Metronidazole As a Disease Modifying Agent in the Ann Rheum Dis: First Published As 10.1136/Ard.51.6.758 on 1 June 1992

758 Annals ofthe Rheumatic Diseases 1992; 51: 758-760 Double blind, placebo controlled study of

metronidazole as a disease modifying agent in the Ann Rheum Dis: first published as 10.1136/ard.51.6.758 on 1 June 1992. Downloaded from treatment of rheumatoid arthritis

D A S Marshall, J A Hunter, H A Capell

Abstract 400 mg three times a day for a further 16 weeks Anecdotal reports suggest that metronidazole if tolerated. Randomisation was double blind may have disease modifying activity in the and did not require any changes to concurrent treatment of rheumatoid arthritis. To assess drug treatment. Patients receiving treatment possible beneficial effects a double blind, with systemic corticosteroids or who had received comparative trial of metronidazole and other disease modifying antirheumatic drugs placebo was performed. Fifty patients with within the preceding three months were active rheumatoid arthritis were randomly excluded. Patients with hepatic or renal insuffi- allocated to receive active drug (n=24) or ciency were not included and nor were patients placebo (n=26) and reviewed at weeks 0, 1, 4, unable to undertake to abstain from all but very 8, 12, 16, and 24. Detailed assessment of drug modest alcohol consumption. Table 1 gives safety, biochemical and haematological details of the patient groups. Informed consent parameters, and efficacy was made at these was obtained from all patients before entry into dates. Dose regimen was 400 mg twice daily the study and permission for the study given by from weeks 0 to eight, increasing to 400 mg the local ethical committees. three times a day from weeks nine to 24 Clinical assessments of duration of morning providedthat no adverse effects were recorded. stiffness, articular index,5 grip strength, pain Most patients were unable to tolerate score (on visual analogue charts), and drug metronidazole because of side effects or lack adverse effects were made at weeks 0, 1, 4, 8, of efficacy, with only five (21%) continuing to 16, and 24. Laboratory tests (full blood count, take the drug at 24 weeks. For those patients urea, electrolytes, liver function tests, erythro- attaining 12 weeks of treatment an overall cyte sedimentation rate, rheumatoid factor titre, improvement in articular index and morning antinuclear factor titre, and C reactive protein) stiffness was found. No improvement in were performed at the same time intervals. laboratory indices of disease activity was Because of reports of peripheral neuropathy as a

seen, however. result of metronidazole treatment, sensory http://ard.bmj.com/ In this study metronidazole did not have function, as determined by light touch, vibration disease modifying properties and was un- sense, and position sense, was tested at each acceptably toxic. clinical assessment. Withdrawal from treatment due to inadequate effect or adverse effects was recorded.

The nitroimidazole group of drugs includes on September 28, 2021 by guest. Protected copyright. some with known disease modifying activity in STATISTICAL ANALYSIS the treatment of rheumatoid arthritis. Leva- Between-group comparisons were made with misole and clotrimazole have both been shown the Mann-Whitney U test and within-group to affect clinical and laboratory indices of comparisons using Wilcoxon matched pairs. disease activity, but toxicity has precluded their Because of the large number of patient with- widespread use in practice.' Occasional reports drawals before completion of the study at 24 have suggested that metronidazole, another weeks, statistical analysis was performed only nitroimidazole with widespread clinical usage as on the differences within or between groups an antibiotic and antiparasitic agent, might also from weeks 0 to 12. Centre for be effective as a second line treatment for Rheumatic Diseases, rheumatoid arthritis.2" This study therefore Glasgow Royal aimed, by means of a placebo controlled, double Results Infirmary, Glasgow, United Kingdom blind trial, at evaluating the effects of metro- Both groups of patients were well matched for D A S Marshall nidazole on the clinical and laboratory indices disease duration, age, sex and disease activity H A Capell of disease activity in rheumatoid arthritis. (tables 1 and 3). At outset there was no signi- Gartnavel General ficant difference in any parameter between Hospital, Glasgow, United Kingdom J A Hunter Patients and methods Fifty patients with clinical signs of active Table I Demographic details of patients. Mean value Correspondence to: (range) is given Dr H A Capell, rheumatoid arthritis and pain or stiffness Centre for Placebo Rheumatic Diseases, inadequately controlled by symptom relieving Metronidazole Glasgow Royal Infirmary, drugs alone were randomly allocated to receive (n=24) (n=26) Glasgow G4 OSF, treatment with either metronidazole 400 mg Age (years) 58 (24-77) 59 (38-81) United Kingdom. Disease duration (years) 9 (1-36) 15 (2-34) Accepted for publication twice daily or identical matching placebo tablets Women/men 18/6 20/6 15 October 1991 for an initial period of eight weeks, increasing to Use ofmetronidazole in the treatment ofrheumatoid arthritis 759

the two groups (Mann-Whitney U test). As can adverse drug reactions. The 'other' causes for be seen from the withdrawal graph (fig 1), only withdrawal included headache, depression, sore 5/24 (21%) and 6/26 (23%) continued to receive throat, colitis, and dyspepsia. drug and placebo respectively until trial Figure 2 shows the variation in clinical and Ann Rheum Dis: first published as 10.1136/ard.51.6.758 on 1 June 1992. Downloaded from completion at week 24. Table 2 gives details of laboratory parameters (expressed as median values) during the study. A significant improvement in articular index and pain score between 1004 Placebo weeks 0 and 12 (AO-12) was seen in the group Metronidazole receiving metronidazole compared with those 80- receiving placebo (p=0-02 and 0-025 respectively). Otherwise, no differences in AO-12 i6 60- between drug and placebo were evident (table 3; Mann-Whitney U test). Within-group analysis (Wilcoxon matched pairs) showed significant = 40- improvement AO-12 in the articular index (p=0 05) for patients receiving metronidazole 20- and significant deterioration AO-12 in both C reactive protein (p=0 045) and visual analogue 0 pain score (p=0 042) for those patients receiving 0 10 20 30 placebo. Weeks Figure 1 Patient withdrawalsfrom study. Discussion This study illustrates the problems of placebo controlled studies of drugs in the treatment of Table 2 Reasons for withdrawal rheumatoid arthritis as most patients withdraw Metronidazole Placebo owing to active disease (indeed, significant (n=24) (n = 26) deterioration was found in some parameters in Lack of efficacy 6 10 the placebo group). Thus, analysis of drug Nausea 7 2 Rash 1 3 effects becomes increasingly difficult owing to Peripheral neuropathy 2 0 the small number of patients remaining in the Other 3 5 Defaulted 1 3 trial. Furthermore, the side effects of metro- Dizziness 1 1 nidazole were found to be intolerable in 12/24 Altered taste sensation 1 0 Dyspnoea 0 1 (50%) patients. Most previous reports of metronidazole as a disease modifying agent in rheu- Total withdrawals 19 (790%o) 20 (77%) matoid arthritis are anecdotal, but its use in http://ard.bmj.com/ *p = 0 008 (improvement)-Wilcoxon = 16 - *p 0.05 (improvement)-Wilcoxon x 144 -|-- Placebo (P) ._ 12 - | Metronidazole (M)

10 - 0 8 .2 on September 28, 2021 by guest. Protected copyright. : 6- 4A +- 0 10 20 30 30 = 0-045 (deterioration)-Wilcoxon I = ,40 701 I*P 0-042 (improvement)-Wilcoxor E C 60 ., 30 0 0. 50- a 20 a-'i6 co_ 4u 10+i 0 10 20 30 tI. 0 10 20 300 c 120 50

(D 80 c E ,Z;60 E 40, co 40. cr C cn 'E 20 0 2 30 0 20 0 10 20 30 No of Weeks Weeks patients: M24 M20 M11 M5 M24 M20 Mul Ms P26 P22 P14 P6 P26 P22 P14 P6

Figure 2 Clinical and laboratory variables (expressed as median values). 760 Marshall, Hunter, Capell

Table 3 Comparison of drug and placebo at weeks 0 and 12. Results are given as there is, however, no known correct dose of median (ranges) metronidazole in the treatment of rheumatoid Metronidazole Placebo Mann-Whitney arthritis. Despite their high dose, side effects AO-12 were not reported as severe, and good subjective Ann Rheum Dis: first published as 10.1136/ard.51.6.758 on 1 June 1992. Downloaded from Week 0 Week 12 Week 0 Week 12 (p value) clinical results were obtained with improvement Haemoglobin (g/l) 122 129 129 124 0 01 in up to 93% of patients. No objective clinical or (72-155) (95-143) (102-148) (78-148) Platelets (x 109/1) 362 363 374 403 NS laboratory data have been published to sub- (205-827) (218-533) (258-601) (271-772) stantiate these findings. A clinical observation ESR' (mm/h) 40 37 35 42 NS (1-113) (1-115) (10-109) (10-96) from these studies is of a Herxheimer reaction CRP' (mg/i) 21 16 24 38 NS occurring in up to 85% of patients and being (10-135) (10-69) (10-140) (10-98) AlI 14 75 85 95 002 associated with a good clinical response to the (0-42) (0-21) (0-23) (0-24) drug. None ofour patients manifested symptoms Morning stiffness (min) 90 60 60 90 NS suggesting a Herxheimer reaction. (0-720) (0-720) (0-720) (0-720) The only previously published placebo con- Grip: leftt 72 64 73 70 NS (36-160) (36-120) (40-156) (40-120) trolled study of metronidazole in rheumatoid Grip: rightt 80 84 74 78 NS arthritis was of 20 patients followed up for six (40-1 16) (44-134) (38-120) (42-105) Pain scoret 62 47 58 5 59 0 025 weeks, and this found no differences in clinical (3-85) (0-76) (18-90) (30-100) or laboratory indices of disease activity. Our 'ESR=erythrocyte sedimentation rate; CRP=C reactive protein; Al=articular index. study followed up patients for a longer period tlOO mm visual analogue chart. and failed to identify a useful role for metronidazole in rheumatoid arthritis.

We thank Sister A Thompson for metrological assessment, Mrs parasitic infestations, bacterial infections, and D McKnight for statistical analysis, and May and Baker inflammatory bowel disease is well documented. Pharmaceuticals for providing metronidazole and placebo. The side effect profile of the drug is also well reported. In this study nausea and vomiting 1 El-Ghobarey A F, Maurikakis M E, McLeod M, et al. affected 7/24 (29%) with peripheral neuropathy Clinical and laboratory studies of levamisole in patients with rheumatoid arthritis. Q J Med 1978; 187: 385-400. occurring in 2/24 (8%) patients, though this 2 Harkness J A L, Griffin A J, Heinrich I, Gibson T, Grahame reversed on stopping the drug. A study of R. A double-blind comparative study of metronidazole and placebo in rheumatoid arthritis. Rheumatol Rehabil 1982; patients with Crohn's disease receiving metro- 21: 231-34. nidazole suggested peripheral neuropathy 3 Pybus P K. Metronidazole in rheumatoid arthritis. S AfrMed 7 1984; 65: 454-6. occurred in 11 of 13, though only six of the 11 4 Pybus P K. Metronidazole in rheumatoid arthritis. S AfrMed were symptomatic.6 Peripheral neuropathy has J 1985; 67: 1039-40. 5 Ritchie D M, Boyle J A, McInnes J M, et al. Clinical studies also been reported in amoebiasis treated with with an articular index for the assessment ofjoint tenderness high dose metronidazole.7 in patients with rheumatoid arthritis. Q J Med 1968; 37: 393-406. Our dose of 400 mg twice daily increasing to 6 Duffy L F, Daum F, Fisher S E, et al. Peripheral neuropathy three times a day is less than that used by some in Crohn's patients treated with metronidazole. Gastro-

enterology 1985; 88: 681-4. http://ard.bmj.com/ authors, who advocate 2 g at night for two 7 World Health Organisation. Amoebiasis and its control. Bull successive nights each week for six weeks4; World Health Organ 1985; 63: 417-26. on September 28, 2021 by guest. Protected copyright.