epidural
Eur J Anaesth, Italy, undergoing lung of the VAS during 192 (60–444) cm hy1 in the blockade in controlling
w7x resection coughing (AUCVAS). epidural group and 228 post-thoracotomy pain, but is
The aim was to (72–456) cm hy1 in the associated with less
Prospective detect a 96-cm hy1 PVB group (Ps0.29) haemodynamic effects
randomized trial difference in the
(level 1b) AUCVAS during the
48 h observation
period with an
expected S.D. ranging
between 60 and
85 cm hy1 accepting a
two-tailed a-error of
5% and a b-error of
5%, for a final power
of 95%

Rescue morphine It was required in four patients

analgesia of epidural group (19%) and
five patients of PVB group
(23%) (Ps0.99)

Reduction in The PaO2 yFiO2 ratio reduced

PaO2 yFiO2 significantly from baseline
values in both groups without
between-group differences

Reduction in The median percentage

systolic blood reduction of systolic arterial
pressure pressure from baseline was –9
(0 to –9)% in group PVB and
(Continued on next page)
 ARTICLE IN PRESS
M. Scarci et al. / Interactive CardioVascular and Thoracic Surgery 10 (2010) 92–96 95

Table 1 (Continued)
Author, date and country Patient group Outcomes Key results Comments
Study type
(level of evidence)

–17 (0 to –38)% in epidural

group (Ps0.02); while clinically
relevant hypotension (systolic
arterial pressure decrease -30%
of baseline) was observed in
four patients of epidural group
only (19%) (Ps0.04)

CI, confidence interval; OR, odds ratio; PVB, paravertebral block; FEV1 , forced expiratory volume in 1 s; VAS, visual analogue pain score; PEFR, peak expiratory
flow rate; AUCVAS, area under the curve of the VAS during coughing; S.D., standard deviation; LA, local anesthetics.

both groups, PVB tended to be associated with higher pain Kaiser et al. w6x looked at pain control, recovery of
scores in the early postoperative period. There was no ventilatory function and pulmonary complications in 30
statistically significant difference between the two modal- thoracic surgery patients undergoing lung resection. PVB
ities. PVB was associated with a reduced incidence of was superior to epidural in the first 24 h postoperatively.
hypotension and pulmonary complications. Statistically significant differences (FVC 46.8% for PVB and
Davies et al. w3x conducted a meta-analysis of 10 random- 39.3% for epidural group P-0.05; FEV1 48.4% in PVB group
ized control trials (RCTs) between 1989 and 2005 including and 35.9% in epidural group, P-0.05) were reached in day
520 patients. There was no significant difference in parav- 2 and continued until day 3. Epidurals were related to a
ertebral and epidural groups for pain scores w95% confi- higher complication rate (atelectasisypneumonia) com-
dence interval (CI): –0.5, 121x, P-0.05 at 8, 24, 48 h. PVB pared to the PVB (2 vs. 0).
was associated with fewer pulmonary complications wodds Casati et al. w7x conducted a prospective, randomized,
ratio (OR) 0.36 w0.14, 0.92xx, urinary retention (OR 0.23 blinded study comparing the efficacy of the PVB vs. epi-
w0.10, 0.51x), nausea and vomiting OR 0.47 (w0.24, 0.53x) dural analgesia in 42 patients undergoing lung resection.
and hypotension, (OR 0.28 w0.2, 0.6x). Thus, PVB provided The main outcome variable was expressed as the area
equally effective analgesia to epidural but with a better under the curve of the VAS during coughing (AUCVAS). The
side-effect profile. aim was to detect a 96-cm hy1 difference in the AUCVAS
Detterbeck w4x reviewed 619 patients from 17 trials. Pain during the 48 h observation period with an expected stan-
control was as good if not better in the paravertebral dard deviation (S.D.) ranging between 60 and 85 cm hy1
group. All studies except one found a better preservation accepting a two-tailed a-error of 5% and a b-error of 5%,
of the forced expiratory volume in 1 s (FEV1). Epidural for a final power of 95%. With these assumptions, the
block was associated with frequent side-effects (urinary AUCVAS over time was 192 (60–444) cm hy1 in the epidural
retention (42%), nausea (22%), itching (22%) and hypoten- group and 228 (72–456) cm hy1 in the PVB group (Ps0.29).
sion (3%) and rarely respiratory depression (0.07%). Addi- Rescue morphine analgesia was required in four patients of
tionally, it prolonged operative time and was associated epidural group (19%) and five patients of PVB group (23%)
with technical failure or displacement (8%). PVB provided (Ps0.99). The PaO2 yFiO2 ratio reduced significantly from
most effective pain-relief with less side-effects. baseline values in both groups without between-group dif-
Richardson et al. w5x conducted a prospective randomized ferences. The median percentage reduction of systolic
study between thoracic epidural and paravertebral bupi- arterial pressure from baseline was –9 (0 to –9)% in group

Best Evidence
vacaine in 100 adult patients. The visual analogue pain PVB and –17 (0 to –38)% in epidural group (Ps0.02); while

Topic
score (VAS) at rest and on cough was significantly lower in clinically relevant hypotension (systolic arterial pressure
the paravertebral group (Ps0.02 and 0.0001, respectively). decrease -30% of baseline) was observed in four patients
Pulmonary function, as assessed by peak expiratory flow of epidural group only (19%) (Ps0.04). Thus, PVB is as
rate (PEFR), was significantly better preserved in the effective as epidural blockade in controlling post-thoracot-
paravertebral group. The lowest PEFR as a fraction of omy pain, but is associated with less haemodynamic
preoperative control was 0.73 in the paravertebral group effects.
in contrast with 0.54 in the epidural group (P-0.004).
Oximetric recordings were better in the paravertebral 7. Clinical bottom line
group (96%) compared to the epidural group (95%)
(Ps0.0001). We conclude that PVB can be at least as effective as
Plasma concentrations of cortisol increased markedly in epidural analgesia. It also has a better side-effect profile
both groups, but the increment was statistically different and a lower complication rate than epidural analgesia.
in favour of the paravertebral group (Ps0.003). Seven
patients in the epidural group became hypotensive com- References
pared to 0 in the paravertebral group. Mean hospital stay
w1x Dunning J, Prendergast B, Mackway-Jones K. Towards evidence-based
was 6.7 days (range 4–11) for the paravertebral group and
medicine in cardiothoracic surgery: best BETS. Interact CardioVasc
6.7 (range 3–16) (PsNS) for the epidural group. PVB is as Thorac Surg 2003;2:405–409.
effective as epidural and better in terms of pulmonary w2x Joshi GP, Bonnet F, Shah R, Wilkinson RC, Camu F, Fisher B, Neugebauer
function, neuroendocrine stress response and side effects. EAM, Rawal N, Schug SA, Simanski C, Kehlet H. A systemic review of
 ARTICLE IN PRESS
96 M. Scarci et al. / Interactive CardioVascular and Thoracic Surgery 10 (2010) 92–96

randomized trials evaluating regional techniques for postthoracotomy pain, pulmonary function and stress response. Br J Anaesth 1999;83:
analgesia. Anesth Analg 2008;107:1026–1040. 387–392.
w3x Davies RG, Myles PS, Graham JM. A comparison of analgesic efficacy w6x Kaiser AM, Zollinger A, De Lorenzi D, Largiadèr F, Weder W. Prospective,
and side-effects of paravertebral vs epidural blockade for thoracotomy. randomized comparison of extrapleural versus epidural analgesia for
A systematic review and meta-analysis of randomised trials. Br J
postthoracotomy pain. Ann Thorac Surg 1998;66:367–372.
Anaesth 2006;96:418–426.
w4x Detterbeck FC. Efficacy of methods of intercostal nerve blockade for w7x Casati A, Alessandrini P, Nuzzi M, Tosi M, Iotti E, Ampollini L, Bobbio A,
pain relief after thoracotomy. Ann Thorac Surg 2005;80:1550–1559. Rossini E, Fanelli G. A prospective, randomized, blinded comparison
w5x Richardson J, Sabanathan S, Jones J, Shah R, Cheema S, Mearns AJ. A between continuous thoracic paravertebral and epidural infusion of
prospective, randomized comparison of preoperative and continuous 0.2% ropivacaine after lung resection surgery. Eur J Anaesth 2006;
balanced epidural or paravertebral bupivacaine on post-thoracotomy 23:999–1004.