1016 ARTHRITIS & RHEUMATISM Vol. 39, No. 6, Jnne 1996, pp 1016-1020 0 1996, American College of Rheumatology

ACUTE FEBRILE TOXIC REACTION IN PATIENTS WITH REFRACTORY RHEUMATOID ARTHRITIS WHO

ARE RECEIVING COMBINED THERAPY WITH METHOTREXATE AND AZATHIOPRINE

RICARDO BLANCO, VICTOR M. MARTINEZ-TABOADA, MIGUEL A. GONZALEZ-GAY,

VICENTE RODR~GUEZ-VALVERDE JOSE ARMONA, JOSE L. FERNhDEZ-SUEIRO, M. CARMEN GONZALEZ-VELA, and

Objective. To assess the frequency and clinical features of an acute febrile toxic reaction (AFTR) in patients with refractory rheumatoid arthritis (RA) re- ceiving combined therapy with methotrexate (MTX) and azathioprine (AZA).

Methods. A cohort of 43 RA patients being treated with MTXIAZA combination therapy were studied. In all of them, RA had been refractory to single-therapy disease-modifying antirheumatic drugs. We analyzed the frequency and clinical features of AFTR, which consisted mainly of the development of fever, leukocy- tosis, and cutaneous leukocytoclastic vasculitis when AZA was added to the MTX regimen.

Results. Four of the 43 patients (9.3%) who had been receiving long-term, well-tolerated treatment with MTX (mean f SD 375.5 f 159.5 days, range 227-561 days) developed AFTR shortly (mean rt SD 25.7 f 13.6 days, range 17-46 days) after the addition of AZA to the regimen. The AFTR resolved rapidly (3 f 1.4 days) after discontinuation of AZA and MTX. In 2 cases, rechal- lenge with AZA and MTX was linked to a new flare of AFTR.

Conclusion. The knowledge of this side effect is particularly important because it mimics a severe infec- tious complication related to immunosuppressive ther- apy, and because rechallenge can produce severe toxi-

Ricardo Blanco, MD, Miguel A. Gonzilez-Gay, MD, PhD: Hospital Xeral-Calde, Lugo, Spain; Victor M. Martinez-Taboada, MD, Jose Armona, MD, Jose L. Fernhndez-Sueiro, MD, M. Carmen Gonzilez-Vela, MD, Vicente Rodriguez-Valverde, MD: Hospital Uni- versitario “M. Valdecilla,” Universidad de Cantabria, Santander, Spain.

Address reprint requests to Ricardo Blanco Alonso MD, Division of Rheumatology, Hospital Xeral-Calde, c/Dr. Ochoa s/n, 27004 Lugo, Spain.

Submitted for publication August 10, 1995; accepted in revised form December 21, 1995.

city. Most of the new combined therapies for RA do not seem to be more toxic than single-drug treatment. Nevertheless, clinicians should be aware of a possible increase in side effects due to drug interactions or some other unidentified mechanism.

Rheumatoid arthritis (RA) is an often-aggressive systemic disease that can result in significant morbidity and mortality. Disease-modifying antirheumatic drugs (DMARDs) used as single agents often fail to control the disease properly. Combined therapy with several of these agents has been shown to be effective and rela- tively safe for the treatment of RA (1,2). Combined treatment with azathioprine (AZA) and methotrexate (MTX) is frequently used, although its therapeutic ad- vantages over single-therapy MTX have not been con- clusively established (3). Wilke and Clough (2) have recently reviewed different combinations of DMARDs. The combination of MTX and AZA seemed to be safe in the 37 patients who received this treatment, and al- though several side effects were described, acute febrile toxic reaction (AFTR) was not reported.

A causative association between toxicity and drug treatment in RA is difficult to prove because of the common usage of many drugs and the absence of specific clinical grounds for a definitive diagnosis of drug toxi- city. However, due to the close temporal relationship between the addition of AZA to the MTX treatment regimen and the fact that single-therapy MTX had been well tolerated in all cases, we attributed the toxic reac- tion to AZA in all of the patients described below. The possible pathogenic mechanisms are discussed, and a review of the literature is presented.

ACUTE TOXICITY IN MTX/AZA-TREATED RA PATIENTS 1017

Table 1. combined therapy with methotrexate (MTX) and azathioprine (AZA)

Characteristics of the 4 patients with rheumatoid arthritis (RA) who developed an acute febrile toxic reaction (AFTR) while receiving

Dosage at onset of Days of treatment at AFTRt onset of AFTR

RA duration Previous DMARD Patient Agehex (years) treatment' MTX AZA MTX AZA

1 14lF I Gold, chloroquine 10 150 45 5 46 2 69/M 13 D-penicillamine, gold 15 100 227 17 3 37lF 4 Gold 12.5 100 561 21 4 42/F 5 Gold, D-penicillamine 7.5 100 259 19

* DMARD = disease-modifying antirheumatic drug. 'r MTX dosage is in mg/week orally; AZA dosage is in mg/day orally.

PATIENTS AND METHODS

Forty-three patients (33 women, 10 men) with RA that was refractory to single-therapy DMARD treatment were studied retrospectively. All of them fullfilled the American College of Rheumatology (formerly, the American Rheuma- tism Association) 1987 revised criteria for RA (4). They had previously received different DMARDs (gold, chloroquine, and/or D-penicillamine) as single therapy. Treatment with MTX was prescribed because of inadequate disease control or side effects with the previous drugs. The MTX dosage was progressively increased to a maximum of 15 mg weekly in all 43 patients.

After at least 6 months of well-tolerated treatment with MTX as single therapy, AZA was added because of inadequate response. We analyzed the frequency and clinical features of A m , which developed when AZA was added to the MTX regimen. The features consisted mainly of fever, skin rash, leukocytosis, and cutaneous leukocytoclastic vasculitis.

RESULTS

Four of the 43 patients (9.3%) receiving long- term, well-tolerated treatment with MTX (mean -+ SD 375.5 f 159.5 days, range 227-561 days) developed AFTR shortly (mean 25.7 f 13.6 days, range 17-46 days) after the addition of AZA to the regimen (Table 1). This side effect resolved rapidly (3.0 f 1.4 days, range 1-4 days) after the discontinuation of AZA and MTX. In 2 cases, rechallenge with AZA and MTX at the same dosage level was linked to a new flare of AFTR. There were no significant differences in MTX or AZA dosage between the 4 patients with systemic toxicity and the other 39 patients without this syndrome. The courses of the 4 patients are described below.

Patient 1. Patient 1, a 74-year-old woman, was diagnosed as having seropositive RA at the age of 67. She was initially treated with various nonsteroidal anti- inflammatory drugs (NSAIDs), and the DMARDs par- enteral gold (discontinued after 5 years due to inefficacy) and chloroquine (discontinued after 7 months due to

skin toxicity). She began taking MTX at an initial dosage of 5 mg/week orally, and the dosage was increased to 15 mg/week without significant improvement. After 13.5 months of MTX treatment, the regimen was changed to M A (150 mg/day orally) plus MTX (10 mghveek). Forty-six days later, she had an acute onset of fever (39"C), chills, joint and muscle pain, and an erythema- tous rash on the palms and dorsa of both hands. Initially, it was believed that she had an infection, and the immunosuppressive treatment was withdrawn and anti- biotic therapy started. Symptoms disappeared within 2 days.

After 10 days of antibiotic therapy, the MTX and AZA were restarted. Similar symptoms reappeared 24 hours later. Skin biopsy of the new cutaneous lesions demonstrated a leukocytoclastic vasculitis. Laboratory tests showed a white blood cell count (WBC) of 20.4 X lO"/liter (81% polymorphonuclear cells, 8% monocytes, 11% lymphocytes), a Westergren erythrocyte sedimen- tation rate (ESR) of 52 mm/hour, and a mild increase in transaminase levels. Symptoms disappeared completely within 4 days of treatment discontinuation.

Patient 2. Patient 2 was a 69-year-old man who had been diagnosed with seronegative RA at age 56. Initially, he was treated with NSAIDs and low-dose prednisone. D-penicillamine and parentera] gold were tried, but were discontinued due to toxicity. He began taking MTX at 7.5 mg/week orally, and the dosage was progressively increased to 15 mg/week. His RA re- mained active, and 7 months later, AZA (100 mg/day orally) was added to the regimen.

Seventeen days later, he was admitted to the hospital because of the sudden development of fever (40"C), chills, nausea, vomiting, worsening of articular symptoms, and a rash on the legs, soles of the feet, and palms of the hands. Laboratory studies showed a WBC count of 19 X lOY/liter (79% polymorphonuclear cells,

1018 BLANCO ET AL

Table 2. with methotrexate and azathioprine"

Clinical features of the 4 rheumatoid arthritis patients who developed an acute febrile toxic reaction while receiving combined therapy

Skin biopsy Patient Temperature Skin lesion findings WBC ESR Other symptoms

1 39°C Erythematous rash L. vasculitis 20.4 52 Chills, jointimuscle pain 2 40°C Erythematous rash L. vasculitis 19 57 Chills, vomiting, jointimuscle pain 3 39°C Purpuric lesions L. vasculitis 22.5 44 Jointimuscle pain 4 40°C Erythematous rash L. vasculitis 2 94 Vomiting, jointimuscle pain

* WBC = white blood cells (X lO'/liter); ESR = erythrocyte sedimentation rate (Westergren; mm/hour); L. vasculitis = leukocytoclastic vasculitis.

6% monocytes, 15% lymphocytes), a mild normochro- mic normocytic anemia, and an ESR of 57 mm/hour. The skin biopsy revealed a leukocytoclastic vasculitis. AZA and MTX were withheld. His systemic manifesta- tions improved 24 hours later, although the skin lesions persisted for 10 days.

Patient 3. Patient 3 was a 37-year-old woman with seronegative RA of 4 years' duration. She was initially treated with NSAIDs and parenteral gold (discontinued after 1 year because of inefficacy). MTX (7.5 mg/week orally) was started, and the dosage was increased to 15 mg weekly.

Eighteen months later, because of persistent ac- tive RA, M A (100 mgiday orally) was added to the MTX (which was then being given at 12.5 mgiweek). Three weeks later, she developed fever (39"C), myalgias, worsening of articular symptoms, and purpuric lesions on her limbs. Skin biopsy revealed a leukocytoclastic vasculitis. Laboratory studies showed a mild normochro- mic normocytic anemia, a WBC count of 22.5 X 109/liter (75% were polymorphonuclear cells, 10% monocytes, 15% lymphocytes), and an ESR of 44 mm/hour. The patient was asymptomatic 72 hours after the treatment was withdrawn.

Patient 4. Patient 4 was a 42-year-old woman in whom seropositive RA had been diagnosed at the age of 38. She had been treated with NSAIDs, low-dose pred- nisone, and parenteral gold for 1 year and, subsequently, D-penicillamine for 6 months, without significant im- provement. She started treatment with MTX, 7.5 mg/ week orally, which was increased to 15 mgiweek.

After 8 months of inadequate control of her RA, AZA (100 mg/day) was added to the MTX (which was then being given at 7.5 mg/week). Nineteen days later, she had an acute onset of fever (40"C), chills, joint and muscle pain, nausea, vomiting, and an erythematous rash on her legs. Skin biopsy revealed a leukocytoclastic vasculitis. The immunosuppressive treatment was with- drawn and amoxicillin was started. The symptoms re- solved after 3 days.

After 7 days of antibiotic treatment, MTX (7.5 mg/week) and M A (100 mg/day) were reintroduced, and 12 hours later she developed the same clinical features described above. The laboratory tests revealed a mild increase in transaminase levels, an ESR of 94 mmihour, and a WBC count of 24 X 109/liter (77% were polymorphonuclear cells, 7% monocytes, 16% lympho- cytes). The symptoms disappeared 4 days after the treatment was discontinued.

DISCUSSION

Azathioprine is an immunosuppressive drug that is widely used in the treatment of autoimmune diseases. Common side effects include gastrointestinal reactions, hematologic reactions, and hepatic toxicity (5,6). AFTR due to AZA treatment has not been commonly re- ported: it has been described in -12% of the patients studied in some small series (6,7), but was not specifi- cally mentioned in a report of a larger series (5).

Saway et a1 ( 8 ) have reviewed the manifestations of reported hypersensitivity reactions secondary to AZA. The time to onset of symptoms following initial exposure ranged from 3 hours to 42 days, with a mean of 14 days. Clinical features mainly included an acute onset of fever, chills, joint and muscle pain, leukocytosis, and cutaneous lesions. Various skin lesions have been re- ported: maculopapular rash was the most frequent, although purpura, urticaria, erythema multiforme, and erythema nodosum-like lesions have been described. All of our patients showed histologic features of leukocyto- clastic vasculitis on skin biopsy (Table 2). Such a vascu- litis has been described, to the best of our knowledge, in only 1 report of AZA hypersensitivity (9).

The clinical picture described above pinpoints a challenging diagnostic problem. In immunosuppressed patients with the aforementioned manifestations, an infectious process is frequently suspected. The differen- tial diagnosis is difficult and must be based on clinical grounds, such as absence of an identified source of

ACUTE TOXICITY IN MTX/AZA-TREATED RA PATIENTS 1019

infection, negative results of microbiological tests, and recovery with MTXIAZA withdrawal. In 2 of our pa- tients, as described in other reports (10,11), an infection was suspected and antibiotic treatment was initially given, and when it was believed that the infection was resolved, therapy with MTX and AZA was reintroduced. As was seen in our patients, a second reaction, generally within a few hours after the AZA is taken and some- times of greater severity than the initial one and occa- sionally life-threatening, may then occur (10-13).

Another diagnostic problem is the discernment of what drug produces the toxic reaction. All of our pa- tients were simultaneously taking NSAIDs, MTX, and AZA. NSAIDs have been frequently identified as the agents responsible for toxic reactions. Also, hypersensi- tivity reactions to MTX have been described, even with leukocytoclastic vasculitis found on skin biopsy (14). When a hypersensitivity reaction is suspected in a pa- tient who is taking several drugs, several issues should be considered in order to attribute the toxic reaction to one drug in particular: clinical features resembling those in a previously reported allergic drug reaction, the frequency of toxic reactions with each drug, the temporal relation- ship of drug exposure to onset of symptoms, and recur- rence after rechallenge with the drug (8). The main clinical features of toxic reactions to NSAIDs, MTX, and AZA are similar.

The temporal relationship between the initial drug exposure and the clinical onset of AFTR in our 4 patients gave us evidence to suspect that AZA was the causative agent. None of these patients had a modifica- tion of their NSAID treatment for at least 4.5 months before the development of AFTR. All of them had been treated with MTX for at least 227 days (Table 1). Moreover, AZA was the last drug introduced, within a mean of 26 days (range 17-46 days). A definitive diagnostic finding may be the recurrence of the symp- toms when patients are rechallenged with AZA, which usually occurs within hours and may be more severe than the initial episode. We suggest that rechallenge should be avoided unless the use of AZA is strictly necessary. Finally, cutaneous vasculitis is well known to occur in patients with RA as a complication of the disease itself. Although the relationship of the cutaneous vasculitis to disease activity in our patients cannot be completely excluded, the overall clinical picture, the close temporal relationship to M A exposure, the resolution of symp- toms after drug withdrawal, and the recurrence of the manifestations after drug rechallenge support the notion that the cutaneous vasculitis was not related to the RA itself.

The pathogenic mechanism of systemic toxicity to AZA is unknown. Some authors have suggested a genetic predisposition associated with the pheno- types HLA-A2, Bw4, Bw6, DR4, and DQw3 (7). As with other drugs, the mechanism is probably immune medi- ated. AZA is composed of 6-mercaptopurine (6-MP) and an imidazole group. Initially, it was postulated that the imidazole molecule, functioning as a hapten, was responsible because rechallenge with 6-MP failed to produce a recurrence of a previous reaction (15). How- ever, in a subsequent case, recurrence of a previous allergic reaction to AZA was produced when 6-MP was given (16).

Hypersensitivity reactions to AZA as single ther- apy have been reported (6-13,15,16). However, among our 74 RA patients treated with AZA as single therapy, we have not seen this side effect. The reported AFTR was seen in 4 of our 43 patients who were taking MTX plus AZA (9.3%); thus, we believe that AZA toxicity may be increased when MTX is being taken simulta- neously. After absorption, AZA is cleaved hepatically into 6-MP, the active metabolite. This undergoes metab- olism along 3 major competing enzymes: hypoxanthine guanine phosphoribosyl transferase, thiopurine methyl- transferase (TPMT), and xanthine oxidase (XO). TPMT activity is controlled genetically, and it is known that very low enzyme activity occurs in 1 out of 300 subjects. These people would be predisposed to the development of AZA toxicity (17,lS). Allopurinol is a well-recognized XO inhibitor, and when it is prescribed simultaneously with AZA, its dosage must be reduced. MTX is a less-recognized XO inhibitor that enhances the bioavail- ability of 6-MP (19). In our patients, low-dose oral MTX was administered concurrently with AZA, potentially increasing its toxicity.

Willkens et a1 (3) concluded, in an elegant study, that combined therapy with MTX plus AZA is superior to single therapy with AZA. They did not describe any case of AFTR among the side effects. Paradoxically, they raised the possibility of protection against AZA toxicity mediated by MTX in patients receiving combined ther- apy with these drugs. We believe our findings may be related to a higher dosage level of these drugs in our patients with AFTR, compared with the group of RA patients studied by Willkens et al.

In conclusion, the differential diagnosis of this syndrome in patients with RA who are undergoing treatment with MTX and AZA presents a number of difficulties. There are significant risks associated with rechallenge with this combined therapy after an episode of systemic toxicity. We suggest that when AZA is

1020 BLANCO ET AL

prescribed simultaneously with MTX, the patient should

the possible occurrence of AFTR.

8. Saway PA, Heck LW, Bonner JR, Kirklin J K Azathioprine hypersensitivity: case report and review of the literature. Am J Med 84:960-964, 1988

9. Bergman SM, Krane NK, Leonard G, Soto-Aguilar MC, Wallin

be closely monitored and the physician made aware of

1.

2.

3.

4.

5.

6.

7.

REFERENCES

McCarty DJ: Personal experience in the treatment of seropositive rheumatoid arthritis with drug used in combination. Semin Arthri- tis Rheum 23:42-49, 1993 Wilke WS, CIough JD: Therapy for rheumatoid arthritis: combi- nations of disease-modifymg drugs and new paradigms of treat- ment. Semin Arthritis Rheum 21:21-34, 1991 Willkens RF, Urowitz MB, Stablein DM, McKendry RJR Jr, Berger RG, Box JH, Fiechtner JJ, Fudman EJ, Hudson NP, Marks CR, Brooks R, Rooney TW, Rubin BR, Schinid FR, Segal AM, Thomas JW, Golstein AG, Yunus MB, Wortmann RL, Sherrer YRS: Comparison of azathioprine, methotrexate, and the combi- nation of both in the treatment of rheumatoid arthritis: a con- trolled clinical trial. Arthritis Rheum 35:849-856, 1992 Arnett FC, Edworthy SM, Bloch DA, McShane DJ. Fries JF, Cooper NS, Healey LA, Kaplan SR, Liang MH, Luthra HS, Medsger TA Jr, Mitchell DM, Neustadt DH, Pinals RS, Schaller JG, Sharp JT, Wilder RL, Hunder GG: The American Rheuma- tism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 31:315-324, 1988 Singh G, Fries JF, Spitz P, Williams CA: Toxic effects of azathio- prine in rheumatoid arthritis: a national post-marketing perspec- tive. Arthritis Rheum 32:837-843, 1989 Kissel JT, Levy RJ, Mendell JR, Griggs RC: Azathioprine toxicity in neuromuscular diseases. Neurology 36:35-39, 1986 Jeurissen MEC, Boerbons AM Th, van de Putte LBA, Kruijsen MWM: Azathioprine induced fever, chills, rash, and hepatotoxicity in rheumatoid arthritis. Ann Rheum Dis 49:25-27, 1990

JD:-Azathioprine and hypersensitivity vasculiti; Ann Intern Med

10. Rosenthal E: Azathioprine shock. Postgrad Med J 62:677-678, 1986

11. Keystone EC, Shabas R: Hypotension with oliguria: a side-effect of azathioprine (letter). Arthritis Rheum 241453-1454, 1981

12. Trotta F, Menegale G, Fiocchi 0: Azathioprine-induced hypoten- sion with oliguria (letter). Arthritis Rheum 25:1388-1389, 1982

13. Cunningham T, Barraclough D, Muirden K: Azathioprine-induced shock. BMJ 283:623-624, 1982

14. Navarro M, Pedragosa R, Lafuerza A, Rubio D, Huguet P: Leukocytoclastic vasculitis after high-dose methotrexate. Ann In- tern Med 105:471-472. 1986

15. Davis M, Eddleston A, Williams R: Hypersensitivity and jaundice due to azathioprine. Postgrad Med J 56274-275, 1980

16. Watts G, Corston R: Hypersensitivity to azathioprine in myasthe- nia gravis. Postgrad Med J 60:326-363, 1984

17. Lennard L, Van Loon JA, Weinshilboum RM: Pharmacogenetics of acute azathioprine toxicity: relationship to thiopurine methyl- transferase genetic polymorphism. Clin Pharmacol Ther 46:149- 154, 1989

18. Kerstens PJSM, Stolk JN, De Abreu RA, Lambooy LHJ, van de Putte LBA, Boerbooms AAMT: Azathioprine-related bone mar- row toxicity and low activities of purine enzymes in patients with rheumatoid arthritis. Arthritis Rheum 38:142-145, 1995

19. Balk FM, Holcenberg JS, Zimm S, Tubergen D, Collins JM, Murphy RF, Gilchrist GS. Hammond D, Poplack DG: The effect of methotrexate on the bioavailability of oral 6-mercaptopurine. Clin Pharmacol Ther 41:384-387, 1987

109:83-84, 1988