Postoperative pain after abdominal hysterectomy: a ... - BJA

British Journal of Anaesthesia Page 1 of 9
doi:10.1093/bja/aet345
BJA Advance Access published October 31, 2013
Postoperative pain after abdominal hysterectomy: a
randomized, double-blind, controlled trial comparing
continuous infusion vs patient-controlled intraperitoneal
injection of local anaesthetic
A. Perniola 1 *,F.Fant 1 , A. Magnuson 2 , K. Axelsson 1 and A. Gupta 1
1 Department of Anesthesiology and Intensive Care and 2 Clinical Epidemiology and Biostatistical Unit, Clinical Reaserch Centre, Örebro
University Hospital, School of Health and Medical Sciences, Örebro University, SE-701 85 Örebro, Sweden
* Corresponding author. E-mail: perniola.d@gmail.com
Editor’s key points
† Effective analgesic
techniques that reduce
opioid consumption may
improve postoperative
recovery.
† There is evidence that
intraperitoneal (i.p.) local
anaesthetics can be
effective.
† This study compared
continuous with intermittent
patient-controlled i.p.
levobupivacaine (PCipA) for
postoperative analgesia.
† PCipA resulted in similar pain
control, reduced opioid
consumption and improved
gastrointestinal function.
† Further work is needed to
evaluate this potentially
beneficial analgesic
technique.
Background. Local anaesthetics (LA) injected intraperitoneally have been found to
decrease postoperative pain. This double-blind randomized study was performed
comparing continuous infusion or patient-controlled intraperitoneal (i.p.) bolus injection
of LA. The primary endpoint was supplemental opioid consumption during the first 24
postoperative hours.
Methods. Two multi-hole catheters were placed intraperitoneally at the end of the surgery in
40 patients undergoing elective abdominal hysterectomy. The patients were randomized into
two groups: Group P: patients self-injected 10 ml of levobupivacaine 1.25 mg ml 21 via the i.p.
catheter as needed, maximum once per hour, and had continuous saline infusion 10 ml h 21
into the second catheter. Group C: patients received a continuous infusion of 10 ml h 21 of
levobupivacaine 1.25 mg ml 21 intraperitoneally through one catheter and 10 ml saline as
bolus as needed via the other. Ketobemidone was administered intravenously as rescue
medication.
Results. Total ketobemidone consumption during 0–24 h was lower in Group P compared with
Group C (mean 23.1 vs 35.7 mg, P¼0.04). No differences in the median pain scores were found
between the groups. Earlier return of gastrointestinal (GI) function was found in Group P vs
Group C (mean 1.5 vs 2.2 days, P,0.01), which also resulted in earlier home-readiness
(mean 1.9 vs 2.7 days, P¼0.04).
Conclusions. A statisticallysignificantopioid-sparingeffectwasfoundwhen patient-controlled
levobupivacaine was administered intraperitoneally as needed compared with continuous
infusion. This was associated with a faster return of GI function and home-readiness. There
was, however, a wide confidence interval in the primary endpoint, opioid consumption.
Keywords: abdominal hysterectomy; local anaesthetics; postoperative pain
Accepted for publication: 11 August 2013
Downloaded from http://bja.oxfordjournals.org/ by guest on January 17, 2014
It is estimated that one woman in 10 undergoes hysterectomy
in Sweden. 1 Abdominal hysterectomy is associated with
moderate-to-severe postoperative pain. 2 Traditional methods
for postoperative pain management include opioids administered
systemically using patient-controlled i.v. analgesia (PCA),
or neuroaxially via epidural or spinal injections. However, pain
relief, specifically on movement, is not always adequately controlled
when using PCA, despite moderate–large doses of morphine.
This is associated with side-effects such as postoperative
nausea and vomiting (PONV), tiredness, pruritus, headache, and
constipation. 3 Therefore, epidural or intrathecal analgesia may
be considered by some to be the gold standard for pain management
after abdominal surgery, and leads to enhanced
and prolonged postoperative analgesia. 4 Although concerns
remain regarding complications aftercentral blocks, specifically
in older patients, 5 recent evidence suggests that these are
extremely rare. 6
There has been recent interest in alternative methods for
analgesia with minimal side-effects and a trend towards
movement from central blocks towards other peripheral and
less invasive methods for pain relief. 7 One such technique is
wound infiltration, intraperitoneal (i.p.) administration of
local anaesthetics (LA), or both. 8 – 11 Continuous i.p. infusion
of LA has been reported to result in a 30–40% morphinesparing
effect and a reduction in postoperative nausea. 12
Several studies have shown that patient-controlled bolus
& The Author [2013]. Published by Oxford University Press on behalf of the British Journal of Anaesthesia. All rights reserved.
For Permissions, please email: journals.permissions@oup.com

BJA
Perniola et al.
administration of LA during neuraxial or peripheral nerve block
gives better postoperative analgesia compared with continuous
infusions. 13 – 17 Therefore, we decided to test the hypothesis
that patient-controlled i.p. analgesia (PCipA) with LA
would reduce pain intensity and thereby supplemental opioid
consumption compared with continuous infusion of LA.
The primary aim was therefore to analyse rescue analgesic
consumption during 0–24 h after operation after abdominal
hysterectomy in patients receiving continuous LA infusion
compared with intermittent patient-controlled i.p. injections.
The secondary outcomes were analgesic consumption during
0–4 h, pain intensity, incidence of nausea and vomiting,
maximum expiratory pressure, time to mobilization, and
length of hospital stay, and also health-related quality of life
at 1 and 3 months.
Methods
This study was registered in an international registry, clinicaltrials.gov
(identification number NCT01492075) before
patient recruitment. The study was performed at the Department
of Anaesthesiology and Intensive Care and the Department
of Gynaecological Surgery, University Hospital, Örebro,
Sweden, and monitored by an external organization not
involved in any way with the study. This unit is a quality-based
organization located within the hospital and initiates and
monitors clinical trials. After approval from the Regional
Ethics Committee, Uppsala, oral and written informed
consent was obtained from 40 patients (ASA status I–II),
aged 40–65 yr undergoing elective abdominal hysterectomy
with or without salpingo-oophorectomy in this randomized,
double-blind study. The exclusion criteriawere: patients undergoing
surgery for suspected gynaecological cancer, patients on
chronic analgesic medication, those with known allergy to LA,
patients participating in another clinical study, and those who
had difficulty in understanding Swedish.
Randomization and blinding
On the day of surgery, the Hospital Pharmacy randomized
patients into two groups (20 in each group) using computergenerated
randomized numbers inserted into sealed opaque
envelopes and marked 1–40. All personnel involved in
patient management were fully blinded to the method of analgesia
until the study was completed.
Anaesthesia and surgery
All patients were premedicated with midazolam 0.1 mg kg 21
orally and paracetamol 1 g was given orally every 6 h with
the first dose at the time of premedication. Anaesthesia was
induced with fentanyl 1–2 mg kg 21 and propofol 1–2 mg
kg 21 intravenously. Tracheal intubation was performed after
muscle relaxation with rocuronium 0.5 mg kg 21 and anaesthesia
was subsequently maintained with sevoflurane 1–3% and
oxygen 33% in air. End-tidal CO 2 was maintained between 4.5
and 5.5 kPa using mechanical ventilation in a low-flow breathing
system with CO 2 absorber. Monitoring included non-invasive
arterial pressure, heart rate, peripheral oxygen saturation,
Page 2 of 9
end-tidal gas monitoring, ECG, and muscle relaxation using
the train-of-four stimulation. Sevoflurane concentration was
adjusted in order to maintain adequate anaesthetic depth,
assessed clinically, and fentanyl was given intermittently intravenously
when required for analgesia during the operation.
Muscle relaxation was reversed at the end of surgery using glycopyrrolate
(0.5 mg) and neostigmine (2.5 mg). Hysterectomy
was performed using either a lower-abdominal midline incision
or a Pfannenstiel incision, depending on the choice of
the operator taking into consideration the size of the uterus
and expected surgical difficulty.
Catheter insertion and postoperative pain
management
Before closing the peritoneum, two multi-hole catheters were
inserted percutaneously by the surgeon; one catheter was tunnelled
from the left side and the other from the right side, about
1–2 cm lateral to the abdominal incision and the tip of the
catheters was placed supravaginally as described previously. 2
The catheters were not fixed intraperitoneally. In all patients,
levobupivacaine 1.25 mg ml 21 (20 ml) was injected subcutaneously
along both sides of the incision before skin closure by the
surgeon; an additional 20 ml was injected via one of the cathetersintothe
peritoneal cavity. The pharmacysent (i) oneblinded
bag, containing 500 ml of levobupivacaine 1.25 mg ml 21 or
placebo (normal saline) and (ii) one elastomeric pump (On-Q
Pain BusterR, ref PS125071; I-Flow Corp., Lake Forest, CA, USA)
containing 500 ml of levobupivacaine 1.25 mg ml 21 or placebo
(normalsaline).Whenthepatientarrivedinthepost-anaesthesia
care unit (PACU), the infusions were set up, connected, and
started by one of the co-authors immediately as described
below. The time of start of infusion was considered t¼0.
Group P: PCipAwith LA. The patients self-administered 10 ml
of levobupivacaine via the first i.p catheter as required
(maximum once per hour) during the first 48 h after the
operation via an electronic pump (Abbot GemStar, Abbot
Laboratories, North Chicago, IL, USA) over 10 min. Normal
saline was infused at 10 ml h 21 via the second i.p catheter
using the elastomeric pump.
Group C: continuous i.p. infusion of LA. Patients received a
continuous infusion of levobupivacaine 10 ml h 21 intraperitoneally
through the first catheter via the elastomeric pump
and saline as bolus injection via the second i.p. catheter
using the electronic pump. After 48 h, the i.p. bolus/infusion
was stopped and both catheters were removed.
Rescue medication and postoperative ward
All patients received 2–4 mg ketobemidone intravenously, an
opioid analgesic equipotent with morphine, 18 intermittently
by the staff nurse if pain was .3 on the numerical rating
scale (NRS) (0, no pain; 10, worst imaginable pain). Analgesics
were administered on patient request, irrespective of whether
the pain was .3 at rest, on sitting, or during coughing. The
patients were observed in the PACU until the day after the
surgery. In the case of severe postoperative nausea or vomiting
(PONV) not relieved by anti-emetics, the nursing staff were
Downloaded from http://bja.oxfordjournals.org/ by guest on January 17, 2014

Intraperitoneal local anaesthetics for pain relief after abdominal hysterectomy
BJA
allowed to administer non-steroidal anti-inflammatory drugs
(NSAIDs) for pain management in order to reduce opioid-induced
PONV.
Recording and measurements
In addition to the routine postoperative protocols, the following
parameters were recorded:
Pain and analgesia: pain at the site of the incision, ‘deep’
(visceral) pain, and pain on coughing at 1, 4, 12, 24, and
48 h were measured using NRS. Rescue analgesic (ketobemidone)
consumption was measured during 0–4, 0–24, and
24–48 h.
Side-effects: nausea, vomiting, or both (0–4, 4–24, and
24–48 h) were recorded on a yes/no score; anti-emetic
consumption during 0–24 and 24–48 h was also recorded.
Sedation (before operation, at 4, 24, and 48 h) was recorded
by nurses using a 0–10 scale (0, awake; 10, verbally arousable).
Expiratory muscle function: this was measured before and
24 and 48 h after the operation using maximum expiratory
pressure (P Emax ) (MicroMedical, Moreton-in-Marsh, UK), in a
similar way as described in previous studies. 2 10 12 P Emax
was measured twice by asking the patient to blow rapidly
using the full abdominal force after a maximal inspiration
while in the semi-reclining position. The best of two measurements
was recorded.
Recovery parameters: the ability to walk with and without
support, time to start drinking, eating, and time to return
of gastrointestinal (GI) function (bowel movement or
passing flatus), postoperative home-readiness, and length
of hospital stay were recorded once a day by nurses in the gynaecological
ward. Standardized home-readiness criteria
were used that have been described previously. 212
LA toxicity: the nurses were informed about the common
signs and symptoms of LA toxicity and the patients told to
inform the nurses of any side-effects and symptoms.
Health-related quality of life: this was measured using the
written form of SF-36 questionnaire before and 1 and 3
months after the operation in both groups and the patients
were asked to return it by post. The SF-36 is a validated
health survey consisting of 36 questions that measure
eight health concepts and has been translated and validated
into the Swedish language. 212 A higher score indicates
an improved level of function.
Statistics
In a previous study, 2 the mean [standard deviation (SD)] total
morphine consumption during 0–24 h was found to be
30 mg (SD 17) in patients receiving continuous i.p. LA. Our
hypothesis was that this could be reduced by 50% to a mean
value of 15 mg (SD 8) in patients receiving PCipA with LA.
The sample size was calculated using the unpaired twosided
t-test. Assuming b¼0.2 (power 80%) and a¼0.05, we
determined that we would require 28 patients (14 per group)
in order to achieve statistical significance for the primary endpoint.
We recruited 40 patients in order to achieve adequate
power in the case of missing data or patient drop out.
The mean and SD are used to summarize continuous variables,
while categorical variables are presented as numbers
(%). Unpaired t-test with 95% confidence intervals (CIs) and
P-values was used to analyse differences between study
groups for morphine consumption and other continuous variables
distributed approximately normally with or without log
transformation. The Shapiro–Wilk test was used to evaluate
normal distribution and if non-normality was found, logtransformation
was performed before statistical evaluation.
The results were then transformed back to original scale and
the effect parameter is therefore the ratio of geometric
means. The results are presented as the mean ratios together
with 95% CI. Owing to multiple comparisons when evaluating
the primary endpoint, 0–24 h ketobemidone consumption, the
Bonferroni–Holm 19 correction was applied. The x 2 test or
Fisher’s exact test when appropriate was used for categorical
variables and odds ratio (OR) with normal approximated 95%
CI was calculated. Owing to some missing values, a mixed
model with autoregressive covariance structure was used to
analyse pain on incision, deep pain, and pain on coughing,
and also for assessment of sedation, P Emax , and when analysing
dimensions of SF-36. Study groups and time points were
fixed effects and patients were random effect in this mixed
model. None of the mixed models used found an overall
statistically significant difference between study groups. The
estimated marginal mean differences between groups at different
time-points with 95% CI and P-values are presented.
When normality assumption was not valid, and log transformation
did not help, sensitivity analysis was performed using the
Mann–Whitney U-test, but because no different conclusions
could be made from the analyses, these results are not presented.
Two-tailed P-value of ,0.05 was considered to be statistically
significant. All statistical analyses were performed
using the SPSS version 17 (Chicago, IL, USA).
Results
Of the 40 patients initially randomized into the study, no
patient was excluded and all patients completed the study
(Fig. 1). The patient characteristic data in the two groups, the
operation performed, and intraoperative fentanyl consumption
are shown in Table 1.
The rescue analgesic ketobemidone consumption was
statistically significantly greater in Group C compared with
Group P during 0–24 h and during 0–4 h after operation
(Table 2). No significant difference in ketobemidone consumption
was seen between the groups during the time period
24–48 h. The number of patients given NSAID as rescue medication
during 0–24 h did not differ significantly between the
groups (Table 2).
The intensity of postoperative pain on the NRS is shown in
Figure 2A–C. In general, the median pain scores were ,5 in
all groups, except during coughing when the median scores
reached ≏7 in the early postoperative period. No significant
differences were seen between the groups in NRS pain score
at any time point during the 48 h study period.
Page 3 of 9
Downloaded from http://bja.oxfordjournals.org/ by guest on January 17, 2014

BJA
Perniola et al.
Total number of patients operated during the study time period:
150 patients
Did not fulfil inclusion
criteria or had one of the
exclusion criteria: 75 patients
Number of patients interviewed
for possible inclusion in the
study: 75 patients
Group P
(Patient-controlled i.p. analgesia):
20 patients
Number randomized into the study:
40 patients
Patients included for primary endpoint:
20 patients
Patients excluded due to missing data for
secondary endpoints in 1–3 patients*
Fig 1 Flow diagram of screened, excluded, and recruited patients. *See Tables for details.
• Refused to participate
• Not able to understand study
protocol because of language limitation:
35 patients
Group C
(Continuous i.p. infusion):
20 patients
Patients included for primary endpoint:
20 patients
Patients excluded due to missing data for
secondary endpoints in 1–3 patients*
Downloaded from http://bja.oxfordjournals.org/ by guest on January 17, 2014
The total volume of levobupivacaine administered after the
first 24 postoperative hours was statistically significantly lower
in Group P (180 ml) than in Group C (240 ml) (P,0.01).
The recovery times and time to home-readiness are shown
in Table 3. A statistically significant difference between the
groups was found in the return of GI function, which was
shorter in Group P compared with Group C. The mean time to
home-readiness was statistically significantly shorter in
Group P compared with Group C, while the length of hospital
stay did not differ significantly between the groups. Expiratory
muscle function, measured as maximum expiratory pressure
(P Emax ), decreased at 4 h compared with preoperative values
in both groups, and gradually recovered over time during
48 h. However, no statistically significant differences were
found between the groups at any time point (Table 3). The incidence
of PONV and the number of patients who received
anti-emetics did not differ between the groups. The sedation
scores between the groups were similar (Table 4) and no differences
were seen between the groups in any of the parameters
of the health-related quality of life (SF-36), or the average score
in SF-36 at 1 or 3 months after surgery (Fig. 3).
Discussion
In this study, we found a statistically significant opioid-sparing
effect when patients self-administered bolus doses of levobupivacaine
compared with those having acontinuous infusion of
LA. Furthermore, there was a faster return of GI function and a
Page 4 of 9

Intraperitoneal local anaesthetics for pain relief after abdominal hysterectomy
BJA
shorter time to home-readiness when intermittent injection of
LA was used.
Pain after abdominal hysterectomy arises from several
structures that are traumatized during surgery 8 and includes
somatic pain from the incision site, and pain from deeper structures
including muscle pain and peritoneal and visceral pain.
Although the magnitude of pain from each component is difficult
to define, pain from the incision site is often relatively mild
in comparison with deeper pain from the muscles and peritoneum.
212 Specifically, pain on mobilization and during coughing
is multifactorial and can be very severe in the early postoperative
period. 20
We studied patients undergoing abdominal hysterectomy
since this is a procedure associated with moderate-to-severe
postoperative pain, specifically in the early postoperative
period. 221 We found a low supplementary analgesic consumption
during the first 24 h after i.p. administration of levobupivacaine
in both groups, confirming our previous studies. 2 12
Table 1 Patient characteristic data, duration of anaesthesia and
operation, and fentanyl consumption are shown as mean (SD)
except age (range). All other data are shown as numbers (ASA
physical status) or n (%) as appropriate. Group P, patient-controlled
i.p. analgesia; Group C, continuous i.p. analgesia
Group P (n520)
Group C (n520)
Age (yr) 48 (38–65) 51 (40–63)
Weight (kg) 73 (12) 68 (9)
Height (cm) 168 (5) 165 (6)
ASA physical status (I/II) 15/5 17/3
Duration of operation (min) 119 (26) 121 (29)
Duration of anaesthesia (min) 156 (27) 156 (29)
Type of operation
Total hysterectomy 12 (60%) 15 (75%)
Salpingo-oophorectomy 8 (40%) 5 (25%)
Type of incision
Pfannenstiel 3 (15%) 4 (20%)
Lower midline 17 (85%) 16 (80%)
Intraoperative fentanyl (mg) 318 (74) 277 (75)
However, we also found that PCipA has an advantage over
continuous infusion in that supplementary analgesic consumption
was further reduced by as much as 35% (from 36 to 23 mg)
during 0–24 h. This was much lower than in one study where the
24 h mean morphine consumption, even with ibuprofen, was
47.3 mg. 21 Although we achieved good analgesia at the incision
site by LA infiltration, several other components of pain may
persist including muscle, peritoneal, and visceral pain. Recent
studies, including two meta-analyses of the literature, have
shown that i.p. administration of LA reduces postoperative
pain after abdominal surgery. 822 Although the mechanism for
improved analgesia with patient-controlled bolus doses could
be speculative, we believe that the analgesic efficacyof continuous
infusions may be partly due to their systemic absorption,
and thereby a central rather than a local effect. Significant
plasma concentration of LA is reached when they are infused
intraperitoneally 2 and this may produce some analgesia by
central mechanisms similar to i.v. infusion. 23 24 In contrast to
continuous infusion, the analgesic efficacy of PCipA may be via
localanti-inflammatoryeffects, 825 and alsobyblockingafferent
sensory nerves transmitting visceral pain through the vagus
nerves. 26 Thus, blocking of visceral pain by the intermittent injection
of i.p. LA may not only result in decreased analgesic consumption,
as in our present study, but even faster recovery of
visceral functions such as the earlier return of bowel function,
possibly due to lower rescue opioid consumption. This is an
interesting hypothesis, supported by some evidence from the
literature. In a recent study, infusion of i.p. ropivacaine after
colectomy improved early surgical recovery. 27 The authors
hypothesized that LA cause a transient chemical afferentectomy
when injected intraperitoneally thereby producing analgesia
and early functional recovery. However, the question
that still remains unanswered from our present study is
whether the analgesic effect of LA after PCipA is a systemic
effect or a local analgesic and anti-inflammatory effect.
In our present study, patients having PCipA with LA selfadministered
180 ml of levobupivacaine while those in the
continuous group received 240 ml during the first 24 h after
operation and yet had lower rescue analgesic consumption,
which may also support a local effect. It is also possible, although
speculative, that 10 ml of LA injection during 10 min
in the PCipA group resulted in a wider spread of i.p. LA resulting
Downloaded from http://bja.oxfordjournals.org/ by guest on January 17, 2014
Table 2 Analgesic consumption. All data are shown as mean (SD) except number of patients n (%) for NSAIDs. Data are analysed by the
unpaired t-test or Fisher exact test as appropriate. Group P, patient-controlled i.p. analgesia; Group C, continuous i.p. analgesia. *A mean ratio is the
difference between two means on a relative scale, that is, mean ratio of 1.56 interprets as 56% higher geometric mean in Group C compared with
Group P. † Following the Bonferroni–Holm correction for multiple comparisons, for details, see text. ‡ Group Cvs P; results are presented as odds ratio
(95% CI)
Group P (n520) Group C (n520) Group C–P, mean
difference (95% CI)
Group C/P, mean
ratio* (95% CI)
P-values
Ketobemidone in mg 0–4 h (range) 13.4 (7.1) (4–28) 21.0 (12.6) (8–59) 7.6 (1.0–14.1) 1.56 (1.09–2.22) 0.015 0.045
Ketobemidone in mg 0–24 h (range) 23.1 (13.8) (8–48) 35.7 (20.2) (10–81) 12.6 (1.5–23.6) 1.58 (1.07–2.32) 0.021 0.042
Ketobemidone in mg 24–48 h 3.1 (4.2) (0–12) 6.3 (8.6) (0–31) 3.2 (21.1–7.5) 1.84 (0.97–3.48) 0.060 0.060
(range)
NSAID (0–24 h) [n (%)] 3 (15%) 4 (22%) 1.4 (0.2–11.1) ‡ 1.00
Corrected
P-values †
Page 5 of 9

BJA
Perniola et al.
A
10
8
Group P
Group C
B
10
8
Group P
Group C
NRS (0 – 10)
6
4
NRS (0 – 10)
6
4
2
2
0
0
1 4 12
Hours
C
NRS (0 – 10)
24
10
8
6
4
2
0
48
1 4 12
Hours
1 4 12
Hours
24 48
Group P
Group C
24 48
Downloaded from http://bja.oxfordjournals.org/ by guest on January 17, 2014
Fig 2 (A–C) Pain at the incision site, deep pain, and pain on coughing. Distributions are shown as box plots and inter-quartile range (IQR). Group P,
patient-controlled i.p. analgesia; Group C, continuous i.p. infusion; NRS, numeric rating scale; circle, outliers. The whiskers represent minimum and
maximum values if outliers are not present.
in a block of more free afferent nerve endings within the peritoneum
compared with a continuous infusion of 10 ml over
60 min. The reason why patients in both groups reported moderate
to severe pain on coughing when they could request
further analgesia (LA or opioids) is unclear. It is possible that
analgesic supplementation was requested only during pain
at rest and not during mobilization/coughing. Additionally, in
our experience, some patients may have severe nausea or
vomiting from opioid supplementation and accept pain on
coughing rather than constant nausea. 28
The health-related quality-of-life questionnaire (SF-36) did
not show any differences between the groups after operation,
which may be not surprising since we did not find any difference
in postoperative pain intensity between the groups. The
faster recovery of GI function in patients receiving LA via
PCipA may have resulted in a shorter time to home-readiness.
The shortened time to recovery of bowel function after patientcontrolled
i.p. administration of LA may be due to the opioid
sparing which was evident in the present study. We found no
major complications when using i.p. LA, including infection,
Page 6 of 9

Intraperitoneal local anaesthetics for pain relief after abdominal hysterectomy
BJA
Table 3 Postoperative functional recovery and respiratory function. All data are shown as mean (SD) unless otherwise shown and analysed by the
unpaired t-test except P Emax which was analysed by using mixed model. P Emax , maximum expiratory pressure; Group P, patient-controlled i.p.
analgesia; Group C, continuous i.p. analgesia; GI, gastrointestinal
Group P (n520) Group C (n520) Group C vs P, mean
difference (95% CI)
P-values
Time to walk with help (h) 21 (4), n¼19 22 (3), n¼19 0.4 (21.8 to 2.8) 0.68
Time to walk without help (h) 23 (6), n¼19 25 (6), n¼18 2.2 (22.1 to 6.6) 0.30
Time to start drinking (h) 16 (7), n¼17 15 (8), n¼17 21.4 (26.8 to 3.9) 0.58
Time to start eating (h) 24 (9), n¼19 26 (8), n¼18 1.7 (24.2 to 7.6) 0.56
Home-readiness (days) 1.9 (0.7), n¼17 2.7 (1), n¼13 0.7 (0.03–1.5) 0.04
Length of hospital stay (days) 3.9 (0.8), n¼20 4.5 (1.3), n¼20 0.6 (20.6 to 1.3) 0.07
Return to GI function (days) 1.5 (0.6), n¼12 2.2 (0.6), n¼18 0.7 (0.2–1.2) ,0.01
P E max (litre min 21 )
Before operation 59 (16), n¼17 60 (14), n¼18
4 h 35 (13), n¼18 34 (16), n¼19 21(210 to 7) 0.72
24 h 42 (16), n¼19 34 (11), n¼20 27(216 to 1) 0.09
48 h 45 (15), n¼19 41 (10), n¼19 24(213 to 5) 0.37
Table 4 Side-effect and complications. All results are shown as number of patients n (%) except sedation which is presented as mean (SD). PON,
postoperative nausea; POV, postoperative vomiting. The x 2 test or Fisher exact test was used as appropriate except sedation which was analysed
using mixed model. NE, not estimated; Group P, patient-controlled i.p. analgesia; Group C, continuous i.p. analgesia. *Group C vs P is presented as
mean difference (95% CI)
Group P (n520) Group C (n520) Group C vs P, odds ratio (95% CI) P-value
Postoperative nausea (PON)
0–4 h 10 (50%) 5 (25%) 0.3 (0.1–1.3) 0.10
4–24 h 2 (10%) 2 (10%) 1.0 (0.1–7.9) 1.00
24–48 h 1 (5%) 0 NE 1.00
Postoperative vomiting (POV)
0–4 h 3 (15%) 1 (5%) 0.3 (0.03–3.1) 0.60
4–24 h 2 (10%) 1 (5%) 0.5 (0.04–5.6) 1.00
24–48 h 0 0 NE
Total PON or POV
0–48 h 11 (55%) 6 (30%) 0.3 (0.1–1.3) 0.11
Anti-emetics given (0–24 h) 16 (80%) 17 (85%) 1.4 (0.3–7.3) 1.00
Anti-emetics given (24–48 h) 4 (20%) 3 (15%) 0.7 (0.1–3.6) 1.00
Sedation (NRS)*
4 h 4.8 (2.5), n¼18 4.4 (2.7), n¼19 20.4 (22.2 to 1.4) 0.65
24 h 4.0 (2.4), n¼19 4.3 (3.2), n¼20 0.4 (21.3 to 2.2) 0.61
48 h 3.5 (3.0), n¼17 3.1 (2.4), n¼17 20.3 (22.1 to 1.6) 0.77
Downloaded from http://bja.oxfordjournals.org/ by guest on January 17, 2014
and no patient had any signs or symptoms of LA toxicity, which
is in accordance with our earlier studies. 212
Study limitations
The patient-controlled i.p. analgesia group received more
intraoperative fentanyl, and therefore, the reduction in the
0–4 ketobemidone consumption in this group may reflect on
the higher intraoperative fentanyl given. Furthermore, the
wide CI of ketobemidone consumption during the first 24 h
suggests caution in concluding that the primary endpoint of
this study was clinically significant. However, the finding that
some secondary endpoints such as return of GI function and
home-readiness favour the patient-controlled i.p. method
would suggest thatthere is some impact of reduced opioid consumption
on recovery parameters. We did not measure plasma
concentration of levobupivacaine in this study, which would be
of interest in understanding the mechanism of action of LA.
The PCipA with LA made it almost impossible to correctly
schedule blood sampling. However, in our earlier study using
higher doses of levobupivacaine via continuous i.p. infusion,
we found a dose-dependent increase in plasma concentration
Page 7 of 9

BJA
Perniola et al.
100
80
60
40
20
0
100
Group P
Group C
Preop
Score
80
60
40
20
1 month after
surgery
but no systemic toxicity due to the low level of free LA in the
plasma. Another limitation of this study was that we did not
verify catheter position by contrast injection. Therefore, we
do not know the exact position of the catheter tip or if there
was catheter migration during mobilization. Finally, we did
not use a patient-controlled analgesia pump intravenously,
which allows the correct assessment of analgesic requirement
after operation since it would have been difficult for patients to
handle two PCA devices, one for i.p. LA and the other for i.v. injection
of opioids as rescue medication.
Conclusions
We found that the patient-controlled i.p. analgesia with LA
resulted in statistically significant lower ketobemidone consumption
compared with continuous infusion both during
the initial postoperative phase and during 0–24 h after
operation. Because of the large confidence intervals in
0
100
80
60
40
20
0
PF RP BP GH VT SF RE MH PCS MCS
Fig 3 Health-related quality of life (SF-36). Data are presented as mean of: physical functioning (PF), role-physical (RP), bodily pain (BP), general
health (GH), vitality (VT), social functioning (SF), role-emotional (RE), mental health (MH); physical component score (PCS), and mental component
score (MCS). Group P, patient-controlled i.p. analgesia; Group C, continuous i.p. infusion. There were missing data for one patient in Group P before
operation. After operation, at 1 month, there were missing data for five and six patients and at 3 months, for three and five patients in Groups P and
C, respectively. See text for details.
ketobemidone consumption, the clinical relevance for this is
uncertain. However, the lower ketobemidone consumption
did translate into quicker return of GI function and earlier
home-readiness. Future studies should focus on identifying
the mechanism of analgesic effect of LA administered intermittently
intraperitoneally.
Authors’ contributions
3 months after
surgery
A.P.: contributed towards designing the study, data collection,
statistical analysis, and writing the manuscript. F.F.: contributed
towards data collection and critical comments in the
writing of the manuscript. A.M.: contributed towards all statistical
analysis and discussion on presentation of results. K.A.:
contributed towards designing the study, data confirmation,
and critical comments in the writing of the manuscript. A.G.:
contributed towards designing, data interpretation, and
writing of the manuscript.
Downloaded from http://bja.oxfordjournals.org/ by guest on January 17, 2014
Page 8 of 9

Intraperitoneal local anaesthetics for pain relief after abdominal hysterectomy
BJA
Acknowledgements
We would like to thank the personnel in the operating theatres
and those in the postoperative and gynaecological wards for
their help and attention during the various phases of this
study. Special thanks to Ingegärd Wilhelmsson for her help
with the data collection and with patient recruitment.
Declaration of interest
None declared.
Funding
This study was supported partly by funds obtained from the Research
Committee, Örebro County Council, Örebro, Sweden.
References
1 Jacobson GF, Shaber RE, Armstrong MA, Hung YY. Hysterectomy
rates for benign indications. Obstet Gynecol 2006; 107: 1278–83
2 Perniola A, Gupta A, Crafoord K, Darvish B, Magnuson A, Axelsson K.
Intraabdominal local anaesthetics for postoperative pain relief
following abdominal hysterectomy: a randomized, double-blind,
dose-finding study. Eur J Anaesthesiol 2009; 26: 421–9
3 White PF. The changing role of non-opioid analgesic techniques in
the management of postoperative pain. Anesth Analg 2005; 101:
S5–22
4 Hein A, Rosblad P, Gillis-Haegerstrand C, Schedvins K, Jakobsson J,
Dahlgren G. Low dose intrathecal morphine effects on posthysterectomy
pain: a randomized placebo-controlled study. Acta
Anaesthesiol Scand 2012; 56: 102–9
5 Moen V, Dahlgren N, Irestedt L. Severe neurological complications
after central neuraxial blockades in sweden 1990–1999. Anesthesiology
2004; 101: 950–9
6 Cook TM, Counsell D, Wildsmith JA. Major complications of central
neuraxial block: report on the Third National Audit Project of the
Royal College of Anaesthetists. Br J Anaesth 2009; 102: 179–90
7 Kehlet H, Liu SS. Continuous local anesthetic wound infusion to
improve postoperative outcome: back to the periphery? Anesthesiology
2007; 107: 369–71
8 Kahokehr A, Sammour T, Soop M, Hill AG. Intraperitoneal local anaesthetic
in abdominal surgery—a systematic review. ANZ J Surg
2011; 81: 237–45
9 Liu SS, Richman JM, Thirlby RC, Wu CL. Efficacy of continuous wound
catheters delivering local anesthetic for postoperative analgesia: a
quantitative and qualitative systematic review of randomized controlled
trials. J Am Coll Surg 2006; 203: 914–32
10 Fant F, Axelsson K, Sandblom D, Magnuson A, Andersson SO,
Gupta A. Thoracic epidural analgesia or patient-controlled local analgesia
for radical retropubic prostatectomy: a randomized,
double-blind study. Br J Anaesth 2011; 107: 782–9
11 Thornton PC, Buggy DJ. Local anaesthetic wound infusion for acute
postoperative pain: a viable option? Br J Anaesth 2011; 107: 656–8
12 Gupta A, Perniola A, Axelsson K, Thorn SE, Crafoord K, Rawal N. Postoperative
pain after abdominal hysterectomy: a double-blind
comparison between placebo and local anesthetic infused intraperitoneally.
Anesth Analg 2004; 99: 1173–9
13 Rawal N, Allvin R, Axelsson K, et al. Patient-controlled regional analgesia
(PCRA) at home: controlled comparison between bupivacaine
and ropivacaine brachial plexus analgesia. Anesthesiology
2002; 96: 1290–6
14 Singelyn FJ, Vanderelst PE, Gouverneur JM. Extended femoral nerve
sheath block after total hip arthroplasty: continuous versus
patient-controlled techniques. Anesth Analg 2001; 92: 455–9
15 Eledjam JJ, Cuvillon P, Capdevila X, et al. Postoperative analgesia by
femoral nerve block with ropivacaine 0.2% after major knee
surgery: continuous versus patient-controlled techniques. Reg
Anesth Pain Med 2002; 27: 604–11
16 van der Vyver M, Halpern S, Joseph G. Patient-controlled epidural
analgesia versus continuous infusion for labour analgesia: a
meta-analysis. Br J Anaesth 2002; 89: 459–65
17 Duncan LA, Fried MJ, Lee A, Wildsmith JA. Comparison of continuous
and intermittent administration of extradural bupivacaine for
analgesia after lower abdominal surgery. Br J Anaesth 1998; 80:
7–10
18 Ohqvist G, Hallin R, Gelinder S, Lang H, Samuelson S. A comparison
between morphine, meperidine and ketobemidone in continuous
intravenous infusion for postoperative relief. Acta Anaesthesiol
Scand 1991; 35:44–8
19 Holm S. A simple sequentially rejective multiple test procedure.
Scand J Stat 1979; 6:65–70
20 Leung CC, Chan YM, Ngai SW, Ng KF, Tsui SL. Effect of pre-incision
skin infiltration on post-hysterectomy pain—a double-blind
randomized controlled trial. Anaesth Intensive Care 2000; 28:
510–6
21 Kroll PB, Meadows L, Rock A, Pavliv L. A multicenter, randomized,
double-blind, placebo-controlled trial of intravenous ibuprofen
(i.v.-ibuprofen) in the management of postoperative pain following
abdominal hysterectomy. Pain Pract 2011; 11:23–32
22 Kahokehr A, Sammour T, Srinivasa S, Hill AG. Systematic review and
meta-analysis of intraperitoneal local anaesthetic for pain reductionafterlaparoscopic
gastric procedures. Br J Surg 2011; 98:29–36
23 Marret E, Rolin M, Beaussier M, Bonnet F. Meta-analysis of intravenous
lidocaine and postoperative recovery after abdominal surgery.
Br J Surg 2008; 95: 1331–8
24 Omote K. Intravenous lidocaine to treat postoperative pain: management.
Novel strategy with a long-established drug. Anesthesiology
2007; 107:5–6
25 Harvey KP, Adair JD, Isho M, Robinson R. Can intravenous lidocaine
decrease postsurgical ileus and shorten hospital stay in elective
bowel surgery? A pilot study and literature review. Am J Surg
2009; 198: 231–6
26 Konsman JP, Luheshi GN, Bluthe RM, Dantzer R. The vagus nerve
mediates behavioural depression, but not fever, in response to peripheral
immune signals; a functional anatomical analysis. Eur J
Neurosci 2000; 12: 4434–46
27 Kahokehr A, Sammour T, Shoshtari KZ, Taylor M, Hill AG. Intraperitoneal
local anesthetic improves recovery after colon resection: a
double-blinded randomized controlled trial. Ann Surg 2011; 254:
28–38
28 Goldstein A, Grimault P, Henique A, Keller M, Fortin A, Darai E. Preventing
postoperative pain by local anesthetic instillation after
laparoscopic gynecologic surgery: a placebo-controlled comparison
of bupivacaine and ropivacaine. Anesth Analg 2000; 91:
403–7
Downloaded from http://bja.oxfordjournals.org/ by guest on January 17, 2014
Handling editor: L. Colvin
Page 9 of 9