EDITOR:

Thoracic epidural analgesia (TEA) is an effective technique in major abdominal surgery, but in daily clinical routine there is still no consensus on whether it should be used intraoperatively or only postoperatively. One benefit of intraoperative use of TEA should be lower analgesic requirements [Reference Katz, Cohen, Schmid, Chan and Wowk1], but the conclusions have been inconsistent [Reference Moiniche, Kehlet and Dahl2,Reference Ong, Lirk, Seymour and Jenkins3]. This double-blind randomized study was designed to compare two strategies that are often used in clinical practice: combined vs. balanced anaesthesia. We focused on the length of postoperative patient-controlled epidural analgesia (PCEA), because PCEA is an expensive form of analgesic treatment with costs proportional to duration [Reference Bartha, Carlsson and Kalman4]. Our aim was to determine whether intraoperative TEA could reduce the time of postoperative epidural treatment.

Ethics Board approval was obtained, and all patients gave written informed consent. All patients scheduled for major abdominal surgery by laparotomy were screened. Inclusion criteria were age above 18 yr, ASA physical status I-III and knowing the German language. Exclusion criteria were contraindications to epidural analgesia or use of opioid analgesics during the past month. Patients were randomized to either intraoperative combined epidural/general anaesthesia (ICEA) or balanced anaesthesia (control, non-ICEA). Postoperatively both groups received PCEA.

All patients were administered midazolam 7.5 or 3.75 mg, according to weight and age, 45 min before transfer. Crystalloids were infused at 10 mL kg⁻¹ over 15 min. Then all patients underwent thoracic epidural catheterization using the midline approach and loss-of-resistance technique. Spinal placement was excluded by aspiration and the injection of a test dose of 2–3 mL bupivacaine 0.5%. Infusion of crystalloids was continued at 10 mL kg⁻¹ h⁻¹. General anaesthesia was induced in both groups with 3.0–5.0 mg kg⁻¹ thiopental, 1.5–3.0 μg kg⁻¹ fentanyl, 0.1–0.2 mg kg⁻¹cis-atracurium or 1.0–1.5 mg kg⁻¹ succinylcholine.

The ICEA group received a 6–8 mL epidural bolus of sufentanil 1 μg mL⁻¹ and ropivacaine 0.16%, 30 min pre-incision. Anaesthesia and analgesia were maintained with isoflurane in oxygen/air and continuous epidural infusion of sufentanil 1 μg mL⁻¹ and ropivacaine 0.16% at 6 mL h⁻¹. If necessary, additional epidural boluses of sufentanil/ropivacaine were administered or the rate of the epidural infusion was modified, to maintain a bispectral index (BIS) of 40–50 and clinical signs of sufficient anaesthesia.

The control group received a combination of isoflurane in oxygen/air and fentanyl intravenously (i.v.). Shortly before the skin incision, 4–6 μg kg⁻¹ fentanyl was injected. During surgery, fentanyl injection was repeated as necessary to achieve a BIS of 40–50 and clinical signs of sufficient anaesthesia. TEA was started with an 8 mL bolus of 1 μg mL⁻¹ sufentanil in ropivacaine 0.16% at the closure of the abdomen. A continuous epidural infusion of the same sufentanil/ropivacaine solution was started immediately at 6 mL h⁻¹.

Postoperatively, both groups received continuous TEA (sufentanil 1 μg mL⁻¹ and ropivacaine 0.16%), starting with 6 mL h⁻¹. Additional patient-controlled analgesia via bolus was possible (4 mL bolus, 20 min lockout). The block was tested and documented immediately after the patients awoke. The 24 h sufentanil/ropivacaine requirement, additional analgesic demand and complications of epidural analgesia were documented twice daily thereafter by staff blinded to the patients’ study assignments. Pain was evaluated every 4 h on a 0–100 numeric rating scale (NRS). A protocol algorithm was used for reducing the flow rate of epidural analgesics and for termination of postoperative PCEA.

It was calculated that 46 patients would be needed in each group, to detect a 1-day reduction in PCEA treatment duration, at a power of 80% and P < 0.05. The data were analysed by χ ²-test, paired two-tailed t-tests, Mann–Whitney rank sum test, Kaplan–Meier technique and log rank test. The analysis was undertaken using SPSS 12.0 (SPSS Inc., Chicago, IL, USA).

We screened 229 consecutive patients, recruited 136 subjects and had 95 complete datasets available (53 controls and 42 ICEA). The two groups did not differ significantly in patient characteristics, ASA classification, pre-existing medical conditions, diagnoses or surgical procedures. There was no difference in duration of surgery (mean 4 h), end-tidal isoflurane concentration at any time, extubation time, fluid intake, fluid loss or mean arterial pressure. The desired BIS of 40–50 was maintained during anaesthesia in both groups. The control group needed a mean of 0.75 mg (2.8 μg kg⁻¹ h⁻¹) i.v. fentanyl during surgery, whereas the ICEA group needed a mean of 39 μg sufentanil epidurally (0.13 μg kg⁻¹ h⁻¹) but no fentanyl.

Intraoperative epidural analgesia did not reduce the duration of postoperative pain treatment compared to balanced anaesthesia with i.v. opioid. The PCEA was stopped after 5 days on average, independent of whether TEA was started early, intraoperatively or postoperatively. The epidural analgesic consumption and supplemental analgesic requirements were identical in both groups. There were no significant differences in any of the four measures of dosage of sufentanil and ropivacaine: total (Fig. 1), basal, additional patient-initiated boluses or doctor-initiated additional boluses. Additional analgesia using another route such as oral or i.v. was necessary only in about 10% of the patients, independent of the study group. The pain intensity was comparably low in both groups. About two-thirds of the patients had no pain most of the day, regardless of whether it was the first or last day of PCEA treatment.

Figure 1

Postoperative daily epidural sufentanil and ropivacaine requirement. Values are means+SD. There were no significant differences in these factors betweeen the two groups. ICEA: intraoperative combined epidural anaesthesia (intervention group); non-ICEA: control group with balanced anaesthesia.

A possible explanation as to why intraoperative TEA was no better than postoperative TEA is that the control group received entirely sufficient intraoperative analgesia. Fentanyl was i.v. administered in our study at the high dosage of 2.8 μg kg⁻¹ h⁻¹. Our results are consistent with those of Burmeister and colleagues [Reference Burmeister, Gottschalk and Freitag5], who used 37 μg kg⁻¹ h⁻¹ sufentanil, thus an equianalgesic dose of 3.7 μg kg⁻¹ h⁻¹ fentanyl [Reference Cherny6]. By contrast, Katz and colleagues applied only about 1.5 μg kg⁻¹ h⁻¹ fentanyl in the control group and found that they did not do as well as the experimental group [Reference Katz, Cohen, Schmid, Chan and Wowk1]. Most other studies have not reported the opioids used in the control group.

There were no severe complications from the epidural technique. Pulmonary and abdominal complications occurred, but there were no advantages for either group. There was no significant difference in ICU duration (mean 1.5 days) or total hospital stay (mean 24 days).

The rate of nausea was 21% in non-ICEA patients vs. 7% in ICEA patients (P = 0.06), and the rate of vomiting was 8% vs. 2% (P = 0.28). Opioids are a risk factor for nausea and vomiting with a strong dose–response relationship [Reference Roberts, Bekker, Carlsen, Moffatt, Slattery and McClure7]. It may be assumed that the equianalgesic dose ratio between the two study groups is 10 : 1 [Reference Cherny6]. Comparing the equianalgesic doses (ICEA 39 μg vs. non-ICEA 75 μg sufentanil for the entire surgery time), the opioid dosage of the ICEA group seems to have been lower, perhaps due to effect of the local epidural anaesthetic.

In conclusion, TEA did not reduce the duration of postoperative PCEA treatment, extubation time, complication rates, readmissions, ICU stay or hospitalization time. The most plausible explanation is that the control group received entirely adequate analgesia, which could not be improved by intraoperative TEA. At least on the basis of our results, there does not appear to be any reason to prefer intraoperative TEA or i.v. opioids, so long as the patient receives sufficient analgesia.