This Editorial View accompanies the following article: Granot M, Lowenstein L, Yarnitsky D, Tamir A, Zimmer EZ: Postcesarean section pain prediction by preoperative experimental pain assessment. Anesthesiology 2003; 98:1422–6.

OPIOIDS are among the most widely used drugs in medicine and are used for the management of acute and chronic pain from a wide variety of disorders. A report in the current issue of Anesthesiology by Aubrun et al.  affords an opportunity to reconsider and reexamine the role of opioids in pain management and also to review the realistic expectations for pain relief and relief of other symptoms for patients receiving opioids for acute pain.

It is worthwhile to begin by defending the uses of these drugs. Opioids are important components of many general anesthetic regimens. They have essential roles in the management of acute pain. For the majority of patients with cancer and sickle cell disease, and for patients in palliative care, opioids are effective in providing pain relief with a tolerable side-effect profile, particularly if mild-to-moderate degrees of sedation are considered acceptable. 1,2The risk of addiction in hospitalized patients who receive opioids for pain is quite low. Contrary to common belief, opioids also provide analgesia for some patients with neuropathic pain 3and even for a subgroup of patients with phantom limb pain after amputation. 4 

Aubrun et al.  examined morphine administration and pain relief in more than 3000 postoperative patients in the PACU, using a standardized titration procedure. Specifically, pain was assessed (using a visual analog scale [VAS] of 0–100) as quickly as possible after arrival in the PACU, and if the reported pain intensity was greater than 30, 3 mg of morphine was given. Assessment and treatment was repeated every 5 min until the reported VAS was 30 or less. Drug administration was stopped before the achievement of adequate pain relief only for respiratory depression (respiratory rate <12 bpm) or other serious events (e.g. , hypotension, vomiting). Note that sedation was not considered a serious side effect. These investigators then examined the relationship between the initial  pain score and amount of morphine required to achieve comfort. They reached several conclusions:

  1. Patients with higher initial VAS pain scores required more incremental doses of morphine to reach an acceptable pain score (VAS ≤30).

  2. There was a marked variation in the number of morphine doses required to produce comfort. The median dose to achieve a VAS of 30 or less was 0.17 mg/kg.

  3. In general, VAS scores of 70 or more should be regarded as indicative of severe pain.

  4. Most patients with a VAS score of 50 will usually require only one additional dose of morphine to report a VAS of 30 or less.

Aubrun et al.  are to be commended for an original examination of an everyday occurrence, opioid titration in the PACU, and for examining the results of a standardized titration procedure in a very large number of patients. Some practitioners may adopt the protocol noting the rapid achievement of pain relief (on average, 25 min for most patients). In addition, Aubrun et al.  examined the quantitative aspects of immediate postoperative pain management and responses to opioids, a process of significant importance to the anesthesiologist, patient, and recovery room nurse.

Several features of the authors’ study design suggest caution in interpretation. The patients underwent a diverse group of operative procedures, and, perhaps as a result, the intraoperative administration of opioids and other agents was not standardized (and it would be difficult to do so for such a patient group). Morphine titration in the PACU reflects a complex, non–steady-state situation, with rapidly decreasing effect-site concentrations of general anesthetics, sedatives, and opioids administered intraoperatively, and a rapid stepwise increase of morphine concentrations as the drug is given at 5 min intervals (as dictated by the authors’ protocol). This situation is, perhaps unavoidable and certainly reflects common practice. However, it is not clear how well their conclusions can be extrapolated to patients receiving opioids for prolonged period of time or to those with other forms of pain.

It seems problematic to define pain severity based on morphine requirement, especially during the later postoperative stages, because a variety of pharmacokinetic, pharmacodynamic, and psychosocial factors can alter morphine dose–effect relationships. The authors assert that a VAS of 70 or more should be regarded as indicative of severe pain, based on the finding that patients whose initial scores decreased in this range required substantially more morphine than those with initial scores of less than 70. This conclusion concurs with previous work by Collins et al. , 5which incorporated a more direct approach: ask patients to give a VAS score and to simultaneously rate their pain as “mild, moderate or severe.” In Collins et al.’  s study, patients reporting severe pain on a categorical scale had a mean 10 cm VAS score of 7.4, whereas 15% of patients reporting “severe pain” on a categorical scale had VAS scores of less than 5.4.

Aubrun et al.  conclude that there is a curvilinear relationship between morphine dosing and VAS scores. More specifically, VAS scores changed little with initial incremental doses, and then decreased rapidly to a value of 3 or less with the final incremental dose. McQuay et al.  depict a different sort of curvilinear relationship between analgesic dosing and effect (fig. 1). A useful indicator of drug effectiveness is the numbers needed to treat  (NNT). For a given binary endpoint (e.g. , success or failure in providing 50% pain relief), NNT expresses how many patients, beyond those who respond to placebo, would have to receive the treatment (e.g. , a certain dose of an analgesic) for one patient to have a successful endpoint. In analgesic studies with 50% pain relief as endpoints, NNTs of less than 2 are indicative of a very good effect. When large numbers of diverse clinical trials are analyzed in this way, two significant conclusions emerge:

Fig. 1. Effect of percent maximum total pain relief (%max TOTPAR) on NNT, illustrating how much more difficult it is to achieve near-complete pain relief than moderate pain relief. (From McQuay H and Moore A: An Evidence-Based Resource for Pain Relief, London, Oxford Publications, 1998, p 142; used with permission of the publisher.)

Fig. 1. Effect of percent maximum total pain relief (%max TOTPAR) on NNT, illustrating how much more difficult it is to achieve near-complete pain relief than moderate pain relief. (From McQuay H and Moore A: An Evidence-Based Resource for Pain Relief, London, Oxford Publications, 1998, p 142; used with permission of the publisher.)

  1. Morphine 10 mg gives NNT values higher than those for maximum recommended doses of nonsteroidal antiinflammatory drugs (NSAIDs). Stated another way, in a diverse group of dental and postoperative trials, morphine 10 mg is less likely to produce 50% pain relief than 800 mg of ibuprofen.

  2. With more demanding criteria for effectiveness (i.e. , 60, 70, or 80% relief), NNTs for opioids, NSAIDs, and other analgesics increase dramatically. That is, a far smaller percentage of patients treated with a fixed dose of morphine (e.g. , 10 mg) achieve 70 or 80% relief compared with the percentage who achieve 50% relief.

The protocol described by Aubrun et al.  did not measure side effects from morphine administration in the PACU. Morphine administration was stopped if respiratory rate or oxygen saturation reached predefined levels. Few patients (2.4%) were excluded from the analysis for severe morphine-related adverse events (allergy, hypotension, vomiting pruritus, or cutaneous rash). This raises the following question: What is the relationship between opioid dosing and the incidence or frequency of opioid-induced side effects? Does it depend on initial pain scores or extent of pain relief with that dose of opioid? It is widely accepted that more severe pain antagonizes the sedative and respiratory depressant effects of opioids. However, it is less clear how the frequencies of a number of other opioid side effects (e.g. , nausea, ileus, itching, urinary retention) vary with both opioid dosing and initial pain severity. An analogous concept to NNT is numbers needed to harm , which is a reasonable way to evaluate dose response for side effects. Plots of numbers needed to harm versus  dose have been used for the assessment of dose-dependence of nausea, vomiting, dizziness, and somnolence in a meta-analysis of studies of tramadol, codeine, and often used combination analgesics. 6 

Opioid side effects exert a major impact on the course of postoperative recovery and limit effective opioid titration in many cases. A variety of strategies have been developed to providing postoperative analgesia while minimizing opioid administration: so-called “opioid-sparing” approaches, including peripheral 7and neuraxial administration of local anesthetics, oral or intravenous formulations of acetaminophen, 8NSAIDs 9or cyclooxygenase-2 inhibitors, 10systemic or epidural administration of N-methyl-D-aspartate antagonists, 11and a variety of other approaches. Most of these approaches have mixed success, depending on numerous factors. 12–14In some cases the same intervention shows opioid-sparing for some procedures or patient groups but not for others. 15One interesting report suggested the possibility that, for some patients, apparent morphine-sparing effects of an NSAID could be an artifact of NSAID-induced reductions in renal clearance of morphine-6-glucuronide, which has analgesic activity roughly similar to that of morphine. 16 

Kehlet et al.  have pioneered studies of approaches to postoperative analgesia and “acute rehabilitation”17that emphasize avoidance of parenteral opioids. 18Major components of this approach include:

  1. preoperative education and changes in the “culture” of postoperative care

  2. minimally invasive surgical techniques

  3. multimodal analgesic approaches with combined use of neuraxial and peripheral regional anesthetic approaches, emphasizing use of local anesthetics

  4. NSAIDs, cyclooxygenase-2 inhibitors, or cortico-steroids

  5. early mobilization, early feeding, early removal of tubes, and supplemental oxygen as needed, particularly during sleep

A major advantage of avoiding opioids in this approach is the ability to minimize postoperative ileus and initiate feeding in the early postoperative period. It should be apparent that any one of these methods of opioid-sparing is useful not as an end in itself, but rather only if it leads to clinically and statistically significant improvements in outcomes, including improved pain scores with rest and with movement, reduced side effects, reduced complications, and improved or accelerated recovery and rehabilitation. Some multimodal interventions produce reductions in opioid use and improved outcomes. 17,19Several interventions that involve analgesic interventions, but no changes in the overall “acute rehabilitation” approach to postoperative mobilization and nutrition, achieve reduced opioid use but with no change in pain scores, side effects, or recovery parameters. 20 

Recently, the Joint Commission on Accreditation of Healthcare Organizations established standards for pain assessment and treatment in healthcare facilities. This well-intended effort probably will have a positive overall impact in terms of improved standards for pain treatment in many clinical settings. A number of surveys suggest that currently, a high percentage of hospitalized postoperative patients continue to experience moderate-to-severe pain. 21For example, even when a protocol somewhat similar to that of Aubrun et al.  was used for patients on surgical wards, the average VAS pain score at rest was still approximately 30 (out of 100), yet the average VAS pain score with activities was 60 after major surgery. 22However, if hospitals, in an effort to comply with the Joint Commission on Accreditation of Healthcare Organizations, try to generate uniformly low pain scores at rest and with movement (i.e. , “all pain scores must be ≤3”), using opioids as the sole or predominant method of analgesia, this would likely result in an increased frequency of side effects. Also problematic are surveys that query “the level of the worst pain experienced after surgery.” Many patients experience brief, severe episodes of acute pain (e.g. , a coughing paroxysm) that cannot be anticipated or well controlled with current opioid regimens. For some patients and some situations, the overall impact on well-being, quality of life, and outcomes may be beneficial. Other patients may prefer moderate pain (e.g. , VAS 4–5) to reduce the severity of dizziness, somnolence, or other side effects. Aubrun et al.’  s study as reported in this issue of the Journal adds to the body of literature that affirms the utility, limitations, and difficulties in titrated dosing of opioids.


Hanks GW, Conno F, Cherny N, Hanna M, Kalso E, McQuay HJ, Mercadante S, Meynadier J, Poulain P, Ripamonti C, Radbruch L, Casas JR, Sawe J, Twycross RG, Ventafridda V, Expert Working Group of the Research Network of the European Association for Palliative C: Morphine and alternative opioids in cancer pain: The EAPC recommendations. Br J Cancer 2001; 84: 587–93
Expert Working Group of the Research Network of the European Association for Palliative C:
Worthington HV, Clarkson JE, Eden OB: Interventions for treating oral mucositis for patients with cancer receiving treatment. Cochrane Database of Systematic Reviews 2002; CD001973
Raja SN, Haythornthwaite JA, Pappagallo M, Clark MR, Travison TG, Sabeen S, Royall RM, Max MB: Opioids versus antidepressants in postherpetic neuralgia: A randomized, placebo-controlled trial. Neurology 2002; 59: 1015–21
Huse E, Larbig W, Flor H, Birbaumer N: The effect of opioids on phantom limb pain and cortical reorganization. Pain 2001; 90: 47–55
Collins SL, Moore RA, McQuay HJ: The visual analogue pain intensity scale: What is moderate pain in millimetres? Pain 1997; 72: 95–7
Moore RA, McQuay HJ: Single-patient data meta-analysis of 3453 postoperative patients: Oral tramadol versus placebo, codeine and combination analgesics. Pain 1997; 69: 287–94
Chelly JE, Greger J, Gebhard R, Coupe K, Clyburn TA, Buckle R, Criswell A: Continuous femoral blocks improve recovery and outcome of patients undergoing total knee arthroplasty. J Arthroplasty 2001; 16: 436–45
Hernandez-Palazon J, Tortosa JA, Martinez-Lage JF, Perez-Flores D: Intravenous administration of propacetamol reduces morphine consumption after spinal fusion surgery. Anesth Analg 2001; 92: 1473–6
Munro HM, Walton SR, Malviya S, Merkel S, Voepel-Lewis T, Loder RT, Farley FA: Low-dose ketorolac improves analgesia and reduces morphine requirements following posterior spinal fusion in adolescents. Can J Anaesth 2002; 49: 461–6
Camu F, Beecher T, Recker DP, Verburg KM: Valdecoxib, a COX-2-specific inhibitor, is an efficacious, opioid-sparing analgesic in patients undergoing hip arthroplasty. Am J Ther 2002; 9: 43–51
Ilkjaer S, Bach LF, Nielsen PA, Wernberg M, Dahl JB: Effect of preoperative oral dextromethorphan on immediate and late postoperative pain and hyperalgesia after total abdominal hysterectomy. Pain 2000; 86: 19–24
Huang JJ, Taguchi A, Hsu H, Andriole GL, Jr, Kurz A: Preoperative oral rofecoxib does not decrease postoperative pain or morphine consumption in patients after radical prostatectomy: A prospective, randomized, double-blinded, placebo-controlled trial. J Clin Anesth 2001; 13: 94–7
Klein JR, Heaton JP, Thompson JP, Cotton BR, Davidson AC, Smith G: Infiltration of the abdominal wall with local anaesthetic after total abdominal hysterectomy has no opioid-sparing effect. Br J Anaesth 2000; 84: 248–9
Kucuk N, Kizilkaya M, Tokdemir M: Preoperative epidural ketamine does not have a postoperative opioid sparing effect. Anesth Analg 1998; 87: 103–6
Green CR, Pandit SK, Levy L, Kothary SP, Tait AR, Schork MA: Intraoperative ketorolac has an opioid-sparing effect in women after diagnostic laparoscopy but not after laparoscopic tubal ligation. Anesth Analg 1996; 82: 732–7
Tighe KE, Webb AM, Hobbs GJ: Persistently high plasma morphine-6-glucuronide levels despite decreased hourly patient-controlled analgesia morphine use after single-dose diclofenac: Potential for opioid-related toxicity. Anesth Analg 1999; 88: 1137–42
Basse L, Raskov HH, Hjort Jakobsen D, Sonne E, Billesbolle P, Hendel HW, Rosenberg J, Kehlet H: Accelerated postoperative recovery programme after colonic resection improves physical performance, pulmonary function and body composition. Br J Surg 2002; 89: 446–53
Kehlet H, Rung GW, Callesen T: Postoperative opioid analgesia: time for a reconsideration? J Clin Anesth 1996; 8: 441–5
Barratt SM, Smith RC, Kee AJ, Mather LE, Cousins MJ: Multimodal analgesia and intravenous nutrition preserves total body protein following major upper gastrointestinal surgery. Reg Anesth Pain Med 2002; 27: 15–22
Tang J, Li S, White PF, Chen X, Wender RH, Quon R, Sloninsky A, Naruse R, Kariger R, Webb T, Norel E: Effect of parecoxib, a novel intravenous cyclooxygenase type-2 inhibitor, on the postoperative opioid requirement and quality of pain control. A nesthesiology 2002; 96: 1305–9
Dolin SJ, Cashman JN, Bland JM: Effectiveness of acute postoperative pain management: I. Evidence from published data. Br J Anaesth 2002; 89: 409–23
Gould TH, Crosby DL, Harmer M, Lloyd SM, Lunn JN, Rees GA, Roberts DE, Webster JA: Policy for controlling pain after surgery: Effect of sequential changes in management. BMJ 1992; 305: 1187–93