`Number 6, June 1995, pp 755-759
`© 1995, American College of Rheumatology
`
`755
`
`CORRELATION BETWEEN GOLD-INDUCED ENTEROCOLITIS AND
`THE PRESENCE OF THE HLA-DRB1*0404 ALLELE
`
`ELLA EVRON, CHAIM BRAUTBAR, STUART BECKER, GA VRIEL FENAKEL, YITZHAK ABEND,
`ZEV STHOEGER, PATRIZIA COHEN, and DAVID GELTNER
`
`Objective. In recen1t years we have treated 4
`rheumatoid arthritis (RA) )Jatients who developed gold(cid:173)
`induced enterocolitis, a well-recognized, although rare,
`complication of chrysotherapy. The aim of the present
`study was to seek any genetic predisposition for this
`complication.
`Methods. HLA DNA typing was done on fresh
`white blood cells from the 4 patients.
`Results. Three of th•e 4 patients (75%) exhibited
`the DRBl *0404 allele, wh•ereas the prevalence of this
`allele among the Ashkena:li Jewish population of RA
`patients without colitis was 9.2% and 10.2% in 2
`different studies.
`Conclusion. The results indicate that the
`DRBl *0404 may be associated with risk for the devel(cid:173)
`opment of gold-induced enterocolitis in this population
`and suggest that HLA DNA typing should be considered
`in Jews who may be undergoing chrysotherapy.
`
`Gold-induced enterocolitis is a well-recognized
`severe complication of chrysotherapy, being fatal in
`approximately one-third of cases (1). Only 28 cases
`have thus far been reported in the literature. We
`describe 4 women with this rare complication who
`were treated in our hospital since 1986. It has previ(cid:173)
`ously been suggested that this complication might be
`more prevalent in Jews (2) and that HLA-B8 and
`DRw3 might predispose to its development (3). Recent
`advances in molecular biology methods of HLA DNA
`typing enabled us to study the correlation between
`gold-induced enterocolitis and HLA-DRB 1 alleles in
`these 4 Jewish patients (Table 1).
`
`Ella Evron, MD, Stuart Becker, MD, Gavriel Fenakel, MD,
`Yitzhak Abend, MD, Zev Sthoeger, MD, Patrizia Cohen, MD,
`David Geitner, MD: Kaplan Hospital, Rehovot, Israel; Chaim
`Brautbar, PhD: Hadassah Medkal Center, Jerusalem, IsraeL
`Submitted for publication October 3, 1994; accepted in
`revised form December 28, 1994.
`
`PATIENTS AND METHODS
`
`HLA DNA typing. Genomic DNA was extracted using
`a salting-out procedure (4), and evaluated as described (5).
`Case reports. Patient 1. Patient 1, a 52-year-old
`Jewish woman with classic seropositive rheumatoid arthritis
`(RA), was treated with intramuscular (IM) aurothioglucose
`along with diclofenac and indomethacin. After 5 months,
`during which she received a cumulative dose of 200 mg of
`aurothioglucose (according to the conventional dosage
`schedule), she developed fever, abdominal pain, and watery
`diarrhea.
`Physical examination findings at this time were un(cid:173)
`remarkable except for tenderness in the abdomen. The
`arthritis was inactive. The erythrocyte sedimentation rate
`(ESR), complete blood cell count (CBC), electrolyte levels,
`and results of kidney and liver function tests were within
`normal limits. The serum protein level was 5. 7 gm/dl, and the
`albumin level was 2. 7 gm/dl. Rheumatoid factor (RF) was
`positive at a titer of 1:2,560 by latex fixation, and negative by
`Rose-Waaler test. Stool cultures for bacteria, parasites, and
`tuberculosis (TB) were negative. Colonoscopy demonstrated
`mildly edematous inflamed mucosa without friability or
`ulceration. Biopsy showed a nonspecific lymphocytic and
`eosinophilic infiltrate. Contrast radiography of the intestine
`showed signs of diffuse and severe ileitis with no involve(cid:173)
`ment of the colon (6).
`Gold salt therapy was discontinued. Treatment with
`300 mg/day intravenous (IV) hydrocortisone and 2 gm/day
`sulfasalazine had no effect on the clinical course. Parenteral
`hyperalimentation was begun, together with British antile(cid:173)
`wisite (dimercaprol). Within a few days marked improve(cid:173)
`ment was noted, with clinical remission occurring after 2
`weeks of British antilewisite (total dose 700 mg). A small(cid:173)
`bowel radiographic series repeated 1 year later showed no
`evidence of disease. Six years after the acute event the
`patient is symptom free. HLA DNA typing revealed that the
`patient carried the DRB1*0405 (homozygous), DQB1*0501,
`and DQB1*0201 alleles.
`Patient 2. Patient 2 was a 46-year-old Jewish woman
`who began treatment with IM aurothioglucose for exacerba(cid:173)
`tion of classic seropositive RA that was not responsive to
`nonsteroidal antiinflammatory drug (NSAID) treatment.
`Four weeks later, after receiving a cumulative aurothioglu(cid:173)
`cose dose of only 115 mg (10 mg, 25 mg, 40 mg, and 40 mg in
`weeks 1, 2, 3, and 4, respectively), she noticed a maculo-
`
`
`
`756
`
`EVRON ET AL
`
`Table I. Features of 4 rheumatoid arthritis (RA) patients with gold-induced enterocolitis*
`
`Involved part
`of the GI tract
`
`Treatment
`
`Outcome
`
`Ileum
`
`Hydrocortisone IV
`300 mg/day,
`sulfasalazine
`orally 2 gm/day,
`BAL IM (total
`700 mg),
`hyperalimentation
`Hydrocortisone IV
`300 mg/day,
`BAL IM (total
`150 mg),
`hyperalimentation,
`antibiotics
`Prednisone 60 mg/
`day,
`hydrocortisone
`IV 300 mg/day,
`antibiotics, BAL
`IM (total I ,000
`mg),
`hyperalimentation
`Terminal ileum Hydrocortisone IV
`and colon
`300 mg/day,
`BAL IM (total
`1,000 mg)
`
`Cured
`
`Died
`
`Cured
`
`Cured
`
`Patient
`
`Agel
`sex
`
`Molecular
`tissue typing
`(allele)
`
`Diagnosis
`
`Gold preparation,
`administration
`route, total dose
`
`52/F DRBI*0405
`(homozygous),
`DQBI*0501,
`DQBI*0201
`
`Seropositive Aurothioglucose, IM,
`200 mg
`RA
`
`Time to
`onset of
`symptoms
`
`5 months
`
`2
`
`3
`
`4
`
`46/F DRBI*0402,
`DRBI*0404,
`DQB1*0302,
`DQBI*0402
`
`40/F DRB1*0404,
`DRBI *IO,
`DQBI*0302,
`DQBI *0501
`
`36/F DRBI*0404,
`DRBI*Ol02,
`DQB1*0402,
`DQB1*0501
`
`Seropositive Aurothioglucose, IM,
`RA
`115 mg
`
`4 weeks
`
`Entire small
`and large
`bowel
`
`Seronegative Aurothioglucose, IM,
`200 mg
`RA
`
`2 months Terminal ileum
`and colon
`
`Seronegative Aurothioglucose, IM,
`RA
`200 mg
`
`1 month
`
`*IM = intramuscular; IV = intravenous; BAL = British antilewisite.
`
`papular rash on her arms, and was having abdominal pain
`with watery diarrhea. Chrysotherapy was stopped but her
`diarrhea continued; she lost 6 kg and was hospitalized for
`evaluation and therapy.
`Physical examination revealed, in addition to the
`maculopapular rash on the arms, abdominal tenderness with
`no signs of peritoneal irritation. There was no evidence of
`joint disease. Initially the ESR, CBC, electrolyte and serum
`protein levels, and liver and kidney function test results were
`normal. The RF titer was 1:96 by latex fixation and I: 160 by
`Rose-Waaler test; stool cultures for bacteria, parasites, and
`TB were negative. Colonoscopy demonstrated severe colitis
`with edema, bleeding, and pus. On colonic biopsies, acute
`and chronic inflammation with infiltration of eosinophils was
`seen, with no granulomas (Figure 1). Small-bowel radio(cid:173)
`graphic series showed narrowing of the entire ileum with
`signs of severe inflammation (Figure 2).
`The patient was treated with total bowel rest and
`high-dose systemic corticosteroids. Fever, abdominal pain
`and tenderness, and diarrhea continued. Treatment with
`parenteral hyperalimentation, British antilewisite (cumula(cid:173)
`tive dose 150 mg), and broad-spectrum antibiotics similarly
`had no effect. Leukopenia (I , 700 white blood cells [WBC]/
`p,l, with 77% neutrophils, 19% lymphocytes), thrombocyto(cid:173)
`penia (32,000 platelets/ p,l), and elevation of liver enzyme
`levels (serum glutamic oxaloacetic transaminase 51 units,
`serum glutamic pyruvic transaminase 158 units, lactate de(cid:173)
`hydrogenase 350 units) appeared. Toxic megacolon, later
`accompanied by shock, necessitated emergency laparotomy.
`
`No bowel perforation was noted, but the entire small and
`large bowel were inflamed. The patient died a few days later,
`in a state of multiple organ system failure. HLA DNA typing
`
`Figure 1. Colonic biopsy of patient 2, showing dilated glandular
`crypts Lined by a flattened epithdium and containing crypt ab(cid:173)
`scesses. Note the mixed inflammatory infiltrate in the suJTounding
`lamina propria (hematoxylin and eosin stained; original magnifica(cid:173)
`tion x 220).
`
`
`
`DRBI *0404 AND GOLD-INDUCED ENTEROCOLITIS
`
`757
`
`with cessation of diarrhea, resumption of oral food intake,
`and clinical remission.
`The patient was discharged a few weeks later, and
`treatment with prednisone 10 mg/day and methotrexate was
`initiated. There was no evidence of ileocolitis during 5 years
`of followup. Results of repeat colonoscopy with biopsies
`were normal, as were the findings of barium enema testing
`and small-bowel radiographic series carried out 5 years after
`the episode. Molecular tissue typing showed DRB 1 * l 0,
`DRB1*0404, DQB1*0302, and DQB1*0501 alleles.
`Patient 4. Patient 4 was a 36-year-old Jewish woman
`who had had seronegative RA for 2 years, with arthritis of
`the hands and deformity of the right hip. After an unsuccess(cid:173)
`ful attempt at treatment with NSAIDs she began treatment
`with intramuscular aurothioglucose, 50 mg/week for I
`month, along with prednisone 15 mg/day. One month later
`fever, abdominal pain, and severe watery diarrhea devel(cid:173)
`oped. The patient was admitted to our unit. Pertinent phys(cid:173)
`ical findings included fever (38°C), mild dehydration, and
`diffuse abdominal tenderness, without signs of peritoneal
`irritation. There was no skin rash. Laboratory studies
`showed normal CBC, ESR, electrolyte levels, and kidney
`and liver function test results. The serum protein level was
`5.3 gm/dl and the albumin level was 3.4 gm/dl. RF was
`absent by Rose-Waaler and latex fixation tests.
`Gold therapy was discontinued. Stool cultures for
`bacteria and parasites were negative. Sigmoidoscopy dis(cid:173)
`closed friable edematous mucosa with no ulcerations. On
`biopsy a chronic, minimal, nonspecific inflammation with an
`inflammatory infiltrate was found. Subsequent small-bowel
`radiographic series disclosed mild rigidity with ulcerations of
`the terminal 12 em of the ileum. Barium enema showed loss
`of haustration of the descending colon.
`The patient was treated with IV fluids and oral
`prednisone 10 mg/day. Severe diarrhea persisted, and IM
`British antilewisite 100 mg twice daily was started on day 4;
`this was discontinued on day 9 because it had no effect. IV
`hydrocortisone 300 mg/day was administered, with gradual
`improvement. The patient was discharged from the hospital
`on day 24 and given prednisone 40 mg/day, which was
`gradually tapered to a maintenance dosage of 10 mg/day.
`Findings on a followup small-bowel radiographic series 2.5
`years later were completely normal. Molecular HLA typing
`showed DRB1 loci with DRB1 *0102 , ORB! *0404 ,
`DQB1*0402, and DQB1 *0501 alleles.
`RA controls. As a control, we used data from 2
`groups of RA patients studied at our institution. In 1 group
`(5), the prevalence of the DRB 1 *0404 allele was 5 of 49
`patients (10.2%), and in the second (7), the prevalence was
`12 of 131 (9.2%) (P < 0.003 for both, by 2-tailed t-test,
`compared with the 75% prevalence in the patients described
`herein).
`
`DISCUSSION
`
`Enterocolitis is a very rare complication of
`chrysotherapy, only 28 cases having been reported in
`the literature to date (2,3 ,6,8,9). Nineteen percent of
`these patients were Jewish, most of them Israeli.
`Different preparations of gold salts have been associ-
`
`Figure 2. Radiograph of the ileum of patient 2, showing narrowing,
`lack of haustration, and clear thickening of the ileal wall and
`surrounding soft tissue.
`
`showed ORB I loci with DRB I *0402, ORB 1 *0404,
`OQB I *0302, and DQB 1 *0402 alleles.
`Patient 3. Patient 3, a 40-year-old Jewish woman
`with a 5-year history of seronegative RA, began treatment
`with gold salts because of exacerbation of joint pain. Two
`months later, after she had received a cumulative dose of200
`mg of aurothioglucose, tenesmus, abdominal pain, and wa(cid:173)
`tery diarrhea with mucus and blood appeared. Treatment
`with gold salts was discontinued and she was admitted to our
`department.
`Physical examination was remarkable only for a pale
`rash on her cheeks and arthritis in her left knee. The ESR
`was normal, the hemoglobin level was 10 gm/dl, and the
`WBC count, platelet count, electrolyte levels, and liver and
`kidney function test findings were normal. The serum pro(cid:173)
`tein level was 4. 7 gm/dl and the albumin level was 2.8 gm/dl.
`RF was negative by latex fixation and Rose-Waaler testing.
`Sigmoidoscopy demonstrated contiguous severe inflamma(cid:173)
`tion extending 40 em from the anus. On colonic biopsies,
`acute and chronic inflammatory changes with edema and
`focal crypt loss, fibrosis, and a few eosinophils were seen.
`Small- and large-bowel radiographic studies demonstrated
`narrowing of a 30-cm section of the terminal ileum with no
`peristalsis, and thickening of descending colon mucosa with
`pseudopolyposis. Gastroduodendoscopy findings were nor(cid:173)
`mal, and biopsies from the second part of the duodenum
`showed lymphoplasmacytic infiltrates.
`Treatment with parenteral hyperalimentation, IV hy(cid:173)
`drocortisone 300 mg/day, broad-spectrum antibiotics, and
`British antilewisite (cumulative dose 1,000 mg) was initiated.
`There was a gradual improvement over a period of 3 weeks,
`
`
`
`758
`
`EVRON ET AL
`
`ated with this complication, implicating gold itself, and
`not other constituents, as the cause ofthe reaction (2).
`Most reported patients received gold salts by intra(cid:173)
`muscular injection, but oral administration has been
`implicated as well (10). The reaction is probably due to
`hypersensitivity rather than direct toxicity, since it is
`not dose related, is usually seen soon after the initia(cid:173)
`tion of gold salt therapy, and sometimes appears along
`with other manifestations of gold toxicity, such as
`rash, neutropenia, and eosinophilia (2,3,8,9).
`The disease varies from a mild, brief diarrheal
`illness (11) to a prolonged, severe disease, leading to
`death in approximately one-third of the patients (2).
`Diarrhea, usually severe and watery and sometimes
`bloody, is a consistent finding. Any part of the gastro(cid:173)
`intestinal tract may be involved. On biopsy, nonspe(cid:173)
`cific inflammation is usually seen, sometimes with
`eosinophilic infiltration. The arthritis often subsides
`completely with the onset of diarrhea (2,8, 12).
`There is no specific treatment for this compli(cid:173)
`cation. Corticosteroids are usually considered ineffec(cid:173)
`tive (2,6). Treatment with the chelating agent British
`antilewisite has been proposed (8,9, 13), but strong
`evidence for its efficacy is lacking. Response to oral
`cromolyn sodium was reported in I case of gold(cid:173)
`induced eosinophilic enterocolitis (12).
`The 4 women with gold-induced enterocolitis
`described herein were all seen at our institution since
`1986. Their average age was 43.5 years. They had all
`received intramuscular injections of aurothioglucose
`for rheumatoid arthritis. The average total dose was
`179 mg, and the mean time to onset of symptoms was
`3 months. The major symptoms were watery diarrhea
`and abdominal pain.
`On physical examination, abdominal tender(cid:173)
`ness was noted in each case. At the time of presenta(cid:173)
`tion with colitis, arthritis had subsided in 3 of the 4
`patients. Laboratory findings including ESR and blood
`counts were usually normal at presentation. Radio(cid:173)
`graphic studies and endoscopy disclosed inflammation
`of the terminal ileum in 1 patient, of the terminal ileum
`and colon in 2 patients, and of the entire small and
`large bowel in another patient. All 4 women received
`corticosteroids as the initial treatment. Patient 4 im(cid:173)
`proved with this treatment, patient 3 improved while
`also receiving British antilewisite and parenteral hyp(cid:173)
`eralimentation, and no improvement with cortico(cid:173)
`steroids was evident in patients 1 and 2.
`The chelating agent British antilewisite might
`have been beneficial in patients I and 3. No effect of
`British antilewisite on patient 4's bowel disease was
`
`noted, and patient 2's condition deteriorated and she
`died 6 weeks after gold therapy was discontinued,
`despite treatment of the enterocolitis. Three patients
`recovered within 2-4 weeks after the cessation of
`chrysotherapy. The 25% fatality rate in our group was
`close to the 26% reported in the literature (2).
`RA patients with HLA-B8/DRw3 antigens have
`been found to be at increased risk for the development
`of proteinuria during gold :salt treatment (14). One of
`our patients (patient 1) was found to also have this
`HLA phenotype, based on the records of tissue typing
`performed many years ago (6).
`In recent years, with the advances in molecular
`biology methods, analysis of HLA allelic variants has
`become possible. We therefore performed HLA DNA
`typing on the 4 patients with gold-induced enterocoli(cid:173)
`tis, according to previously described techniques (5).
`As a control, we used 2 groups of RA patients who
`were recently tested in our laboratory by the same
`methods (5,7). In both of those studies the prevalence
`of the DRB 1 *0404 allele was significantly lower than in
`the patients described herein. The remarkable finding
`that 3 of the 4 patients carried the HLA-DRB I *0404
`allele may support the contention that some patients
`have a genetic predisposition to drug allergy, specifi(cid:173)
`cally to gold salts. The advantage of identifying these
`people in advance is evident.
`The possibility ofCrohn's disease always arises
`as the main differential diagnosis of gold-induced en(cid:173)
`terocolitis. However, the distinct association of
`Crohn's disease with HLA class II DRI/DQ5 genes
`(15), together with the temporal association with gold
`salt therapy and the clinical course, makes the diag(cid:173)
`nosis of gold-induced enterocolitis more likely in our
`patients.
`The correlation between the classic RA pheno(cid:173)
`type HLA-DR4 and the DRB 1 *0404 allele is also not
`frequent. To the best of our knowledge this is the first
`time a correlation between hypersensitivity to gold
`salts and the DRB 1 *0404 allele has been described. If
`this correlation is confirmed in other RA patients with
`gold toxicity, it might enable the identification of a
`subgroup of patients in whom such therapy should be
`avoided.
`
`REFERENCES
`
`I. Stefano PM, Murray NE, William FF: Gold induced ulcerative
`proctitis. J Rheumatol 14:142-144, 1987
`2. Jackson CW, Habobi NG, Whorwell PJ, Schofield PF: Gold
`induced enterocolitis. Gut 27:452-456, 1986
`
`
`
`DRB 1 *0404 AND GOLD-INDUCED ENTEROCOLITIS
`
`759
`
`3. Sukenik S, Hirsch M, Yanai-Inbar I, Krawiec J, Freund B,
`Horowitz Y: Enterocolitis wmplicating chrysotherapy. Isr J
`Med Sci 18:1040-1043, 1982
`4. Miller AS, Dykes DD, Polesky HF: A simple salting out
`procedure for extracting DNA from human nuclear cells. Nu(cid:173)
`cleic Acids Res 161:1215-1218, 1988
`5. Gao X, Gazit E, Livneh A, Stastny P: Rheumatoid arthritis in
`Israeli Jews: shared sequences in the third hypervariable region
`of DRBI alleles are associat,~d with susceptibility. J Rheumatol
`18:801-803, 1991
`6. Geitner D, Sternfeld M, B'ecker SA, Kori M: Gold induced
`ileitis. J Clin Gastroenterol 8:184-186, 1986
`7. Gao X, Brautbar C, Gazit E, Segal R, Naparstek Y, Livneh A,
`Stastny P: A variant of HLA-DR4 determines susceptibility to
`rheumatoid arthritis in a subset of Israeli Jews. Arthritis Rheum
`34:547-551, 1991
`8. Siegman-Igra Y, Yaron M, Siletzki M, Schysman E, Gilat T:
`Colitis and death following gold therapy. Rheumatol Rehabil
`15:245-247, 1976
`
`9. Kaplinsky N, Pras M, Frankl 0: Severe enterocolitis compli(cid:173)
`cating chrysotherapy. Ann Rheum Dis 32:574-577, 1973
`10. Jarner D, Nielsen AM: Auranofin (SK F 39162) induced entero(cid:173)
`colitis in rheumatoid arthritis. Scand 1 Rheumatol 12:254-256,
`1983
`II. Kirkham B, Wedderburn L, MacFarlane DG: Gold induced
`colitis. Br J Rheumatol 28:272-273, 1989
`12. Martin OM, Goldman JA, Gilliam S, Nasrallah SM: Gold
`induced eosinophilic enterocolitis: response to oral cromolyn
`sodium. Gastroenterology 80:1567-1570, 1981
`13. Martindale E: The Extrapharmacopoeia. 29th edition. Edited by
`JEF Reynolds. London, Pharmaceutical Press, 1989
`14. Wooley PH, Griffin J, Panayi GS, Batchelor JR, Welsh KI,
`Gibson TJ: HLA-DR antigen and toxic reaction to aurothioma(cid:173)
`late and 0-penicillamine in patients with rheumatoid arthritis. N
`Engl J Med 300:300-302, 1980
`15. Yang H, Rotter JI, Toyoda H, Landers C, Tyan D, McElree
`CK, Targan SR: Ulcerative colitis: a genetically heterogeneous
`disorder defined by genetic (HLA Class II) and subclinical
`(ANCA) markers. J Clin Invest 92:1080-1084, 1993
`
`