Interpleural analgesia for attenuation of postoperative pain after hepatic resection


  • Presented in part at the Australian & New Zealand College of Anaesthetists’ Annual Scientific Meeting, May 2009, Cairns, Australia.

Correspondence to: Dr Laurence Weinberg


We performed a prospective randomised trial to evaluate the analgesic efficacy of interpleural analgesia in patients undergoing hepatic resection. The control group (n = 25) received multimodal analgesia with intravenous morphine patient-controlled analgesia; in addition, the interventional group (n = 25) received interpleural analgesia with a 20-ml loading dose of levo bupivacaine 0.5% followed by a continuous infusion of levobupivacaine 0.125%. Outcome measures included pain intensity on movement using a visual analogue scale over 24 h, cumulative morphine and rescue analgesia requirements, patient satisfaction, hospital stay and all adverse events. Patients in the interpleural group were less sedated and none required treatment for respiratory depression compared to 6 (24%) in the control group (p< 0.01). Patients in the interpleural group also had lower pain scores during movement in the first 24 h. Patients’ satisfaction, opioid requirements and duration of hospital stay were similar. We conclude that continuous interpleural analgesia augments intravenous morphine analgesia, decreases postoperative sedation and reduces respiratory depression after hepatic resection.

Derangements in coagulation after hepatic resection have favoured systemic multimodal analgesia in preference to neuraxial techniques as the preferred method for effective delivery of postoperative analgesia in many institutions. Several studies report the use of interpleural local anaesthetic solutions as an effective technique for analgesia following unilateral upper abdominal surgery [1–19]; however, it is unknown if interpleural analgesia is effective for patients receiving a right subcostal surgical incision with a midline abdominal incision extension (reverse-L incision), a commonly used surgical incision for patients undergoing this procedure. For patients undergoing hepatic resection we think that interpleural analgesia is a beneficial analgesic adjunct to systemic morphine patient controlled analgesia. However, there are no studies to support its use in this setting. Interpleural analgesia is considered less effective for patients receiving midline incisions as contralateral phrenic, somatic, and visceral nerve afferents that innervate the upper central abdomen may not be blocked by the interpleural solution. Therefore we hypothesised that for patients undergoing elective open hepatic resection receiving a reverse-L incision, a continuous interpleural infusion of local anaesthetic solution in combination with multimodal analgesia using intravenous morphine patient-controlled analgesia (PCA) provides more effective postoperative analgesia and fewer complications than multimodal analgesia with morphine PCA alone.


We performed a prospective randomised study from January 2007 to December 2008 after approval from our hospital research ethics committee. We studied patients of ASA physical status 1–4 undergoing elective hepatic resection requiring a right subcostal rooftop incision with a midline extension (reverse-L incision). Written informed consent was obtained at the pre-operative admission clinic 1–2 weeks before surgery. Exclusion criteria included patients with chronic opioid use (≥ 1−1 intravenous or ≥ 3−1 oral morphine for a period > 1 month), known allergy or intolerance to morphine or local anaesthetic solutions, pre-operative coagulopathy, moderate/severe renal impairment (serum creatinine > 200 μmol.l−1), or contra-indication to interpleural catheter insertion (local skin site infection, active pleural disease with or without pulmonary fibrosis, history of spontaneous pneumothorax, bullous emphysema).

For postoperative analgesia patients were randomly assigned to one of two groups: a control group receiving intravenous morphine delivered via a PCA system (PCA group); and an interventional group consisting of intravenous morphine delivered via a PCA system in combination with continuous interpleural analgesia with local anaesthetic solution consisting of L-bupivacaine 0.125%– the Interpleural group. Before the study, a random-number table was electronically generated specifying the group to which each patient would be assigned. For each patient, an opaque envelope containing the group assignment was prepared, sealed and sequentially numbered. On the morning of surgery the anaesthetist opened the envelope and allocated the patient into one of the two groups. All medical, surgical, physiotherapy, nursing and acute pain services staff managing the patients in the postoperative period were unaware of patients’ participation in this clinical trial.

Anaesthesia was conducted by a group of specialist anaesthetists using a protocol designed to standardise intra-operative opioid use and patient care. Induction of anaesthesia consisted of a balanced technique combining intravenous midazolam 0.02–0.03−1, propofol 1–3−1 and a non-depolarising neuromuscular blocker. Anaesthesia was maintained with desflurane at inspired concentrations of 0.5–1.5 MAC, in a 50% oxygen-air balance. Remifentanil 0.1–0.3 μ−1 was infused continuously for intraoperative analgesia. Routine monitoring included continuous electrocardiography, pulse oximetry, capnography, invasive arterial blood pressure, central venous pressure, urine output and core body temperature. Intraoperative normothermia was maintained with warm fluids and a forced-air warming device. In accordance to standard anaesthetic technique at our institution for this operation, during the prehepatic resection phase and for the duration of hepatic parenchymal resection, the administration of fluids was reduced and central venous pressure was maintained at < 6 mmHg. After hepatic resection, an infusion of warm fluids was administered to render patients euvolaemic. Urine output was maintained at > 0.5−1, and systolic blood pressure was maintained within 20% of the pre-operative value. During surgical closure of the abdomen both groups of patients received intravenous paracetamol 1 g and a loading dose of intravenous morphine 0.1–0.2−1 titrated to achieve a respiratory rate of < 15 breaths.min−1.

Before tracheal extubation, patients assigned to the Interpleural group were turned to the left lateral decubitis position. While breathing spontaneously, an interpleural catheter was inserted using a ‘passive loss of resistance’ or ‘negative pressure’ technique [20]. The site of insertion of the catheter was 8–10 cm from the spinous processes on the posterior midline of the back, above the seventh or eighth intercostal space. Following skin puncture with a 16-G Tuohy needle, sterile saline was used to ‘fill’ the Tuohy needle shaft. The needle was advanced through the intercostal space and then through the parietal pleura. Puncture of this layer was sometimes identified with a ‘click’. Upon entering the interpleural space, saline in the Tuohy needle shaft was ‘sucked in’ (due to the negative interpleural pressure), confirming the correct postion. An epidural catheter was threaded 8 cm into the interpleural space in a medial and cephalad direction. The catheter was aspirated to exclude intravascular placement and a test dose of 2 ml levo-bupivacaine 0.5% was administered. The patient was then placed in supine position and a loading dose of 20 ml levo-bupivacaine 0.5% with adrenaline 1:200 000 was administered, as previously described [1, 2, 6–8, 13, 16, 19]. To ensure even and optimum spread of the local anaesthetic solution, the patient was left in the supine position for 10 min [20]. The trachea was extubated and the patient transferred to the post-anaesthesia care unit (PACU). A continuous infusion of levo-bupivacaine was commenced at 12.5 mg.h−1. Patients therefore received levo-bupivacaine 300 mg per day (excluding initial loading dose), a 24-h dose range that has been shown to be effective in other studies [1, 9, 18].

In the PACU both groups of patients received an intravenous morphine PCA infusion attached to the side arm of their intravenous infusion line. Standard instructions on the use of the PCA device were reinforced, and the device was set to deliver a 1-mg bolus with a 5-min lockout, and a maximum dose of 12 mg.h−1 (standard practice at our institution). Patients were reviewed in the postoperative period by a dedicated acute pain service to ensure their analgesic regimen was optimal. In the absence of any postoperative liver function test derangements, all patients received paracetamol 1 g every 6 h for the first 24 h. Patients who developed breakthrough pain were encouraged to use their morphine PCA. Rescue analgesia was controlled and protocol-driven by a dedicated acute pain service: a visual analogue score (VAS) > 5 at rest or movement was managed with additional boluses of morphine, maintaining a respiratory rate > 8 breaths.min−1. After 30 min, if pain was still not controlled, intravenous tramadol 2−1 was added; if no improvements were reported within an hour, a standardised continuous ketamine infusion was commenced at 0.1–0.2−1. Lastly, if VAS scores remained > 5, intercostal nerve blocks were administered. The interpleural catheter was managed according to our hospital’s acute pain management guidelines.

The primary end point measured was pain during movement assessed on a VAS using a standardised 0–100 mm line, ranging from 0 (no pain), to 100 (the most severe pain), over a 24-h period. The VAS score was averaged over the 24-h period and the mean reported. Each patient had the same number of measurements. The first VAS measurement was made on arrival in PACU and then every 10 min. Once transferred to the ward or high dependency unit, VAS measurements for pain at rest and on movement were made every 3 h. Secondary end points included cumulative morphine consumption (recorded every 6 h), requirements for additional rescue analgesia, and complications of opioid analgesia (sedation score, respiratory depression, VAS scores for nausea and pruritus, return of bowel function). Complications rates and outcomes were collected for 30 days after surgery. A modified sedation score between 0 and 3 (0 = awake and alert, 1 = mild drowsiness/eyes open to speech, 2 = very drowsy/eyes open when shaken, 3 = somnolent or unrousable) was recorded [21]. Respiratory depression was defined as a respiratory frequency < 8 breaths.min−1, oxygen saturation < 90% or Paco2 > 9.3 kPa requiring medical intervention. Satisfaction score for pain control on a five-point scale was recorded before discharge from hospital. In addition, duration of hospital stay and any postoperative adverse event were recorded. The following symptoms of levobupivacaine toxicity were recorded: dizziness; perioral numbness; muscle weakness; muscle twitching; and confusion.

Previous surveillance data for patients undergoing open hepatic resection at our institution indicated that the mean (SD) 24-h VAS score for pain with movement was 60 (25). Twenty-five patients per group thus provided an 80% power for detecting a 30% difference in pain scores at a significance level of 0.05. Continuous variables were compared using the t-test or the Mann–Whitney test depending on observed distribution of the data. Categorical variables were analysed using the chi-squared or the Fisher’s exact test. Serial measurements were analysed using repeated measures ANOVA; any significant overall differences between groups was investigated with t-tests at individual time intervals and the Bonferroni correction was performed on the raw p value. A p value < 0.05 was defined as a significant difference. All analysis was performed using Stata™ version 10 software (Statacorp, College Station, TX, USA).


Sixty-five patients consented for this study (Fig. 1). Fifteen were excluded because of an inappropriate surgical incision (conversion of a right subcostal ‘rooftop’ incision to full midline incision) or because the procedure was abandoned due to inoperable disease. Fifty patients were randomised: 25 patients in the Interpleural group and 25 in the PCA group. The groups were similar with regard to baseline characteristics, surgical pathology, co-existing morbidity, and ASA physical status (Table 1). The most frequent indication for surgery was metastatic colorectal disease. Baseline intra-operative variables, including duration of surgery, size of the hepatic resection, intra-operative blood loss, transfusion requirements and intra-operative fluids administered were similar in both groups and summarised in Table 2.

Figure 1.

 Trial CONSORT diagram.

Table 1.   Characteristics of patients receiving interpleural analgesia with morphine patient controlled analgesia (Interpleural group) or morphine patient controlled analgesia alone (PCA group). Values are number, mean (SD), median (IQR [range]) or number (proportion).
 Interpleural groupPCA group
 (n = 25)(n = 25)
  1. *Addison’s disease, thalassaemia minor, diabetes mellitis.

Sex: M:F10:1511:14
Age; years56.0 (43–72 [19–84])58.0 (38–66 [23–77])
Weight; kg64 (13.7)76.5 (17.8)
Body mass index; kg.m−221.7 (3.9)25.8 (6.9)
 Cardiovascular disease8 (32%)6 (24%)
 Pulmonary disease3 (12%)2 (8%)
 Renal impairment1 (4%)1 (4%)
 Other diseases*3 (12%)6 (24%)
 Malignant:benign22 (88%):3 (12%)24 (96%):1 (4%)
 ASA grade
  11 (4%)2 (8%)
  221 (84%)20 (80%)
  33 (12%)3 (12%)
Table 2.   Baseline intra-operative variables of patients receiving interpleural anal-gesia with morphine patient controlled analgesia (interpleural group) or morphine patient controlled analgesia alone (PCA group). Values are mean (SD), or number (proportion).
 Interpleural groupPCA group 
 (n = 25)(n = 25)p value
  1. CVP, central venous pressure.

Duration of surgery; min271 (82.1)326 (130.9)0.08
CVP; mmHg
 Baseline8 (3.5)8 (3.8)0.44
 Pre-hepatic resection5 (3.6)7 (2.2)0.06
 Post-hepatic resection5 (3.1)5 (2.4)0.67
Lowest temperature; °C35.2 (0.5)35.1 (0.6)0.85
Intra-operative fluids
 Crystalloids; ml
  Hartmann’s solution2032 (1110.3)2138 (869.1)0.70
 Colloids; ml484 (556.3)870 (927.4)0.08
  Gelofusine350 (444.8)550 (653.5)0.21
  4% Albumin134 (290.3)300 (595.1)0.22
Blood loss; ml928 (1130.6)1145 (861.8)0.74
Transfusion requirements
 Blood4 (16%)6 (24%)0.73
  Volume transfused; ml131 (383.0)100 (195.4)0.77
 Fresh frozen plasma1 (4%)0 (0%)0.11
 Cryoprecipitate1 (4%)0 (0%)0.11
 Platelets1 (4%)0 (0%)0.1
Hepatic resection weight; g337 (404.0)309 (311.8)0.80

During the first 24 h postoperatively, the only significant difference in resting pain scores between the groups was found at 6 h (p = 0.01) (Fig. 2). However, pain intensity using the VAS score was less on movement in the Interpleural group compared with the PCA group for the first 24 postoperative hours (p = 0.02). The greatest difference in VAS score on movement was also found at 6 h postoperatively where the mean VAS on movement was 42 mm in the Interpleural group and 61 mm in the PCA group (difference 18 mm; 95% CI 4–32 mm; adjusted p = 0.04) (Fig. 3). At 24 h postoperatively, mean (SD) VAS scores during movement remained lower in the Interpleural group compared with the PCA group 44 (22) vs 51 (18) mm, respectively; p = 0.23). However after 48 h, VAS for pain at rest and during movement was similar in both groups. At 24 h postoperatively the cumulative mean morphine consumption was similar in the Interpleural PCA groups (Table 3). Patients in the PCA group were more sedated (p = 0.05) with 6 (24%) patients requiring medical intervention for respiratory depression compared with no patients in the Interpleural group (difference 24%; 95% CI 6–43%; p = 0.01).

Figure 2.

 Visual analogue scores for pain at rest (0–100 mm scale) vs time over the first 48-h period in patients receiving interpleural analgesia with morphine PCA (inline image) or morphine PCA alone (inline image). Data presented as mean and standard deviation. *p = 0.01.

Figure 3.

 Visual analogue scores for pain during movement (0–100 mm scale) vs time over the first 48-h period in patients receiving interpleural analgesia with morphine PCA (inline image) or morphine PCA alone (inline image). Data presented as mean and standard deviation. **p = 0.04.

Table 3.   Analgesic requirements, satisfaction levels, recovery profiles and adverse events in patients receiving interpleural analgesia with morphine patient controlled analgesia (interpleural group) or morphine patient controlled analgesia alone (PCA group). Values are mean (standard deviation), or number (proportion).
 Interpleural groupPCA group 
 (= 25)(= 25)p value
  1. *Myocardial infarction (ECG changes and myocardial enzyme elevation ((troponin I concentration > 0.06 mg.l−1), pulmonary oedema with CT/scintigraphic changes, arrhythmia (ECG changes requiring medical treatment or cardioversion).

  2. †Respiratory rate < 8 breaths.min−1 requiring medical intervention.

  3. ‡Elevated temperature with radiographic changes.

  4. §Renal failure requiring renal replacement therapy.

Cumulative morphine; mg
 24 h42.4 (27.4)44.7 (24.0)0.70
 48 h79.9 (53.0)85.6 (42.5)0.68
Patients requiring rescue analgesia
 24 h6 (24%)10 (40%)0.23
  Tramadol6 (24%)8 (32%)0.53
  Ketamine06 (24%)0.02
 48 h8 (32%)11 (48%)0.25
  Tramadol8 (32%)10 (40%)0.56
  Ketamine1 (4%)6 (24%)0.01
Analgesic side effects at 24 h
 Nausea13 (52%)17 (68%)0.38
 Vomiting2 (8%)7 (28%)0.14
 Pruritus2 (8%)2 (8%) 
 Dizziness2 (8%)2 (8%) 
 Peri-oral numbness/tingling00 
Patients requiring antiemetics
 24 h13 (52%)8 (32%)0.15
 48 h18 (72%)9 (36%)0.01
Postoperative intake; h
 Oral sips16 (12.8)20.5 (12.9)0.29
 Free fluids34 (28.2)30.0 (15.4)0.50
 Light diet48 (32.2)41 (22.4)0.30
Time to mobilise; h25 (14.6)24 (11.6)0.84
 Satisfied23 (92%)22 (88%)0.85
 Dissatisfied1 (4%)1 (4%) 
 Unable to recall1 (4%)2 (8%)1.00
Length of hospital stay; days8.4 (3.5)9.0 (4.8)0.62
Adverse events
 Cardiovascular complications*1 (4%)2 (8%)1.00
 Respiratory depression†060.02
 Pneumonia‡3 (12%)4 (16%)1.00
 Wound healing complications00 
 Renal complications§1 (4%)00.11

Ten (40%) patients in the PCA group required rescue analgesia compared with six (24%) patients in the Interpleural group (Table 3; p = 0.22). Requirements for the amount and type of rescue analgesia are summarised in Table 3. Nausea and vomiting, pruritus, satisfaction scores, time to mobilisation and duration of hospital stay were similar in both groups. There were no reported symptoms of local anaesthetic toxicity. Except for respiratory depression, other postoperative complications, adverse events and duration of hospital stay were comparable between the groups (Table 3).


We performed a prospective study evaluating the analgesic efficacy of continuous postoperative interpleural analgesia in patients undergoing open hepatic resection. We sought to define the analgesic efficacy of interpleural analgesia in this setting, evaluate the complications of interpleural and opioid analgesia, and determine complications rates and outcomes 30 days after surgery. We found that interpleural analgesia augmented PCA analgesia on movement in the first 24 h after surgery. Importantly, in the Interpleural group postoperative sedation and respiratory depression were decreased.

Patients receiving continuous interpleural analgesia experienced superior analgesia and had fewer opioid related side effects than those with morphine PCA alone. Analgesia remained superior in the Interpleural group until 24 h postoperatively; however at 48 h, pain scores were almost identical in both groups. The greatest difference in analgesia efficacy, at 6 h postoperatively, is most likely to be attributable to the initial interpleural loading dose of local anaesthetic. Our findings are consistent with other studies confirming the analgesic efficacy of interpleural analgesia after abdominal surgery involving surgical flank or unilateral incisions [1, 4–6, 8, 9, 14, 16]. The mechanism of action of interpleural analgesia is likely to be due to the diffusion or bulk flow of local anaesthetic into the subpleural space with a concomitant multiple segmental intercostal nerve block [22, 23].

Although the requirements for overall analgesia were similar in both groups, the type of rescue analgesia differed. Patients in the PCA group were more likely to require ketamine (24% vs 0%) in the first 24 h. Only one patient in the Interpleural group required ketamine for rescue analgesia and this was on the second postoperative day when we demonstrated the interpleural block to be ineffective. Patients receiving morphine PCA alone were more sedated and twice as likely to experience respiratory depression than patients in the Interpleural group. These adverse events were not associated with prolongation of hospital stay. The opioid sparing effect of ketamine and the higher incidence of excess sedation in the PCA group may explain why cumulative morphine requirements were similar in both groups. In the Interpleural group, the need for supplementary morphine was still of considerable magnitude, a consistent finding seen in other studies examining the efficacy of interpleural analgesia [4, 9, 15]. This may be due to patients’ experiencing pain on the contralateral side, or due to different nerve pathways innervating the upper abdomen, including phrenic, somatic, and visceral nerve afferents – some of which may not be blocked by the interpleural solution [14].

Numerous studies demonstrate an acceptable safety profile with bupivacaine administered via the interpleural route [6, 8, 12–14, 16, 19]. In our study we used a conservative dosing regime and found no reported symptoms consistent with local anaesthetic toxicity. Although uncommon, central nervous system toxicity has been demonstrated [24, 25] and as with other forms of regional anaesthesia, systemic toxicity is a potential risk. There were no pneumothoraces detected in this study, and the incidence of cardiovascular, renal, and wound healing complications were similar in both groups.

The optimum dose, volume, concentration and method (intermittent bolus vs continuous infusion) of the local anaesthetic solutions is yet to be defined, with considerable variation in current practice. The addition of adrenaline does not appear to affect the pharmacokinetics of bupivacaine delivered via the interpleural route, but its use may be associated with lower peak plasma concentrations [26]. However, independent of the method used to deliver interpleural analgesia (intermittent boluses vs continuous infusion), analgesia appears to be efficacious for unilateral abdominal incisions. After liver resection, the possibility of extending the analgesic benefits of interpleural analgesia beyond 24 h with either a more concentrated solution of local anaesthetic or an intermittent bolus technique warrants further investigation.

There are a number of limitations in this study. Whilst there are sufficient patients to detect a clinically important difference in the primary endpoint, we are unable to detect the true incidence of complications from interpleural analgesia due to the relatively small sample size. Plasma levo-bupivacaine levels were not evaluated as they have been well reported in numerous other studies [1, 6, 8, 12, 13, 16, 19, 26]. Although we found no reported symptoms of local anaesthetic toxicity, it is possible that plasma levels may have demonstrated concentrations above the accepted safe range. This is a single-centre study, which limits the external validity of our findings. However, our hospital has all the typical characteristics of a tertiary institution in a developed country and a recent comparative study confirmed that its patients and their outcomes were identical to those of other tertiary hospitals in Australia [27]. Finally, after careful consideration with our research ethics committee we felt it ethically undesirable to have a placebo-controlled arm due to the potential risks of the invasive intervention without any clinical benefit. The study is therefore non-blinded. We chose not to tape a placebo catheter to the skin of patients in the control group. In order to minimise bias between the two groups, medical and nursing staff, as well as the acute pain services staff managing these patients in the postoperative period were not aware of patients’ participation in this study. Postoperative pain management for all patients undergoing hepatic resection at our institution is managed with a protocol-driven, controlled and standardised analgesic regimen,described above. This regimen is the current acute pain service management protocol for all patients undergoing hepatic resection at our institution.

Our study also has several strengths. This is the first study evaluating the analgesic efficacy of interpleural analgesia in patients undergoing elective hepatic resection. The inclusions of patients with multiple co-morbidities make the findings of this study more consistent with the experience of day-to-day clinical practice. By defining the complication rate in these patients, we have defined both a need for improved care and the necessary background for the power calculations needed to design further interventional trials. Balancing the advantages of optimal analgesia against the risks associated with current techniques requires a body of clinical evidence that does not yet exist in patients undergoing hepatic resection. Although continuous neuraxial analgesia is an accepted technique for patients undergoing hepatic resection [28–32], severe derangements of coagulation can occur even in patients with normal pre-operative coagulation and liver function profiles [33–36]. Some institutions have questioned the safety of neuraxial techniques for hepatic resection [37–41] and systemic opioids remain the mainstay of postoperative analgesia in many centres. This study demonstrates that interpleural analgesia is an effective alternative to neuraxial techniques in this setting.

In conclusion, our evaluation of interpleural analgesia in patients undergoing elective hepatic resection has identified several important findings. We found that interpleural analgesia is a useful analgesic adjunct in postoperative multimodal analgesia after open hepatic resection, particularly on movement during the first 24 h post-operatively. Respiratory complications occur frequently in patients receiving intravenous morphine PCA alone, findings that are of particular interest as the higher rates of such complications provide good opportunities to evaluate strategies aimed at improving perioperative care. We have demonstrated that interpleural analgesia is a feasible alternative to using indwelling neuraxial catheter techniques for postoperative analgesia. Interpleural analgesia augmented PCA analgesia, decreased postoperative sedation and reduced the incidence of respiratory depression.


We gratefully acknowledge a research grant received from the Australian & New Zealand College of Anaesthetists. No competing interests declared.