Are paravertebral block and thoracic

epidural analgesia comparable for Francis Bonnet1, Girish Joshi2 and Rajesh
Shah3 on behalf of the PROcedure-SPECific
post-thoracotomy pain relief? postoperative pain managemenT (PROSPECT)
Working Group

A systematic review
1
Department d'Anesthésie-Réanimation, Hôpital Tenon, Paris, France; 2Department of
Anesthesiology and Pain Management Southwestern Medical School, Dallas, TX, USA;
3
Wythenshawe Hospital, Wythenshawe, Manchester, United Kingdom

Background and Goals Materials and Methods

l PROSPECT: web-based clinical decision support programme Figure 1
(www.postoppain.org), which aims to formulate robust
evidence-based recommendations for procedure-specific Inclusion criteria:
postoperative pain management. l Qualitative and
l Randomised controlled trials of PV (local
Systematic literature
l Initiated by an expert Working Group of surgeons and review using the quantitative (meta-analysis,
anaesthetic (LA)) or TE (LA + opioid) vs. control, where possible) analyses
anaesthesiologists. Cochrane protocol;
or PV vs. TE, at comparable times of administration
l The aim of this systematic review, which was part of a MEDLINE and EMBASE of postoperative analgesic
in each group and in adult thoracotomy. and recovery outcomes.
(1966–May 2004), the
l Postoperative pain scores (converted to
larger review investigating postoperative pain management
following thoracotomy (total included studies = 169; total Cochrane library, and l Significant outcomes:
excluded studies = 163), was to compare the efficacy and secondary literature. VAS 0–100 mm) at rest, on coughing and/or p<0.05.
safety of paravertebral block (PV) and thoracic epidural (TE) on movement.
analgesia for the management of post-thoracotomy pain.

Results and Discussion

PV (LA) vs. control Figure 2 Other outcomes, TE (LA) vs. PV (LA)
l There was a significant benefit of PV (LA) treatment over
PV (LA) vs. control
No significant effect Meta-analyses:
l Compared with TE (LA), PV (LA) was associated with
control for reducing VAS pain scores at rest, see Figure Number of studies Significant reduction of pain at rest
2; the effect was evident from the early postoperative 10 Significant reduction of pain on coughing
period (0–6 h) through to the third postoperative day. Significant reduction of pain on movement decreased incidence of:

l PV (LA) treatment was also significantly superior to

8 – hypotension (three studies, OR 0.07, p=0.003)
control for reducing VAS pain scores on coughing1 and – urinary retention (two studies, OR 0.28, p=0.01)
6
on movement2 at the time points recorded, see Figure 2. – nausea (two studies, OR 0.26, p=0.01)
4 See Table 1 for details.
Meta-analyses:
l PV (LA) was superior to control for reducing pain scores
2 Table 1
at rest on: Outcome Incidence
0
0–6 h 8–12 h Day 1 Day 2 Day 3 PV (LA) TE (LA)
– day 1 (four studies, including two arms of one study,
WMD -12.28 mm, p=0.0007) Time postoperatively
Hypotension 0/20 patients8 1/20 patients8
– day 3 (three studies, including two arms of one Figure 3 0/10 patients9 6/9 patients9
study, WMD -18.25 mm, p=0.009) TE (LA + opioid) vs. control 0/46 patients10 7/49 patients10
l There were no significant differences between groups: No significant effect
Urinary retention 1/10 patients9 6/9 patients9
Significant reduction of pain at rest
– at 8 h (three studies, including two arms of one Number of studies
Significant reduction of pain on coughing 5/46 patients10 11/49 patients10
study, WMD -11.89 mm, p=0.11) 7
Significant reduction of pain on movement
PONV Nausea Nausea
6
– on day 2 (four studies, including two arms of one
4/15 patients11 7/15 patients11
study, WMD -7.13 mm, p=0.36) 5
2/46 patients10 10/49 patients10
4
TE (LA + opioid) vs. control 3
Vomiting Vomiting

l There was a significant benefit of TE (LA + opioid) 2

1/14 patients11 0/1511
treatment over control for reducing VAS pain scores at 2/46 patients10 7/49 patients10
1
rest and on coughing, see Figure 3.
l This effect was evident at most time points recorded, l Two10,12 out of four studies found significantly improved
0
0–6 h 8–12 h Day 1 Day 2 Day 3
with the exception of Day 2 and 8–12 h for pain at rest pulmonary function in the PV (LA) group compared with
Time postoperatively
and on coughing, respectively; at those time points, half the TE (LA) group.
l The other two studies8,11 reported no significant differences
of the studies showed a significant benefit of TE (LA +
Figure 4a
opioid) treatment over control.
between groups for pulmonary function outcomes.
l TE (LA + opioid) treatment was also significantly superior TE (LA ± opioid) vs. PV (LA ± opioid)
to control for reducing VAS pain scores on movement at Mean/median VAS pain (paravertebral) A recent systematic review comparing the analgesic
all time points recorded. 100 efficacy and side-effects of paravertebral vs. epidural
l One study that did not specify the time of assessment
Bimston 1999, 0 h rest, NS
3
De Cosmo 2002, 1 h movement, blockade also found that the two techniques were
80 no p-value
found no significant difference between groups for pain comparable for reducing pain scores, but PV (LA ±
on movement. Dhole 2001, 0 h cough, NS
60 opioid) was associated with improvements in pulmonary
Kaiser 1998, day 0 rest, NS
Matthews 1989, 4 h rest, NS
function and a reduction in side-effects, compared with
Meta-analyses: 40
Perttunen 1995, 1 h cough, NS TE (LA ± opioid)13.
l TE (LA + opioid) was superior to control for reducing 20 Richardson 1999, 4 h cough, NS
pain scores at rest: Wedad 2004, no mean or median

– at 12 h (one study plus two arms of one study,

0
0 20 40 60 80 100
Conclusions
WMD -16.14 mm, p<0.0001) Mean/median VAS pain (epidural)
This systematic review found that:
l PV (LA) and TE (LA + opioid) were both effective
– on day 1 (three studies plus two arms of one study,
WMD -12.66 mm, p<0.00001) Figure 4b
for reducing pain after thoracotomy compared with
– on day 2 (three studies plus two arms of one study, Mean/median VAS pain (paravertebral) control.
l When comparing PV (LA) with TE (LA), there was no
WMD -7.44 mm, p<0.00001) 100

– on day 3 (three studies, WMD -8.20 mm, 80

Bimston 1999, 24 h rest, p=0.02
overall benefit of either technique for reducing pain
p<0.00001) De Cosmo 2002, 24 h movement, NS
scores.
l PV (LA) was associated with fewer side-effects and
60 Kaiser 1998, day 1 rest, NS
Matthews 1989, 24 h rest, NS
TE (LA ± opioid) vs. PV (LA ± opioid) 40 Perttunen 1995, 24 h cough, NS improved pulmonary function compared with TE (LA).
l TE (LA) and PV (LA) were largely comparable for Richardson 1999, 24 h cough, p=0.0001
l There were very few studies comparing PV (LA) with
20 Wedad 2004, no mean or median
reducing VAS pain scores both at rest and on coughing. TE (LA + opioid), which currently prevents evaluation
l There were no studies assessing VAS pain scores on 0 of the ‘gold standard’ for post-thoracotomy pain.
movement for this particular comparison. 0 20 40 60 80 100

l TE (LA + opioid) is commonly used in clinical practice

4
and it would be valuable to compare this technique with
PV (LA).
l Only three studies included opioid in the LA solution in
one or both groups5–7; these studies showed mixed
results.
l In the early postoperative period, PV (LA) and TE (LA)
showed comparable effects on pain scores, see Figure
Mean/median VAS pain (epidural)
References
NB: Bimston et al.,5 De Cosmo et al.6 and Kaiser et al.7 included 1. Bilgin M, et al. J Thorac Cardiovasc Surg 2003; 126: 1580–3.
opioid in the LA solution in one or both groups; in addition,
2. Carabine UA, et al. Reg Anesth 1995; 20: 412–7.
the time of administration for the thoracic epidural and 3. Licker M, et al. Br J Anaesth 2003; 91: 525–31.
4. Savage C, et al. Chest Surg Clin N Am 2002; 12: 251–63.
paravertebral blocks differed in the De Cosmo et al.6 study, with 5. Bimston DN, et al. Surgery 1999; 126: 650–6; discussion 656–7.
epidural analgesia being started pre-operatively; larger circles 6. De Cosmo G, et al. Acta Medica Romana 2002; 40: 340–7.
represent studies with greater patient numbers than studies
7. Kaiser AM, et al. Ann Thorac Surg 1998; 66: 367–72.
8. Dhole S, et al. J Cardiothorac Vasc Anesth 2001; 15: 288–92.
represented by smaller circles
9. Matthews PJ, et al. Br J Anaesth 1989; 62: 204–5.
10. Richardson J, et al. Br J Anaesth 1999; 83: 387–92.

l Quantitative analysis of pain score data was not

4a; studies including opioid in the LA solution showed 11.
12.
a tendency towards higher pain scores in the PV group. 13.
performed because there was heterogeneity in study
l Mean or median pain scores were comparable overall design between studies that presented data suitable
between TE (LA ± opioid) and PV (LA ± opioid) on for meta-analysis.
day 1/at 24 h, see Figure 4b.
Perttunen K, et al. Br J Anaesth 1995; 75: 541–7.
Wedad M, et al. Egypt J Anaesth 2004; 20(1):67–72.
Davies RG, et al. Br J Anaesth 2006; 96: 418–26.
Acknowledgement
PROSPECT is supported by Pfizer.