Elderly Patients

Elderly patients present several pain management problems. First, relatively little attention has been paid to the topic of geriatric pain control in medical or nursing texts (Ferrell, 1991). This is ironic because elderly people often suffer acute and chronic painful diseases, have multiple diseases, and take many medications (From the NIH, 1979). They may have more than one source of pain and an increased risk for drug-drug as well as drug-disease interactions (Kane, Ouslander, and Abrass, 1989). It has been estimated from population studies that the prevalence of pain is two-fold higher in those over age 60 (250 per thousand) compared with those under 60 (125 per thousand) (Crook, Rideout, and Browne, 1984). Among institutionalized elderly, the prevalence may be over 70 percent (Ferrell, Ferrell, and Osterweil, 1990). Indeed, more than 80 percent of elderly people suffer various forms of arthritis, and most will have acute pain at some time (Davis, 1988). Many elective or emergent operations are performed in the elderly to correct orthopedic problems (e.g., fractures, degenerative joint disease). Acute and/or postoperative severe pain related to cancer and its treatment are also more common in the elderly (Foley, 1985). Other acutely painful conditions that affect the elderly disproportionately include herpes zoster, temporal arteritis, polymyalgia rheumatica, and atherosclerotic peripheral vascular disease (From the NIH, 1979).

Second, pain assessment may present unique problems in elderly patients. They often report pain very differently from younger patients due to physiologic as well as psychological and cultural changes associated with aging (Fordyce, 1978). Institutionalized elderly are often stoic about pain (Foley, 1985). Age-associated changes in acute pain perception have long been of interest. Elderly patients often demonstrate altered presentations of common illnesses including "silent" myocardial infarctions and "painless" intra- abdominal emergencies (Bayer, Chada, Farag, and Pathy, 1986 ; Bender, 1989). Whether these clinical observations are the result of age-associated changes in pain perception remains to be explained. The widespread belief among clinicians that aging results in increased pain thresholds may be a myth. A variety of experiments using heat, pressure, or electrical current have not disclosed a trend regarding age-associated changes in either pain threshold or tolerance. It should be noted that the clinical relevance of these studies remains questionable, since experimentally induced pain may not be analogous to clinically experienced pain (Harkins, Kwentus, and Price, 1984).

Cognitive impairment, delirium (common among acutely ill frail elderly), and dementia (occurring in as many as 50% of institutionalized elderly) represent serious barriers to pain assessment for which no solution exists in the literature. Whether behavioral observations (e.g., agitation, restlessness, groaning) are sensitive and specific for pain assessment among the demented elderly remains to be shown. Traditional approaches, including the use of visual analog scales, verbal descriptor scales, and numerical scales, have not been psychometrically established in this population. Moreover, a high prevalence of visual, hearing, and motor impairments in the elderly may impede the universal use of such scales by clinicians. Preliminary reports from ongoing work among the nursing home population suggest that many patients with moderate to severe cognitive impairment are able to report acute pain reliably at the moment or when prompted, although pain recall and integration of pain experience over time may be less reliable. If these early observations prove correct, pain assessment among this population may require frequent monitoring. Monitoring may have major implications for quality assurance, quality of care, and quality of life among this large population of elderly people (Ferrell, and Ferrell, personal communication: Work in progress on the epidemiology of pain among community nursing homes, July 1991).

Third, elderly people, especially the frail and old-old (those over 85) are at particular risk for both under- and overtreatment. Unfortunately, few studies of analgesic dosage requirements are performed in the elderly, and most studies have systematically excluded all potential subjects over 65. Age- related observations are extremely variable among elderly people. Indeed, the variance in measurements of most physiologic and pharmacologic parameters increases with age in cross-sectional studies (Kane, Ouslander, and Abrass, 1989). Age-related changes in pharmacokinetics and pharmacodynamics contribute to a variety of adverse drug effects that have been reported in the elderly.

The widespread belief among clinicians that aging results in increased pain thresholds may be a myth.

Non-opioid analgesic drugs, including NSAIDs and acetaminophen, are effective and appropriate for a variety of pain complaints in the elderly. However, it is recognized that the risk for gastric and renal toxicity from NSAIDs is increased among elderly patients, and unusual drug reactions including cognitive impairment, constipation, and headaches are also more common in the elderly population (Roth, 1989). If gastric ulceration is a particular concern, coadministration of misoprostol or use of "platelet- sparing" NSAIDs should be considered as a way to lessen the risk of gastrointestinal bleeding.

Opioid analgesic drugs are effective for the management of acute pain in most elderly patients. Cheyne-Stokes breathing patterns are not unusual during sleep in the elderly and need not prompt discontinuation of opioid analgesia unless such analgesia clearly is associated with unacceptable degrees of arterial oxygen desaturation (< 85%). PCA has been shown in at least one study to be safe and effective for postoperative pain relief among selected patients (Egbert, Parks, Short, and Burnett, 1990). If PCA is used, careful titration of dosage is necessary to avoid undesirable effects due to drug accumulation or from a decrease in the arousal effect as painful stimuli subside later in the patient's course. Elderly people are more sensitive to the analgesic effects of opioid drugs as they experience a higher peak and longer duration of pain relief (Bellville, Forrest, Miller, and Brown, 1971 ; Kaiko, 1980 ; Kaiko, Wallenstein, Rogers, Grabinski, and Houde, 1982). They are also more sensitive to sedation and respiratory depression probably as a result of altered distribution and excretion of the drugs. This is especially true in opioid-naive patients. Caution is required in the use of longer acting drugs such as methadone for this reason (Ferrell, 1991).

Elderly people, in general, have increased fat-to-lean body mass ratios and reduced glomerular filtration rates (Kane, Ouslander, and Abrass, 1989). Opioids produce cognitive and neuropsychiatric dysfunction through poorly defined mechanisms that in part include the accumulation of biologically active metabolites such as morphine-6-glucuronide or normeperidine (Wood, and Cousins, 1989). Opioid dosage titration should take account of not only analgesic effects but also side effects that extend beyond cognitive impairment. These side effects may include urinary retention that looms as a larger threat in elderly males with prostatic hypertrophy, constipation and intestinal obstruction, respiratory depression, or exacerbation of Parkinson's disease. The management of nausea using phenothiazines or antihistamines is fraught with problems, because elderly people are exquisitely sensitive to anticholinergic side effects including delirium, bladder and bowel dysfunction, and movement disorders (Ferrell, 1991).

Local anesthetic infusions may result in cognitive impairment if significant blood levels are reached. Yet prior to that point, orthostatic hypotension may result from sympathetic blockade and clumsiness may ensue from partial motor or sensory anesthesia. Thus, appropriate precautions should be taken, such as help with ambulation or the use of side rails at night.

Finally, attitudes among health care professionals, the lay public, and patients themselves may impede appropriate care. Many members of all three groups consider acute and chronic pain a part of normal aging (Ferrell, Ferrell, and Osterweil, 1990 ; Ferrell, 1991).

Patients Who Are Known or Suspected Substance Abusers

Management of acute pain in the substance abuser is a difficult but increasingly common clinical problem. Substance abusers experience traumatic injuries (see section on Patients with Shock, Trauma, and Burns) and a variety of health problems more often than the general population. Often during their postoperative care issues arise that prompt the staff to request consultation with specialized "pain teams." For example, the question of possible withdrawal from preexisting opioid use may be raised because sympathetic nervous system stimulation (restlessness, tachycardia, sleeplessness) may be caused by either undertreated pain or opioid abstinence. The related issue of risk of development of substance abuse behaviors in opioid-naive patients given opioid analgesics postoperatively appears to be small based on survey data by Porter and Jick (1980), who found that only four such instances of iatrogenic drug abuse occurred in approximately 12,000 patients screened through the Boston Collaborative Drug Study.

A variety of reports, increasingly frequent in number, have addressed the difficult questions surrounding postoperative opioid use in substance abusers, and several recommendations have recently emerged from reviews of this literature (Portenoy, and Payne, in press). First, every effort should be made to define the mechanism of the pain and to treat the primary problem. Infection, tissue ischemia, or a new surgical diagnosis in the postoperative period may require specific measures such as antibiotic therapy, fasciotomy, or re-operation rather than an increase in the opioid dose. Attention to the primary cause of pain symptoms may reduce greatly the requirement for and negotiations about opioid analgesics.

Second, clinicians should distinguish between the temporal characteristics of the abuse behavior. For example, a distant history of substance abuse might predispose to re-emergence of substance abuse behaviors with the stress of surgery and postoperative pain but may not require treatment approaches different from those appropriate for nonaddicted patients. The implications are different for patients with a recent history of active drug abuse who may require higher than usual starting doses of opioids and who may not have acquired an ability to set limits on their drug use.

Third, one should follow relevant pharmacological principles of opioid use. For example, treatment with an opioid agonist-antagonist should not be started in the patient who enters tolerant to opioid agonists such as methadone. Mixed agonist-antagonists may precipitate withdrawal if given in this setting. Loading doses of opioids will be required in normal patients as well as in substance abusers to reduce the intensity of postoperative pain to acceptable levels. PCA is being used with increasing frequency for many patients, including known substance abusers. This mode of opioid delivery can be utilized safely in substance abusers when appropriate lockout intervals and hourly dosage limits are programmed, and when the device is "tamper- resistant," so that the patient cannot reprogram the pump or remove any drug. If used for an opioid-tolerant patient, PCA doses must be increased to achieve the same analgesic effect as for opioid-naive or nonaddicted patients.

Fourth, just as for other patients, non-opioid therapies should be given concomitantly with or even to replace opioids. Such therapies include NSAIDs, local anesthetic solutions given via catheter into the epidural space or surrounding peripheral nerves, cryoanalgesia, TENS, and nondrug therapies. Appropriate use of non-opioid therapies frequently will reduce the dosage requirement for opioid analgesia.

Fifth, specific drug abuse behavior in the postoperative patient should be recognized and dealt with firmly. Such behavior includes tampering with PCA machines (or other drug delivery devices), hoarding of oral doses of opioid analgesics, or attempting to self-inject the melted contents of capsules or siphoned infusion solutions. At that point, limits of analgesic dosages and expected patient behavior should be made clear to the patient in a frank discussion that also considers the medical, ethical, and legal consequences to the patient and physician if drug abuse behaviors continue. For the most part, security measures already in place (locked closets, and inventory checks each nursing shift, antitamper features on PCA machines) will frustrate such attempts. A toxicological screen may be ordered on an inpatient's urine or blood specimens to confirm or exclude a diagnosis of surreptitious drug use (e.g., cocaine administration).

In the outpatient setting, clear instructions (preferably written out and copied into the patient's record) should be offered regarding doses and frequency of medication and the number of days the prescription is expected to last. The addition of a random urine testing procedure to outpatient medication contracts should be considered in all outpatients with a known history of substance abuse who are given opioid analgesics for pain. Opioid medications should be prescribed only by one physician, and attempts to circumvent this restriction or to falsify prescriptions should not be tolerated. The claim of needing additional medication to make up for lost or stolen controlled substances should be accompanied by documentation that the patient has reported this to the police.

Sixth, caregivers should set limits to avoid excessive negotiation about drug selections or choices. For example, it is not appropriate to depart from an institutional policy for morphine use for postoperative analgesia just because of a substance abuser's request for meperidine, as long as the patient has no history of adverse reaction to morphine. Once every effort has been made by clinicians to adjust a patient's opioid regimen in light of the extent and site of the operation, the patient's prior tolerance to opioids, and clinical response to initial analgesic titration, it is not unreasonable to adhere to this regimen with few, if any, changes. When possible, the treatment plan should include clear criteria (e.g., number of days postoperatively, ability to ambulate) by which opioid doses will be tapered and stopped. Within such a treatment plan, requests for higher dosage by a patient who can walk easily or spends much of the time resting comfortably may appropriately be denied. Consultation should be obtained with appropriate services early in a difficult patient's course. Consultation should help develop a unified multidisciplinary plan that includes, usually, psychiatric, psychological, and substance abuse expertise. Medical input can be obtained beyond the scope of the immediate caretakers to evaluate these and other problematic behaviors. Medical and/or neurologic input can be valuable when assessing neurologic symptoms such as seizures in patients who have recently discontinued alcohol or barbiturate use.

These guidelines will aid the clinician in distinguishing "drug seeking" behaviors from "pain avoidance" behaviors. Often patients who are undertreated for pain and who voice displeasure are labeled as "addicts." This phenomenon has been called "pseudo-addiction" (Weissman and Haddox, 1989). Careful and objective assessment and reassessment of the postoperative patient with an active or previous substance abuse history will minimize the chance of a clinician being "duped" into providing inappropriate opioids but still provide the patient with legitimate pain complaints the opportunity to obtain meaningful and safe pain relief with opioids.

Patients with Concurrent Medical Conditions

Postoperative pain frequently must be treated in patients who have concurrent medical conditions for which they are taking one or more medications. The medical condition per se or the medications taken for it may influence the choice of analgesic and its dosage. These medical conditions are frequently chronic, and it may not be possible to discontinue problematic medications because of surgery. The most common medications or classes of medications that produce clinically significant drug interactions with opioid analgesics include alcohol and any central nervous system depressants, phenytoin, and monoamine oxidase inhibitors. Drugs whose primary site of action lies outside the central nervous system (e.g., antibiotics such as rifampin) also may interact with opioid analgesics.

Coexisting conditions themselves influence the type and doses of opioid analgesics and the relative risks of pain treatment in the postoperative period. For example, patients with chronic pain who have been treated recently with opioids (e.g., patients with cancer, and sickle cell disease) will usually require higher-than-the-recommended starting doses to overcome opioid tolerance. Coagulopathy, neutropenia, and sepsis may contraindicate the use of epidural catheters or other regional anesthetic techniques in which the risks of bleeding or "seeding" of infection are increased.

Drug pharmacokinetics may change following surgery because of changes in drug absorption and distribution caused by alterations in cardiac output, venous capacitance, extravascular fluid shifts ("third spacing"), and changes in protein binding. Fever and sepsis in the postoperative period may affect drug disposition, as do shock, trauma, and burns. In addition, patients may not attain clinically effective plasma concentrations of opioids following intramuscular and subcutaneous injections due to pharmacokinetic alterations.

The major factor to consider in selecting analgesics for patients with concurrent medical conditions is whether the disorder produces either hepatic or renal impairment. Most analgesics are metabolized by the liver or kidney, so that any impairment of function in these organs influences the pharmacokinetics of the analgesic. The net result can be drug accumulation. Therefore, caution is essential when using opioids in patients with altered hepatic or renal function. Morphine is metabolized in the liver, and the parent compound, along with the metabolites, is excreted through the kidney. Acute or chronic hepatic failure (e.g., viral hepatitis or cirrhosis) appears to lower plasma clearance of morphine, prolong the terminal elimination half-life, and increase oral bioavailability (Hasselstrom, Eriksson, Persson, Rane, Svensson, and Sawa, 1990). Even mild renal failure, such as that associated with a decline in glomerular filtration rate with aging, can impede excretion of the metabolites of many opioids, resulting in clinically significant narcosis and respiratory depression (Sear, Hand, Moore, and McQuay, 1989). Physiological alterations during surgery (e.g., changes in regional blood flow to the liver or kidneys, hepatic enzyme activity, enterohepatic circulation, or hormonal responses) may also alter drug metabolism and excretion. Meperidine, pentazocine, and propoxyphene have increased bioavailability, prolonged half- lives, and decreased systemic clearance and thus accumulate in hepatic and renal dysfunction. Doses of these drugs must be decreased appropriately. In contrast, the disposition and elimination of methadone are not significantly altered in patients with chronic liver disease.

Renal excretion is a major route of elimination for pharmacologically active opioid metabolites: norpropoxyphene, normeperidine, morphine-6- glucuronide, and dihydrocodeine. Elimination is decreased in patients with renal failure, and doses must be lowered or given less frequently.

For individual patients, it is difficult to predict the degree of impairment of metabolism or excretion of the most commonly used analgesics from either the clinical condition or laboratory indicators of hepatic or renal function because many factors impinge on clinical response. A lowered initial dose, careful titration of the opioid to desired effect, and ongoing monitoring of clinical response, level of consciousness, and respiratory effort are indicated. Continuous infusions and opioid administration around-the-clock at conventional intervals can result in accumulation of the parent compound or clinically active metabolites. To avoid undertreatment of pain during as- needed dosage schedules, the patient can be assessed at regular intervals, and if stable, a dose of opioid can be offered. PCA does not necessarily protect against accumulation of opioids and respiratory depression (Covington, Gonsalves-Ebrahim, Currie, Shepard, and Pippenger, 1988). In renal failure, especially, a decreased dose and prolonged lockout interval may be required. Non-opioid analgesics are often contraindicated in patients with hepatic or renal dysfunction.

Other diseases that may influence the control of postoperative pain include psychiatric illnesses requiring tricyclic antidepressants and monoamine oxidase inhibitors for treatment; neurologic disorders; pulmonary diseases; and acute and chronic infections.

Patients with respiratory insufficiency and those with chronic obstructive pulmonary disease, cystic fibrosis, and neuromuscular disorders affecting respiratory effort (e.g., muscular dystrophy, myasthenia gravis) are vulnerable to the respiratory depressant effects of opioids. However, when a patient splints his or her respiratory effort because of uncontrolled pain, that also can impair gas exchange. Therefore, careful planning is required to provide effective and safe postoperative analgesia. Unless specific contraindications exist, the use of non-opioid analgesics can be optimized in this group of patients.

However, severe postoperative pain may not be adequately controlled with just these agents. Epidural opioids have a lesser effect on pulmonary function than do systemic opioids (Bromage, Campoersi, and Chestnut, 1980). Other regional anesthetic techniques can also be applied in this setting as a way to lower the dosage of systemic opioids required for satisfactory postoperative analgesia. If epidural analgesia cannot be used, small doses of opioids given frequently, continuous infusion, or PCA may provide smoother control of pain with less impact on respiratory effort at the time of peak effect. Whichever method is used to administer systemic opioids, a low initial dose is recommended; later doses can then be titrated to the desired effect. Appropriate monitoring of respiratory rate and effort and adequacy of gas exchange is necessary. Oximetry may be useful in selected cases.

Neurologic disorders can influence postoperative pain management if they: 1) produce weakness of the respiratory muscles (e.g., amyotrophic lateral sclerosis, poliomyelitis); 2) impair alertness and mental function so that the sedative effects of opioids are exaggerated, and pain cannot be assessed easily; and 3) cause seizures requiring use of chronic anticonvulsant medications that may interact with analgesics. The first two circumstances have been addressed in the discussion above and other sections of this Guideline. Phenytoin, a very commonly used anticonvulsant, increases the biotransformation of meperidine, causing faster elimination and necessitating increased doses of this analgesic (Foley, and Inturrisi, 1987).

Patients with psychiatric illnesses taking anxiolytics or other psychoactive drugs must be carefully evaluated for drug interactions between the psychotropic and pain medications they take. Because both opioids and psychotropic drugs generally have sedative effects, it is not uncommon for these effects to be additive when the drugs are combined. Further, the tricyclic antidepressants, clomipramine and amitriptyline, may increase morphine levels as measured by an increase in bioavailability and the half-life of morphine (Ventafridda, Ripamonti, DeConno, Bianchi, Pazzuconi, and Panerai, 1987). Of particular importance is avoiding meperidine in patients receiving monoamine oxidase inhibitors. Severe adverse reactions, including death through mechanisms that mimic malignant hyperthermia, have been reported when these drugs have been used together (Armstrong, and Bersten, 1986 ; Foley, and Inturrisi, 1987).

Patients treated with drugs for cardiovascular and metabolic disease frequently must continue their drugs throughout the intra- and postoperative period. Fortunately, severe interactions between these drugs and opioids are unlikely.

Alcoholics who must have surgical procedures should be maintained on benzodiazepines or alcohol throughout the intra- and postoperative period to prevent a withdrawal reaction or delirium tremens.

Clinicians must remain aware that patients in the categories discussed in this section may not respond as expected to medications administered for symptom control following surgery. Careful assessment and reassessment of patients' responses to analgesics and dose titration to response are always necessary. The concomitant use of nonpharmacological treatments as adjunctive therapy of postoperative pain is also strongly recommended.

Patients with Shock, Trauma, and Burns

Victims of injury frequently present in a state of cardiovascular or respiratory instability that mandates immediate life-saving procedures (e.g., endotracheal intubation, defibrillation, cut-down, and chest tube insertion) without analgesia. The trauma patient is usually young (58.4%), frequently male (72.8%), and commonly (51.2%) has used alcohol or drugs prior to injury (Soderstrom, Trifillis, Shankar, Clark, and Cowley, 1988).

Beecher (1959) was the first to point out the difficulties in providing analgesia via the intramuscular route following a significant burn or injury. He noted the variability in absorption from site to site and prolonged absorption times in soldiers with shock. He also provided the basis for modern-day pain control after burns or injuries by suggesting the exclusive use of the intravenous route. Most authorities now recommend incremental small intravenous doses of an opioid analgesic (morphine) carefully titrated to cardiovascular and respiratory stability. Concern for cardiorespiratory instability is particularly important in the first hour after injury. Any analgesic therapy also must allow for continuous monitoring of neurologic status after a head injury and neurovascular status after limb injury.

Once the patient is resuscitated and requires definitive surgical procedures, analgesia should be provided as outlined in this guideline for the various operative sites. The use of NSAIDs in the trauma patient remains controversial. They are undoubtedly of value in the patient with minor trauma, but the risk of excessive bleeding and gastric stress ulcers may prohibit their use following closed head injury, burn injury, or other multisystem injuries. When not contraindicated by sepsis, coagulopathy, or cardiorespiratory instability, the use of regional anesthetic approaches may be beneficial as described earlier for particular operative sites. For example, discomfort and splinting due to flail chest injury may improve with epidural analgesia, and borderline perfusion of an injured limb can increase with a sympathetic blockade by an epidural local anesthetic. On the other hand, surgical evaluation must always take priority over analgesic titration in the face of sudden increases in pain (e.g., extremity swelling) or somnolence (e.g., from an expanding subdural hematoma).

The serious burn injury will require very special pain control after the initial resuscitation. The myth that "third degree burns don't hurt" unfortunately still serves as a basis for widespread institutional denial of pain assessment and treatment for burned patients (Atchison, Osgood, Carr, and Szyfelbein, 1991). Pain control is essentially absent from current reviews of burn management, scientific programs of national burn associations, or funding agendas of the Federal government or major private burn treatment organizations, much as pediatric pain and cancer pain were a decade ago. In reality, after a brief (hours-long) period of endogenous analgesia evoked by the stress of immediate burn injury, pain is often severe and intermittently excruciating for months during burn dressing changes, skin grafts, reconstructive surgery, or other interventions related to needs for prolonged ventilation or intravascular access. Nonviable, insensate tissue is always surrounded by regenerating areas from which pain may arise, considering that viable perfused tissue typically forms the inner margin of an excision. Altered pharmacokinetics and pharmacodynamics in the burn patient, who may be intubated, splinted, and unable to articulate pain, further combine to render pain management an individualized challenge. The almost universal presence of hypotension and vasodilation with or without sepsis generally precludes the use of spinal or epidural routes for pain control until the burn wound is closed. While some authors have described analgesic regimens for burn dressing changes that call for nonnarcotic analgesics, such as ketamine or nitrous oxide, these approaches are best reserved for unusual or refractory instances because of side effects such as dysphoria (Dripps, Eckenhoff, and Vandam, 1982) or bone marrow depression (Skacel, Hewlett, Lewis, Lamb, Nunn, and Chanarin, 1983), respectively. More typically, high doses of opioids are required to bring pain under control. Even then, there may be pain that is relatively refractory to opioid use, particularly if the burn site is deep or extensive. A morphine infusion alone is inadequate to produce anesthesia, such as for operative procedures or prolonged ventilation, since awareness often persists. Recent clinical studies suggest that damage to underlying nerves may account for the opioid-resistant quality of pain after severe burns (Choiniiere, Melzack, and Papillon, 1991 ; Atchison, Osgood, Carr, and Szyfelbein, 1991), and that the continuous infusion of low doses of lidocaine -- known to lessen neuropathic pain in other settings -- may be a useful analgesic option in patients with burns (Jonsson, Cassuto, and Hanson, 1991). For these reasons, and also because fear and anxiety are an almost universal response to burn injury and trauma of any kind, sedatives such as benzodiazepines are useful to supplement opioid analgesics. In addition, cognitive-behavioral strategies such as relaxation, imagery, and hypnosis have been described by burn survivors as very helpful. The large full-thickness burn with its consequences of pain, separation from family and job, and (frequently) disfigurement, is usually accompanied by depression that may require drugs and, in turn, influence analgesic effects.

Patients Who Have Procedures Outside of the Operating Room

Thousands of patients undergo painful procedures each day outside of operating rooms in emergency departments, clinics, wards, and intensive care units. Analgesia issues outside the operating room also broadly apply to patients who have ambulatory surgical procedures, after which same-day discharge is expected. In any of the above settings, many procedures can be safely performed under local infiltration or regional anesthesia or by adopting behavioral, nondrug strategies, but systemic analgesia is often required to provide optimal pain control.

Only when immediate treatment of cardiorespiratory instability is required, or if a competent patient declines treatment, should analgesia be withheld for a painful procedure. The presence of a condition that could eventually result in cardiovascular, hemodynamic, neurologic, or pulmonary instability (e.g., femur fracture, pneumothorax, skull fracture) is not an absolute contraindication to systemic analgesia, although careful titration and monitoring must be provided. Though pain control may not be needed for certain procedures (e.g., diagnostic computerized imaging, intravenous pyelogram, or ultrasound examination), providing analgesia is likely to enhance the accuracy of these studies by reducing patient writhing or restlessness because of pain.

No anesthetic or analgesic agent should be used unless the clinician understands the proper technique of administration, dosage, contraindications, side effects, and treatment of overdose. As described earlier, the intravenous route is the preferred delivery mode because of its rapid onset and easy and reliable dosing. Using an intravenous route sidesteps the pain and the unpredictable absorption, onset, and duration of action associated with intramuscular or subcutaneous injections. An intravenous cannula may be placed painlessly following intradermal injection of 1 ml of 1-2% lidocaine through a 27-30 gauge needle. Most often, intravenous titration of an opioid like morphine, with observation for 5-10 minutes between doses, will provide safe and adequate analgesia. Intravenous morphine doses may range from 1 to 10 mg depending on the age, weight, pain intensity, opioid tolerance, and nature of the procedure to be done. Dose titration must be continued throughout the procedure, since pain may break through, for example, during reduction of a joint or vigorous probing of an abscess.

Contraindications to opioid analgesia include altered sensorium, lung disease, pregnancy near term, or an inability to monitor and manage side effects such as respiratory depression in the setting where care is given. Since respiratory depression is strongly correlated with the degree of sedation, stimulation of the patient as well as the administration of small doses of naloxone (e.g., 0.04 mg), may be adequate to reverse mild degrees of hypoventilation. Of course, assisted ventilation by bag and mask, or (ultimately) endotracheal intubation and repetitive naloxone dosing, may be required to reverse more severe degrees of respiratory depression. If such respiratory depression does occur, the patient should be observed until well after the naloxone effect has worn off (usually after 1 hour). Nausea, bradycardia, and hypotension are other side effects to watch for in the clinic, ward, or emergency department.

Other opioids may be used in place of morphine. Meperidine is suitable for brief, titrated dosing but not for prolonged use. Fentanyl may be used in small doses (25 [mu]g increments in the above example) but carries a higher risk than morphine or meperidine of inducing chest wall rigidity that must be immediately managed by administering a quick-onset muscle relaxant and supporting ventilation. Apart from chest wall rigidity, any opioid may trigger an acute Parkinson's-like syndrome particularly in the elderly or in patients with Parkinson's disease under medical therapy. Some mixed agonist- antagonists have the advantage that they produce lesser degrees of biliary or ureteral smooth muscle spasm, but they also may precipitate a withdrawal syndrome in patients habituated to opioid agonists such as methadone or heroin (or in other patients taking opioids for chronic pain).

No anesthetic or analgesic agent should be used unless the clinician understands the proper technique of administration, dosage, contraindications, side effects, and treatment of overdose.

NSAIDs currently have a limited role in the management of pain during brief, painful procedures, but two other nonnarcotic agents have proven useful when administered in monitored settings by trained personnel. Intravenous ketamine has a rapid onset of action and produces a state of conscious sedation in which patients respond to verbal commands and maintain airway reflexes but experience analgesia. Possible side effects include dysphoria, tachycardia, increased salivary and tracheal secretions, and myocardial ischemia in patients with preexisting cardiac disease. Inhalation of a nitrous oxide:oxygen mixture can provide prompt anxiolysis and moderate analgesia. As a precaution, the patient should breathe through a face mask that he or she is holding, so that the mask will drop away if the patient becomes somnolent. Appropriate precautions should be taken to prevent environmental contamination with nitrous oxide (i.e., scavenging system), to avoid the possible inhalation of pure nitrous oxide without oxygen, and to withhold nitrous oxide in cases of altered sensorium, entrapped air such as pneumothorax or pulmonary blebs, bowel obstruction, air embolism, chronic pulmonary disease, or suspected decompression sickness.

Benzodiazepines may be valuable adjuncts to opioids in this setting. Although they lack analgesic properties for treatment of pain due to acute tissue injury, benzodiazepines diminish skeletal muscle spasm (e.g., during orthopedic reduction), reduce anxiety, and in higher doses, provide amnesia. Coadministration of an opioid and a benzodiazepine carries a substantially higher risk of inducing respiratory depression than administration of either drug individually, so particular vigilance is necessary. Typically, in a 70-kg adult, midazolam is used in incremental doses of 1 mg intravenously. Other agents such as phenothiazines (as antiemetics) or antihistamines (because of their weak sedative, and analgesic properties) are useful in individual cases.

Regardless of the analgesic or adjuvant given, patients should be monitored closely according to institutional standards. Such standards may include continuous observation of the electrocardiogram, frequent recording of heart rate, blood pressure, and respiratory rate, and pulse oximetry. Considering the risks associated with opioid, benzodiazepine, and other analgesic use, patients should not be left unattended between successive doses of these agents and should be watched for at least 30 minutes after the completion of outpatient procedures for which intravenous analgesia has been provided. In a transient care setting, patients should not be discharged until they are awake and can converse and ambulate. Once discharged they should be accompanied by an adult for at least two half-lives of the agents used (e.g., at least 6 hours for morphine) and should be advised not to drive an automobile or operate dangerous machinery until it is likely that all medication effects are resolved (usually 24-48 hours). Documentation of monitoring during the procedure, observation prior to discharge, and discharge instructions should be part of the patient's permanent record.

At any site where painful procedures may be performed, equipment should be available to promptly treat any untoward effects of the analgesics selected. Apart from monitoring devices, such equipment includes supplemental oxygen, devices to maintain airway patency (e.g., oral, and nasal airways, face masks, endotracheal tubes, laryngoscopes, and a bag-valve device), suction, drugs for resuscitation (e.g., atropine, naloxone), and a defibrillator. Most important, there must be present on site a physician or other provider skilled in resuscitation, particularly airway management.