The stomach is not an important site for absorption. Most substances are absorbed from the small intestine because of its very large surface area. However, the rate of gastric emptying is the rate-limiting step for the absorption of many orally administered substances, including acetaminophen (paracetamol). Postoperative delays in gastric emptying have been documented in several groups of patients, including those undergoing oesophagogastrectomy , cardiopulmonary bypass  and heart-lung transplantation . However, normal gastric emptying has been noted in other groups of patients after operation, including those who have undergone cholecystectomy receiving bupivacaine epidural anaesthesia , orthopaedic surgery with spinal anaesthesia  and general anaesthesia for minor surgical procedures . There is no information in the literature concerning gastric emptying in post-thoracotomy patients receiving fentanyl-bupivacaine epidural infusions for post-operative analgesia. Gastric emptying after thoracotomy has implications for anaesthetic practice: many patients are receiving long-term anti-anginal, anti-dysthythmic and other oral medication. Furthermore, a delay in gastric emptying leads to an increased volume of gastric contents, with an increased risk of pulmonary aspiration, increased likelihood of nausea and vomiting and a delay in restarting oral feeding . Fentanyl-bupivacaine epidural infusions are effective for the control of pain after thoracotomy and are widely used for this purpose . The aim of the present study was to determine the influence of a bupivacaine-fentanyl epidural analgesia on gastric emptying until the second postoperative day after thoracotomy in patients undergoing lung resection.
In a prospective repeated measures study, the effect of fentanyl-bupivacaine epidural analgesia at the mid-thoracic level on gastric emptying was assessed in 11 patients undergoing thoracotomy for lung resection. The local research Ethics Committee approved the study and written informed consent was obtained from 11 patients scheduled to undergo a thoracotomy with lung resection who fulfilled the study criteria (Table 1). No patient taking paracetamol or medication known to affect gastric motility was included in the study. The day of surgery was designated as 'day 0', the first postoperative day as 'day 1' and the second postoperative day as 'day 2'.
Under local anaesthesia, an intravenous (i.v.) cannula was sited on day 0 before surgery (study period 'preoperative'). After collecting a 5 mL blood sample (T0) patients received 1.5 g soluble paracetamol (Sterwin Medicines, Guildford, UK; batch no. SB59) in 100 mL water. Patients remained in bed reclined at approximately 15° for 120 min. The cannula was reserved for the study and was not used for drug or fluid administration. Further samples were collected into lithium heparin bottles 10, 20, 30, 45, 60 and 120 min after the administration of paracetamol.
The process of administering paracetamol and collecting blood samples was repeated on day 0, 4 h after the operation (study period '4h') and on the morning of day 2 (study period day 2). At the 4 h and day 2 study periods, an additional blood sample was taken at T0 from each patient and used to measure plasma fentanyl concentration.
Patients were premedicated with diazepam 5 or 10 mg orally immediately after the last preoperative blood sample (T120) had been taken. When the patients arrived in the anaesthetic room, venous and radial arterial cannulae and a thoracic epidural catheter were inserted under local anaesthesia. The thoracic epidural catheter was sited in all patients between the T5 and T7 interspaces. After administering a 3 mL test dose of 0.5% bupivacaine with 1:200 000 epinephrine, patients received a 0.1 mL kg−1 epidural bolus of a solution containing 10 μg mL−1 fentanyl in 0.1% bupivacaine. Operative and postoperative analgesia was provided by an epidural infusion initially at 0.1 mL kg−1 h−1. The rate was then adjusted as clinically appropriate between 1 and 10 mL h−1. Patients received a standard general anaesthetic. Propofol was used for the induction of anaesthesia. Neuromuscular blockade was achieved and maintained with atracurium. A fibreoptic bronchoscope aided the accurate placement and the maintenance of the correct positioning of a double-lumen endobronchial tube. Cefotaxime 1 g was given as an antibiotic prophylaxis. Anaesthesia was maintained with isoflurane in oxygen. Intraoperative arterial pressure was maintained to within 20% of preoperative values by altering the inhaled concentration of isoflurane, or by administering i.v. fluids and phenylephrine boluses as appropriate. Patients received neostigmine and glycopyrrolate to antagonize residual neuromuscular blockade at the end of surgery and were extubated in the operating theatre. Deviations from the planned surgery were noted, including extension of the planned resection, extension below the diaphragm and vagotomy or vagal damage. Any additional drugs given, including β2-agonists and antagonists, or any adverse events were noted, as was the time of admission to the recovery room.
In all patients, the effectiveness of epidural analgesia was ascertained in the recovery room by noting the loss to cold sensation (using ice) over the chest wall and the area of the thoracotomy incision. If the extent of block was inadequate for postoperative analgesia, a further 5 mL bolus of the epidural solution was administered epidurally. If this failed to provide adequate analgesia, the patients were with-drawn from the study.
Patients who fulfilled the postoperative criteria (Table 1) were studied 4 h after surgery (study period 4 h). Subsequently patients were fasted from the night of day 1 and were then studied again on the morning of the next day (study period day 2). Prochlorperazine and ondansetron were used as antiemetics, if required. No patient received metoclopramide, erythromycin, naloxone, cisapride or dopamine postoperatively. Additional paracetamol or anticholinergic agents were not administered to patients.
Serum from the blood samples was analysed for paracetamol concentration using enzymatic cleavage of the amide bond and spectrophotometric analysis of the p-aminophenol moiety when it reacted with o-cresol in ammoniacal copper. This assay is specific for the parent compound and does not detect paracetamol metabolites (predominantly glucuronide, sulphate, cysteine and mercapturate conjugates) .
Plasma fentanyl concentration was measured using a radioimmunoassay method.
The maximum serum concentration (cmax) was documented and used as the measure of gastric emptying for liquids. The time taken to reach the maximum serum concentration (tmax) was recorded. The area under the paracetamol concentration-time curve at 120 min (AUC120) was calculated using the linear trapezoidal rule. A power calculation based on previous publications showed that seven patients would be required for a 50% change in cmax at P = 0.05 and 80% power.
Data were analysed using SAS for Windows, v.6.12, and repeated measures for analysis of variance was performed with corrections for multiple comparisons. P < 0.05 was taken to indicate a significant difference between the pre- and postoperative periods.
Eleven patients were recruited into the study. Two patients were excluded postoperatively, one due to vomiting and the other due to incomplete blood sampling on day 2. No patient was excluded postoperatively due to inadequate epidural analgesia.
Patient characteristics and operative details for the remaining nine subjects are summarized in Table 2. In no patient was the surgery extended below the diaphragm. No patient underwent a vagotomy. There was no visual indication of any perioperative vagal damage in any of the patients. Table 3 and Figure 1 summarize the paracetamol results for the three study periods. The mean (SD) plasma fentanyl concentrations at 4 h and day 2 were 0.8 (0.3) and 1.6 (0.4) ng mL−1 respectively. The mean volume of the epidural solution infused intraoperatively and until the end of study period day 3 was 189.3 (SD 42.9) mL.
The causes of postoperative delays in gastric emptying are multifactorial and previous studies provide conflicting results. Although postoperative delays in gastric emptying for liquids and drug absorption are common in many groups of surgical patients, in some studies involving abdominal  or pelvic surgery , it has not been possible to differentiate the effect of epidural analgesia from the effect of surgery itself on the gut. There is no reference in the literature concerning postoperative gastric function in patients undergoing thoracotomy and lung resection with or without epidural analgesia.
Gastric emptying can be measured directly by several methods, many of which are invasive or require the administration of radioactive isotopes . However, indirect measurement of gastric emptying for liquids by paracetamol absorption correlates well with more direct measurement of gastric emptying [11,12] and has been used by a number of investigators [2,4,11-14].
Paracetamol, a weakly acidic para-aminophenol derivative with a pKa = 9.59, is rapidly absorbed from the small intestine by passive diffusion, and its serum concentration can be easily and accurately assayed. Insignificant paracetamol absorption occurs in the stomach and it does not undergo significant first-pass metabolism or enterohepatic recirculation . Thus, gastric emptying is the rate-limiting step for the absorption of paracetamol. In the present study, the parameters measured were cmax (maximum serum concentration of paracetamol), tmax (time taken to reach cmax) and AUC120 (area under the paracetamol concentration-time curve at 120 min, reflecting the total dose of drug absorbed in that time). AUC120 was selected as a parameter for comparison between study periods rather than for single sample measurements at fixed time intervals as rapid absorption or emptying could have resulted in a sampling during the downstroke of the curve and slow absorption during the upstroke . Soluble paracetamol was chosen to avoid the stages of drug disintegration and dissolution. A standard formulation was used to avoid previously noted differences in the pharmacokinetics of different drug formulation . Patients acted as their own preoperative controls to avoid any intersubject variability in paracetamol absorption observed before .
The process of gastric emptying is variously described as linear  or monoexponential [6,11]. Many drugs used perioperatively, such as metoclopramide, erythromycin, naloxone, cisapride or dopamine, affect gastric emptying. These drugs were not used in the present study. Patients were not included in the study if they were taking drugs known to affect gastric motility, and no such drug was administered postoperatively. Prochlorperazine and ondansetron were used as antiemetics if required. Additional paracetamol was not administered to patients in the postoperative period.
All patients were premedicated with diazepam 5 or 10 mg orally. Propofol was used for induction of anaesthesia. Benzodiazepines  and light propofol sedation  have no significant effect on paracetamol absorption. Gastric emptying and paracetamol absorption are increased if the subjects are ambulatory . In the present study, patients were not ambulant during the course of the sampling. The use of anticholinergic drugs or vagotomy at the time of operation can cause decreased gastric emptying. The only anticholinergic used was a single dose of 0.5 mg glycopyrrolate administered to antagonize residual neuromuscular paralysis at the end of operation. None of the patients underwent a vagotomy, and observable intraoperative vagal damage did not occur.
It was not possible to establish the exact cause for the delay in gastric emptying seen in the patients. The trauma of surgery could have contributed, but we consider a major effect to be unlikely as major, non-abdominal, injuries have minimal effect on gastric emptying  and delays in gastric emptying are not seen after all surgery . The role of general anaesthesia per se on gastric emptying is difficult to elucidate and, in the absence of systemic opioids and the use of nitrous oxide, may have no significant effects . We also avoided the use of any drugs perioperatively that were known to affect gastric motility.
The stomach receives its innervation through branches of the vagus nerve and T6-T10 sympathetic nerves . The main effect of a fentanyl-bupivacaine epidural infusion on gastric emptying is likely to be an interaction between the sympathetic blockade caused by the local anaesthetic increasing the rate of gastric emptying and the gastric hypomotility caused by the opioid fentanyl. It could be expected that epidural local anaesthetics alone would hasten gastric emptying by blocking the sympathetic supply to the stomach. However, it has been shown previously in human volunteer studies that thoracic epidural analgesia using bupivacaine to block the sympathetic efferents to the stomach (T6-T10) does not influence gastric emptying [20,21].
It is known that systemically administered opioids have a major influence on gastric emptying. Systemic opioids were not used in the present study. However, epidural opioids can cause a greater delay in gastric emptying when compared with systemic opioids, possibly due to an additional central effect [20,22]. It is speculated that the epidural administration of fentanyl combined with its subsequent systemic absorption would have contributed largely to the delayed gastric emptying seen in the present study. Pharmacokinetic studies of epidurally administered opioids such as fentanyl and alfentanil reveal that plasma opioid concentrations after large bolus doses and infusions administered epidurally are similar to plasma levels produced by i.v. infusions [23,24]. The mean plasma fentanyl levels of 0.8 ng mL−1 at 4 h and 1.6 ng mL−1 at day 2 in patients are similar to those reported in patients receiving a fentanyl infusion for postoperative analgesia .
A delay in gastric emptying after operation can lead to several complications, including nausea and vomiting, reflux or regurgitation with aspiration and resulting respiratory complications and delayed drug absorption . It has been suggested that delays in gastric emptying will have little effect on the total amount of drug absorbed but will effect the rate of absorption [11,16]. In the present study, there was not only a delay in reaching cmax, but also a significant decrease in AUC4h and AUCday2. Therefore, there was not only a delay in gastric emptying when compared with preoperative values, but also it affected adversely the amount of paracetamol absorbed at 120 min.
Drugs that are normally absorbed rapidly after oral administration and for whom the rate of absorption is important include analgesics and antidysrhythmics . Our practice with postoperative lung resection patients includes reintroduction of enteral liquids and medication on the first postoperative night, and many of our patients are receiving regular antidysrhythmic and anti-anginal agents. Drugs such as calcium channel-blocking drugs are normally rapidly absorbed after oral administration with clinical effects being manifest within 30-60 min of oral absorption . We could expect a diminished effect of these drugs in the postoperative period when gastric emptying is delayed. Similarly, delayed gastric emptying retards digoxin absorption, and in about 10% of subjects, a substantial fraction of the orally administered drug is converted to inactive products by intestinal microorganisms before absorption . It could be envisaged that delayed absorption would lead to increased destruction of digoxin in these patients. Certain other drugs may be destroyed largely by remaining in the stomach in contact with low gastric pH before intestinal absorption . Moreover, gastric hypomotility may result in orally administered drugs being vomited. This has important consequences in the light of recent evidence indicating an increased incidence and duration of postoperative tachydysrhythmia and myocardial ischaemia in patients undergoing thoracotomy, associated with an adverse outcome .
Can this delay be reduced in future? The use of plain bupivacaine epidural blockade alone, even though having little or no effects on gastric emptying, may cause problems with suboptimal analgesia with low concentrations or cardiovascular instability at higher concentrations . A possible solution is to use a lower concentration of fentanyl in the epidural infusion solution, although pain due to inadequate analgesia may, in itself, cause a delay in gastric emptying. The optimal concentration of fentanyl in 0.1% bupivacaine for postoperative analgesia remains to be established. Thoracic epidural analgesia still remains the gold standard for pain control in our group of patients. Other analgesic techniques used to manage post-thoracotomy pain, such as patient-controlled analgesia (PCA) using opioids or paravertebral blockade (often needing supplementation with a PCA system ), are unlikely to offer substantial advantages with regards to gastric emptying. The present anaesthetic technique, including antagonism of neuromuscular blockade with both neostigmine and an anticholinergic agent, would not seem to be amenable to any major changes. An increase in gastric emptying may be enhanced by the use of prophylactic prokinetic drugs, such as cisapride and metoclopramide. The use of low-dose naloxone to abolish the effect of fentanyl on gastric emptying is unlikely to be successful as it has been shown both to cause and reverse delays in gastric emptying [6,10] and to reverse epidural opioid analgesia .
In summary, it has been shown that in patients undergoing thoracotomy for lung resection and receiving thoracic fentanyl-bupivacaine epidural analgesia, gastric emptying as demonstrated by paracetamol absorption is delayed. This delay persists for at least the second postoperative day, and is clinically important as the absorption of many orally administered drugs is likely to be delayed during this period. The study indicates that physicians need to be aware that the desired effect of orally administered drugs during this period may not be achieved in this group of patients. Further studies are needed to provide evidence that altering our present anaesthetic technique for thoracotomy will prevent this delay.
Financial support for the study was from Cardiothoracic Centre NHS Trust. Dr P. Browning, Principal Scientist, Cardiothoracic Centre NHS Trust, Liverpool, helped with the paracetamol assay. Dr M Jackson, Manager of Research, Development and Clinical Audit, Cardiothoracic Centre NHS Trust, Liverpool, helped with the statistical analysis.
1. Finley FJ, Lamy A, Clifton J, Evans KG, Fradet G, Nelems B. Gastrointestinal function following esophagogastrectomy for malignancy. Am J Surg
2. Goldhill DR, Whelpton R, Winyard JA, Wilkinson A. Gastric emptying in patients the day after cardiac surgery. Anaesthesia
3. Au J, Hawkins T, Venables C, et al.
Upper gastrointestinal dysmotility in heart-lung transplant recipients. Ann Thorac Surg
4. Thorn SE, Wattwil M, Naslund I. Postoperative epidural morphine, but not epidural bupivacaine, delays gastric emptying on the first day after cholecystectomy. Reg Anesth
5. Petring OU, Dawson PJ, Blake DW, et al.
Normal postoperative gastric emptying after orthopaedic surgery with spinal anaesthesia and i.m. ketorolac as the first postoperative analgesic. Br J Anaesth
6. Nimmo WS. Effect of anaesthesia on gastric motility and emptying. Br J Anaesth
7. Romer HC, Russell GN. A survey of the practice of thoracic epidural analgesia in the United Kingdom. Anaesthesia
8. Product information, Paracetamol (Acetaminophen) Assay Kit, 20-K8001/60-K8002. Cambridge Life Sciences Plc, Ely, UK, 1994.
9. Geddes SM, Thorburn J, Logan RW. Gastric emptying following caesarean section and the effect of epidural fentanyl. Anaesthesia
10. Nimmo WS. Gastric emptying and anaesthesia. Can J Anaesth
11. Heading RC, Nimmo J, Prescott LF, Tothill P. The dependence of paracetamol absorption on the rate of gastric emptying. Br J Pharmacol
12. Clements JA, Heading RC, Nimmo WS, Prescott LF. Kinetics of acetaminophen absorption and gastric emptying in man. Clin Pharmacol Ther
13. Petring OU, Flachs H. Inter- and intrasubject variability of gastric emptying in healthy volunteers measured by scintigraphy and paracetamol absorption. Br J Clin Pharmacol
14. Tarling MM, Toner CC, Withington PS, Baxter MK, Whelpton R, Goldhill DR. A model of gastric emptying using paracetamol absorption in intensive care patients. Intensive Care Med
15. Finnegan BA, Jyn CA. Paracetamol kinetics and gastric emptying. Br J Anaesth
16. Orme M. Drug absorption in the gut. Br J Anaesth
17. Hammas B, Hvarfner A, Thorn SE, Wattwil M. Propofol sedation and gastric emptying in volunteers. Acta Anaesthesiol Scand
18. Steinbrook RA. Epidural anesthesia and gastrointestinal motility. Anesth Analg
19. Bonica JJ. Autonomic innervation of the viscera in relation to nerve block. Anesthesiology
20. Thoren T, Wattwil M. Effects on gastric emptying of thoracic epidural analgesia with morphine or bupivacaine. Anesth Analg
21. Thoren T, Wattwil M, Jarnerot G, Tanghoj H. Epidural and spinal anesthesia do not influence gastric emptying and small intestinal transit in volunteers. Reg Anesth
22. Thorn SE, Wattwil M, Lindberg G, Sawe J. Systemic and central effects of morphine on gastroduodenal motility. Acta Anaesthesiol Stand
23. Coda BA, Brown MC, Schaffer R, et al.
Pharmacology of epidural fentanyl, alfentanil and sufentanil in volunteers. Anesthesiology
24. Coda BA, Brown MC, Risler L, Syrjala K, Shen DD. Equivalent analgesia and side effects during epidural and pharmacokinetically tailored intravenous infusion with matching plasma alfentanil infusion. Anesthesiology
25. Sandler AN, Stringer D, Panos L, et al.
A randomized, double-blind comparison of lumbar epidural and intravenous fentanyl infusions for postthoracotomy pain relief. Analgesic, pharmacokinetic, and respiratory effects. Anesthesiology
26. Murad F. In: Gilman AG, Rall TW, Nies AS, Taylor P, eds. Goodman and Gilman's The Pharmacological Basis of Therapeutics.
London, UK: Pergamon, 1990: 778.
27. Hoffman BF, Bigger JT Jr. In: Gilman AG, Rall TW, Nies AS, Taylor P, eds. Goodman and Gilman's The Pharmacological Basis of Therapeutics.
London, UK: Pergamon, 1990: 828.
28. Benet LZ, Mitchell JR, Sheiner LB. In: Gilman AG, Rall TW, Nies AS, Taylor P, eds. Goodman and Gilman's The Pharmacological Basis of Therapeutics.
London, UK: Pergamon, 1990: 6.
29. Groves J, Edwards ND, Carr B, Sherry KM. Perioperative myocardial ischaemia, heart rate and arrhythmia in patients undergoing thoracotomy: an observational study. Br J Anaesth
30. Mahon SV, Berry PD, Jackson M, Russell GN, Pennefather SH. Thoracic epidural infusions for post-thoracotomy pain: a comparison of fentanyl-bupivacaine mixtures vs. fentanyl alone. Anaesthesia
31. Richardson J, Sabanathan S, Jones J, Shah RD, Cheema S, Mearns AJ. A prospective, randomized comparison of preoperative and continuous balanced epidural or paravertebral bupivacaine on post-thoracotomy pain, pulmonary function and stress responses. Br J Anaesth
32. Wang J, Pennefather S, Russell G. Low-dose naloxone in the treatment of urinary retention during extradural fentanyl causes excessive reversal of analgesia. Br J Anaesth