Jay Thomas MD, PhD
Charles F. von Gunten MD, PhD
Optimal treatment of pain requires an understanding of the pathophysiology underlying the pain (ie, whether the pain is nociceptive, neuropathic, or mixed). Moreover, since pain is subjective, clinicians need to have an understanding of what the patient is experiencing emotionally. One important part of a multifaceted treatment program is pharmacologic intervention. The World Health Organization (WHO) has developed a useful three-step ladder that helps organize the pharmacologic approach to pain management (Figure 3-1):
Using clinical judgment, the ladder can be entered at any step. For example, if a patient has a broken bone, the clinician need not start at Step 1 and wait until pain control fails before moving on to the next step. It is also important to remember that even if pain is sufficiently severe to require medications from a higher step, combination therapy with medications from lower steps may still be used synergistically. Finally, clinicians may use adjunctive medications to optimize pain control at each step.
Step 1: Treating Mild Pain
Nonsteroidal Anti-inflammatory Drugs
The predominant action of nonsteroidal anti-inflammatory drugs (NSAIDs) is to inhibit the enzyme cyclooxygenase (COX), which mediates the conversion of arachidonic acid to prostaglandins and thromboxanes. Isozymes of this enzyme exist. COX-1 is expressed constitutively in many tissues and regulates gastric cytoprotection, renal autoregulation of blood flow, platelet aggregation, and vascular homeostasis. COX-2 is constitutively expressed in a few tissues, such as the central nervous system (CNS), bone, and kidney, but predominantly is induced in inflammatory states. A third isozyme, COX-3 is an RNA splice variant of COX-1 and appears to be localized predominantly in the CNS but is also present in the heart. Its clinical significance remains unclear.
NSAIDs impact pain processing in two known ways. First, peripheral nociceptors (afferent sensory nerve fibers that signal pain) are sensitized by inflammation and in turn augment inflammation. For example, in the presence of inflammation, a normally silent peripheral C-fiber may start firing in response to a mild stimulus, and its response to a normal noxious stimulus is enhanced. In addition, the activated nociceptor may release inflammatory mediators that maintain or strengthen the inflammatory milieu. By reducing inflammation, NSAIDs decrease this peripheral sensitization and neurally mediated augmentation of inflammation. Second, COX is present in the spinal cord and has been implicated in CNS events that lead to central sensitization. In an experimental example of this central sensitization known as “wind up,” a repetitive peripheral noxious stimulus can lead to central spinal changes that in turn lead to peripheral
hyperalgesia and allodynia. Hyperalgesia is a state where a noxious stimulus is perceived as more intense than it normally would. Allodynia is a state where anon-noxious stimulus, such as light touch, is perceived as painful. NSAIDs have experimentally been shown to prevent this central sensitization and enhancement of pain.
Figure 3-1. A 3-step ladder developed by The World Health Organization to help organize the pharmacologic approach to pain management. On a numeric scale, mild pain is rated between 1 and 3; moderate pain, between 4 and 6; and severe pain, between 7 and 10. Adjunctive medications may be added to any step.
NSAIDs include nonselective COX inhibitors (eg, aspirin, ibuprofen, naproxen) and COX-2 selective inhibitors (eg, celecoxib, rofecoxib, valdecoxib), which have a 200- to 300-fold greater inhibitory effect on COX-2 than COX-1. Rofecoxib and valdecoxib are no longer available in the United States due to cardiac side effect concerns.
There may also be nonprostaglandin-mediated effects of NSAIDs. Studies indicate that NSAIDs may decrease neutrophilendothelial cell interaction and may also decrease the production of nitric oxide. However, the clinical significance of these effects is unknown.
There are multiple classes of NSAIDs listed in Table 3-1. They include acetic acids, fenamates, naphthylalkanones, oxicams, propionic acids, and the COX-2 selective inhibitor. All agents inhibit COX but there may be differences in other pharmacodynamic properties that may help explain the variability in individual response.
NSAIDs in general are well absorbed and have high oral bioavailability. Oral forms typically reach peak effect for analgesia in 1 to 3 hours. NSAIDs are metabolized by the liver. Half-lives are variable. They all have a ceiling effect for efficacy but the risk of side effects continues to escalate with increasing dose.
Some NSAIDs are formulated as an elixir or suppository. These formulations can facilitate dosing when patients, especially those receiving palliative care, have difficulty swallowing pills. When formulations are not available commercially, pharmacists often compound an alternate formulation. In some circumstances, palliative care physicians have taken advantage of the fact that the oral and rectal routes of administration for many medications have similar pharmacokinetics. Often, oral pills can be used rectally with good effect. In fact, some oral time-release formulations retain their long-acting properties when used rectally, eg, morphine sulfate.
The authors recommend the use of nonselective NSAIDs as first-line agents for several reasons. First, there is no clear analgesic benefit of COX-2 selective NSAIDs over nonselective agents. Second, COX-2 selective agents may increase the risk of cardiovascular events, and even low-dose aspirin negates the gastroprotective advantage of COX-2 selective agents. If gastric protection is needed with nonselective NSAIDs, a proton pump inhibitor or misoprostol is effective prophylaxis. Whether nonselective NSAIDs significantly increase the risk of cardiovascular events requires further study. Individual risk-benefit analysis must guide prescribing practice.
Typically, doses are started low. When steady state is reached after 3 to 5 doses, doses can be titrated up to maximum recommended doses limited by either achieving an effect or side effect. NSAIDs, independent of class, appear to be equally efficacious as analgesics but an individual patient may respond variable to them. If one NSAID fails at maximal doses, it is reasonable to try another agent in a different class. Typical dosing of NSAIDs is shown in Table 3-1.
Especially when used long term, frequency of dosing may affect patient compliance; therefore, once or twice a day dosing may be advantageous. Renal function should be monitored after initiating an NSAID as well as intermittently thereafter if the NSAID is to be given long term.
COX-1 is involved in gastric protection. Therefore, COX-1 inhibition by nonselective COX inhibitors increases the risk of peptic ulcer, and COX-2 selective NSAIDs have less risk of gastrointestinal tract toxicity. However, studies have shown that concurrent use of even low-dose aspirin obviates the COX-2 selective gastrointestinal advantage. Furthermore, studies have shown that the use of either a proton pump inhibitor or misoprostol, a synthetic
analogue of prostaglandin E1, in conjunction with a non-selective NSAID can significantly protect against ulcer formation.
Table 3-1. Prescribing Guidelines for Nonsteroidal Anti-Inflammatory Drugs.
Platelet COX-1 is responsible for thromboxane A2 generation. Thromboxane A2 mediates platelet activation and aggregation. Thus, in general, nonselective NSAIDs (nonacetylated NSAIDs, such as salsalate, are exceptions) increase the risk of bleeding, while COX-2 selective inhibitors have no antiplatelet activity and no effect on bleeding risk. However, recent clinical trials have implicated some of the COX-2 selective inhibitors in increased risk of cardiovascular events. A plausible explanation for this phenomenon is that COX-2 selective inhibitors reduce endothelial cell prostaglandin I2 (prostacyclin) production but leave platelet prothrombotic thromboxane A2 production unaffected.
A nonselective COX inhibitor, naproxen, has also recently been implicated in increased risk of cardiovascular events, although the strength of this association is unclear. The National Institutes of Health (NIH) stopped a large trial designed to determine whether celecoxib or naproxen versus placebo decreased the risk of developing Alzheimer's disease. Without releasing exact numbers, the NIH stated naproxen increased the risk of cardiovascular events 50% over placebo.
To clearly resolve the cardiovascular effects of both nonselective and COX-2 selective inhibitors, further studies are required.
In normal circumstances, glomerular perfusion is not dependent on prostaglandins. However, in cases of chronic renal insufficiency and prerenal conditions, such as volume depletion, liver failure, and congestive heart failure, glomerular
perfusion is dependent on prostaglandin-mediated vasodilation. NSAIDs, by inhibiting this vasodilation, can decrease glomerular filtration rate and worsen renal function. Systemically, NSAID inhibition of vasodilation leads to increased vascular tone. This effect elevates blood pressure and can worsen preexisting heart failure. Both nonselective and COX-2 selective NSAIDs can adversely affect renal and systemic hemodynamics.
Acetaminophen's mechanism of action remains controversial. It has no peripheral anti-inflammatory effects. Its analgesic and antipyretic effects are believed to be centrally mediated. As mentioned previously, a COX-1 RNA splice variant termed “COX-3” has been identified in the brain. Acetaminophen is active as an inhibitor of this enzyme. Mouse studies have shown acetaminophen to reduce brain prostaglandin levels in a parallel with analgesia. Mice altered to lack either COX-1 or COX-2 showed this effect to be dependent on the COX-1 gene, which is needed to produce COX-3.
Acetaminophen is 60 to 90% orally bioavailable. Its onset of action is 15 to 30 minutes and peak serum levels (Cmax) are reached in 40 to 60 minutes. The half-life is about 2 to 4 hours. Acetaminophen is extensively metabolized in the liver. Importantly, about 10% of it is converted to a highly reactive toxic metabolite that is normally inactivated by glutathione. When glutathione stores are depleted, the toxic metabolite can cause severe hepatotoxicity. Given this dose-dependent toxicity, acetaminophen also has a ceiling effect.
Typical dosing is 500 to 1000mg orally every 4 to 6hours. However, maximal dosing should not exceed 4 g/d due to the risk of hepatotoxicity. This maximal amount should be further reduced for those with underlying liver disease or who consume three or more alcoholic drinks per day. Acetaminophen is available in tablet, capsule, elixir, and suppository forms.
Other than dose to dependent hepatotoxicity as described above, acetaminophen is generally well tolerated. Long-term therapy at high doses can lead to nephrotoxicity.
Chandrasekharan NV et al. COX-3, a cyclooxygenase-1 variant inhibited by acetaminophen and other analgesic/antipyretic drugs: cloning, structure, and expression. Proc Natl Acad Sci U S A. 2002;99:13926.
Ghilardi JR et al. Constitutive spinal cyclooxygenase-2 participates in the initiation of tissue injury-induced hyperalgesia. J Neurosci. 2004;24:2727.
Silverstein FE et al. Gastrointestinal toxicity with celecoxib vs nonsteroidal anti-inflammatory drugs for osteoarthritis and rheumatoid arthritis: the CLASS study: a randomized controlled trial. Celecoxib Long-term Arthritis Safety Study. JAMA. 2000;284:1247.
Step 2: Treating Moderate Pain
On this step of the ladder, Step 1 medications (NSAIDs and acetaminophen) are commonly combined with opioids. Table 3-2 lists these agents as well as their prescribing categories.
Opioids themselves have no theoretical ceiling effect. However, by virtue of their formulation, these combined agents have a ceiling effect imposed by their Step 1 component.
The opioids used in combination with NSAIDs and acetaminophen include tramadol, codeine, hydrocodone, and oxycodone. The less potent opioids codeine and tramadol are discussed in this section and the more potent opioids will be addressed below in Step 3: Treating Severe Pain.
Codeine, as all clinically useful opioid analgesics, eventually acts at µ opioid receptors, which are located in the brain and in the spinal cord. (See the Pharmacodynamics section under Step 3: Treating Severe Pain for a more detailed discussion of the action of µ-receptors.)
Codeine is predominantly a prodrug of morphine. Liver metabolism via the cytochrome P450 system enzyme CYP2D6 leads to activation. Patients who lack this enzyme (approximately 5 to 10% of whites) or who have concomitant inhibitors of it (such as fluoxetine or paroxetine) derive little analgesia from codeine. Orally, it is 40% bioavailable and reaches peak effect in about 1 hour. Half-life is 2.5 to 3.5 hours.
The usual dose of codeine alone, a schedule II drug, is 30 to 60 mg orally every 4 hours. Fixed combinations of codeine (15, 30, or 60 mg) with acetaminophen (300 mg) are available and are schedule III. Because some patients cannot activate codeine and there are drug interactions that weaken its efficacy, codeine preparations are not front-line Step 2 agents.
Other Step 2 opioid combinations contain hydrocodone or oxycodone, which are roughly equipotent. The authors consider these medications equally
efficacious, but since hydrocodone preparations are schedule III and oxycodone preparations are schedule II, hydrocodone preparations are more commonly prescribed. The choice of prescribing an agent containing acetaminophen or ibuprofen depends on whether inflammation is present, in which case an NSAID would be preferred, or depends on the side effects that each agent can cause.
Table 3-2. Medications Used in Step 2: Treating Moderate Pain.
Because the acetaminophen amount varies in different formulations, the prescriber must ensure that the total acetaminophen content does not exceed a toxic level, taking into account not only the Step 2 agent but also any other acetaminophen preparations a patient may be taking.
Codeine and other opioids share the same set of side effects. For a full discussion of these effects, please refer to the section Common Side Effects in the following section Step 3: Treating Severe Pain.
Tramadol is a weak µ-receptor agonist. It has several orders of weaker magnitude affinity for the µ-receptor than morphine. However, in addition, neuronal serotonin release is enhanced while also inhibiting serotonin and norepinephrine reuptake. This reuptake inhibition is similar mechanistically to the tricyclic antidepressants. Quantitatively, reuptake inhibition is also 1 to 2 orders less in magnitude than tricyclic antidepressants. Tramadol has also been cited to have anti-inflammatory activity that is independent of COX inhibition. It is hypothesized that tramadol's weak effects synergize to make it a clinically useful analgesic. By virtue of its multiple mechanisms of action, tramadol may be useful for mild to moderate nociceptive and neuropathic pain. A clinical trial demonstrated its efficacy in treating diabetic neuropathy.
Tramadol is 75% orally bioavailable. The liver metabolizes it to an active metabolite, O-desmethyltramadol, which has increased activity over its parent compound. Time to peak plasma tramadol concentration is about 2 hours and its half-life is 6 hours.
The typical dose for tramadol is 50 to 100 mg orally every 4 to 6 hours. The maximum recommended dose is 400 mg/d. Although the risk of addiction may be lower than with opioids, there is still some risk. However, in general, the risk of addiction to opioids when used for pain is overstated. Physical dependence does occur and withdrawal may occur with rapid cessation.
Since tramadol is an opioid with its own inherent adjunctive properties, it could be considered a first-line Step 2 agent when mild to moderate pain has a neuropathic component. Because of its multiple, synergistic properties, tramadol may be effective with fewer opioid side effects than a pure µ-receptor agonist titrated to equal efficacy. For example, tramadol may cause less constipation than an equally analgesic dose of an oxycodone-containing product.
CNS effects, such as dizziness and somnolence, and gastrointestinal effects, such as constipation and nausea, are the most commonly cited side effects. Maximal dosing is limited due to concerns for lowering seizure thresholds.
Harati Y et al. Double-blind randomized trial of tramadol for the treatment of the pain of diabetic neuropathy. Neurology. 1998;50:1842.
Step 3: Treating Severe Pain
Clinically used opioid analgesics are agonists at µ-receptors. They include morphine, oxycodone, hydromorphone, fentanyl, and methadone. In the brain, these µ-receptors are located in areas such as the periaqueductal gray, known to be involved in mediating pain. In the spinal cord, they are located in the dorsal horn where small fiber pain afferents synapse.
Mu receptors are transmembrane proteins that are coupled to G-proteins. Presynaptically, opioid binding can lead to blockage of calcium channels and thus a decrease in the release of neurotransmitters thus damping pain signaling. Postsynaptically, opioid binding can lead to increased potassium conductance that hyperpolarizes the neuron and makes it less likely to fire to transmit a pain signal.
In addition to µ-receptor agonism, methadone uniquely has two other pharmacodynamic effects. First, it is an N-methyl-D-aspartate (NMDA) receptor antagonist. At the spinal cord level, the NMDA receptor is involved in central facilitation in “wind up” and neuropathic pain. In this state, pain may be refractory to even high-dose opioids. Inhibition of the NMDA receptor can block this “wind up” and increase the efficacy of opioids at the µ receptor. Second, methadone can block presynaptic serotonin reuptake.
Older equianalgesic tables indicated methadone was roughly equianalgesic with morphine. However, empirically, when patients who are taking high-dose opioids are rotated to methadone, it is observed that a much lower dose of methadone is effective than the dose calculated
from traditional equianalgesic tables. This increased efficacy is attributed to the synergism of methadone's multiple pharmacodynamic properties.
The concept that the dose of a medication must be increased over time to maintain the same pharmacodynamic effect is called tolerance. Opioid tolerance at the cellular and molecular level overall remains an enigma. To explain tolerance, researchers have invoked processes such as receptor downregulation, receptor desensitization, NMDA receptor upregulation, as well as others. Tolerance to opioid effects varies. For example, tolerance never seems to develop to the constipating effects of opioids whereas tolerance rapidly develops to respiratory depression. Tolerance to analgesia develops in animals, but empirically in humans, chronic stable pain is often well treated with stable doses of opioids.
Among opioids, tolerance also varies. If a patient who is tolerant to one opioid is switched to another opioid, it is observed that there is incomplete cross-tolerance (Table 3-3). The second opioid is more effective than would be expected from equianalgesic conversion calculations. There is some tolerance due to the effects of the first opioid but it is incomplete. The practical effect of incomplete cross-tolerance is that when rotating to a new opioid the calculated equianalgesic dose must be reduced by 25 to 50% to have a similar pharmacodynamic effect. When patients remain in pain or are having intolerable opioid side effects, this phenomenon can be used in a process called opioid rotation. By switching to an alternative opioid, analgesia may be enhanced and side effects may be reduced due to incomplete cross-tolerance.
It is important to divide opioids into two types—hydrophilic and lipophilic-when discussing the pharmacokinetics of opioids. The major differences between hydrophilic and lipophilic opioids are their pharmacokinetic and metabolite profiles.
Morphine, codeine, hydrocodone, oxycodone, and hydromorphone are examples of clinically useful hydrophilic opioids. The hydrophilic opioids share a similar pharmacokinetic profile.
Table 3-3. Opioid Rotation and Incomplete Cross-Tolerance.
Hydrophilic opioid oral bioavailability ranges from 35 to 70%. There is an extensive hepatic first-pass effect. Because of this effect, conversion from oral to parenteral dosing requires reduction by approximately a factor of three. For example, 30 mg of oral morphine would be converted to 10 mg of IV morphine.
Morphine has an active metabolite, morphine-6-glucouronide (M6G) that is even more potent than morphine itself. M6G must be cleared renally. When creatinine clearance is compromised, M6G may accumulate and cause opioid neurotoxicity (eg, myoclonus, delirium, seizure). It is believed the other hydrophilic opioids may also have renally cleared metabolites that can cause toxicity with accumulation.
For the short-acting hydrophilic opioids, the time to maximal serum concentration (Cmax) depends on the dosing route: orally, Cmax is 60 minutes; SBcutaneously, Cmax is about 30 minutes; intravenously, Cmax is about 6 minutes. The half-life of the hydrophilic opioids is approximately 4 hours. Steady-state levels are reached after 4 to 5 half-lives; thus, steady-state levels will be reached in 16 to 20 hours.
These short-acting opioids may be subject to the bolus effect. Patients may experience side effects when serum levels are maximal yet later experience recurrent pain as trough levels are approached before the next scheduled dose. Either continuous infusion or long-acting opioids are needed to avoid this bolus effect.
Examples of long-acting forms of the hydrophilic opioids include morphine (eg, MS Contin, Kadian), oxy-codone (eg, Oxycontin), and hydromorphone exist. Depending on the formulation, the half-life is 12 to 24 hours. At steady state, peak and trough effects are blunted thus avoiding the bolus effect. Long-acting formulations also improve patient compliance by reducing dosing frequency, reducing pill burden, and reducing sleep interruptions from pain or dosing.
Examples of the major lipophilic opioids are fentanyl and methadone. Because of their affinity for lipids, fentanyl and methadone have high bioavailability, and they rapidly cross the blood-brain barrier. The liver metabolizes fentanyl and methadone, but there are no known active or toxic metabolites (unlike the hydrophilic opioids).
The noninvasive forms of fentanyl include a trans-dermal patch and a transmucosal lozenge. The trans-dermal delivery system establishes equilibrium with the subcutaneous tissue and systemically delivers a defined amount per hour. Its bioavailability approaches 100%. Available patches deliver 12.5, 25, 50, 75, and 100 mcg/h. After placement of a patch, 12 to 16 hours are needed to reach clinically significant levels. During this time, other short-acting opioids are needed to maintain analgesia. The patch is typically replaced every
3 days, although some patients may require a change every 2 days.
The lozenge contains fentanyl in a candy matrix that is applied by twirling against the buccal mucosa until it is consumed. It is available in 200, 400, 600, 800, 1200, and 1600 mcg doses. It must be uniquely titrated for each patient's pain. The effective dose has no correlation with the oral morphine equivalent dose a patient is currently receiving. The typical starting dose for fentanyl lozenge is 200 mcg. If pain is not relieved in 15 minutes, a second 200-mcg lozenge is consumed. If this controls the pain, the appropriate dose is 400 mcg. If this does not control the pain, titration is resumed at the next episode of pain starting with 400 mcg and repeating the above procedure. Approximately half of the bioavailable amount is absorbed transmucosally and has kinetics of action similar to the intravenous route. Onset of action is within 5 to 10 minutes. The other bioavailable half is swallowed and has the kinetics of the oral route. Overall, peak serum levels are achieved in 20 to 40 minutes. In addition to its rapid onset of action, it also has a relatively rapid offset of 1 to 3 hours. Given this kinetic profile, it is advantageous for short-lived breakthrough pain. Other short-acting opioids with longer half-lives may still have significant serum levels even after the short-lived breakthrough pain has diminished. In this situation, a patient has a relative opioid excess and attendant opioid side effects, eg, lethargy.
Methadone has a long and variable half-life that can range from 8 to 72 hours. Thus, it may take from 1 to 15 days for steady state to be reached. Careful individual titration is necessary to avoid accumulation over time. Methadone is typically taken every 8 hours, but some patients may only need to take it once or twice daily. Methadone has some drug interactions. Carba-mazepine, phenobarbital, phenytoin, and rifampin can increase methadone metabolism; whereas, amitriptyline and cimetidine can decrease its metabolism. Methadone may also increase zidovudine levels.
Table 3-4 outlines prescribing guidelines for opioids. When rapidly titrating opioids to treat uncontrolled pain, it is best to use short-acting agents on an as-needed basis until the pain is controlled and daily requirements have been established. This method also works best when clinically significant renal insufficiency is present. Renally cleared active opioid metabolites may accumulate, but patients will integrate this fact in their dosing as needed. The time to Cmax of the route the opioid is being administered determines how often titration can be performed. For example, the time to Cmax for an oral hydrophilic opioid is about 1 hour. Therefore, if a patient still has significant pain 1 hour after an oral dose of a hydrophilic opioid, there should be no expectation of better analgesia by waiting because serum levels will only fall with the passage of time. Thus, it is rational and safe to give an appropriate amount of a hydrophilic opioid orally on the hour until pain is adequately controlled.
Table 3-4. Opioid Titration Guidelines.
If the route is intravenous, Cmax is approximately 6 to 10 minutes. Thus, doses could be repeated every 10 minutes until pain is tolerable. If pain is still severe at the Cmax, the short-acting opioid could be safely doubled without fear of respiratory depression. Similarly, if pain is still severe at steady state, the total daily opioid dose can also be safely doubled. For example, if a patient takes 30 mg of immediate-release morphine orally, yet remains in severe pain 1 hour later at Cmax, doubling the oral morphine dose to 60 mg would be clinically safe as titration to tolerable pain control is pursued. Once patients are no longer opioid naïve, it is the relative change in dose that matters, not the absolute values.
Once 24-hour opioid requirements are established, the dosing is converted to a long-acting regimen. This conversion enhances compliance by reducing the burden of taking pills frequently, by eliminating the bolus effect of short-acting agents, and by allowing for sleep that is uninterrupted by pain or dosing. If at steady state, pain remains at a mild to moderate level, the 24-hour dose can be increased 25 to 50%. If pain remains at a moderate to severe level, the 24-hour dose can be increased 50 to 100%. A 100% increase represents a doubling of the 24hour dose and is safe in the context of continued pain.
When titrated as above, long-acting hydrophilic opioids are still safe and most cost-effective for treating patients with clinically significant renal insufficiency. When renal function is changing, there may be a role for opioids without active or toxic metabolites that must be renally cleared, such as methadone or fentanyl. However, hydrophilic opioids may still be useful if dose or frequency of dosing is decreased. Sometimes, returning to “as-needed” dosing is effective for analgesia and avoiding
opioid side effects as renal function is declining and patients are approaching death.
When the oral route cannot be used for long-acting agents, there are still several options that can be used before resorting to continuous parenteral administration. First, the fentanyl transdermal patch can be used. Next, some long-acting hydrophilic opioid preparations are capsules containing small time-release granules. The capsules can be opened and the granules can be put down an enteral feeding tube (eg, Kadian). Finally, as mentioned previously, some oral long-acting formulations can be used rectally (eg, MS Contin).
For pain that breaks through a basal opioid regimen, clinicians can give 5 to 15% of the 24-hour requirement as a breakthrough dose. Again, since hydrophilic opioids have a Cmax of about 1 hour, this breakthrough dosing is safely given up to every hour if pain persists. If pain is persistent, requiring multiple breakthrough doses per day, the total opioid use (basal plus breakthrough) per day can be totaled and divided into dosing of a long-acting agent. For example, if a patient is taking 60 mg of oral morphine sulfate every 12 hours and needs 12 extra 10-mg doses of immediate-release morphine sulfate in a day, the basal regimen can be adjusted to long-acting morphine sulfate 120 mg orally every 12 hours.
This kind of persistent breakthrough pain must be distinguished from incident breakthrough pain that is incited by a particular event. For example, for a patient with an acute vertebral compression fracture, there may be little pain when lying at rest but there may be severe pain on weight bearing to go to the bathroom. Although multiple doses of breakthrough medication may be required per day, these doses would not be rolled into the basal opioid regimen. Doing so would give a relative excess of opioid when the patient is pain free when lying and not enough opioid for the acute pain exacerbation when standing.
When converting from one opioid to another, equianalgesic conversion tables guide dosing (Table 3-5). It is important to note that these conversions are only guidelines and clinical judgment is needed to individualize dosing for patients. Moreover, because of the phenomenon of incomplete cross-tolerance (see Table 3-3), calculated equianalgesic doses must be reduced by 25 to 50% for equal effect. Sometimes, when pain is not well controlled and opioids are being rotated, clinicians purposely do not take into account incomplete cross-tolerance in order to have a net increase in opioid effect. It is comforting to know that for opioid tolerant patients, even two-fold differences in dosing of an opioid will not cause life-threatening complications.
Conversion from other opioids to methadone requires special consideration (see Table 3-5). Methadone, as discussed previously, has multiple pharmacodynamic actions, making it more potent than predicted from traditional equianalgesic tables. In cases of neuropathic or opioid-resistant pain, methadone needs to be uniquely titrated based on the total oral morphine equivalent dose a patient is currently receiving. The higher the oral morphine equivalent dose, the more potent methadone may be and the conversion must be adjusted accordingly. For example, at an oral morphine equivalent dose of
300 mg/d, a conversion factor of 5 would be used, yielding a methadone dose of 60 mg/d. At an oral morphine equivalent dose of 1000 mg/d, a conversion factor of 15 would be used, yielding a similar methadone dose of ~ 67 mg/d.
Table 3-5. Equianalgesic Dosing Guidelines for Chronic Pain.
Given methadone's long and variable half-life, it is not typically used to titrate acutely for severe pain. Therefore, patients are often taking high doses of other opioids before being converted to methadone. This conversion requires careful attention. If methadone potently relieves pain, a patient may be left with a relative excess of the original opioid. In theory, this excess could suppress respirations since pain is no longer present as an antidote.
There are many conversion protocols in use. Bruera et al have published a conservative conversion to methadone over 3 days. First, the targeted methadone dose per day is calculated as above. Then,
During this conversion, the authors recommend using the original opioid for breakthrough pain at the original breakthrough dose. If patients have significant pain relief early in the conversion, the original opioid can continue to be tapered, but the methadone dose can be maintained without further up-titration. Again, due to methadone's long and variable half-life, patients must be carefully monitored for lethargy as an early sign of accumulation. Typically, this occurs within 3 to 5 days of initiating long-term therapy, but may occur later. If it is noted, the methadone should be held and then restarted at a lower or less frequent dose. Due to the complexity of methadone dosing, expertconsultation may be indicated.
Methadone may also be started as the initial opioid, particularly when neuropathic pain is present. In opioid naïve patients with moderate pain, 5 mg of oral methadone 2 or 3 times daily is a reasonable starting dose. Doses can be titrated up every 3 to 5 days as indicated, again monitoring for signs of accumulation. The authors use a short-acting hydrophilic opioid for breakthrough pain during titration.
All the opioids share similar side effects (Table 3-6); the common include nausea, constipation, and altered cognition (eg, sedation, mental clouding). Although respiratory depression is much feared, it is not common when opioids are dosed appropriately. This statement is especially true when opioids are titrated in the presence of pain, which is a powerful antagonist to respiratory depression. The other less common side effects include dysphoria, delirium, myoclonus, seizures, pruritus and urticaria, and urinary retention.
Table 3-6. Opioid Side Effects.
Side effects typically occur at the time of opioid initiation and at times of dose increments. At any given opioid dose, tolerance to the side effect may develop but is variable. Tolerance to respiratory depression happens quickly. Tolerance to nausea and cognitive changes typically occurs within a few days to 1 week. Unfortunately, tolerance to constipation never develops.
To avoid having to discontinue opioid therapy, side effects can be treated. Nausea is usually well controlled with an antidopaminergic antiemetic. Since gastric motility is slowed by opioids, a particularly effective agent is metoclopramide, which has promotility effects in addition to antidopaminergic effects. Sedation and mental clouding may respond to stimulants such as methylphenidate or modafinil. Since tolerance to constipation never develops, a bowel regimen should be instituted at the same time opioids are initially prescribed. Typically, a stimulant laxative, such as senna, is combined with a stool softener, such as docusate, to treat opioid-induced constipation. New agents to treat opioid-induced constipation are in clinical trials and may be available in the near future. These agents are peripherally acting opioid antagonists that do not cross the blood-brain barrier. Therefore, they do not negate the opioids' central analgesia but can reverse the peripheral constipating effects.
Dysphoria and delirium may be managed by opioid rotation, reducing the opioid dose by adding an adjunctive agent, or adding a psychoactive agent to treat symptoms (eg, an antipsychotic to treat delirium). Opioids can directly cause mast cell degranulation independently of IgE, resulting in pruritus and urticaria. Opioid
rotation and antihistamines can be useful. Myoclonus and seizures indicate neurotoxicity and opioids should be rotated, potentially to opioids lacking active or toxic metabolites that need to be renally cleared. Urinary retention can be treated with a catheter, and opioid rotation can be attempted.
Morley JS et al. Low-dose methadone has an analgesic effect in neuropathic pain: a double-blind randomized controlled crossover trial. Palliat Med.2003;17:576.
Thwaites D et al. Hydromorphone neuroexcitation. J Palliat Med. 2004;7:545.
Waldhoer M et al. Opioid receptors. Annu Rev Biochem. 2004;73:953.
Table 3-7 lists the typical dosing for adjunctive medications, including antidepressants, anticonvulsants, sodium channel blockers, NMDA receptor antagonists, α2-agonists, and corticosteroids.
Tricyclic antidepressants were the first antidepressants found to be effective for neuropathic pain. The analgesic effect has been separated from the antidepressant effect. Consistent with this observation is that doses effective for analgesia are typically lower than doses required for depression. Amitriptyline is the best studied of the tri-cyclic antidepressants. It blocks both serotonin and nore-pinephrine reuptake. There is also evidence that it can act as an NMDA receptor antagonist.
The tricyclic antidepressants vary in their anticholinergic effects. Amitriptyline is the most potent anticholinergic, while nortriptyline and desipramine have the least effect.
Serotonin-norepinephrine reuptake inhibitors (SNRIs) are also effective for neuropathic pain. The US Food and Drug Administration (FDA) has approved duloxetine for diabetic neuropathy. Studies on venlafaxine indicate efficacy for neuropathic pain as well.
Selective serotonin reuptake inhibitors (SSRIs) that block presynaptic reuptake have had varied success in neuropathic pain. In randomized controlled trials, fluoxetine was no better than placebo, but citalopram and paroxetine showed some efficacy.
In general, acute pharmacokinetics are not as important for this class since acute analgesia is not expected. Doses are titrated up over time as tolerated and to efficacy. Typically, analgesia occurs within 1 week once an effective dose has been reached, but it may take weeks to titrate to this level. Tricyclic antidepressants have long half-lives and can be taken once per day, often at bedtime. Duloxetine also has a long half-life and can be taken once per day. Venlafaxine is typically taken 2 to 3 times a day, but once an effective dose is found, there is a long-acting form that can be dosed daily.
Overall, tricyclic antidepressants and the newer SNRIs appear more effective than SSRIs. There have been no head-to-head comparisons of tricyclic antidepressants with SNRIs. Tricyclic antidepressants are more cost effective, but because they have more side effects, they may be less well tolerated than SNRIs.
Tricyclic antidepressants are prescribed using the adage, “Start low and go slow.” Amitriptyline, nortripty-line, and desipramine are started at 10 to 25 mg orally at bedtime and titrated up to about 100 mg/d. Titration occurs every few days to 1 week as tolerated until efficacy or side effects limit dosing. Analgesia ensues within 1 week of attaining an effective dose.
The dose of duloxetine approved by the FDA for treating diabetic neuropathy is 60 mg orally once a day. If side effects are experienced, the dose can be reduced and titrated up as tolerated. Venlafaxine can be started at 75 mg/d orally in two or three divided doses. The dose can be titrated up by 75 mg about every 4 days until efficacy or a side effect is reached. Typically, effective analgesic doses range from 75 to 225 mg/d. There is an extended-release once-a-day formulation that can reduce the burden of pills.
The anticholinergic properties of tricyclic antidepressants induce dry mouth, constipation, urinary retention, and sedation. Many of the side effects wane over time. The sedating properties can be advantageous when insomnia is present. Tricyclic antidepressants should not be used in patients with narrow-angle glaucoma. At the lower doses that are effective for analgesia, levels need not be monitored and cardiovascular side effects are uncommon. However, especially for geriatric patients and patients with known cardiac problems, orthostatic hypotension and cardiac conduction abnormalities should be monitored. Overdose of tricyclic antidepressants can be lethal, so prescribers must remain vigilant for signs of suicidal ideation.
The SNRIs and the SSRIs are well tolerated overall. Headaches, gastrointestinal upset, and sexual dysfunction are the most common side effects reported.
Anticonvulsants are effective neuropathic pain medications most likely by virtue of their membrane stabilizing
properties. Although not definitively known, gabapentin and pregabalin (recently FDA approved) probably act by binding to a calcium channel subunit that appears to be upregulated in nerves in certain neuropathic pain states. Carbamazepine, oxcarbazepine, and lamotrigine appear to inhibit sodium channels. Valproic acid, in addition to inhibiting sodium channels, may also enhance levels of the inhibitory neurotransmitter GABA. Topiramate, in addition to inhibiting sodium channels, may also enhance GABA activity and inhibit an NMDA receptor.
Table 3-7. Prescribing Guidelines for Adjunctive Medications.
Acute analgesic effects are not expected from this class of drugs. Gabapentin has variable absorption that decreases as the dose increases. For example, 300 mg of oral gabapentin three times daily is about 60% bioavailable; whereas, 1200 mg orally three times daily is only about 33% bioavailable. Pregabalin has an oral bioavailability of about 90%. Of note, both gabapentin and pregabalin are primarily excreted unchanged renally. Therefore, in renal insufficiency, their dosage must be modified. Carbamazepine has many potential drug interactions that must be monitored. The newer anticonvulsants, especially gabapentin and pregabalin, tend to have fewer interactions than the older anticonvulsants.
Gabapentin is commonly considered the first-line anticonvulsant. It is well tolerated, does not require that serum levels be monitored, and has few drug interactions. It is usually started at low dose and titrated to effect. The minimal effective dose is 900 mg/d, but doses have been titrated up to 4500 mg/d. A common mistake is discontinuing gabapentin for lack of efficacy before titrating up to clinically effective levels. The disadvantages of gabapentin include its variable absorption and time needed to titrate to effect.
Pregabalin shares gabapentin's advantages, but also is more potent and has predictable bioavailability. These characteristics make it easier and faster to titrate to effect. In clinical studies, pregabalin was titrated to effect in about 1 week, whereas gabapentin titration required about 4 weeks. The typical dosage for pregabalin starts at 25 to 50 mg orally three times daily and can be titrated up to 200 mg orally three times daily.
Table 3-7 lists the typical dosages for the other anti-convulsants.
Headache, dizziness, ataxia, and nausea are common side effects seen among anticonvulsants. Somnolence and dizziness are the most common side effects associated with gabapentin and pregabalin therapy. These effects can usually be controlled by titrating up slowly and by habituation over time.
In addition, carbamazepine can also cause the syndrome of inappropriate antidiuretic hormone (SIADH), hepatitis, and bone marrow suppression, so appropriate laboratory tests should be performed. Oxcarbazepine, a metabolite of carbamazepine, is better tolerated overall than carbamazepine but can still cause hyponatremia. In addition to the above common side effects, valproic acid can induce thrombocytopenia. Topiramate can block carbonic anhydrase lowering serum bicarbonate levels, which should be monitored.
Sodium Channel Blockers
Lidocaine, a nonselective sodium channel blocker, is effective in neuropathic pain syndromes such as diabetic neuropathy and postherpetic neuralgia, and there are case reports of effectiveness in cancer pain. Researchers have identified sodium channels on damaged nerves and dorsal root ganglion cells that fire spontaneously after damage.
Systemic lidocaine can suppress this ectopic, spontaneous firing at a concentration that does not affect normal nerve and cardiac conduction. This suppressive ability may at least partially explain nonselective sodium channel blockers' utility in neuropathic pain. An oral congener of lidocaine, mexiletine, is presumed to operate similarly. Systemic lidocaine has been used as a predictor of response to oral mexiletine but the usefulness of this practice has not been well substantiated.
Lidocaine can be given parenterally; a topical 5% patch is also available. Lidocaine 5% patches do not have significant systemic absorption in usual clinical applications. It is metabolized by the liver and has a half-life of ~100 minutes.
Mexiletine has oral bioavailability approaching 90%. It is metabolized by the liver and has peak serum levels in 2 to 3 hours. Half-life is about 10 to 14 hours.
Parenteral lidocaine has been used to treat diabetic neuropathy and postherpetic neuralgia in small trials. Based on preliminary observations, parenteral lidocaine may quickly control neuropathic or opioid-refractory cancer pain and provide a window of opportunity for other agents to be titrated to effective levels. The authors challenge opioid-refractory patients with a lidocaine dose of 1 to 2 mg/kg given intravenously over 20 minutes. As soon as 30 minutes after administration, pain relief is measured. If pain is improved, a continuous lidocaine infusion is started at 1 mg/kg/h. Steady-state levels are checked 8 to 9 hours later; the infusion is adjusted based on efficacy and side effects to a level between 2 mg/L and 5 mg/L. The authors do not use cardiac monitoring in a hospice population. Moreover, there is a good safety record in published small trials. However, larger trials are needed to substantiate the efficacy and safety of parenteral lidocaine. Currently, parenteral lidocaine is best used in consultation with a specialist.
Lidocaine patches are applied over the painful area and left in place for 12 hours. Studies have used up to 3 patches left in place for 24 hours with good efficacy and no increase in side effects.
Mexiletine is usually started at 150 mg/d orally for 3 days, titrated to 300 mg/d orally for another 3 days, and then titrated to a dose of 10 mg/kg.
Systemic lidocaine at therapeutic levels (2 to 5 mg/L) is well tolerated with the most frequent complaints being somnolence and dizziness. However, it does have a relatively narrow therapeutic window. Therefore, serum levels should be monitored. Above 8 mg/L, myoclonus may occur, and at higher levels the risk of seizure (>10 mg/L) and cardiovascular collapse (>25 mg/L) increases. Topical lidocaine patches are well tolerated. In typical usage, clinically significant serum levels are not a concern. Mexiletine can cause gastrointestinal upset in up to 40% of patients, limiting its clinical usefulness.
NMDA Receptor Antagonists
As discussed previously, the NMDA receptor is involved in the spinal process of “wind up” and is believed to be involved in the generation of neuropathic pain and opioid tolerance. Inhibition of the NMDA receptor can have potent analgesic effects. Clinically available NMDA receptor antagonists that have been reasonably studied include methadone, dextromethorphan, and the dissociative anesthetic, ketamine. They all have approximately the same affinity for the NMDA receptor. For best efficacy, it is likely that NMDA receptor antagonists should be used in conjunction with opioids.
As discussed earlier, methadone has a long and variable half-life, requiring slow titration. Dextromethorphan is available orally in short- and long-acting forms. In the short-acting form, it has an onset of action of 15 to 30 minutes. Ketamine is available as a parenteral solution that has also been used orally. There is a significant first-pass effect when taken orally. The liver metabolizes ketamine to norketamine. Norketamine is equipotent with ketamine as an analgesic but only one-third as potent as an anesthetic. Orally, ketamine has an onset of action of 30 minutes.
In chronic severe pain, when there is time for titration, the authors recommend methadone. It provides both µ receptor agonism and NMDA receptor antagonism. Dosing based on previous opioid levels is shown in Table 3-5.
In crescendo pain where there is no time to titrate medications slowly, ketamine has a kinetic advantage. In palliative care populations, ketamine has been used par-enterally at low dose with good effect. Dosing is usually started at 0.1 to 0.2 mg/kg/h parenterally and titrated to effect. At low dose, the risk of psychotomimetic effects is reduced. If they appear, low-dose benzodiazepines are usually able to control the negative effects.
Dextromethorphan has had mixed success in the literature, and its dose has been limited by side effects. Typical dosing in the literature ranges from 20 mg orally three times daily to 90 mg orally three times per day.
Methadone shares the opioid side effects already discussed. Dextromethorphan and ketamine can cause dysphoria, hallucinations, somnolence, and dizziness.
Clonidine and dexmedetomidine are α2-agonists that are effective for both nociceptive and neuropathic pain. Tizanidine is another α2-agonist used in spasticity but has not been well studied otherwise as an analgesic. They have CNS and peripheral nervous system effects. In the spinal cord, α2-agonists have effects similar to the opioids but act through a different receptor, thus potentially providing additive effects. Specifically, they alter calcium and potassium conductance. Presynaptically, they decrease neurotransmitter release; and postsynaptically, they hyperpolarize the neuron making it less likely to fire. α2-agonists also have a sympatholytic effect that may be mediated both spinally and at postganglionic nerve terminals with the net effect of decreasing catecholamine release. This decreased sympathetic outflow may help in certain forms of sympathetically driven neuropathic pain, such as complex regional pain syndromes.
Clonidine is available as an oral agent and a transdermal patch. Orally, it is 75 to 100% bioavailable, and the patch is 60% bioavailable. Dexmedetomidine is available as a parenteral solution, but the buccal and oral routes of administration have also been studied.
Dexmedetomidine needs further study to be useful routinely. Clonidine is typically started at 0.1 mg/d orally and titrated to efficacy or intolerable side effects. To limit systemic effects, α2-agonists are often used intraspinally, but this technique requires specialist assistance and is beyond the scope of this chapter.
Clonidine and dexmedetomidine share hypotension and bradycardia as potential side effects. Clonidine tends to cause more dry mouth and somnolence.
Corticosteroids are potent anti-inflammatory drugs. They include hydrocortisone, prednisone, methylprednisolone, and dexamethasone. They bind a cytosolic receptor that translocates to the nucleus and alters transcriptional regulation. One subsequent effect is to suppress the action of nuclear factor κ B, which induces many inflammatory cytokines. Corticosteroids are often used at supraphysiologic doses, above what should be needed for receptor-mediated effects. Researchers hypothesize that there may be a direct effect of corticosteroids dissolved in membranes.
Corticosteroids reduce pain in several ways. As stated previously, inflammation sensitizes some nociceptors. Corticosteroids, by reducing inflammation, can reduce pain. Second, neural compression causes pain. By decreasing inflammation and edema, such as peritumoral edema, corticosteroids relieve nerve compression and pain. Finally, studies have shown that corticosteroids can decrease the spontaneous firing of sodium channels in neuromas. This suppression may be an example of a direct membrane effect.
The corticosteroids differ in their mineralocorticoid effect, which affects salt retention. Dexamethasone has the least mineralocorticoid effect, and therefore is often used when patients are hypoalbuminemic with fluid third spacing.
Corticosteroids have high oral bioavailability and can also be administered parenterally. Their plasma half-lives are short, but with the exception of hydrocortisone, their duration of action is long, allowing for once a day dosing.
Especially in patients with advanced medical illness potentially facing the end of life, dexamethasone is often the first-line corticosteroid due to its minimal miner-alocorticoid effect and its long duration of action that supports once a day dosing. As a pain adjunct, doses range from 4 mg/d to 20 mg/d. Dexamethasone can be given orally, rectally, intravenously, and subcutaneously. Typically, doses are started at high levels to determine whether there is an effect. If there is no clinical benefit in 1 to 2 days, the corticosteroid can simply be discontinued without fear of adrenal suppression. If there is a clinical benefit, the dose can be tapered down to the minimally effective dose. In this population, long-term sequelae are typically not relevant.
In other inflammatory pain syndromes, corticosteroids may have a role, but due to long-term sequelae, this role is usually time-limited.
Patients who have been taking the equivalent of 20 mg/d of prednisone for more than 3 weeks should be assumed to have suppression of the hypothalamic-pituitary-adrenal axis. Hyperglycemia and corticosteroid-induced psychosis can occur early after starting corticosteroids.
Longer-term sequelae include osteoporosis, Cushing syndrome, cataracts, peptic ulcer, and myopathy.
Overall Prescribing Recommendations for Adjunctive Medications & Their Combinations
The literature provides little guidance on the optimal use of adjunctive medications and their combinations (see Table 3-7). However, a recent randomized, double-blind, active placebo-controlled, crossover trial of neuropathic pain demonstrated that the combination of morphine and gabapentin provided better analgesia with fewer side effects at lower doses than either agent alone. This study indicated opioids were effective treatment for neuropathic pain but highlighted the fact that combination therapy can be synergistic for analgesia and a reduction in side effects. Further studies are needed to quantitatively assess other adjunctive combinations.
For moderate to severe neuropathic pain, the authors recommend methadone for consideration as a first-line therapy. Its multiple pharmacodynamic properties make it an effective analgesic. It is an opioid with its own adjunctive properties. In addition to efficacy, major advantages of methadone are decreased pill burden, long-lasting effects, and cost effectiveness. Its major disadvantages include its slow titration, complicated opioid conversion calculation, and potential for accumulation due to its long and variable half-life. Prescribers should consider expert consultation until they are knowledgeable in the use of methadone.
Gabapentin and pregabalin are also recommended for consideration as first-line adjunctive medications. The above-cited study provides evidence of its useful combination with opioids. Their major advantages include a good side effect profile, little drug interaction, and no need to monitor serum levels. The major disadvantage of gabapentin is its variable absorption that worsens with increased dosage.
Although it has been poorly studied in the literature, the authors recommend further combinatorial therapy for resistant pain syndromes. Anecdotally, in severe cancer pain syndromes, the authors have effectively combined µ-receptor agonists, NMDA receptor antagonists, neuron-specific calcium-channel blockers, sodium channel blockers, tricyclic antidepressants, and anti-inflammatory drugs for optimal pain control. Clearly, there is a need for more evidence to guide clinical practice. However, in its absence, the principal of combining analgesics that may work through different pathways to produce synergism is rational.
Arnold LM et al. A double-blind, multicenter trial comparing duloxetine with placebo in the treatment of fibromyalgia patients with or without major depressive disorder. Arthritis Rheum. 2004;50:2974.
Devor M et al. Corticosteroids suppress ectopic neural discharge originating in experimental neuromas. Pain. 1985;22:127.
Galer BS et al. The lidocaine patch 5% effectively treats all neuropathic pain qualities: results of a randomized, double-blind, vehicle-controlled, 3-week efficacy study with use of the neuropathic pain scale. Clin J Pain. 2002;18:297.
Gilron I et al. Morphine, gabapentin, or their combination for neuropathic pain. N Engl J Med. 2005;352:1324.
Rosenstock J et al. Pregabalin for the treatment of painful diabetic peripheral neuropathy: a double-blind, placebo-controlled trial. Pain. 2004;110:628.
Schulte H et al. The synergistic effect of combined treatment with systemic ketamine and morphine on experimentally induced windup-like pain in humans. Anesth Analg. 2004;98:1574.
Thomas J et al. Intravenous lidocaine relieves severe pain: results of an inpatient hospice chart review. J Palliat Med. 2004;7:660.