Continuous infusion of local anaesthetic following laparoscopic hysterectomy—a randomised controlled trial




To estimate whether a continuous infusion of intraperitoneal local anaesthetic for 48 hours following laparoscopic hysterectomy reduced the need for opioids delivered with a patient-controlled analgesia pump.


Double-blind randomised placebo-controlled trial.


District general hospital in the UK.


Women undergoing a laparoscopic hysterectomy for a benign indication.


Women were randomised to receive either 0.5% levobupivicaine or 0.9% normal saline via an ON-Q elastomeric pump for 48 hours postoperatively. The amount of opioids used via the patient-controlled analgesia pump was recorded and pain was measured using an 11-point Box Scale.

Main outcome measures

The primary outcome was the amount of patient-administered morphine used over the first 48 postoperative hours. Secondary outcomes were length of hospital stay, oral analgesia use and level of patient-reported pain.


Sixty women participated and completed the trial. There was no difference (= 0.59) in the median amount of patient-administered morphine used between the levobupivicaine (23 mg) and placebo (18.5 mg) groups; median group difference 3.0 (95% CI −7.0 to 14.0). There was also no difference in the length of hospital stay with 40% of the treatment group remaining in hospital >48 hours compared with 30% of the placebo group (= 0.08). Pain scores at all postoperative time points remained similar, with a median group difference in pain scores of 1.0 (95% CI −1.0 to 2.0) at the end of the first postoperative day.


Continuous infusion of 0.5% levobupivicaine into the peritoneal cavity following laparoscopic hysterectomy does not have any opioid-sparing effects.


The increasing use of a laparoscopic approach to surgery allows a shorter hospital stay, with a concomitant requirement for adequate pain relief. The use of incisional and intraperitoneal local anaesthetic (LA) provides the potential for good pain relief, minimising opioid use and allowing early hospital discharge. The evidence for its effectiveness is conflicting, however, leading to the need for study designs that take account of the timing, dosage and site of administration of LA. Randomised controlled trials (RCTs) have shown that greater pain relief can be achieved through the use of pre-incisional LA at laparoscopic port sites as opposed to post-incisional LA administration.[1] Similarly, in a randomised trial of patients undergoing laparoscopic cholecystectomy, benefits were demonstrated for using preoperative intraperitoneal LA to suppress nociceptive stimulation and activation of pain pathways.[2] This finding has been supported in a meta-analysis of RCTs, where intraperitoneal LA before dissection in laparoscopic cholecystectomies provided greater pain relief than when the LA was given during or following surgery.[3]

Although there appears to be an analgesic benefit from early administration of intraperitoneal LA, particularly when given before any dissection, there is less certainty about its continued use in the postoperative period. Indeed, a systematic review and meta-analysis of 44 RCTs of postoperative LA infusion revealed substantial heterogeneity in the data, possibly due to inadequate sample size, different dosages and varying types of surgical procedures.[4]

A small RCT of 30 patients undergoing laparoscopic gastric banding reported a reduction in self-reported pain scores among patients receiving postoperative intraperitoneal LA infusion compared with those in the placebo group, although they failed to demonstrate any reduction in morphine use during the first 48 postoperative hours.[5] Where patient-controlled administration (PCA) of analgesia was used in an RCT of women undergoing abdominal hysterectomy, women receiving intraperitoneal LA infusion were found to use less opioids in the 24 postoperative hours than those in the placebo group, although they reported similar levels of pain.[6] Although these two small studies point towards possible opioid-sparing or analgesic benefits associated with postoperative intraperitoneal LA, their use in laparoscopic pelvic surgery has yet to be established.

The aim of the trial was to measure if the amount of self-administered morphine used via a PCA pump was reduced by a clinically significant amount in women who had a continuous local anaesthetic infusion into the peritoneum, compared with a control group who were given 0.9% normal saline.


A single-centre, double-blind, randomised placebo-controlled trial was conducted. Participants were recruited from Ashford and St Peter's Hospitals NHS Foundation Trust in the UK. Women were eligible to participate if they were undergoing a total laparoscopic hysterectomy or laparoscopically assisted vaginal hysterectomy and were excluded if they were having surgery for suspected cancer; had a reported allergy to local anaesthetic agents or morphine; or were unable to give informed consent. A laparoscopically assisted vaginal hysterectomy was defined as a laparoscopic hysterectomy where the pedicles up to and including the uterine artery were ligated laparoscopically. In a laparoscopically assisted vaginal hysterectomy the uterosacral pedicle was ligated vaginally and, in addition, the vault was closed vaginally as well. With a total laparoscopic hysterectomy all the pedicles were ligated laparoscopically and the vaginal vault were closed laparoscopically. All participants provided written informed consent.

Eligible participants were randomly assigned in a 1:1 ratio to a continuous infusion of either 0.5% levobupivicaine or 0.9% normal saline. Group allocation was performed using computer-generated random numbers, which were placed in sealed, opaque, serially numbered envelopes, which were opened at the end of the operation. Randomisation was conducted by a statistician who was otherwise not involved in the study. The envelopes were opened by the anaesthetist who prepared the solution, but who was not otherwise involved with the collection or analysis of the data.

At the end of the operation, all women had a multi-hole infusion catheter attached to an ON-Q pump inserted percutaneously through a 2-mm suprapubic incision into the pelvis under direct laparoscopic vision. The catheter tip was guided so that the infusion part of the catheter was placed in the peritoneal cavity at the vaginal vault. The catheter was kept in place with steri-strips and a dressing to the skin, and the ON-Q pump was attached to the patient's clothing. The ON-Q pump is an elastomeric pump with a capacity to hold 100 ml of fluid and releases this at a rate of 2 ml/hour. A bolus of 4 ml of the infusion was delivered via the catheter before the infusion was commenced. The participants and surgeon were blinded to the type of solution being administered.

All operations were carried out by an experienced surgeon (minimum grade ST7 trainee in Obstetrics and Gynaecology). All women had a general anaesthetic with no additional intrathecal opiates and were given intravenous antibiotic prophylaxis. The type of anaesthetic and analgesia given intraoperatively was at the discretion of the anaesthetist. An intraumbilical 12-mm incision was made to insert the laparoscope, and two 5-mm lateral ports, one in each iliac fossa, were used. Skin incisions were closed with vicryl rapide™ (polyglactin 910) or monocryl (poliglecaprone 25). The vaginal vault was closed laparoscopically or vaginally according to the surgeon's preference.

A PCA pump with morphine was set up in recovery. The pump contained 1 mg of morphine per ml and released 1 ml each time it was activated with a 5-minute lock out period before a second dose of morphine could be released. The PCA pump was continued for as long as clinically indicated and was discontinued before discharge. Women were discharged when clinically appropriate and according to our usual hospital procedure. The ON-Q pump infusion was provided for a total of 48 hours following surgery. Women staying in hospital for less than 48 hours were given verbal and written instructions on catheter removal.

Additional oral analgesia was administered by the nursing staff on the ward at the women's request. A total dose of oral analgesia over the 48 postoperative hours was calculated for each woman. Each of the following was considered to be one dose; 1 g of paracetamol, 50 mg diclofenac, 400 mg ibuprofen, 50 mg tramadol, 2 tablets of co-dydramol, and 20 mg oramorph.

The primary outcome measure was the amount of self-administered morphine used in the first 48 hours following a laparoscopic or laparoscopically assisted vaginal hysterectomy. The secondary outcome measures were pain intensity, duration of hospital stay, and use of oral analgesia in the first 48 hours postoperatively. Pain was assessed using an 11-point Box Scale (BS-11), with 0 representing no pain and 10 the worst imaginable pain. Pain was measured using a self-administered questionnaire on day 1 in the evening, day 2 in the morning, afternoon and evening, and on day 3 in the morning and afternoon, and finally, on removing the pump. If women were discharged home within the 48 hours of pain measures, they were given the questionnaire to complete at home and asked to return them by post.

We calculated that a sample size of 60 women would provide 80% power to detect a 30% reduction in morphine use in the 0.5% levobupivicaine group, with a two-sided α level of 0.05. The sample size calculation was based on a previous RCT of levobupivicaine versus normal saline in women undergoing a total abdominal hysterectomy.[6] In this study women receiving levobupivicaine used a mean of 30.5 mg (SD 15.4) of ketobemidone as rescue analgesia in the first 24 postoperative hours, compared with the placebo group who used a mean of 44.3 mg (SD 21.6).

Data were analysed through intention-to-treat and IBM SPSS version 19 was used for the analysis. Fisher's exact tests were used to investigate differences in proportion for independent categorical data, Mann–Whitney U-tests were used to calculate differences in medians and the Hodges–Lehmann test was used to estimate the median group differences and the 95% confidence intervals (95% CI). A Student's t-test was used to calculate differences in means.


Sixty-five women were invited to participate and 60 (92%) agreed to participate over a 12-month period commencing in February 2011. Of the 60 women, 30 received an infusion of 0.9% normal saline and 30 received 0.5% levobupivicaine. All questionnaires were completed at 48 hours (Figure 1). Despite randomisation, women in the treatment arm were less likely to have a contraindication to taking nonsteroidal anti-inflammatory drugs (97% versus 67%, < 0.01) (Table 1). All other baseline measures were similar in both groups. There was no difference in the type of anaesthetic or intraoperative analgesia given between the two groups (Table 2).

Table 1. Participant demographic data
 0.5% levobupivicaine (n = 30)0.9% normal saline (n = 30)
  1. NSAID, nonsteroidal anti-inflammatory; TLH, total laparoscopic hysterectomy.

Median age in years (range)44 (32–57)45 (32–73)
Parity2 (0–4)2 (0–5)
TLH (%)9 (30)8 (27)
NSAID tolerance (%)29 (97)20 (67)
Table 2. Anaesthetic type and intraoperative analgesia
Anaesthetic used0.5% levobupivicaine (%)0.9% normal saline (%)P value
  1. Group differences analysed with Fisher's exact test.

  2. a

    The number who did not receive pethidine is too small for valid statistical comparison with those who received pethidine.

Fentanyl (n = 40)22 (73)18 (60)0.21
Alfentanil (n = 6)3 (10)3 (10)1.00
Morphine (n = 50)25 (83)25 (83)1.00
Parecoxib (n = 23)13 (43)10 (33)0.60
Pethidine (n = 59)29 (97)30 (100) a
Paracetamol (n = 41)21 (70)20 (69)1.00
Diclofenac (n = 13)5 (17)8 (27)0.53
Figure 1.

Consort flow diagram.

The main indication for hysterectomy was heavy uterine bleeding (67%) followed by pelvic pain and dyspareunia (47%). There were no differences between the two study groups in terms of indication for hysterectomy or the type of hysterectomy (total or laparoscopically assisted vaginal). Operative details, additional surgical interventions and complications were similar between the two groups (Table 3). Thirteen women had an operative complication. Of these, three had an estimated blood loss of 1 l, six were converted to a laparotomy because it was not technically possible to complete procedures laparoscopically, two returned to theatre because of concerns regarding bleeding (one of whom had a laparoscopy, and no evidence of the initially suspected postoperative bleeding) and two had a bladder injury.

Table 3. Operative details, additional surgical interventions and complications
 0.5% levobupivicaine (%)0.9%normal saline (%)P value
  1. Group differences analysed with a Fisher's exact test.

Ovaries removed16 (52)15 (50)1.00
Pelvic side wall endometriosis removed6 (20)1 (3)0.10
Uterosacral endometriosis removed5 (17)2 (7)0.42
Rectovaginal endometriosis removed4 (13)00.11
Indwelling urinary catheter inserted19 (63)11 (37)0.07
Vaginal pack inserted01 (3)1.00
Abdominal drain inserted2 (7)2 (7)1.00
Operative complications9 (30)4 (13)0.21

Women in the treatment group self-administered a median of 23 mg (interquartile range 11.5–39.25 mg) of morphine compared with a median of 18.5 mg (interquartile range 7.0–45.25 mg) in the placebo group over the first 48 hours (median group difference 3.0; 95% CI −7.0 to 14.0; = 0.59) (Table 4).

Table 4. Analgesia use in the first 48 postoperative hours
 Group allocationMedian group differencea (95% CI)P valueb
0.5% levobupivicaine0.9% normal saline
  1. IQR, interquartile range.

  2. a

    Hodges–Lehmann median difference.

  3. b

    Mann-Whitney U-test.

  4. c

    Student's t-test.

Morphine (mg) via PCA Median (IQR)23 (11.75–39.25)18.5 (7.0–45.25)3.0 (−7.0 to 14.0)0.59
Paracetamol (g) Median (IQR)7.0 (4.0–7.25)4.0 (3.0–7.0)1.0 (0.0–3.0)0.15
NSAI doses Median (IQR)2.0 (0–5.0)0.2 (0–0.25)1.0 (0.0–2.0)0.01
   Mean differencec (95% CI) 
Total dose of oral analgesia Mean (SD)9.07 (4.14)6.87 (3.78)2.2 (0.15–4.25)0.036

Additional oral analgesic consumption (Table 4) was greater in the treatment group (mean 9.07 doses) compared with the placebo group (mean 6.87 doses) (mean group difference 2.2 doses; 95% CI 0.15–4.25; = 0.36) but there were no group differences in the amount of pain women reported in the first 48 hours (Table 5), with a median group difference in pain scores of 1.0 (95% confidence interval −1.0 to 2.0) at the end of the first postoperative day.

Table 5. Pain scores in the first 48 postoperative hours
Pain scores (0–10)0.5% levobupivicaine Median (IQR)0.9% normal saline Median (IQR)Median group differencea (95% CI)P valueb
  1. a

    Hodges–Lehmann median difference.

  2. b

    Mann–Whitney U-test.

Day of operation evening5 (3–7)4 (0.75–7)0.0 (−1.0 to 2.0)0.59
1st day post operation morning3.5 (0–6)4.0 (1–7)0.0 (−2.0 to 1.0)0.75
1st day post operation afternoon4.0 (1–6)3.5 (1–6)0.0 (−2.0 to 2.0)0.87
1st day post operation evening4.0 (2–6)3.0 (1–6)1.0 (−1.0 to 2.0)0.54
2nd day post operation morning2.0 (1–5)2.0 (0–5)0.0 (−1.0 to 2.0)0.68
2nd day post operation evening2.0 (0–5)2.0 (0–5)0.0 (−1.0 to 1.0)0.93

The length of hospital stay was not affected by having a local anaesthetic infusion. Twelve (40%) women in the treatment arm remained in hospital for more that 48 hours, compared with nine (30%) who received normal saline (= 0.08).


Main findings

We found that a continuous postoperative infusion of intraperitoneal levobupivicaine conferred no advantage compared with placebo, with women using equal amounts of self-administered morphine, reporting similar levels of postoperative pain and experiencing similar lengths of hospital stay. It was surprising to find that women receiving the infusion of levobupivicaine requested greater amounts of oral analgesia in the first postoperative 48 hours than women receiving the placebo infusion. This suggests that an intraperitoneal infusion of levobupivicaine induces more pain than one of normal saline. The only explanation that we can offer here is that levobupivicaine has a pH of 4–6.5 and being acidic, it is possible that it induces peritoneal pain.

Strengths and limitations

The strengths of our study lay in it being a sufficiently powered, prospective double-blind RCT. Although there have been many studies on the effectiveness of postoperative local anaesthetic infusions via indwelling catheters, this is the first trial to consider their use in laparoscopic gynaecological surgery. Our use of patient-administered analgesia as the primary outcome measure has also allowed us to overcome some of the design difficulties reported in previous trials and our complete patient follow up increases the accuracy of our findings.

The limitations of the study, however, are that we powered the study to detect a 30% difference in self-administered analgesia use between the intervention and placebo groups and therefore smaller effect sizes will not be detected. Nevertheless, it could be argued that the cost of using infusion catheters warrants a reasonably moderate reduction in analgesia use, particularly when previous studies have not found their use to be effective in decreasing length of stay in hospital. Our relatively small sample size, however, has meant that we ended up with more women in the intervention arm being intolerant of nonsteroidal anti-inflammatory drugs.

A further limitation in our design is that it was a single-site study, which reduces the generalisability of the results. We also used a nonvalidated method to calculate the total oral drug dose used, largely because there were no satisfactory validated measures available. Although this was a secondary outcome measure, it would be useful to validate it for use in future studies. Although our failure to standardise the anaesthetic, surgical procedure and postoperative pain management creates some potential biases, it has the advantage of reflecting the ‘real world’ scenario.


Our findings do not fully support those reported by Sherwinter et al.[5] in what appears to be the only other placebo-controlled trial of intraperitoneal LA following laparoscopic surgery.[5] In this trial of 30 patients undergoing laparoscopic gastric banding, the authors report a significant reduction in the amount of self-reported pain experienced by patients receiving a peritoneal infusion of bupivicaine, although, similarly to our findings, patients' morphine requirements and hospital length of stay were not reduced. As the authors suggest, their failure to find a reduction in morphine use may have arisen from their absence of PCA, which makes the results difficult to interpret.

We also found that our results are not in total agreement with the few RCTs of postoperative intraperitoneal infusion of LA following open surgery. Similarly to our findings, Gupta et al.[6] reported similar self-reported pain levels in women receiving an intraperitoneal infusion of LA for 24 hours following total abdominal hysterectomy and those receiving a placebo. In contrast to our findings, the authors report a reduction in PCA opioid use for up to 24 hours by women receiving the LA compared with those receiving the placebo. One possible explanation for this is that a postoperative intraperitoneal LA infusion may be more effective at relieving local incisional pain than visceral pain, making it less effective following laparoscopic surgery. Indeed, a systematic review and meta-analysis of intraperitoneal LA agents given following open surgery suggests that there may be some analgesic benefits,[7] although the quality of the studies and variability in anaesthetic methods included in this review makes the findings very difficult to interpret.

Given that we found a lack of analgesic benefit from an intraperitoneal infusion of bupivacaine, it is unsurprising that we did not find a reduction in hospital length of stay.


Intraperitoneal infusion of LA following laparoscopic hysterectomy does not offer any analgesic or opioid-sparing benefits and may result in women requesting more oral analgesia than those receiving an infusion of normal saline. It would appear, therefore, that the benefits associated with intraperitoneal local anaesthetic agents are restricted to their use before surgery and there does not seem to be any additional advantage to continuing an infusion following surgery.

Disclosure of interests

We have no conflicts of interest to declare.

Contribution to authorship

VA, JTW and KB designed the study. VA, JTW, FZ and SB collected the data. VA and KB were responsible for data analysis. VA, JTW and KB wrote the article. All authors interpreted the data and reference articles. All authors reviewed and approved the final version of the paper.

Details of ethics approval

The protocol was approved by the London–Surrey Borders Research Ethics Committee. REC Reference 10/H1109/69; date of approval 2 October 2010.


No funding was obtained for this study.


We thank all of the women who participated in the study.

The pursuit of the best postoperative pain control

  • T Tulandi

  • Department of Obstetrics and Gynecology, McGill University, Montreal, QC, Canada

Mini commentary on ‘Continuous infusion of local anaesthetic following laparoscopic hysterectomy—a randomised controlled trial’

Postoperative pain is most intense in the first 2–3 days after surgery and over half of women rate this pain as moderate to severe (Apfelbaum et al. Anesth Analg 2003;97:534e540). Reducing the invasiveness of surgery may help to minimise postoperative pain. For example, laparoscopic hysterectomy has been shown to be associated with less postoperative pain than traditional abdominal hysterectomy (Garry et al. BMJ 2004;328:129). Despite this potential advantage, laparoscopic surgery is still associated with postoperative pain, which is due to the surgical incisions themselves, the creation of a pneumoperitoneum, induced stretching of the abdominal cavity, dissection of the abdominal viscera, subsequent peritoneal inflammation and, in some cases, direct nerve injury. Peritoneal manipulation induces visceral stimulation sending a signal through the afferent pain pathway sensitising dorsal horn neurons.

In this issue of the journal, Andrews et al. (BJOG 2014; DOI: 10.1111/1471-0528.12610) attempted to attenuate this pain response by using a continuous intraperitoneal infusion of local anaesthetic after laparoscopic hysterectomy. Women were randomised to receive either 0.5% levobupivacaine or physiological saline for 48 hours postoperatively. Compared with the control group, the amount of pain and the quantity of morphine used by the women were comparable. Unexpectedly, the intervention group required more analgesia in the first 48 hours after surgery and the authors postulated that this was a result of the acidity of bupivacaine. These findings are supported by another randomised study of laparoscopic hysterectomy where 100 mg of bupivacaine diluted in 100 ml of normal saline was instilled into the peritoneal cavity at the completion of the procedure (Arden et al. J Minim Invasive Gynecol 2013;20:620–6). Compared with the placebo group, the amount of postoperative pain and the amount of opioid use were similar. It seems that intraperitoneal analgesia administered at the completion of surgery is ineffective.

An alternative method to decrease postoperative pain is ‘pre-emptive anaesthesia’. Local anaesthesia is administered before tissue injury occurs and this may reduce pain by blocking receptor activation and decreasing the production or activity of pain neurotransmitters. In a meta-analysis of randomised trials, decreased analgesic use was reported but the postoperative pain scores in women who received pre-emptive local anesthetic wound infiltration were similar to scores in the control group. (Ong et al. Anesth Analg 2005;100:757). It seems that the effect of local pre-emptive analgesia at best is modest but better than that of local injection of analgesia at the completion of surgery. Systemic pre-emptive analgesia appears to be more promising. Several agents including clonidine, ketorolac, ibuprofen and gabapentin have been shown to reduce postoperative analgesic requirement by at least one-third in the first 24 hours after surgery.

Alleviating postoperative pain is important if we are to enhance recovery and minimise morbidity post surgery. Randomised studies evaluating strategies to reduce postoperative pain after laparoscopic surgery are still needed. At present, postoperative pain control appears to be best achieved by the administration of systemic pre-emptive analgesia followed by narcotics and nonsteroidal anti-inflammatory drugs in the postoperative period.

Disclosure of interests

T. Tulandi is on the advisory board for Actavis Inc.