Jump to ContentJump to Main Navigation
Palliative Care Perspectives$

James L. Hallenbeck

Print publication date: 2003

Print ISBN-13: 9780195165784

Published to Oxford Scholarship Online: November 2011

DOI: 10.1093/acprof:oso/9780195165784.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2017. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 26 February 2017

Pain Management

Pain Management

(p.36) 4 Pain Management
Palliative Care Perspectives

James L. Hallenbeck

Oxford University Press

Abstract and Keywords

This chapter examines pain-management issues in palliative care. Pain is the most common complaint of terminally ill patients and it has been estimated that between 85 to 95 percent of pain syndromes can be adequately palliated using relatively simple techniques. Despite this, pain is often under-treated. The chapter discusses different types and classifications of pain and explains the factors that should be considered in the development of pain management strategy.

Keywords:   pain management, palliative care, terminally ill, pain syndromes, disease management

Always the same. Now a spark of hope flashes up, then a sea of despair rages, and always pain; always pain, and always the same.

Leo Tolstoy, The Death of Ivan Ilych

All of us experience pain in life. People look to clinicians for relief from pain when it becomes difficult to bear. Our duty to alleviate the suffering engendered by pain harkens back to the very roots of what it means to be a healer. Recent advances in the understanding and treatment of pain allow us to fulfill this obligation to our patients far better than we ever could before. Unfortunately, new pain relief methods are too often underused or poorly used.

Pain is the most common presenting complaint to physicians in North America, and I suspect this is true in other regions of the world.1 It has been estimated that 85% to 95% of pain syndromes, including severe forms, such as cancer-related pain, can be adequately palliated using relatively simple techniques.2 However, pain is often under-treated. In one study of cancer patients at a famous cancer center, as many as 50% of cancer patients suffered unrelieved pain.3 Such undertreatment of pain is not isolated to cancer. The SUPPORT study demonstrated that 50% of the 9105 patients studied were estimated by surviving relatives as having moderate or severe pain 50% of the time or more in the last three days of life.4 A study of the treatment of nonmalignant pain in (p.37) 49,971 nursing home patients found that 25% of patients with daily pain received no analgesics whatsoever. Advanced age (〉85), male sex, cognitive impairment, and being a member of a racial minority were statistically significant risk factors for receiving no analgesics.5

Classification of Pain

Acute Pain

We all have experienced acute pain. Bee stings, bumped knees, and bone fractures are simple examples. Most acute pain serves a clear purpose: some problem needs to be addressed. Acute pain is characterized by help-seeking behavior. In most cases people cry out and move about in a very obvious manner. Physiologic responses to acute pain include tachycardia, tachypnea, and sweating due to discharge in the sympathetic nervous system. It is easy to recognize and empathize with acute pain. It is practically automatic. We wince if we see severe, acute pain and respond with our own “sympathetic” discharge.

The treatment of acute pain can be difficult in that the intensity of pain may change dramatically over a short period of time. Physicians may have trouble adjusting pain medications rapidly enough to match the level of pain being experienced because pain intensity tends to escalate and decrease swiftly. Both under-and overtreatment can easily occur. Undertreatment risks excessive suffering. Overtreatment poses real medical risks. Thus, as acute pain changes rapidly, treatment of such pain requires frequent reassessment of the patient's status in order to avoid extremes of under-and overtreatment.

Chronic Pain

Chronic pain is very different from acute pain. It serves no biological purpose. While the suffering engendered may be as great as is that in acute pain, it is subjectively experienced and objectively displayed in a very different way. For reasons not well understood, chronic pain is characterized by physical and mental withdrawal. Vegetative signs very similar to those found in depression, such as anorexia, anhedonia, lethargy, and sleep disturbance are often present. Chronic pain frequently coexists with depression, making it difficult at times to distinguish between the two. Obvious displays of distress, as are found in acute pain, are usually absent. Chronic pain is very difficult to recognize. Even when recognized we tend not to experience the same intense, visceral empathy that arises so easily in the presence of acute pain. Lack of recognition of chronic pain and difficulty empathizing with it are major barriers to successful treatment.

It is difficult to judge by observation alone the degree of chronic pain suffered. Both family members and professional health care workers tend to miss (p.38) the mark. Tragically, the correlation between a patient's and another's assessment of pain intensity is poorer at higher degrees of pain. In one study using visual analog scales (VAS) from 0 to 10 to measure pain intensity, correlation between the observers' estimated intensity score and that reported by the patient was worst at high levels of pain, scores of 7 to 10 (severe to unendurable).6 This stands in dramatic contrast to our experience with acute pain; the more severe the acute pain the easier it is for us to recognize. From this and other studies it has been concluded that for patients with chronic pain, we cannot simply “see” if a patient is in pain. We are, in effect, “color-blind” to chronic pain. We also cannot judge the degree of pain by measures such as noting how calm or disturbed a patient appears. In a manner of speaking we suffer a disability. As with any disability, we must try to find ways to compensate. In order to determine how much pain a patient is in, we have to ask. If possible, it is advisable to use a scale of pain intensity to communicate intensity of pain. Numeric scales from 0 to 5 or 0 to 10 or visual analogue scales, some with pictures reflecting varying degrees of distress, are commonly used. These allow a better assessment of pain intensity and a more accurate measure of changes with therapeutic intervention.

Is it reasonable or necessary to measure pain on a routine basis? I would argue that the evidence is overwhelming that people simply lack the proper receptors to detect chronic pain (much like we cannot see blood pressure). It would be wonderful if all clinicians routinely asked patients about their pain. However, the evidence is also strong that clinicians have resisted efforts to improve pain assessment less formally. However, is such measurement scientific?

Pain management would be infinitely easier if we had something like an O2 saturation meter to measure pain—something like the tricorder in Star Trek movies. Clinicians prefer hard data and become uncomfortable with subjective reports. In an absolute sense we cannot know if one patient's 7 to 10 pain is the same as that of another. However, studies have demonstrated that the individual patient is consistent in reporting pain scores. That is, intra-rater reliability has been validated.7 I think of pain scores as imperfect tools to compensate for our collective disability.

Even so, what if people lie? In hospice literature it has been said that pain is whatever the patient says it is. This seems a bit simplistic to me. While it is true that we are color-blind to chronic pain, it is also true that some patients may be less than honest about their pain. Patients may fabricate or exaggerate symptoms for psychological reasons or secondary gain. Patients with very real chronic pain may also learn to exaggerate their pain and become demanding, as they believe physicians will not otherwise take them seriously, a process called “pseudoaddiction.”8 Paradoxically, this may arouse suspicion in the practitioner that the pain is not “real.” There is no easy way to tell what is real. However, common sense and a trusting relationship between provider and patient go a long way. As a general (p.39) rule, if the complaint of pain is plausible and if there are no very good reasons for doubting the patient, believe it. My philosophy is that ties go to the patient. Of two possible “sins” in pain management, the sin of ignoring real pain seems greater than does the sin of occasionally being fooled by a patient.

Types of Pain

There are two major types of pain, nociceptive and neuropathic. Distinguishing between them is important because the causes and treatments are different. Ideally, the causes of both types of pain will be identified and treated, resulting in pain relief. Unfortunately, it is often the case that cure is impossible and palliation is necessary.

Nociceptive (Tissue) Pain

Nociceptive pain results from tissue damage. Intact neurons dutifully report damage, and pain is experienced. Nociceptive pain can be subdivided into somatic and visceral (gut) pain. Nociceptive pain can be experienced as sharp, dull, or aching. There may be radiation of the pain, especially visceral pain, but it will not be in a direct nerve distribution. For example, gallbladder pain can radiate to the scapula. Nociceptive pain is generally responsive to NSAIDs (nonsteroidal anti-inflammatory drugs) and opioids. Conditions associated with inflammation, bone pain, and joint disease are particularly responsive to NSAIDs.

Neuropathic (Nerve) Pain

Neuropathic pain may occur when there is either damage to or dysfunction of nerves in the peripheral or central nervous system. Faulty signals are sent to the brain and experienced as pain. Neuropathic pain can be either peripheral (outside the central nervous system) or central in origin. Examples of neuropathic pain include diabetic neuropathy, trigeminal neuralgia, postherpetic zoster pain (peripheral pains), and the thalamic pain syndrome (a central pain). Neuropathic pain frequently coexists with nociceptive pain. Examples include trauma that damages tissue and nerves, burns (that burn skin as well as nerve endings), and external nerve compression. Examples of the latter include tumor nerve compression and sciatica from herniated discs pressing on nerves.

(p.40) Neuropathic pain is often described as having a burning or electrical quality. It may feel like a shock or lightning bolt. Sometimes stimuli that usually do not cause pain, such as light touch, may elicit a paroxysm of pain. A light stroke of the cheek that results in the sudden pain of trigeminal neuralgia is an example of this type of pain. Sometimes patients do not describe the sensation as being “painful” but rather as feeling unpleasantly strange or tingly, like an arm feels when it wakes up from “going to sleep.” This is called a dysesthesia. Diabetic neuropathy commonly results in this type of sensation.

Neuropathic pain in the peripheral nervous system frequently follows a nerve distribution. This distribution may replicate a particular nerve, as in sciatic pain or trigeminal neuralgia, or may represent the distribution of terminal nerve endings, as in the stocking-glove distribution of peripheral neuropathies.

Neuropathic pain is relatively resistant to NSAIDs and opioids, although they may be helpful in certain cases. The other major classes of medications useful for neuropathic pain, tricyclic antidepressants, anticonvulsants, and sodium channel blockers, will be discussed later.

Evaluation of Pain

As discussed earlier, pain is a complex and personal experience. It is affected by physiological, psychological, and spiritual factors. The evaluation of pain must consider these factors and their interactions that result in the experience of pain. A useful mnemonic in evaluating pain(s) is the acronym: NOPQRST.9

N: Number of Pains

Although we tend to speak of a patient's pain as an overall experience, in fact, many patients have more than one pain. It would be more appropriate to speak of a patient's pains. These should be individually evaluated.

O: Origin of Pain

Understanding the cause of a particular pain is immensely helpful. Removal of the underlying cause may eliminate the pain. Even if this is not possible, understanding the origin of the pain may help with consideration of specific therapeutic options.

P: Palliate and Potentiate

What makes the pain better or worse? Do certain activities or body positions alleviate or worsen the pain? How have previously tried medications affected the pain (partial relief, no relief, etc.)? This may provide an important clue as to the type of pain being experienced, if in doubt. While the focus of this text is on pharmacologic therapy, it is important to point (p.41) out that nonpharmacologic interventions can have a significant (positive or negative) effect on pain.

Nociceptive pain tends to worsen when stress or pressure is applied to an affected area. Neuropathic pain may be “set off” when usually non-painful stimuli, such as a light touch, temperature change, or even air movement, provoke a “reverb”-like phenomenon, with paroxysms of pain.

While patients may be able to speak about the effect of medications or body positions, they may be less able to comment on the effect of psychological or spiritual factors. The clinician should be aware that depression, anxiety, confusion, and spiritual distress may all contribute significantly to the experience of pain. If these conditions are present, treating them may result in significant palliation.

Q: Quality

A great variety of words are used to describe pain. Nociceptive pain may be sharp, dull, stabbing, or pressurelike. Neuropathic pain descriptions often have an electrical quality: burning, lancinating, buzzing, tingling, zapping, and lightninglike.

R: Radiation

Both nociceptive and neuropathic pains may radiate, although we usually associate radiation with neuropathic pain. Neuropathic pain tends to radiate in a distribution that follows nerves. Classic examples include trigeminal neuralgia and herpes zoster pain. The stocking glove distribution of peripheral neuropathies, as in diabetes, also follows a pattern of terminal nerve endings. Nociceptive pain radiates in less obvious ways. Thus, pericarditis, for example, may radiate to the scapula. Cardiac pain may radiate to the arm(s) or neck.

S: Severity/Suffering

Severity. As mentioned above, it is impossible to accurately gauge the severity of chronic pain by observation alone. You have to ask. The use of pain scales is strongly recommended when patients report having pain. Some patients respond better to certain scales than others. Some scales use nonnumerical images such as faces or colors to represent the range of distress.

Having said this, even numerical scales communicate limited data. On a scale of 0 to 10, is one person's 7 the same as another's? How many of us have really experienced level 10 pain? What is the worst pain we could endure? Hopefully, we will never find out.

If nothing else, pain scales seem to communicate the patient's urgency in wanting their pain addressed. On a scale of 0 to 10, pains of 1 to 3 are often well tolerated. Patients often decline additional interventions at these levels. Levels 3 to 6 usually indicate that some intervention is desired, although it is (p.42) not an emergency. Pain at 7 to 10 is a serious problem. Not only is the pain seriously distressing, there is usually a fear that it will become a 10 and thus “out of control.” Pain at level 10 (or greater) is perceived as an emergency by the patient and should usually be treated as an emergency by the clinician. Such pain is overwhelming. Having said this, some patients will report that pain management is “adequate” with a score as high as 7 to 10. Others will want urgent treatment when the scores rise from 1 to 2.6. Understanding how that patient interprets the pain score is most important. It is strongly advised that the examiner assess whether a given score means treatment is adequate for the individual patient. Individual patients tend to be consistent over time.

Suffering. What impact is the pain having on the patient? The impact may be an internal experience, such as depression or a thought of suicide, or may directly affect the patient's functioning. Sleep disturbances, difficulty walking, inability to work, and impairment of the activities of daily living may all reflect the pain experience. As obvious as this may seem, I am struck by how often we forget to ask how pain (or other symptoms) affect a person's life. Perhaps it is because we assume that pain is simply awful—what more do we need to know? However, hearing how a person is suffering with pain (both the nitty-gritty—“I can't work”—and the deep issues—“I wonder why God did this to me”) helps us understand and empathize with what the patient is going through. Personally, I have trouble relating to a number, but if a patient can begin talking to me about how life has changed for them, then I feel I can gain a small glimpse of their experience.

T: Timing and Trend

Timing. Pain is rarely the same at all times. Pain has a pattern over time. Later, I will explain how matching the patient's pain pattern with therapeutic interventions (pattern matching) enables one to maximize therapeutic efficacy and minimize side effects, especially when treating nociceptive pain. Acute pain comes on rapidly and usually dissipates rapidly. Most chronic pain has a base and occasional spikes of incident pain, which may be predictable or unpredictable. Both need to be addressed. For example, some men may experience predictable trigeminal neuralgia only when shaving. Bed-bound patients often experience pain predictably with turning or cleaning. Wounds may hurt during dressing changes.

Trend. Pain often has momentum. It is very difficult to get a handle on rapidly escalating pain. Therapy is harder and suffering appears to be greater when the trend is worsening. This is true both physiologically and psychologically. Physically, we now know that escalating pain can “rev-up” in the central nervous system, amplifying painful stimuli and resulting in stronger pain signals. Specific receptors in the spinal cord such as those for N-methyl-D-aspartate (NMDA) are involved in this process, and blockading such (p.43) receptors can be useful in resistant pain syndromes. Psychologically, patients are very aware of their pain trend. If the pain is worsening, patients understandably project into the future that it will become worse and even unendurable. This projection itself contributes to the pain experience and may be communicated as a higher pain score. Likewise, if the trend is good, patients may be able to tolerate more physical pain at any given moment as they project into a more pleasant future. Pain is certainly experienced in the present but is understood in terms of the past and the future.

Perhaps a personal story will help illustrate this point. A few years ago while having a dental cavity filled, I tried to distract myself from the pain I was experiencing by asking myself, “On a scale of 0 to 10, how much pain do you have,” as I had asked countless patients. “Hmm, perhaps a 3 or a 4,” I thought. With each little whine of the drill I thought, “Only a few moments more, then it will be over.” I was to raise my hand if I had too much pain, but I could bear it a bit longer. And then it was over. I wondered—what if that pain had gone on forever? What if I had known it would not only go on, but get worse and worse? My hand would have been up in a flash. The idea of that pain going on and getting worse would have been too much to bear.

Pain Management Strategy

Having assessed the patient's pain, a strategy for management should be developed. The discussion that follows emphasizes opioids because these are so commonly used in palliative care. However, this is not to suggest that opioids are more or less appropriate in any individual case. Sound clinical judgment must be used in selecting specific agents.

What Nonpharmacologic Approaches to Pain Should Be Adopted?

Although the emphasis here is on the pharmacology of pain management, the clinician should also consider other interventions in developing a strategy. How does the patient's psychological state affect his or her pain? Is the patient depressed, anxious, or confused? How does the patient relate to his or her pain? Some patients want all pain to be abolished. Others may want some pain to remain. (As one cancer patient put it, “If I didn't feel some pain, how would I know what that cancer is doing in there.”) Some may see the pain as something to be conquered. Some may see it as something to be accepted. A thorough discussion of the psychological and spiritual aspects of pain is beyond the scope of this text. Often, assistance from others—psychologists, psychiatrists, social workers, and chaplains—will be necessary if proper care is to be delivered.

(p.44) A variety of medical interventions other than medications may also be extremely useful. Radiation therapy and chemotherapy may help alleviate pain in patients with certain cancers. Nerve blocks, trigger-point injections, and (rarely) surgical approaches may also be useful. Physical therapy, occupational therapy, and massage therapy may help in certain cases. Experts in these areas should be consulted, as needed.

Principles in Choosing Medications

  1. 1. Avoid specific toxicities. In choosing among possible medications, an otherwise useful drug might be excessively toxic for a particular patient. A patient with thrombocytopenia, for example, would be a poor candidate for a traditional nonsteroidal anti-inflammatory drug (NSAID), because such drugs interfere with platelet aggregation.

  2. 2. Look for “two-fers.” When possible, identify agents with additional effects that might be beneficial—two for the price of one. Anticonvulsants, for example, might be particularly useful in a patient with a seizure disorder who also had neuropathic pain. In contrast, one might choose a tricyclic antidepressant (TCA) for neuropathic pain in a depressed patient.

  3. 3. Think about who will be administering the medicine. A medication that requires injection might be very appropriate in a hospital or nursing home but difficult to administer at home. Competent patients administering their own medications may be better able to manage short-acting pain medications on an as-needed basis. In contrast, a demented patient with pain cared for in a nursing home or at home by family will probably receive inadequate analgesia when treated q4 h prn, as family and staff may not assess pain regularly (especially at night) and the patient may be unable to advocate for him-or herself. Long-acting preparations of both NSAIDs and opioids may be more appropriate in such a situation.

  4. 4. Consider the drug delivery route of administration. Possible routes of therapy include oral, enteral tube, percutaneous and parenteral intravenous (IV), intramuscular (IM), and subcutaneous (SC). (See later discussion of routes of delivery)

  5. 5. Identify the patient's pain pattern and perform pattern matching with your therapy.

Pattern Matching

Management of pain is optimized when therapy overlaps the patient's pattern of pain. This maximizes analgesia while minimizing side effects. In using opioids for therapy when pain increases, so should the drug dose. Similarly, when (p.45) pain lessens, the drug dose should be decreased. Pain itself can counteract certain opioid side effects. In particular, sedation and respiratory depression are significantly blocked by pain. Thus, the goal in using opioids is to have pain signals and opioid signals neutralize each other.

Acute pain, with a pattern of rapid escalation and de-escalation, requires short-acting opioids and careful titration if pain is to be adequately managed and side effects avoided (Fig. 4.1).

Chronic pain typically has both a background “noise” of pain with intermittent spikes of incident, or breakthrough, pain. The general strategy for such pain is to use a long-acting agent to manage the background basal pain and short-acting opioid doses as needed for breakthrough pain (Fig. 4.2).

While these are common patterns, the patient's individual pain pattern should be considered. For example, a patient may complain of pain only at night. This pattern should generate a “differential diagnosis” that may lead to important changes in therapy. This pattern may reflect pain worsened by lying down. Perhaps the patient is unable to get needed pain medications at night, as he or she is dependent on others, family, or nursing staff who may be less responsive during this time. Maybe he or she is no longer distracted, as in the daytime, which increases an awareness of pain. Each of these underlying causes would require a different approach.

Let us review the classes of common analgesics before getting into a more in depth discussion of routes of therapy and dosing strategies for opioids (Fig. 4.3).

                   Pain Management

Figure 4.1. Acute pain pattern matching. Analgesia is maximized and side effects minimized when the rise and fall of the blood level of an analgesic closely overlaps the temporal pattern of a patient's pain.

                   Pain Management

Figure 4.2. Chronic pain pattern. Generally, a long-acting medication for the baseline pain that is always present and a short-acting medication that rapidly peaks in tandem with an acute pain spike are needed.

Classes of Analgesics

Nonsteroidal Anti-Inflammatory Drugs (NSAIDs)

NSAIDs are a mainstay in the management of mild to moderate nociceptive pain. As mentioned above, they are particularly useful in inflammatory states and in pain involving the musculoskeletal system.10 It is beyond the scope of this book to discuss the large number of NSAIDs available today. Rather, some guidelines for choosing and using them are offered.

  1. 1. Consider drug half-life and frequency of administration. Short-acting agents such as ibuprofen may be preferable for pain that arises intermittently and is of short duration. Such agents can be given on an as needed basis. For patients with chronic pain that requires round-the-clock analgesia, a longer-acting agent such as naproxen, which may allow more convenient dosing, may result in better patient compliance and improved analgesia.

  2. 2. It is a mistake to consider NSAIDs as necessarily less toxic than opioids. NSAIDs may cause upper intestinal symptoms such as heartburn, nausea, or vomiting in 10% to 20% of patients. Significant upper GI bleeding from either gastritis or duodenal ulceration can occur, as can nephrotoxicity. Bronchospasm may be precipitated in sensitive asthmatics, as with aspirin. NSAIDs can cause altered mental status, especially in the elderly and frail. Most NSAIDs inhibit platelet aggregation. This is particularly a risk in patients who receive anticoagulants and in patients with thrombocytopenia.

  3. (p.47) 3. For some patients NSAIDs may be as effective or more effective than are opioids in relieving pain. In such cases it may be a mistake to withhold this class of medication for fear of some of these side effects. Rather, additional steps may be necessary to minimize the risk. Misoprostol, a prostaglandin El analog, can significantly reduce the risk of gastritis and gastric ulceration due to NSAIDs. Misoprostol is also effective against duodenal peptic ulceration. H2 blockers, such as ranitidine, have been found to reduce the risk of duodenal ulcers for patients treated with NSAIDs. Standard doses of H2 blockers such as famotidine 20 mg BID do not protect against NSAID-related gastric ulcers. However, one study with higher dose famotidine, 40 mg BID, did show a protective effect.11 Identification and eradication of H. pylori infection may lessen the risk of bleeding with NSAIDs in those infected for both duodenal and gastric ulcers. When there is concern over renal function, careful monitoring of blood chemistries may allow early detection of adverse effects. If deterioration in renal function is detected, discontinuation of the NSAID usually results in a gradual return to baseline function. Use of COX2 inhibitors (see below) may also help minimize risk.

COX2 Inhibitors

Recently, a new class of NSAIDs has become available, the COX2 inhibitors.12 Selective blockade of cyclooxygenase permits analgesia and an anti-inflammatory effect while minimizing the risk of GI toxicity and bleeding. Although experience in the use of these drugs in palliative care is limited, the following should be helpful in considering when to use one of these agents.

                   Pain Management

Figure 4.3. Pain exacerbation at night. What is the “differential diagnosis” for this kind of pain?

  1. (p.48) 1. They are relatively expensive. This makes them considerably more expensive than naproxen, for example. However, in absolute terms, if an NSAID is truly indicated and significant contraindications exist for standard NSAIDs, the difference in price may be worth it.

  2. 2. They appear to be equi-efficacious with other NSAIDs but with no evidence of superior analgesia.

  3. 3. Evidence is strong that the risk of GI bleeding is significantly reduced with these agents.13 When an NSAID is truly needed and there is concern about GI bleeding, the total cost may be lowered by using one of these agents rather than by adding a relatively expensive prophylactic drug such as misoprostol or a pump-inhibitor.

  4. 4. Evidence is strong that platelet aggregation is not inhibited by these agents. Thus, in patients with thrombocytopenia, defective platelets (uremia), and bleeding tendencies these drugs may offer a strong advantage.

Opioids—A Misunderstood Class of Drugs

Many misconceptions exist relative to the use of opioids. These misunderstandings have resulted in global underuse of these medications for pain relief. It is a mistake to extrapolate from the effects of opioids on addicts to the effects of opioids on patients in pain. Opioids used for pain relief act very differently on the body and mind than do opioids used by people not in pain. Most of the horrors observed in addicts do not occur in patients who take opioids for pain relief. The presence of pain itself changes the effect of opioids. When we administer opioids for pain relief, we do so considering the opioid to be an antidote to the poison of pain. Many are not aware that pain reciprocally acts, in part, as an antidote to certain of the toxic effects of opioids. Ideally, the effects of pain and opioids cancel each other out, allowing pain relief with minimal toxicity.

Pain also seems to block much of the euphoric effect of opioids. Contrary to popular thinking, the risk of turning a nonaddict into an addict when treating pain is very, very small to nonexistent. In one review by Porter, 11,882 patients without a prior history of addiction who were treated for pain with opioids were followed to determine how many developed addictive behavior. Only four did so.14 A more recent study reviewed statistics on the medical use of opioids and (p.49) the incidence of opioid abuse from 1990 to 1996 in the United States. It found that while medical use of opioids (except for meperidine) had increased dramatically over that period of time (morphine use up 59%), this trend was not paralleled by increased opioid abuse. It concluded, “The trend of increasing medical use of opioid analgesics to treat pain does not appear to contribute to increases in the health consequences of opioid analgesic abuse.”15 Physical dependence does occur with repeated dosing. However, such dependence does not result in craving for the drug, associated with deleterious life consequences (a definition of addiction). It is notable that many. drugs cause similar dependence without addiction—meaning that sudden discontinuation of the drug results in adverse physiological effects. Beta blockers, clonidine, and steroids, among others, can result in physical dependence.

If pain is skillfully treated with opioids, adverse complications such as respiratory depression and significant sedation rarely occur. No good evidence base exists for the common belief that the proper use of opioids (compared to the use of opioids by addicts) in the treatment of pain hastens the chance of death. Mild sedation and nausea may occur with initiation of opioid therapy or with subsequent increases in dosing. These side effects relate more to increases in opioid dosing and serum levels than to the absolute amounts of opioid administered. They usually subside after steady state levels are achieved—usually in a day or two.

While sedation and nausea tend to resolve over time as steady serum levels are obtained, such is not the case with constipation. As a rule, patients receiving chronic opioid therapy require continuous laxative therapy. This need for laxatives should be anticipated at the outset of therapy and not when constipation becomes a problem, as it almost inevitably does. (See the section on constipation in Chapter 5.) Constipation is caused by opioids binding mu receptors in the intestinal tract. Patients on very high doses of opioids do not necessarily require higher laxative doses to compensate than do those on lower doses. Orally administered opioids may result in somewhat more severe constipation than does transdermally administered fentanyl, perhaps because of more concentrated mucosal exposure.16,17

Many physicians fail to prescribe opioids when indicated because of fear that they may be criticized or, worse, have their licenses revoked for improperly (p.50) prescribing narcotics. However, with proper documentation of relevant history, physical exam, diagnosis, indication for therapy, and response to therapy, there is little to fear. Triplicate prescriptions, where still required, and local institutional policies regarding order and prescription renewal are significant barriers to proper opioid use. However, these should not limit the use of opioids when they are truly indicated.

Finally, some may be reluctant to use opioids, as they may wish to save the “big guns” for later. While tolerance to opioids does exist, it is not a major problem and usually occurrs in patients who require large doses of opioids. For most patients tolerance is not a problem at all. Many patients are successfully maintained on a steady dose of opioids for months to years. When patients complain of increased pain on opioids, it usually reflects worsening underlying disease. Equally important, there is no absolute dose ceiling for opioids (unlike acetaminophen, aspirin, and NSAIDs). Some patients may require and do quite well on the equivalent of more than 100 mgs of IV morphine an hour.

Specific Opioids

Available medications can be classified by their durations of action. Short-acting opioids can be administered alone or in combination with another analgesic, most commonly acetaminophen. Long-acting medications either have a long serum half-life or a delivery vehicle that allows less frequent administration. A distinction between “strong” and “weak” opioids is less useful. Opioids are similarly efficacious when given in adequate doses.

Short-acting agents are useful for the treatment of rapidly changing pain, such as acute pain and bursts of pain that occur with chronic pain. A short half-life allows rapid titration both up and down. The disadvantages of these agents also relate to their short half-lives. Frequent administration, if given around the clock, can make compliance difficult. Rapid changes in serum drug levels occur. Certain side effects, such as sedation, nausea, and euphoria, are more strongly related to changes in serum levels than to absolute blood levels. Thus, when taken intermittently patients may experience proportionately more side effects with these agents than with comparably dosed long-acting agents.

Long-acting agents either have long half-lives (methadone) or mechanisms for sustained delivery of a short-acting agent (sustained release morphine, hydrocodone, or oxycodone and fentanyl patches). Sustained-release preparations (p.51) have the advantage of allowing more frequent titration of doses (as the agents' pharmacologic half-lives are short) with the convenience of infrequent dosing. Sustained-release preparations may be safely adjusted after 24 hours. They also may result in smoother opioid blood levels, thereby minimizing side effects.


Morphine can be given as soluble tablets, as an elixir, as rectal suppositories, subcutaneously (SC), intravenously (IV), and intramuscularly (IM). While the elixir form is convenient for patients unable to swallow pills, its bitter taste is a problem for some. Morphine has a strong “first pass effect” when given orally. Thus, the oral to parenteral potency ratio is given as 6:1 for a single dose. With chronic administration, this first pass effect lessens, and the oral to parenteral potency ratio is 3:1. It is metabolized in the liver and excreted with active metabolites by the kidneys. The metabolite morphine 6-glucuronide's half-life is particularly prolonged in renal failure. Thus, kidney failure can result in significant drug accumulation of this active metabolite.18 Although controlled studies comparing different opioid use in renal failure have not been done, many experts recommend that morphine be avoided in such cases.19 Morphine can cause histamine release, resulting in itching or, rarely, more severe allergic symptoms, although the extent to which this occurs more frequently than with other opioids is debated.9, 20 21 22 As Katcher notes, the mechanism of morphine-related itching may not always be due to histamine release.23

Although histamine release can be a mechanism of true allergy to morphine, more common and clinically important is a “psychological allergy.” Having vomited in the past with morphine or simply out of a fear of the drug's reputation, certain patients are very resistant to the use of morphine. In such cases alternatives should be tried.

Tablet sustained-release (SR) morphine is regular morphine given in a slowrelease pill form. Because the morphine is in a special waxlike matrix, these pills cannot be crushed, which limits their use in patients who cannot take pills. The lowest dose available is 15 mg. They can be given two to three times a day. The great advantage of these agents is that they combine the best of both worlds: a pharmacologically short serum half-life and twice daily or TID administration. This allows rapid titration of dose. Slow release morphine is relatively inexpensive compared to short-acting morphine.

Kadian is a new form of SR morphine in capsule form with morphine placed inside very small polymeric beads (20 to 100 mg).24 It is possible to give Kadian capsules only once a day, although patients may require BID dosing. Kadian is expensive but may offer an advantage when enteral morphine administration is desired and patients are unable to swallow pills. The capsule can be opened and (p.52) sprinkled onto food, such as applesauce, or the beads may be flushed through a gastric-tube. Because of the expense involved, for stable pain methadone, which can be administered as a liquid, is generally preferred in such situations (see below).

If a patient taking slow release morphine becomes nauseated for whatever reason or is unable to take the medication orally, SR morphine can be administered rectally with approximately 90% to 100% of oral efficacy.9 This may be of help in an emergency. Because of possible discomfort, long-term SR morphine via the rectum is not usually recommended.

Generally speaking morphine is the drug of choice for parenteral administration unless side effects become apparent or very high doses (usually greater than 30 mg/h) are to be given subcutaneously (as the volume of administration may become too great for efficient SC administration).


Hydromorphone is similar to morphine in duration of action but is approximately five to seven times as potent. When administered orally it can be given in pill form, as an elixir, or as a rectal suppository. An advantage of hydromorphone is that the name does not evoke the strong emotional reactions of morphine.

Hydromorphone is the drug of choice for subcutaneous administration of a high dose of opioid, as it can be delivered in a very concentrated form. It is also the drug of choice when using parenteral opioids in a patient with renal failure or when rotating a patient who is experiencing agitation and myoclonus off morphine.

A slowrelease formulation of hydromorphone is available in some countries (not yet the United States). It may serve as a useful alternative to morphine and oxycodone. Angst demonstrated that this agent is effective, although the time to peak-effect is prolonged, approximately 9–12 hours, significantly longer than sustained release morphine or oxycodone preparations (3–6 hours).25, 26


Short-acting forms of oxycodone (without acetaminophen) are now available in pill and liquid form. Oxycodone is approximately 1.5 times as potent on a milligram (p.53) to milligram basis compared to chronic use morphine. It is more expensive than morphine.

SR oxycodone is quite similar to SR morphine in dosage and dosing intervals. SR oxycodone 10 mg ql2h is roughly the same as two Percocet tablets (5mg oxycodone/325mg acetaminophen strength) spread over 12 hours. Histamine release appears to be less of a problem than with morphine. Oxycodone should be considered for patients who experience pruritus on morphine and when an oral SR agent is needed in the presence of significant renal failure. It has been argued that oxycodone produces fewer mental status changes than does morphine. This may be a reason for using the drug when mental status changes are attributed to morphine.27 Other studies have not demonstrated significant differences in side effects between long-acting oxycodone and morphine.28 However, large controlled trials of the two situations in which an advantage for oxycodone over morphine has been suggested (in significant renal failure and in patients with delirium or at very high risk for delirium) have not been performed.

SR oxycodone can offer a great psychological advantage in overcoming “opioidphobia” in patients and health care providers. It may be easier to achieve compliance if it is explained that it is the same active ingredient found in Percocet. The oxycodone is just “spread out” over 12 hours. Unfortunately, there have been reports in the United States of SR oxycodone becoming a drug of choice for opioid abusers, who crush the drug and then take it. Pharmacologically, I know of no reason why oxycodone would have any more or less addictive potential than does any other opioid. Thus, a new wave of opioid-phobia is arising in response to this recent fad of addicts, which may inappropriately limit the use of this agent when otherwise indicated. SR oxycodone is significantly more expensive than is SR morphine. Its use instead of morphine should be justified on a case-by-case basis.


Methadone has a long half-life, which allows it to be used on a BID or TID schedule for most patients. It can be given in liquid form and is thus useful in patients who have an enteral route of administration when they are unable to take pills or have a feeding tube. Methadone is cheap. The major problem with methadone is its long half-life. It cannot be rapidly titrated up or down, as it takes days to achieve stable serum levels. Thus, methadone is not indicated in the treatment of rapidly changing pain. Low-dose methadone can be an excellent and (p.54) inexpensive agent for patients with stable, low-grade, chronic pain. As a rule of thumb, the dose of methadone should not be increased any more rapidly than every three days to avoid possible stacking of doses. In addition to direct opioid receptor agonist effects, methadone blocks n-methyl-D-aspartate (NMDA) receptors, which may be helpful in refractory pain syndromes.29, 30

Considerable controversy has arisen regarding conversion ratios for methadone. Recently experts have noted a useful principle: the higher the dosage of the opioid being converted to methadone, the lower the conversion methadone dose should be.30, 31 As summarized by Ripamonti, “The results of our study confirm that methadone is a potent opioid, more potent than believed. Caution is recommended when switching from any opioid to methadone, especially in patients who are tolerant to high doses of opioids.”31

Fentanyl Patches

Fentanyl patches may be useful for patients with chronic, stable pain who cannot use the oral route for opioid administration. Patches should not be used for acute pain or in very unstable situations (rapidly escalating or de-escalating pain). Fentanyl has a very short half-life. Transcutaneous patches allow a noninvasive, nonenteral route of opioid delivery, which can be very useful. Serum blood levels remain stable, allowing consistent analgesia and minimizing side effects. Fentanyl patches are expensive. The lowest-dose patch, 25 micrograms per hour, changed every 72 hours, is too strong for some milder chronic pain syndromes. The patches work poorly on very hairy or oily skin. Some patients become allergic to the patches. Absorption increases with increased skin temperature, as with fever, which can be a problem for patients with temperature spikes.32 Additionally, there are anecdotal reports of decreased absorption (and analgesia) with hypothermia, as is commonly present in the dying process. The clinical significance of these temperature effects is debatable. Further research is needed. Although the drug has a short half-life, it takes approximately 12 to 16 hours to build a reservoir of the drug in the subcutaneous fat when initiating therapy and somewhat longer for the drug to wear off when discontinuing the patch. Caution should be exercised (p.55) during initiation and discontinuation of therapy with this drug. On initiation care must be taken to ensure adequate analgesia until the patch “kicks in.” If a long acting opioid, such as SR morphine, is given at the time the patch is placed, this is usually adequate, as this dose will last approximately 12 hours. Otherwise, short-acting agents, such as subcutaneous injections, must be given. Patches cannot be cut in half to reduce the dosage.


Recently available tramadol is an opioidlike drug that binds mu receptors, as do opioids. This drug is unique in that it also blocks reuptake of serotonin and norepinephrine in the CNS. Such reuptake inhibition is a coanalgesic effect to opioid receptor blockade. This inhibition is believed to be partially responsible for the analgesic properties of the drug. Initially, it was thought to have less potential for abuse and to cause fewer side effects. However, with more widespread use concerns have arisen as to abuse potential, and other common opioid side effects have been observed. This drug is more expensive than are traditional agents, and thus routine use is not encouraged. It may be considered in cases with special concerns regarding constipation or respiratory depression.33

Opioid Combinations

Combination medications, such as acetaminophen with codeine, oxycodone, or hydrocodone, are similar in terms of duration of action. They differ in terms of opioid potency and coanalgesic dose (acetaminophen or aspirin). Oxycodone (in Percocet, Percodan) appears roughly equipotent and similar in action to hydrocodone (in Vicodin), although good bioequivalent studies for hydrocodone have not been done. However, Vicodin contains 500 mg of acetaminophen, compared to 325 mg in Percocet. These medications are effective when a short-acting agent is needed for mild to moderate pain. Their use is limited by the addition of acetaminophen or aspirin, which are toxic in high doses. Acetaminophen and aspirin as antipyretics may also mask fevers in cases when it would be important to identify a fever, such as immunosuppressed states and postoperatively. Various formulations are available in both pill and elixir forms. Generic formulations, when available, are less expensive.

Acetaminophen with codeine is probably the most commonly used combination analgesic. Codeine is both a direct analgesic, weakly binding mu receptors, and a prodrug. It is converted into morphine in the liver. Patients deficient in a converting enzyme, CYP2D6 (10% of population) or those taking inhibitors of this enzyme, (examples are quinidine, cimetidine and fluoxetine), may not achieve analgesia because of an inability to convert codeine to morphine.34 Such patients may experience relief with other agents, such as acetaminophen with oxycodone.

(p.56) Perhaps the biggest problem associated with these drugs is that many physicians will not prescribe stronger opioids when needed if the patient is receiving a maximum dosage of a combination drug. Whether this is out of ignorance, fear, or, at times, laziness is hard to say. I think of this barrier to better pain relief as “the combo wall.” However, the barrier exists only in the mind of the prescribing physician. It is very easy to get over this wall—just use an opioid without acetaminophen. For example, if using oxycodone with acetaminophen, simply give pure oxycodone in either a long-or short-acting formulation and continue to titrate up the drug as needed. In the United States certain combo drugs such as hydrocodone and acetaminophen have a lower level of regulatory control. In some states, special prescriptions, triplicates, are required when prescribing a pure opioid. Although using triplicates, where necessary, is a nuisance, it seems inexcusable to deny patients pain relief if needed because of the hassle involved in filling out these forms.

Meperidine (Demerol)

Meperidine has few, if any, advantages over morphine. Meperidine is different in subtle ways from other opioids (greater smooth muscle relaxing effects related to anticholinergic effects, less anti-tussive effect), but these offer no definitive advantages. If given, meperidine should be given IM or IV, not SC, as it is irritating.9 Meperidine has a toxic metabolite, normeperidine, with a longer half-life than meperidine, which can cause altered mental status, myoclonus, and seizures. This is particularly true in the presence of decreased renal function. Meperidine should not be used for more than 72 hours, not for chronic pain, and never in the presence of renal failure.

Propoxyphene (Darvon, Darvocet)

I can think of no reason why propoxyphene-containing agents would be drugs of choice. The analgesia provided is not superior to that of other agents, and there may be a greater potential for side effects associated in part with the accumulation of norprorpoxyphene, a metabolite that can cause cardiac toxicity.35 The half-life of propoxyphene is 12 to 15 hours and is 30 to 36 hours for norpropoxyphene. Because of these prolonged half-lives, it takes two to three days to achieve steady state serum levels.9 Thus, there is a potential for dose-stacking, as with methadone.

Medications for Neuropathic Pain

As a rule, treatment of neuropathic pain is more difficult than is treatment of nociceptive pain. Referral to a specialist should be considered if the basic measures (p.57) outlined below are not successful. Physicians should also be reminded that some neuropathic pain (especially when mixed with nociceptive pain) may respond to the agents discussed above, NSAIDs and opioids. However, these agents are less effective for pure neuropathic pain than they are for nociceptive pain.

Three major classes of medication are commonly used in the treatment of neuropathic pain: antidepressants, especially tricyclics; anticonvulsants, especially gabapentin and carbamazepine; and sodium channel blockers, especially mexiletine. A primary care provider may initiate tricyclics, carbamazepine, or gabapentin. Most primary care providers should seek consultation before using other anticonvulsants or sodium channel blocking agents. As a special case, neuropathic pain caused by tumor-related nerve compression is often relieved by steroids by alleviating swelling around the tumor, thereby reducing compression and pain. Dexamethasone 4–8 mg qd. is often effective.


The analgesic properties of these agents appear to relate primarily to their ability to block neuronal reuptake of serotonin and norepinephrine in the CNS. The anticholinergic properties of tricyclics do not appear to have analgesic effects. It is therefore curious that the most anticholinergic agent of available tricyclics, amitriptyline, is used most frequently by physicians for neuropathic pain. This is probably because it is the oldest agent. Agents with less anticholinergic effects, such as nortriptyline and desipramine, should be considered. (Note: these agents may still cause hypotension and should not be used in patients with heart block.) Tricyclics have been used particularly for dysesthetic and constant neuropathic pain, such as diabetic neuropathy. However, they can be used successfully for more paroxysmal pain, such as trigeminal neuralgia. There appears to be a dose-response curve in the use of these agents. While therapy should generally be initiated at low doses (10–25 mg for most tricyclics), dosage should be slowly raised to therapeutic levels (usually 75 mg) or until unacceptable side effects appear if pain relief is inadequate.9

Newer antidepressants, such as serotonin reuptake inhibitors (fluoxetine, sertraline), have been inconsistent in their analgesic properties and are not recommended for analgesia. The reason for this lack of analgesia is puzzling and not known.


Most anticonvulsants have been found to have some analgesic effect on neuropathic pain.36 The most studied is carbamazepine. It has been most commonly used in paroxysmal pain syndromes, such as trigeminal neuralgia, although efficacy (p.58) has also been demonstrated in more constant pain syndromes, such as diabetic neuropathy. In choosing between a tricyclic and carbamazepine as a first-line agent, it would be reasonable to consider the “two-for-the-price-of-one” principle. Depressed patients with either steady or paroxysmal pain may do better with a tricyclic (barring contraindications). A patient with seizures may better benefit from carbamazepine.

Carbamazepine can be sedating. Patients treated with it should have complete blood counts (CBC) and liver function tests monitored, as blood dyscrasias and LFT abnormalities can occur. These generally resolve with discontinuation of the drug. In patients prone to blood dyscrasias or liver abnormalities, carbamazepine should be used with caution, if at all. A new anticonvulsant, gabapentin, is increasingly used in the treatment of neuropathic pain.3738 39 It is remarkably well tolerated, with some evidence of anxiolytic effects, and therefore may be helpful for anxious patients.40, 41 Unlike carbamazepine, it lacks significant drug-drug interactions. Patients may experience nausea, especially with rapid dose escalation, or dizziness with higher doses. I have noticed a tendency recently for physicians to prescribe gabapentin in very low doses, 100 mg qhs, for example. Gabapentin is an expensive drug. It makes no sense to prescribe this drug in dosages that virtually prohibit any efficacy. When used, the dosage should be titrated up to therapeutic levels and the patient observed for a clinical response. If unresponsive, the medication should be discontinued.

Routes of Opioid Administration

I turn now to a more detailed discussion of the use of opioids. As mentioned in the section on pain strategy, having considered specific side effects to be avoided and “two-fers,” the route of therapy is usually a key consideration in choosing a particular agent.

Oral Route

Generally, the oral route is the preferred route of administration. It is easy to use, and medications are generally cheaper. The oral route also maximizes patient autonomy. It works poorly if patients have trouble swallowing, are intermittently nauseated, or need rapid onset of analgesia. Patients on oral opioids who have nausea and vomiting may have a particularly difficult time reaching a steady state of opioid blood level, which perpetuates nausea. Confused patients and patients dependent on others for administration may have difficulty complying with oral regimens, especially if short-acting agents are relied upon. (p.59) Short-acting oral agents, which generally take an hour to reach peak effect, do not work as fast as parenteral agents, which may be a disadvantage if rapid analgesia is desired. This delayed time to peak effect risks slow titration to adequate analgesia or (if additional doses are given prior to peak effect) a “stacking” of multiple doses, resulting in overdosage. On the other hand, use of oral agents, especially long-acting opioids, may avoid possible toxicities associated with rapid increases in blood opioid levels. Some patients may prefer equianalgesic parenteral doses of drugs, believing a shot or injection is more potent. As injections usually take more work on the part of the administrator or nurse, a preference for parenteral injections may reflect a desire for more hands on care. I have noticed this tendency particularly in fearful and isolated dying patients.

Enteral Tubes

Nasogastric (NG), percutaneous endoscopically placed gastrostomy (PEG) tubes, and jejunal (J) tubes, if already present, may also be used for drug delivery. These tubes overcome the inability of patients to swallow. More noxious agents, such as concentrated liquid morphine, which is bitter, can be easily administered. Absorption of some drugs may be limited if the patient is vomiting or has significant intestinal obstruction. Probably the greatest problem associated with tubes is the administration of long-acting opioids. SR opioids, such as SR morphine and SR oxycodone, cannot be crushed, as to do so would release the total drug in a short-acting and excessively strong form. Methadone in liquid form or Kadian (morphine in polymeric bead form, which is expensive) can be used if administration of a long-acting agent by tube is desired.

Transdermal Route

Currently, the only major opioid available by the transdermal route is fentanyl. The transdermal route is useful when the enteral route cannot be used. Nauseated patients, patients with poor compliance, and patients unable to swallow are all potential candidates. It takes at least 12 hours (12 to 22 h) for fentanyl to work by the transdermal route. When removed, serum levels fall, on average, 50% in 17 hours.42 Because of this slow onset and offset of serum levels, fentanyl is useful only for stable, chronic pain and should not be used to treat acute pain. Transdermal fentanyl also works poorly for patients who have very high opioid needs (generally above 500 mcg/h—five 100 mcg/h patches), as patients tend to have difficulty tolerating more than five patches. Transdermal fentanyl is relatively expensive compared to long-acting oral opioid preparations.

(p.60) Transmucosal

Fentanyl is now available as a lozenge that can be administered through the buccal mucosa. Although expensive, this approach allows rapid onset of analgesia (about 20 minutes) without parenteral administration.43


Used primarily in the treatment of dyspnea, opioids administered by a nebulizer can allow rapid peak blood levels, comparable to parenteral administration.44 Bioavailability via this route is a subject of debate. It would be safest to assume a high degree of bioavailability and then titrate the medication up, based on patient response. Only IV preparations are used for aerosols, with similar peak levels to those obtained using IV administration. Morphine is most commonly used. Care must be taken as, morphine can theoretically release histamine locally, causing bronchospasm, although the clinical significance of this is debatable. I have not seen a patient in which this has happened. All patients I have treated via this route tolerated morphine via other routes without pruritus. It may be wise to give patients who have not previously taken morphine a test dose via another route before giving aerosolized morphine.

Parenteral Administration of Opioids

Parenteral administration of opioids should be considered when:

  1. 1. Other routes of administration, especially oral, are not feasible.

  2. 2. Rapid dose titration is desired.

  3. 3. Very high doses are required.

  4. 4. Pain is very unstable, requiring rapid adjustments up and down.

Intravenous Route

The IV route allows the most rapid administration of opioids. This may be useful if rapidly titrating doses is necessary or in treating acute pain. IV administration of opioids offers little, if any, advantage over the SC route. In a prospective cross-over study patients were given IV and SC morphine infusions. There was no significant difference in perceived analgesia between the two routes or in side effects.45 For practical purposes IV and SC doses are equivalent. The IV route does have a slightly faster onset of action by a few minutes. For very rapid dose titration this may be an advantage. The IV route may be used if an IV is otherwise necessary or if long-term IV access is available—via a (p.61) MediPort or percutaneous intravenous catheter (PIC) line, for example. Rarely should an IV be placed simply for the management of chronic pain. The SC route is safer and better tolerated and provides equivalent analgesia.

Intramuscular Route

Opioids may be given via the IM route, although in the majority of cases the SC route is less painful and allows adequate absorption. Of commonly used agents, only meperidine (Demerol) must be given IM, as SC administration is irritating.9 Patients who require large doses of morphine—may also prefer the IM to the SC route, because injection of large volumes may be irritating.

Subcutaneous Route

Opioid administration via the SC route is generally preferred to the IM. It is useful when short-acting agents need to be administered infrequently. For example, many patients who die over 24 to 48 hours and cannot tolerate oral opioids may be adequately managed with SC morphine injections alone or in combination with a nonoral basal medication, such as a fentanyl patch. The SC route is also frequently used for long-term parenteral administration of opioids.

I am amazed and appalled by the resistance of the medical community to the use of this route for both injections and infusions of opioids. SC injections of opioids (into fat) hurt much less than those into deeper, more tender muscle. Consider the irony. The whole idea is to relieve pain. Why would clinicians unnecessarily use a more painful route of administration? Demonstrating the power of culture, the environment, and resistance to change, I have noticed some residents I have trained continue to order IM injections when they leave our palliative care ward. When asked about this practice, some have guiltily admitted that they are tired of being hassled when ordering opioids via the SC route by other members of the health care team, attending physicians, nurses, and pharmacists, who question this behavior. While I am sympathetic to the pressure they may experience, courage is encouraged. Almost all of us will be in an emergency room somewhere, someday and require an opioid injection. Personally, I would like a doctor with the courage to do what is right for me (and is less painful), despite what others might say. (This anecdote also highlights the importance of improving education across the spectrum of health care disciplines. It is not enough just to train physicians.)

Continuous SC infusion of opioids (possible with all but meperidine) offers the following advantages:

  1. 1. Ease and comfort of starting infusion: SC infusion needles are quickly and easily inserted into SC fat, either in the abdomen or thigh. Unlike IVs, it is impossible to miss. Insertion is far less painful.

  2. (p.62) 2. Lower infection rate: Relative to IVs, complications, especially infections, are far less common. SC needles are often changed every three days, but may stay in for up to a week.

  3. 3. Greater freedom of movement: SC pumps are often small and may be worn on a belt. As the arms are not used, patients may use their arms more freely, maximizing their independence of movement.

Problems with subcutaneous infusions

The major problem with SC infusions is that irritation of the subcutaneous tissue is often volume related. As a rule of thumb, if more than 3 cc's per hour are infused, irritation and pain may be experienced.

A problem can arise when high doses of opioids are needed and the SC route is desirable. It may be difficult to administer morphine, for example, if more than 30 to 40 mg per hour are required. Switching to the more potent hydromorphone, which maybe delivered in concentrations as high as 10 mg/cc (equivalent of 50 to 70 mg of morphine/cc), will allow most patients to be successfully treated using the subcutaneous route.

Principles of Basal and Breakthrough (Incident) Drug Dosing

Having determined the opioid to be used and the route of administration, the next questions that usually arise are, “What basal (around-the-clock) drug dose should I prescribe, and what breakthrough (short-acting) drug should I prescribe for breakthrough, or incident, pain?” Much of the art of good pain medication prescription lies in calculating the relative basal and breakthrough doses.

Most patients who have chronic pain require around-the-clock drug dosing. How does one determine the relative basal and breakthrough doses? The following principle is helpful: the more chronic the pain, the more one should rely on the basal dose; the more acute (or unstable) the pain, the more one should depend on the breakthrough, or short-acting, dose.46

In treating chronic, relatively stable pain, a good rule of thumb is that the breakthrough dose should be at approximately that dose that will double the serum opioid level over the basal level when the peak effect has been obtained. This is because a 100% increase (doubling the dose) is usually quite safe. The total (p.63) dosage of breakthrough medication taken in 24 hours should rarely exceed the 24 hour basal dose in chronic, stable pain. If it does exceed this (or whenever very frequent dosing is given), the basal dose usually needs to be increased. Breakthrough doses should be administered at intervals such that the peak effect of the drug given via a certain route occurs before the next possible breakthrough dose. It is unnecessary and undesirable to schedule breakthrough doses based on the duration of action of the drug. For short-acting oral opioids, such as morphine, the peak effect occurs in approximately one hour. The duration of action of oral morphine in most patients is four hours. Generally, oral morphine as a breakthrough drug should be dosed ql-2 hours, not q4 hours. SR oral opioids usually show a peak effect in three to six hours. IV peak effects correlate with lipid solubility. Peak effect is usually seen in 10 to 15 minutes for morphine. For SC/IM administration, peak effect may occur somewhat later than that for IV. Transmucosal fentanyl demonstrates peak plasma levels in approximately 20 minutes.

The importance of this principle can be appreciated by considering “standard medical practice” in many hospitals and nursing homes. Most clinicians still order oral short-acting opioids every four to six hours PRN pain. Consider this practice more closely. Most short-acting oral opioids do last approximately four hours, assuming normal pharmacokinetics. Dosing every six hours makes no sense, as the drug has worn off significantly by that time. If the patient required exactly the prescribed opioid dose every four hours to relieve pain and if nursing staff administered the opioid exactly at four hours, then adequate analgesia would be achieved. However, q4h dosing effectively prohibits drug titration against pain, given the drug lasts only four hours. Nurses (and other care providers) often cannot respond immediately at four hours to administer a new dose. Delays of 30 to 60 minutes are not uncommon. During this time the drug may wear off. The peak effect of the new dose will not occur for another hour. The chance of inadequately treated pain in this gap increases substantially. Of course, patients should not generally require opioids as frequently as every one to two hours. If this does occur, it means that either the drug dose is too low or adjustments must be made in basal drug dosing. I suspect the bad habit of prescribing short-acting oral opioids q4–6h arose both from a misunderstanding that stressed dosing based on duration of action, not peak effect, and from concerns about acetaminophen or aspirin toxicity if combination drugs were being used. The misunderstanding can be addressed through education. (p.64)

In very unstable, or acute, pain situations it is often appropriate to use no basal dose or to rely primarily on short-acting breakthrough doses. As pain stabilizes and comes under control (and if pain persists in a chronic form), the daily dosage of the short-acting agent used can be of help in calculating a new basal drug dose. The key to appropriate management is flexibility and frequent reevaluation.

When raising the basal dose, it is important to remember to increase by percentage, not milligram. At a minimum an increase should be 25% of the prior dosage. Commonly, the dose is increased by 25% to 100%. In higher dose ranges, physicians tend to underdose because they often increase dosages by milligrams. For example, many physicians are quite comfortable increasing morphine from 2 to 4 mg per hour (a 100% increase of 2 mgs) but will seriously underdose in increasing a patient from 20 mg to 22 mgs (a 10% increase of 2 mgs)

Many times I have made the mistake of dutifully increasing the basal drug dose while ignoring the breakthrough dose. For example, 10 mg of morphine may have been a reasonable breakthrough dose when the patient was on 30 mg of sustained-action morphine ql2h, but it is clearly inadequate if the basal dose has risen to 90 mg q!2h. It is important to remember to adjust breakthrough doses in parallel to basal doses.

Parenteral Basal/Patient Controlled Analgesia Dosing

Most infusion systems allow settings for both basal and patient controlled analgesia (PCA) opioid doses. Basal doses are set as X mg/h. PCA doses are ordered as Y mg qZ minutes. For SC systems, breakthrough intervals should be no more often than q15 minutes. Every 20 to 30 minutes may be preferable to ensure adequate drug delivery before giving an additional dose.

(p.65) Once a steady-state basal dosage has been determined, the PGA dosage and interval can be calculated. For stable, chronic pain in accord with the palliative care note above, a good dose is often one that results in a doubling of the serum opioid blood level with peak effect. Dosing intervals should be set such that stacking of PCA doses does not occur. Thus, a patient on 6 mg basal SC morphine might have 2 mg q20 minutes ordered for a PCA dose. In contrast, if a patient is unable to push the PCA button (due to illness or altered mental status, such as dementia), either the basal rate must be emphasized or nursing staff (or family) instructed to assess and administer the PCA dose as needed.

As is true for oral breakthrough doses, patients usually need occasional PCA boluses, but generally these should be less than three to five per day. Such incidents may be anticipated (for example, before nursing care or dressing changes) or unanticipated. If more boluses than this are required, pain should be reassessed and consideration given to raising the basal dose.

It is usually safe to increase the basal dose administered over 24 hours by the amount of PCA doses given over the prior 24 hours. For example, a patient on 6 mg/h basal morphine infusion and a PCA of 2 mg q20 minutes has required 24 injections for a total of 48 mgs of PCA dosing. Dividing this total by 24, this is the equivalent of 2 mg per hour. Thus, at a minimum, the basal dose should be increased by 2 mg/h, from 6 to 8 mg per hour (if pain was well controlled with these PCA doses). If pain was not well controlled with this combination of basal and PCA dosing, a higher basal dose, 10–12 mgs, for example, may be needed.

If no PCA boluses are required and the patient is pain free, the basal dose may be too high. This should especially be considered in a patient who appears sedated or talks with slurred speech. If opioid excess is suspected, holding administration of the opioid for a couple of hours and then lowering the dose will usually suffice. Rarely is naloxone administration required. Naloxone should generally be reserved for patients who are beginning to show signs of respiratory depression or hemodynamic compromise, usually associated with bradycardia. Even here, in all but the most extreme cases, consideration should be given to administering small boluses, 0.1–0.2 mg (1/4–1/2 an ampule) incrementally as needed to avoid complete opioid reversal and concomitant pain exacerbation and chemical withdrawal.

Parenteral Basal: Pca Dosing in Unstable and Acute Pain

Unstable pain changes rapidly up and down. Opioids require frequent adjustment. Rapidly lessening pain may be seen in most acute pain syndromes, in which pain naturally decreases over time. Pain following surgery or trauma is usually of this nature. Chronic pain, such as is found in many cancers, may also rapidly decrease in certain situations, such as when a painful focal metastasis is radiated.

(p.66) When pain is either rapidly increasing or decreasing, proper opioid therapy requires less reliance on basal opioid doses and more on PGA doses. In the extreme, many surgeons treat postoperative pain exclusively with PGA doses. Such therapy has the advantage of minimizing the chance of opioid excess that would result from unnecessary amounts being administered to a patient with lessening pain and a decreasing need for opioids. Excessive reliance on PGA dosing risks making the patient overly dependent on pushing the PGA button on time. Such patients may state that pain is well controlled while awake, but when they fall asleep (and thus cannot push the button) they may suddenly awaken in pain and play “catch-up” with PGA doses. Sleep is thus disturbed, which is an impediment to healing.

Often, what works best in such situations is the use of a PCA: basal hourly ratio that is greater than the 1:1 ratio described for chronic, stable pain. Such a ratio will allow both rapid titration up (in the case of increasing pain) or down.

All machines with PCA: basal drug administration capabilities I have used have a means of recording how many PGA doses were given, when they were given, and often how many attempts were made to administer a dose (a crude measure of desperation). Machines may have a button labeled “history” that allows the clinician to scroll through the PGA history over the past several hours. This data should be used in making adjustments as discussed above.

Conversion among Different Opioids

One would think that converting from one opioid to another would be a simple thing. In theory it is. Just find the right ratio, do the math, and voila! However, experts disagree on what the appropriate ratios are.30, 47 Recent studies have challenged traditional conversion ratios, which historically were often based on single dose comparisons rather than chronic dosing. Proper conversion is dependent upon a variety of factors—drug dosage, cross-tolerance (or lack thereof) among opioids, and physiologic differences in drug metabolism. As if this were not enough, the skill of converting opioids requires more than the use of simple conversion rates. The conversion process must take into account such factors as the amount of residual drug in the patient's system and the time to achieve steady-state blood levels with the new drug as well as individual patient responses during the conversion process. Pereira and Anderson's articles, referenced above, offer excellent recent reviews of controversies in this area for those who wish to pursue further reading. Here, I offer some principles that should help guide conversion efforts:

  • When converting from one chronically (around-the-clock) administered opioid to another, first calculate the 24-hour equivalent of the drug from which you are converting. Then convert to the 24-hour equivalent of the new drug (or (p.67) same drug via a different route of administration) using published conversion ratios.48

  • Calculate the dosing interval for the new drug.

  • Divide the 24 hour dose as appropriate for the new dosing interval. (For example, for ql hour dosing, divide by 24; for q12 hour dosing, divide by 2.)

  • Round off this value (up or down, based on factors such as the quality of pain control at that time and breakthrough drug use.) This dose and dosing interval can serve as the target for the new drug, but not usually as the initial order in the conversion process.

  • Account for residual drug in the patient's system (if any) during the conversion process! This is particularly important if converting from long-acting oral opioids or fentanyl patches. Overdosage may occur if an equivalent dose, based on a conversion table, is used while the old opioid is still in the patient's system. Initiation of the new drug, especially the basal dose, may be delayed if significant residual drug is in the body.

Example. Mr. Smith had been taking sustained-release oral morphine 60 mg q12. His family just managed to get him to take his last oral dose two hours ago. He is admitted to the hospital and can no longer take pills. His pain is well controlled. You wish to start him on a SC (or IV) infusion of morphine. How do you convert to parenteral morphine?

  1. 1. Old 24 hour oral morphine dose = 120 mg.

  2. 2. Conversion tables show that the oral to parenteral ratio for morphine is 3:1. Therefore, divide 120 mg by 3 to obtain the 24-hour equivalent of parenteral morphine = 40 mg parenteral morphine per 24h.

  3. 3. Basal infusions of morphine are written qlh. Therefore divide 40 mg by 24 = 1.66 mg/h.

  4. 4. As his pain is well controlled, round-down 1.66 to 1.5 mg/h IV or SC. This is the target basal dose.

  5. 5. As approximately 10 hours of sustained-release morphine is in the patient's system, this basal dose should be started in approximately 10 hours. Until then breakthrough doses (approximately 1 mg q30 minutes) may be used. This is the initial drug order.

  6. 6. If the patient begins to need frequent breakthrough doses before 10 hours pass, a low basal dose, 0.5–1.0 mg/h, may be initiated based on the reported pain score and breakthrough drug usage. This is the process of adjusting the initial order in the direction of the target order.

• The initial order for the new opioid should have a relatively low basal dose and relatively high breakthrough dose. In calculating a new basal dose, where controversy exists as to the correct conversion value, it is safest initially to use (p.68) a conversion value that will result in a more conservative (low) new drug dose relative to the old dose. Based on the estimated wash-out period for residual opioids and the individual patient's response in the conversion process, gradually increase the new basal dose, if necessary, based on reported pain scores and the use of breakthrough doses. For example, in converting from oral morphine to oral hydromorphone, it would be better to use the more conservative ratio (from Table 4.1) of 180 mg morphine: 45 mg hydromorphone. However, if converting/rom hydromorphone to morphine, it is safer to use 180 mg morphine: 60 mg hydromorphone, as this will result in a lower initial basal morphine dose.

The reviews cited above demonstrate that particular caution should be used in converting from one opioid to another at high doses.30, 47 As opioids may differ significantly in terms of their mechanisms of action (for example, methadone's NMDA antagonism) and the degree of cross-tolerance and metabolism, conversion tables may be inaccurate for calculating true equivalent doses, which risks overdosage with the new drug (or occasionally underdosage). This has been found to be particularly true in converting from high doses of opioids such as morphine and hydromorphone to methadone; appropriate morphine-to-methadone ratios tend to be much higher than commonly published ratios.31, 49 That is, a lower methadone dose than is suggested by most published tables is appropriate.

TABLE 4.1. 24-Hour Drug Equivalencies of Selected Opioids



Morphine sulfate, parenteral

60 mg IM, SC, IV

Morphine sulfate, oral

180 mg PO (chronic use)


10–40 mg PO

Hydromorphone, parenteral

9–12 mg IM, SC, IV

Hydromorphone, oral

45–60 mg PO


120 mg PO


50–100 mcg/h patch (change q72h)


1200 mg PO


No consensus on equivalent dose

All equivalencies are approximations, not starting doses. There is no universal agreement on equivalent doses. Individual dose adjustment is essential!

(p.69) Recent articles have demonstrated the complexities of opioid conversion and argue against simplistic reliance on tables. Review of these articles is strongly recommended. 30, 47 These values are presented as crude guidelines for conversion. As discussed in the text, as important or more important than simply converting milligrams or micrograms is using a protocol that allows adjustment of conversion doses based on individual patient characteristics and responses over time.

Using above values to convert between drugs.

  1. 1. Calculate current drug 24-h dose (dose times number of times given per day).

  2. 2. Multiply this current 24-h dose times the ratio of 24-h equivalent dose of new drug over 24-h equivalent of old drug. This gives 24-h dose of new drug (equivalent doses from table above).

                       Pain Management

  3. 3. Divide new 24-h drug dose by number of times drug to be given per day. This gives new individual drug dose.

  4. 4. Order new individual drug dose to be divided per the dosing interval determined above. This is the target dose.

  5. 5. Accounting for residual drug in the system, if any, increase new drug toward this target, adjusting as necessary.

Example. Convert sustained-release morphine to parenteral hydromorphone (using conservative conversion—may need to increase hydromorphone later).

Current order reads: Sustained-release morphine 60 mg BID.

  1. 1. Current 24-h dose is 60 × 2 = 120 mg.

  2. 2. 120 mg (old 24-h oral morphine dose) times 9 mg (hydromorphone, parenteral) divided by 180 mg (morphine, oral) = 6 mg/24 h (new dose hydromorphone, parenteral).

  3. 3. Divide this 24-h dose by dosing interval—24—for ql hour basal dose = 0.25 mg/h (hydromorphone, parenteral). This is the target dose.

  4. 4. If residual drug is in the patient's system, initially use prn hydromorphone doses. Increase the basal dose based on patient response and prn usage.


                   Pain Management

Figure 4.4. Example conversion calculation. Writing in drug name, route, and time intervals in calculations helps avoid serious conversion errors.

Methadone. Considerable controversy has arisen regarding conversion ratios for methadone. Experts have noted a useful principle: The higher the dosage of the opioid being converted to methadone, the lower the conversion methadone dose that should be used. Thus, if a patient were on a high dose of morphine, in converting to methadone I would initially use a conservative 180 mg oral morphine: 10 mg of methadone conversion ratio from Table 4.1. As summarized by Ripamonti, “The results of our study confirm that methadone is a potent opioid, more potent than believed. Caution is recommended when switching from any opioid to methadone, especially in patients who are tolerant to high doses of opioids.”31 I strongly advise using low equivalent doses of methadone initially, gradually increasing the dose while using more liberal doses of breakthrough opioid in the conversion process.

Hydromorphone, parenteral. Controversy exists as to the potency ratio of hydromorphone compared to morphine. Earlier studies suggested that parenteral hydromorphone was seven times as potent as parenteral morphine. More recent studies suggest a 5:1 relative potency.50 The oral: parenteral ratio for hydromorphone is 5:1 in most tables.

Oxycodone. While American conversion tables have listed oxycodone as being equianalgesic to oral morphine, there is some debate as to equivalency. Some have listed oxycodone as being two times as potent as morphine (Canada). A recent study suggested that oxycodone may be approximately 1.5 times as strong as morphine (Bruera, 1998). Thus, 120 mg of oxycodone is suggested here.

Fentanyl. Fentanyl comes in patches from 25 to 100 ug/h. Patches are changed every 72 hours. A 25 ug/h patch is approximately equal to 50 to 75 mg of oral morphine over 24 h. Levy suggests a simple rule of thumb: that the Fentanyl patch strength (in mcg/h e.g., 25, 50, 100, etc.) is equal to the SR morphine dose given BID.34 Thus, by this rule a 150 mcg/h patch is equivalent to (p.71) 150 mg SR morphine ql2h. Jannsen, the manufacturer of fentanyl patches, suggests a more conservative dosing schedule in converting to fentanyl patches. Jannsen's dosing table does not allow for a simple conversion value. They suggest a relatively higher patch strength at low oral morphine equivalents (25 meg patch for 45–134 mg morphine/24h—consistent with Levy's rule) and a relatively lower patch strengths at higher morphine doses (200 meg patch at 675–764 mg morphine/24h—not consistent with Levy's rule). (See manufacturer's product information for details). Their product information stresses that conservative conversion values used for converting to fentanyl patches may result in drug overdosage if used to convert from patches to other opioids, highlighting the danger of a simplistic use of conversion tables.

Codeine. Although this is the classic conversion value listed in tables, codeine is metabolized into morphine in the liver via the P450 system. About 10% of the population have trouble doing this, and many drugs—for example, Cimetidine and SSRIs—may inhibit this, resulting in even lower potency. Thus actual potency is highly variable.

Hydrocodone. Most conversion tables do not list an equivalent dose for hydrocodone. I have found hydrocodone:oral morphine ratios that range from 6:1 to 75:1 Thus, no recommendation can be given. My impression is that Vicodin (hydrocodone) and Percocet (oxycodone) are very similar in potency.

Special note to teachers of palliative care

Please learn from my mistake! For years I lectured residents and students on principles of pain management along the lines above. Trainees were inspired to go forth and do good deeds. They applauded, I basked in a teacher's glory, and nothing changed. In my teaching I had stressed attitudes (treating pain is important) and knowledge but had underweighted skill training. I found, for example, that most trainees could not use even a basic opioid conversion table (let alone deal with such subtleties as are discussed above). Most residents carried an opioid conversion table in their “peripheral brain”—a small handbook or, recently, a PDA. However, when I asked them to demonstrate the skill of using the conversion table to switch from one opioid to another, they could not. They got confused converting from one drug dosed ql2h to another drug dosed TID or qlh. Conversion tables listed an equivalent dose of fentanyl ql hour, but they could not figure out how this related to the fact that the fentanyl patch is changed q72h. Being unskilled in using tables, they would not use them, fearful of making a mistake (and potentially killing a patient). Thus, it is critical that trainees demonstrate the skills outlined above. Ideally, the practice of such skills will occur in real life under appropriate supervision (as is the case for other medical skills). Barring this, it is strongly recommended that trainees practice skills such as opioid (p.72) conversion using cases in groups or via self-directed learning. Teachers may wish to make up their own cases and ask trainees to perform certain skills (such as writing appropriate opioid orders using a conversion table), or trainees may wish to work independently.


Pain is a ubiquitous aspect of human experience and epitomizes human suffering. Historically, we have done a poor job of addressing pain. Mercifully, in the majority of cases it can be treated using relatively simple approaches. With a little work we can significantly improve our ability to alleviate the suffering associated with pain.



☼ This rule of thumb assumes simple escalation of methadone. Various schemes for converting other opioids to methadone have been described, including gradual daily decreases in prior opioid doses and daily equivalent increases in methadone over a period of days. (As described in a talk by L. Friedman, “Using Methadone,” American Academy of Hospice and Palliative Medicine National Conference, June 2001).

☼ Patient controlled analgesia (PCA) refers to a particular type of breakthrough dosing used when a predetermined parenteral drug dose is injected following the push of a button. In reality not all “PCA” breakthrough doses are controlled or administered by patients, who may be physically or mentally impaired. Clinicians or families may administer the breakthrough dose in such cases, which is still often (erroneously) called the “PCA” dose. Thus, in practice parenteral breakthrough doses, when programmed by a pump are often referred to as PCA doses, regardless of who administers the dose.

(1.) Good, M. et al., eds. Pain as Human Experience. 1992, University of California Press: Berkeley, p. 1.

(2.) Portenoy, R. K. and P. Lesage. Management of cancer pain. Lancet 1999; 353(9165): 1695–700.

(3.) Bonica, J. and J. D. Loeser. Medical evaluation of the patient with pain. In: Bonica J., C. Chapman and W. Fordyce, eds. The Management of Pain. 1990, Lea & Febiger: Philadelphia, pp. 563–79.

(4.) SUPPORT. A controlled trial to improve care for seriously ill hospitalized patients. The study to understand prognoses and preferences for outcomes and risks of treatments (SUPPORT). JAMA 1995; 274(20): 1591–8.

(5.) Won, A. et al. Correlates and management of nonmalignant pain in the nursing home. SAGE Study Group. Systematic assessment of geriatric drug use via epidemiology. J Am Geriatr Soc 1999; 47(8): 936–12.

(6.) Grossman, S.A. et al. Correlation of patient and caregiver ratings of cancer pain. J Pain Symptom Manage 1991; 6(2): 53–7.

(7.) McDowell, I. and C. Newell. Measuring Health: A Guide to Rating Scales and Questionnaires, 2nd ed. 1996, Oxford University Press: New York, pp. 355–79.

(8.) Weissman, D. E. and J. D. Haddox. Opioid pseudoaddiction—An iatrogenic syndrome. Pain 1989; 36(3): 363–6.

(9.) Twycross, R. Pain Relief in Advanced Cancer. 1994, Churchill Livingstone: London, pp. 240, 288, 325, 353, 409–16.

(10.) Jenkins, C. A. and E. Bruera. Nonsteroidal anti-inflammatory drugs as adjuvant analgesics in cancer patients. Palliat Med 1999; 13(3): 183–96.

(11.) Hudson, N., A. Taha, et al. Famotidine for healing and maintenance in nonsteroidal anti-inflammatory drug-associated gastroduodenal ulceration. Gastroenterology 1997; 112: 1817–22.

(12.) Hawkey, C. Cox-2 Inhibitors. Lancet 1999; 353: 307–14.

(13.) Silverstein, F. E. et al. Gastrointestinal toxicity with celecoxib vs nonsteroidal anti-inflammatory drugs for osteoarthritis and rheumatoid arthritis: The CLASS study: (p.73) A randomized controlled trial. Celecoxib Long-term Arthritis Safety Study. JAMA 2000; 284(10): 1247–55.

(14.) Porter, J. and H. Jick. Addiction rare in patients treated with narcotics [letter]. N Engl J Med 1980; 302(2): 123.

(15.) Joranson, D. E. et al. Trends in medical use and abuse of opioid analgesics. JAMA 2000; 283(13): 1710–4.

(16.) Allan, L. et al. Randomized cross-over trial of transdermal fentanyl and sustained release oral morphine for treating chronic non-cancer pain. BMJ 2001; 322(7295): 1154–8.

(17.) Radbruch, L. et al. Constipation and the use of laxatives: A comparison between transdermal fentanyl and oral morphine. Palliat Med 2000; 14(2): 111–9.

(18.) Mercandante, S. The role of morphine glucuronides in cancer pain. Palliat Med 1999; 12: 182–9.

(19.) Abrahm, J. A Physician's Guide to Pain and Symptom Management in Cancer Patients. 2000, Johns Hopkins University Press: Baltimore, pp. 135–6.

(20.) Grossman, S. A. et al. Morphine-induced venodilation in humans. Clinical Pharmacol Ther 1999; 60: 554–60.

(21.) Warner, M.A. et al. Narcotic-induced histamine release: A comparison of morphine, oxymorphone, and fentanyl infusions. J Cardiothorac Vase Anesth 1991; 5(5): 481–4.

(22.) Doenicke, A. et al. Intravenous morphine and nalbuphine increase histamine and catecholamine release without accompanying hemodynamic changes. Clin Pharmacol Ther 1995; 58: 81–9.

(23.) Katcher, J. and D. Walsh. Opioid-induced itching: Morphine sulfate and hydromorphone hydrochloride. J Pain Symptom Manage 1999; 17(1): 70–2.

(24.) Broomhead, A. et al. Comparison of a once-a-day sustained-release morphine formulation with standard oral morphine treatment for cancer pain. J Pain Symptom Manage 1997; 14(2): 63–73.

(25.) Angst, M.S. et al. Pharmacodynamics of orally administered sustained-release hydromorphone in humans. Anesthesiology 2001; 94(1): 63–73.

(26.) Hanks, G. and N.I. Cherny. Opioid analgesic therapy. In: D. Doyle, G. Hanks, et al., eds. The Oxford Textbook of Palliative Medicine. 1998, Oxford University Press: New York, pp. 331–5, 338–9.

(27.) Maddocks, I. et al. Attenuation of morphine-induced delirium in palliative care by substitution with infusion of oxycodone. J Pain Symptom Manage 1996; 12: 182–9.

(28.) Bruera, E. et al. Randomized, double-blind, cross-over trial comparing safety and efficacy of oral controlled-release oxycodone with controlled-release morphine in patients with cancer pain. J Clin Oncology 1998; 6: 3222–9.

(29.) Mercadante, S. et al. Switching from morphine to methadone to improve analgesia and tolerability in cancer patients: A prospective study. J Clin Oncol 2001; 19(11): 2898–904.

(30.) Pereira, J. et al. Equianalgesic dose ratios for opioids. A critical review and proposals for long-term dosing. J Pain Symptom Manage 2001; 22(2): 672–87.

(31.) Ripamonti, C. et al. Switching from morphine to oral methadone in treating cancer pain: What is the equianalgesic dose ratio? J Clin Oncol 1998; 16(10): 3216–21.

(32.) Southam, M. Transdermal fentanyl therapy: System design, pharmacokinetics and efficacy. Anti-Cancer Drugs 1995; 6(suppl 3): 26–34.

(33.) Grond, S. et al. High-dose tramadol in comparison to low-dose morphine for cancer pain relief. J Pain Symptom Manage 1999; 18(3): 174–9.

(p.74) (34.) Levy, M.H. Pharmacologic treatment of cancer pain. New Engl J Med 1996; 335(15): 1124–32.

(35.) Ulens, C., P. Daenens, et al. Norpropoxyphene-induced cardiotoxicity is associated with changes in ion-selectivity and gating of HERG currents. Cardiovasc Res 1999; 44(3): 568–78.

(36.) Tremont-Lukats, I.W., C. Megeff, et al. Anticonvulsants for neuropathic pain syndromes: Mechanisms of action and place in therapy. Drugs 2000; 60(5): 1029–52.

(37.) Hemstreet, B. and M. Lapointe. Evidence for the use of gabapentin in the treatment of diabetic peripheral neuropathy. Clin Ther 2001; 23(4): 520–31.

(38.) Garafalo, E. Gabapentin for the symptomatic treatment of painful neuropathy in patients with diabetes mellitus: A randomized controlled trial. JAMA 1998; 280: 1831–6.

(39.) Kishore, A. and L. King. Fast Fact and Concept #49; Gabapentin for neuropathic pain. End of Life Care Education Project, http://www.eperc.mcw.edu/, 2001.

(40.) Miller, L.J. Gabapentin for treatment of behavioral and psychological symptoms of dementia. Ann Pharmacother 2001; 35(4): 427–31.

(41.) Pande, A.C. et al. Placebo-controlled study of gabapentin treatment of panic disorder. J Clin Psychopharmacol 2000; 20(4): 467–71.

(42.) Hanks, G. and N.I. Cherny. Opioid analgesic therapy. In: D. Doyle, G. Hanks, et al., eds. The Oxford Textbook of Palliative Medicine. 1998, Oxford University Press: New York, p. 340.

(43.) Farrar, J. et al. Oral transmucosal fentanyl citrate: Randomized, double-blinded, placebo controlled trial for treatment of breakthrough pan in cancer patients. J Natl Cancer Inst 1998; 90: 611–6.

(44.) Chandler, S. Nebulized opioids to treat dyspnea. American Journal of Hospice & Palliative Care 1999; 1999(16): 418–22.

(45.) Nelson, K. et al. A prospective, within-patient, crossover study of continous intravenous and subcutaneous morphine for chornic cancer pain. J Pain Symptom Manage 1997; 13: 262–7.

(46.) Portenoy, R.K., ed. Cancer pain management: Update on breakthrough pain. Seminars in Oncology 1997; 24(suppl): 116.

(47.) Anderson, R. et al. Accuracy in equianalgesic dosing: Conversion dilemmas. J Pain Symptom Manage 2001; 21(5): 397–406.

(48.) Gordon, D.B. Opioid equianalgesic calculations. Journal of Palliative Medicine 1999; 2(2): 209–18.

(49.) Ripamonti, C. et al. Equianalgesic dose/ratio between methadone and other opioid agonists in cancer pain: Comparison of two clinical experiences. Ann Oncol 1998; 9(1): 79–83.

(50.) Lawlor, P., K. Turner, et al. (1997). “tio between morphine and hydromorphone in patient with cancer pain: a retrospective study.” Pain 72(1–2): 79–85.