Abstract: The purpose of this review is to present relatively detailed information on the characteristics of tricyclic antidepressants, mainly amitriptyline hydrochloride and doxepin hydrochloride, for use as an integral part of the safe and effective management of fibromyalgia and, to a lesser extent, other chronic pain syndromes. Data sources include MEDLINE searches in English, relevant reference books and textbooks, my personal database and library, as well as personal clinical experience. I discuss these data with regard to the pharmacologic characteristics, mechanisms of action, adverse effects, and precautions involved with the use of tricyclic antidepressants. Additional information is given on drug selection and dosage titration. Much emphasis is placed on the fact that while tricyclic antidepressants play a major role in the management of fibromyalgia and other chronic pain syndromes, lifestyle alterations (e.g., physical reconditioning and exercise), as well as behavior modification, are also vital to a successful outcome in management.
Copyright © 1996 American Medical Association.
Tricyclic antidepressants (TADs), especially amitriptyline hydrochloride, imipramine hydrochloride, and doxepin hydrochloride, as well as other antidepressants (ADs), have been found to be useful and effective in the treatment of fibromyalgia (FM) and other chronic pain syndromes (CPSs), both with and without depression; however, none has yet been formally approved for these indications without associated depression. It is clear that most patients (less than or equal to 87%) with FM and other CPSs have an associated, although usually mild element of depression.1,2 This, of course, raises the question does pain produce depression, or could it be the other way around? In a controlled study, Frank et al3 found that low-dose TADs were effective in the adjunctive relief of the pain of rheumatoid-arthritis either with or without depression, but low-dose TADs had no significant effect on depression when it was present.
In FM and CPSs, TADs are given in low doses and are usually safe, although they are not without significant side and adverse effects. Appropriate use of TADs requires a good understanding of their pharmacologic features and actions, careful selection of the drug to be used, slow adjustment of the dosage given, and close attention to the time of day the medication is taken. Failure to consider any of these factors can make the difference between a satisfactory response and failure. Also, as patients gain sophistication and a better understanding of medicines, especially patients with chronic pain, one must spend more time (or provide handouts) explaining how and why these drugs work. Physical reconditioning and exercise are essential parts of the therapeutic program for FM and usually for CPSs as well.
Although there is much research — much of it controlled — that supports the use of low-dose TADs in the treatment of FM,4-9 their use in treating other CPSs has been much more extensively documented, and this has long antedated their use in treating FM. Cyclobenzaprine hydrochloride (Flexeril) has a chemical structure that is similar to that of TADs, and it also seems to work in treating FM10; however, treatment with cyclobenzaprine has not seemed to work in all studies.11 By usual definition, chronic pain has to be present for at least 6 months to qualify as a CPS.
Chronic pain syndromes may be of two types. The first is usually the result of an identifiable, ongoing illness (e.g., an injury, arthritis, or even cancer) where the cause is evident and rationally adequate to explain the perceived pain. The use of TADs here is as adjuvant therapy. The other type of CPS results from chronic pain of an unidentifiable or seemingly inadequate cause. This latter type of pain is generally what is meant by CPS and is usually refractory to standard pain therapy, including, eventually, narcotics. This is also called a learned pain response or syndrome and seems to involve a self-perpetuating alteration in pain perception and an amplified response to that perceived pain to produce suffering behavior.12
There are numerous reports of the successful use of TADs and other ADs in the management of both types of CPS. Detailed discussion of them is beyond the scope of this report; however, there are many good reviews and books on the subject.2,13-18 The discussion here is focused mainly on the treatment of FM, although much of it is applicable to the treatment of other CPSs as well.
Pharmacokinetics and Pharmacodynamics of TADs
Absorption, distribution, metabolism, and excretion of TADs are highly variable among individuals and may vary as much as 50-fold for both amitriptyline and doxepin.19 Some of this variability seems to be genetically controlled but, in any case, requires close attention to the initial drug treatment, dosage, and rapidity of altering the dosage.
The psychopharmacologic effects of TADs are quite different in normal as opposed to depressed individuals. While they cause mood elevation in depressed patients, they do not produce euphoria in normal persons but rather a mild dysphoria and more intense and noticeable anticholinergic effects.
The pharmacokinetics and pharmacodynamics of TADs have a great influence on how they are best used. Their onset of action of a single dose is variable but slow and usually ranges from 2 to 3 hours. A peak plasma concentration is achieved between 3 and 8 hours. A plasma half-life ranges from 10 to 20 hours for a single dose. Metabolism occurs in the liver with excretion by the kidneys. Many TADs have active metabolites (e.g., nortriptyline hydrochloride from amitriptyline and nordoxepin from doxepin) that may further prolong their action — in some cases, up to several days.
Unlike the nearly immediate anticholinergic and sedative effects, maximum antidepressant clinical effectiveness is considerably delayed, requiring 3 to 6 weeks.19 Patient response to a given low dosage that is taken every night for sleep or pain modification is also quite variable and delayed. The full sleep-modifying and analgesic effect at a given nightly dose requires some 1 to 3 weeks (similar to, but still shorter than the time that is needed to obtain the full antidepressant effect at higher dosages).2,4,6,14
Fortunately, some tolerance develops to most of the anticholinergic effects within a few weeks.19 On the other hand, tolerance to the antidepressant and sedative effects does not occur even after years of use.20 Tolerance to the analgesic effect may develop in some patients.6,17,21
Tricyclic antidepressants tend to improve all the characteristic features of the FM sleep disorder.22 The quality and depth of sleep do seem more important than does the simple duration. Regular sedatives or hypnotics are ineffective and even counterproductive in treating FM and CPSs.
Mechanism of Action
This group of drugs has been available for nearly 40 years; yet, their exact mechanism of antidepressant action is still not known.23 There is at least equal uncertainty as to the mechanism of their analgesic and sleep-modifying actions.24 Much has been written on this topic, mainly on the effects of TADs on neurotransmitters and their receptors. 2,14,25,26 Henriksson 29 recently wrote an excellent and exhaustive review on the origin and pathogenesis of FM, including the morphological, biochemical, physiological, and psychological aspects, as well as some hypotheses. Henriksson29 summarized by saying, “The conclusion is that no one single mechanism can explain FMS (fibromyalgia syndrome) and is thus in that sense a compromise.”
Many authorities believe serotonin activity and receptor-reuptake blockade may be most closely related to the analgesic effects of TADs since the ones that affect serotonin most seem to work best for this purpose. Part of their pain modification also seems to come from an increase or potentiation of endogenous opioids (endorphins) in the brain.15,26 There is, however, no real consensus on these views, and the real mechanism(s) may be for entirely different reasons that we do not yet understand.
If serotonin production and uptake blockade are crucial to the analgesic effect of TADs and to the treatment of the FM sleep disorder, then amitriptyline and doxepin may be uniquely (or at least better) suited for this purpose since amitriptyline has a serotonin-norepinephrine effect of 9: 1.2,23 This apparent uniqueness may also be partly because amitriptyline has been used and studied for this indication much more than any other TAD.14 Doxepin has also been extensively studied and found to be especially effective in treating CPSs, including rheumatic diseases,17 but not specifically in treating FM. It is curious that, in diabetic neuropathy, fluoxetine hydrochloride (Prozac) — although a strong serotonin-reuptake inhibitor — had little or no effect as an analgesic, while desipramine hydrochloride (Norpramin) had virtually no serotonin effect but strongly inhibited the reuptake of norepinephrine and was a good analgesic.31 Wolfe et al32 found fluoxetine was also ineffective in treating FM in a controlled study; however, Goldenberg et al33 Very recently did another controlled study in which they used fluoxetine and amitriptyline separately and in combination for treating FM, and they found fluoxetine about as effective as amitriptyline, but the combination was more effective than either alone.
The antidepressant and the analgesic and sleep-modifying effects seem to be quite different from each other.14,25 Although depression and pain seem to be chemically linked,25 all TADs (and other ADs for that matter) are about equally effective against depression despite the extremely different other actions that they have. On the other hand, for analgesia and sleep modification, they are as different in efficacy as their different actions.14
There are few absolute contraindications to the use of TADs. Known allergic hypersensitivity to one of them is a contraindication for that drug and a relative contraindication or considerable caution in the use of others. Use of TADs with monoamine oxidase inhibitors is contraindicated; use of TADs is also contraindicated in children younger than 12 years in most instances and in children younger than 6 years in all cases. Also, use after a recent myocardial infarction is relatively contraindicated.
Adverse and Side Effects and Drug Interactions
Tricyclic antidepressants act primarily on the central nervous system by inhibiting the uptake of biogenic neurotransmitter amines. Most TADs also have generalized, wholebody effects. Most have variable but significant sedative action. They also have variable anticholinergic effects. Additional and potentially dangerous related adverse effects include postural hypotension, and — most dangerous and even rarely fatal — precipitation or aggravation of some cardiac arrhythmias and increase in conduction blocks. The cardiovascular adverse effects occur mostly with antidepressant doses of amitriptyline.
In many cases, they are dose-related, and generally, these adverse effects most commonly occur with amitriptyline or imipramine. Many are rare even at antidepressant doses, and few are a significant problem with low-dose therapy for FM. Even so, one needs to be aware of them and of the possibility of them occurring when TADs are used.
Putting It All Together
By using the foregoing information, one may devise a practical and effective management plan for treating FM. I am presenting the one that I use and have found to be most useful. Certainly, there are many other options and alternatives. Similar, although less-detailed, plans for treating FM have been proposed by several authorities, including Goldenberg35 and Bennett et al.36
The first step in the proper use of TADs in treating either FM or other CPSs is to be reasonably certain of the diagnosis. The 1990 criteria for the diagnosis of FM according to the American College of Rheumatology,37 Atlanta, GA, have greatly simplified and permitted improved confidence in this diagnosis. The diagnosis of other CPSs is still one of exclusion, and a thorough effort needs to be made to exclude or find other distinct and otherwise treatable causes of pain.
Once one is reasonably confident of the diagnosis, patient education becomes very important. Many patients are enormously relieved simply to learn that they really do have a true illness to explain their symptoms. One must also explain the vagaries, beneficial and adverse effects, as well as the method of using a TAD. Being able to tell patients that TADs are nonaddicting and even minimally habituating is a large bonus.
Patients who are already taking narcotics must be weaned off them either by the physician or, in extreme and refractory cases, by a carefully selected pain center or clinic. Withdrawal is best accomplished by establishing a “contract” with the patient at the outset as to how and when narcotics will be stopped entirely. The same is true for sedatives or hypnotics. Most patients with FM are referred to a rheumatologist because treatment with nonsteroidal anti-inflammatory drugs has been ineffective; their use is controversial in this condition. The consensus of most rheumatologists and authors is that, when used alone, nonsteroidal anti-inflammatory drugs are of little if any benefit in treating FM; however, in a controlled study, naproxen sodium was additive to the effect of amitriptyline, although naproxen was no better than placebo when it was given alone.9 A similar result was seen in a study of ibuprofen and alprazolam.38 Nonsteroidal anti-inflammatory drugs are usually used in low, analgesic doses as adjunctive therapy for FM by most rheumatologists.
The choice of a TAD and the dosage usually can be amitriptyline, 25 mg to be taken 2 to 3 hours before bedtime. In this way, the medication has time to begin working before the patient retires so he or she is drowsy at bedtime or shortly after going to bed; also, most of its sedative effect will be worn off by the end of an 8-hour sleep. Because of prolonged and variable effects, some patients will need to take the drug even earlier, and uncommonly, some will find it works better if they take it later or even at bedtime. It is important to remember that there is a 1- to 3-week delay in achieving the maximum sleep modification and analgesic effect at any given dose. Consequently, upward or downward dosage changes should not be made too rapidly unless adverse effects are extreme.
A small but important group of patients will find even 25 mg of amitriptyline is either too sedating or it may have intolerable anticholinergic side effects. It may produce a hangover on the next day or cause them to feel “brain-dead” as some have put it (i.e., dysphoric). Because of this, it is well to advise all patients in advance of this possibility and, if it happens, to counter it by dividing the 25-mg tablet in half with a razor blade or commercial tablet cutter. If this still proves to be too much after trying it for several nights, drug therapy for such patients should be switched to doxepin, which has a chemical and pharmacologic action profile that is very similar to that of amitriptyline; however, doxepin has less anticholinergic and other undesirable side effects. Also, doxepin comes in a liquid form. I (and many authorities on FM and CPSs) start therapy for patients with doxepin hydrochloride at a 10-mg dosage and, if need be, decrease it to only one half of one of these capsules as well. Patients who divide tablets or capsules often complain of the terrible taste of TADs and a numb mouth and tongue from the potent local anesthetic action of many TADs. This can be avoided by instructing patients to wrap the half tablet in a small ball of bread and swallow the ball and medication whole, without letting the raw drug come in contact with the tongue or oral mucosa.
Very few patients still will be unable to tolerate even 5 mg of doxepin hydrochloride; for these patients, the use of doxepin hydrochloride (10-mg/ml liquid concentrate) is indicated. Therapy is started with one to five drops 2 to 3 hours before bedtime. I have yet to see a patient who could not tolerate the 0.67 mg in one drop (15 drops per milliliter), although it seldom has any beneficial effect either. Because of this, patients should be instructed to increase the dose dropwise, by adding one or two drops every 1 to 2 weeks until a tolerable and effective level is reached, usually around five to 10 drops, or until adverse effects again become intolerable.
If neither amitriptyline nor doxepin is effective or if either produces intolerable side effects even in minute dosages, another milder TAD (e.g., nortriptyline or desipramine 31) or possibly a newer, different class of AD may be tried. For vascular headaches, Diamond39 recommended the use of trazodone hydrochloride (Desyrel), a nontricyclic AD with strong sedating properties but with minimal anticholinergic effects and only mild to moderate cardiotoxicity, which may be used when regular tricyclic agents are poorly tolerated or contraindicated. Diamond39 recommended a 50- to 100-mg dose of trazodone hydrochloride to be taken “at bedtime” to a maximum 200- to 250-mg dose. I have had little success and, hence, little experience with trazodone in treating FM. Alprazolam has also been used in a controlled study of FM and was found to be helpful.38
An additional problem arises with shift workers who must sleep at odd hours. The problem is even worse when they must sleep on a revolving or intermittent schedule, such as hospital workers, firefighters, and police often must do. These people do have a reason for concern about taking any psychoactive drug while they are on duty or taking a sedative drug when they may have to return to duty unexpectedly. Fortunately, this is not a very common problem. Cognitive and motor impairment definitely can result from the use of TADs, but these effects also seem to be highly variable and individualized.40
It is very important to remember that, at best, only about 50% of patients with FM can be expected to show a striking response to TADs.41 Many more patients will get more modest and only partial relief, while a small, significant, but unfortunate group will show little or no response to any of the ADs.15 As previously mentioned, a few patients may experience an apparent tolerance to a TAD.6,17,21 Fortunately, such cases are unusual, but if the problem should occur, the dosage of their TAD may be increased somewhat or therapy with another TAD or other AD may be tried for these patients.
Physical Reconditioning, Exercise, and Other Lifestyle Modifications
The entire use of TADs in the management of FM and CPSs presupposes their use in conjunction with maximum yet safe physical reconditioning and exercise.2,42
The very old (i.e., circa World War II) but still outstanding Royal Canadian Air Force Exercise Plan paperback book 43 gives an inexpensive, elegantly structured, graded exercise program that includes separate plans for men and women. There are, of course, many other alternatives. Spas and health clubs give a structured approach, and their cost is additional motivation to use them. I believe swimming (especially in a warm pool) is particularly beneficial. As there is an increase in exercise and improvement in physical conditioning, patients can often decrease or stop taking their TAD. In any case, an attempt should be made to decrease or stop their treatment with TADs at least after 6 months; if this is unsuccessful, the attempt should be repeated at about 6-month intervals.
Patients should be taught that “pain behavior” itself can be modified as a form of alternative or supplemental therapy that is often overlooked. This is the major area of expertise of most good pain clinics. It involves such techniques as biofeedback, coping skills, breathing, relaxation, meditation, pacing, and others, as well as exercise.12,44
Finally, other lifestyle modifications often include encouragement to lose weight, along with consideration of other dietary measures (e.g., intake of fat, cholesterol, salt, fluids, sugars, and alcohol). Vitamin and other “health food” supplements are usually not necessary, and I believe they are mostly a waste of money.
The therapy for FM is, at best, usually less than spectacular, and most patients continue to have symptoms for years despite all that may be done. On the other hand, the conditions of most patients will improve with decreased pain and increased function if they follow a good combined therapeutic program.
Accepted for publication October 19, 1995.
This work was supported in part by a research grant from Physicians Associated, Chartered, of Overland Park, Kan.
No funding was supplied by the manufacturers of any of the drugs discussed, nor do any other potential conflicts of interest exist with them.
The author thanks Daniel J. Stechschulte, MD, and H. B. Lindsley, MD, for reviewing the manuscript and offering valuable comments and suggestions. The author also especially thanks Richard Kammer, chief librarian at the Medical Staff Library, Columbia Overland Park Regional Medical Center, for exceptional assistance in literature searches and data retrieval, as well as the librarians at Dykes Medical Library, University of Kansas Medical Center, Kansas City.
Reprints: Robert Godfrey, MD, Division of Allergy, Clinical Immunology, and Rheumatology, 4035 Wescoe Pavilion, University of Kansas Medical Center, 39th and Rainbow Boulevard, Kansas City, KS 66160-7317.
Hendler N. The anatomy and psychopharmacology of chronic pain. J Clin Psychiatry. 1982;43: 15-20.
Tollison CD, Kriegel MA. Selected tricyclic antidepressants in the management of chronic benign pain. South Med J. 1988;81:562-564.
Frank RG, Kashani JH, Parker JC, et al. Antidepressant analgesia in rheumatoid arthritis. J Rheumatol 1988;15:1632-1638.
Tollison CD. Antidepressant use in patients with chronic pain. Drug Ther. November 1990:50-57.
Miller DR, Seifert RD. Management of fibromyalgia, a distinct rheumatologic syndrome. Clin Pharmacol. 1987;6:778-786.
Pilowsky I, Halleff EC, Bassett DL, Thomas PG, Penhall RK. A controlled study of amitriptyline in the treatment of chronic pain. Pain. 1982;14; 169-179.
Carette S, McCain GA, Bell DA, Fam AG. Evaluation of amitriptyline in primary fibrositis. Arthritis Rheum. 1986;29:655-659.
Jaeschke R, Adachi J, Guyatt G, Keller J, Wong B. Clinical usefulness of amitriptyline in fibromyalgia: the results of 23 N-of-1 randomized controlled trials. J Rheumatol 1991;18:447-451.
Goldenberg DL, Felson DT, Dinerman H. A randomized, controlled trial of amitriptyline and naproxen in patients with fibromyalgia. Arthritis Rheum. 1986;29:1371-1377.
Bennett RM, Gatter RA, Campbell SM, Andrews RP, Clark SR, Scarola JA. A comparison of cyclobenzaprine and placebo in the management of fibrositis: a double-blind controlled study. Arthritis Rheum. 1988;31:1535-1542.
Reynolds JW, Moldofsky H, Saskin P, Lue FA. The effect of cyclobenzaprine on sleep physiology and symptoms in patients with fibromyalgia. J Rheumatol 1991;18:452-454.
Fordyce WE. Cognitive and behavioral treatment of chronic pain and disability. In: Foley KM, Payne RM, eds. Current Therapy Series: Current Therapy of Pain. Philadelphia, Pa: BC Decker Inc; 1989:7-13.
Foley KM, Payne RM, eds. Current Therapy Series: Current Therapy of Pain. Philadelphia, Pa: BC Decker Inc; 1989.
Goodman WK, Charney DS. Therapeutic applications and mechanisms of action of monoamine oxidase inhibition and heterocyclic antidepressant drugs. J Clin Psychiatry. 1985;46:6-22.
Getto CJ, Sorkness CA, Howell T. Antidepressants and chronic nonmalignant pain: a review. J Pain Symptom Manage. 1987;2:9-18.
US Pharmacopeial Conventions Drug Information Division Advisory Panels. Drug Information for the Health Care Professional (USP DI). 12th ed. Rockville, Md: US Pharmacopeial Conventions Inc; 1992:IA.
Magni G. The use of antidepressants in the treatment of chronic pain. Drugs. 1991;42:730-748.
Watson CPN. Antidepressant drugs as adjuvant analgesics. J Pain Symptom Manage. 1994;9: 392-405.
Baldessarini RJ. Drugs and the treatment of psychiatric disorders. In: Gilman AG, Rall TW, Nies AS, Taylor P, eds. Goodman and Gilman’s The Pharmacological Basis of Therapeutics. 8th ed. Elmsford, NY: Porgamon Press Inc; 1990:404.
Ayd FJ Jr. Five to fifteen years’ maintenance doxepin therapy. Int Clin Psychopharmacol 1986; 1:53-65.
Carrette S, Bell MJ, Reynolds WJ, et al. Comparison of amitriptyline, cyclobenzaprine, and placebo in the treatment of fibromyalgia. Arthritis Rheum. 1994,37:32-40.
Hauri PJ, Esther MS. Insomnia. Mayo Clin Proc. 1990;65:869-882.
Richelson E. Antidepressants and brain neurochemistry. Mayo Clin Proc. 1990;65:1227-1236.
Yunus MB. Towards a model of pathophysiology of fibromyalgia: aberrant pain mechanisms with peripheral modulation. J Rheumatol. 1992; 19:846-849.
Feinmann C. Pain relief by antidepressants: possible modes of action. Pain. 1985;23:1-8.
Sacerdote P, Brini A, Mantegazza P, Panerai AE. A role for serotonin and beta-endorphin in the analgesia induced by some tricyclic antidepressant drugs. Pharmacol Biochem Behav. 1987; 26:153-158.
Rudorfer MV, Potter WZ. Antidepressants: a comparative review of the clinical pharmacology and therapeutic use of the ‘newer’ versus the ‘older’ drugs. Drugs. 1989;37:713-738.
Potter WZ, Rudorfer MV, Manji H. The pharmacologic treatment of depression. N Engl J Med. 1991;325:633-642.
Henriksson KG. Chronic muscular pain: aetiology and pathogenesis. Baillieres Clin Rheumatol 1994;8:703-719.
American Medical Association Department of Drugs, Division of Drugs, and Toxicology. Drug Evaluations Annual 1991 (AMA DE). Chicago, Ill: American Medical Association; 1991.
Max MB, Lynch AS, Muir MJ, Shoaf SE, Smoller B, Dubner R. Effects of desipramine, amitriptyline and fluoxetine on pain in diabetic neuropathy. N Engl J Med. 1992;326:1250-1256.
Wolfe F, Cathey MA, Hawley DJ. A double-blind controlled trial of fluoxetine in fibromyalgia. Scand J Rheumatol 1994;23:255-259.
Goldenberg DL, Mayskly M, Mossey C, Ruthazer R, Schmid C. The independent and combined efficacy of fluoxetine and amitriptyline in the treatment of fibromyalgia. Arthritis Rheum. 1995;38 (suppl 9):S229. Abstract.
Hayes JR, Bojrab SL, McCarthy MC. Gastrointestinal effects of tricyclic antidepressants: Olgivie’s syndrome. Psychosomatics. 1987;28:442-443.
Goldenberg DL. Review of the role of tricyclic modications in the treatment of fibromyalgia syndrome. J Rheumatol 1989;16(suppl 19):137-139.
Bennett RM, Smythe HA, Wolfe F. Recognizing fibromyalgia. Patient Care. July 15, 1989:60-83.
Wolfe F, Smythe HA, Yunus MB, et al. The American College of Rheumatology 1990 criteria for the classification of fibromyalgia. Arthritis Rheum. 1990;33:160-172.
Russell IJ, Fletcher EM, Michalek JE, McBroom PC, Hester GG. Treatment of primary fibrositis/ fibromyalgia syndrome with ibuprofen and alprazolam: a double-blind, placebo-controlled study. Arthritis Rheum. 1991;34:552-560.
Diamond S. Vascular headache. In: Foley KM, Payne RM, eds. Current Therapy of Pain. Philadelphia, Pa: BC Decker Inc; 1989:181.
Payne R. Medication-induced performance deficits: analgesics and narcotics. J Occupat Med. 1990;32:362-369.
Richlin DM. Nonnarcotic analgesics and tricyclic antidepressants for the treatment of chronic nonmalignant pain. Mt Sinai J Med. 1991;56: 221-228.
McCain GA, Mai FM, Halliday PD. A controlled study of the effects of a supervised cardiovascular fitness training program on the manifestations of primary fibromyalgia. Arthritis Rheum. 1988;31:1135-1140.
Royal Canadian Air Force Staff. Royal Canadian Air Force Exercise Plan for Physical Fitness. New York, NY: Simon & Schuster Inc Publishers; 1978.
White KP, Nielson WR. Cognitive behavioral treatment of fibromyalgia syndrome: a follow-up assessment. J Rheumatol 1995;22:717-721.
From the Division of Allergy, Clinical Immunology, and Rheumatology, Department of Medicine, University of Kansas School of Medicine, Kansas City.
Copyright * 1996 American Medical Association. Copyright * 1997 Information Access Company.