The Proteus Theory of Rheumatoid Arthritis

Table of Contents

1. The Theory in Plain Words
2. Molecular Mimicry and the DR4 Link
3. The Antibody Findings
4. The Urine Studies, Including the One That Found No Difference
5. Starch, Vegetarian Diets and What People Report
6. Minocycline, Doxycycline and the Tetracycline Trials
7. Sulfasalazine and Its Antibacterial Side
8. Antibiotic Use and Infection Records in Large Populations
9. The Wider Microbiome and Mucosal Research
10. What Is Not Known, and Who Disagrees
11. Key Research Papers
12. Connections
13. Featured Videos

The Theory in Plain Words

The idea started with a blood test. In 1985 Ebringer's group reported in The Lancet that antibodies to Proteus mirabilis were higher in 30 people with rheumatoid arthritis being treated with gold than in 24 people with active ankylosing spondylitis, 28 with inactive ankylosing spondylitis and 41 healthy people. The comparison germ, a Klebsiella, gave high antibody levels only in the active ankylosing spondylitis group. That pairing of one disease with one germ became the group's central argument for both conditions. The ankylosing spondylitis half of the argument has its own page: Starch-Free Diet and the Ebringer Hypothesis.

Over the next three decades the authors built this into a chain of steps. In plain terms, their statements run as follows:

• The trigger. Their 2006 review proposes that sub-clinical Proteus urinary tract infections, ones a person may not notice, are the main triggering factors.
• The mistaken identity. Parts of two Proteus proteins resemble parts of human molecules, so antibodies made against the germ may also bind human tissue.
• The damage. These cross-reacting antibodies are described as cytopathic, meaning harmful to cells, which the authors say helps keep the disease going.
• The link to modern blood tests. The two Proteus sequences contain paired arginine building blocks that the enzyme peptidyl arginine deiminase can act on. The authors offer this as the explanation for the early appearance of anti-citrullinated protein antibodies in patients.

A 2010 paper lays this out as ten "Popper sequences", named after the philosopher Karl Popper, and concludes that rheumatoid arthritis "is most probably caused by Proteus upper urinary tract infections". A 2014 review, titled "Rheumatoid arthritis is caused by a Proteus urinary tract infection", describes six stages and reports raised antibodies from 16 countries (the 2010 paper counted 14). A 2017 review by the same group describes Proteus mirabilis as the main microbial culprit and molecular mimicry as the most likely mechanism.

The authors' treatment proposals are add-ons to standard care. The 2014 review argues for early anti-Proteus antibiotics as well as biological agents. The 2007 review proposes antibiotic and dietary measures together with current treatments. The 2006 review proposes antibiotics, vegetarian diets and a high intake of water and fruit juices such as cranberry juice, in addition to current treatments. The 2007 review also states that prospective long-term studies of antimicrobial measures are required to establish a benefit.

Molecular Mimicry and the DR4 Link

Molecular mimicry means that a short stretch of a germ's protein looks enough like a stretch of a human protein that the immune system may confuse the two. The theory rests on two such look-alikes.

The first involves genes. Rheumatoid arthritis is found more often in people carrying the tissue types human leukocyte antigen DR1 and some subtypes of human leukocyte antigen DR4 (HLA-DR4). A short run of amino acids in these types, written EQRRAA, marks susceptibility; the idea that such types share a common risk segment was named the shared epitope hypothesis in 1987. In 1992 Ebringer's group reported a similar six-letter stretch in the Proteus haemolysin protein, ESRRAL.

The second involves cartilage. In 1995 the same group compared a Proteus urease sequence, IRRET, with a sequence in type XI collagen, LRREI. Computer modelling showed similar shapes for both pairs. Antibodies to a peptide containing ESRRAL, to the haemolysin and to the urease were higher in people with rheumatoid arthritis than in people with ankylosing spondylitis and in healthy controls. Three control peptides and the haemolysins and urease of other bacteria showed no such rise.

The 2010 paper adds that blood from patients harmed sheep red cells coated with the self-peptides, a laboratory result. A separate group in Poland reported in 2020 that antibodies purified from patients' blood against a Proteus mirabilis O3 surface molecule (lipopolysaccharide) reacted with type I collagen more strongly than those from healthy donors, and that this cross-reaction tracked levels of anti-citrullinated protein antibodies (r = 0.7). This is a different molecule and a different collagen from the original proposal, and those authors suggest less specific antibodies were involved.

The genetic link has been examined directly in patients, with results that do not line up neatly:

• A 1992 family study found antibody levels only non-significantly higher in 42 DR4-positive patients than in 15 DR4-negative patients with active disease. Thirty people with active disease did have higher antibodies than their healthy siblings with identical tissue types.
• A 2005 study of 246 people with early inflammatory arthritis detected no association between anti-Proteus antibodies and the shared epitope.

The Antibody Findings

Raised antibodies to Proteus are the most repeated finding behind the theory, and groups outside London have reported them too.

• Bermuda and England, 1995. 34 people with active disease in Bermuda and 34 in England had raised anti-Proteus antibodies compared with local healthy controls. Escherichia coli and three normal bowel bacteria showed no such rise. The authors note that the study could not separate a link with the disease from a link with disease activity.
• Japan, 1997. Among 50 patients with rheumatoid arthritis, IgG antibodies to Proteus mirabilis were raised in those with active or probably active disease compared with ankylosing spondylitis patients and healthy controls.
• Chandigarh, India, 1997. Among 70 patients, those with active disease had higher antibody levels than controls, who were 18 people with osteoarthritis and 82 healthy people. Those with inactive disease did not differ from controls.
• Early arthritis, 2005. In 246 people with inflammatory arthritis of less than a year, IgM and IgA antibodies to Proteus were higher in the 30% who had rheumatoid-factor-positive rheumatoid arthritis. IgG did not differ, and IgM antibodies to Escherichia coli were also raised in that group. The authors state that the role of either germ requires further investigation.
• Newcastle, 1992. Among 142 patients and 121 healthy same-sex siblings, antibody levels were higher in active disease than in inactive disease or people without the disease. Over time, in 36 patients, changes in antibodies moved with changes in C-reactive protein, a blood marker of inflammation (r = 0.52). The authors wrote only that these observations "require an explanation".

Two findings complicate this picture. The 2010 paper states that the antibodies are disease specific, with none found in other conditions; a 1988 study, however, found raised Proteus antibodies in coeliac disease as well as rheumatoid arthritis. And a 1994 Newcastle study read the antibodies differently. Seventy patients with raised C-reactive protein had higher titres than 17 without it and 36 controls, but neither a snapshot nor follow-up over time found a link with clinical measures of disease activity. Its authors suggest the antibody may act as an acute phase protein, part of the body's general response to inflammation, and that its binding to Proteus may be irrelevant.

The Urine Studies, Including the One That Found No Difference

If a urinary infection drives the disease, the germ would be expected to turn up more often in patients' urine. Three studies looked for it directly and did not agree. The germ's usual role in the urinary tract is covered on the Urinary Tract Infections page.

• London, 1997. Among 89 patients, Proteus mirabilis grew from the urine of 63% of women (47 of 75) and 50% of men (7 of 14), against 32% of healthy women (38 of 119) and 11% of healthy men (13 of 115). In 20 patients, higher blood antibodies went with higher urine counts (r = +0.714).
• Dundee, 1999. Urine was infected in 25 of 76 patients (33%) and 2 of 48 matched controls (4%). Proteus mirabilis was found in 52% of the patients' infected urines, usually in numbers the authors call insignificant (below 10,000 per mL). Antibodies were raised in both blood and urine. No antigen specific to patients could be detected.
• Newcastle, 1994. The researchers were unable to isolate Proteus from the urine or faeces of patients more often than from controls, and finding the germ did not correlate with blood antibodies. Urinary antibodies were present equally often in people with and without the disease. The authors state that the study "does not support a role for Proteus" in rheumatoid arthritis.

A related record study looked at urinary infection more generally. In 120 women with rheumatoid arthritis, recurrent urinary infection was more common in the 31 who also had Sjögren's syndrome (11 of 31) than in the 89 who did not (4 of 89). The authors suggest a local lack of protective secretions or other immune mechanisms may be responsible. That abstract does not report which germs were involved.

Starch, Vegetarian Diets and What People Report

The low-starch idea comes from the ankylosing spondylitis side of Ebringer's work. A 1996 paper argued that bowel microbes depend on dietary starch and described a diet with less "bread, potatoes, cakes and pasta". It reported lower total IgA in healthy people and patients, and less inflammation and fewer symptoms in people with ankylosing spondylitis, and stated that the diet's role "requires further evaluation". That paper concerns ankylosing spondylitis and Klebsiella in the bowel, not rheumatoid arthritis and Proteus in the urinary tract.

What practitioners and patients report. Some practitioners and patients report using low-starch or starch-free eating for rheumatoid arthritis and describe less pain and stiffness. These are personal reports, not measured results. A Europe PMC title search made for this page found no trial of a low-starch diet in rheumatoid arthritis, and the abstracts of the theory's own reviews read for this page name vegetarian diets, water and fruit juices, not starch reduction.

The diet findings that do connect to Proteus come from Norway, and they concern fasting and vegetarian eating:

• Blood from 53 patients in a controlled trial of fasting followed by a one-year vegetarian diet was tested in 1995. Anti-Proteus levels fell at every time point in the vegetarian group, did not change significantly in those eating an ordinary mixed diet, and fell most in those who responded well to the diet. Antibodies to Escherichia coli hardly changed. The fall in anti-Proteus levels tracked the fall in a disease activity score.
• A second 1995 study tested urine from 22 people sent to a health farm for various reasons, who fasted for 7 to 10 days and then ate a vegan diet. Both Proteus mirabilis and Escherichia coli grew more slowly in urine collected after 8 days. After 18 days growth was still reduced, although not significantly for Escherichia coli. This was a laboratory test of urine, not a measure of infection or of arthritis.

Other diet research on the disease is covered on Diet and Rheumatoid Arthritis.

Minocycline, Doxycycline and the Tetracycline Trials

Antibiotics from the tetracycline family have been tested in rheumatoid arthritis in controlled trials. These trials are often mentioned alongside the theory, but none of the abstracts below reports measuring Proteus, so they test the drugs rather than the germ.

• Minocycline, 26 weeks, 1994. The trial gave up to 200 mg a day or placebo to 80 people with active disease. The minocycline group improved more, with a pronounced change in laboratory measures and a less impressive one in clinical measures. Side effects attributed to the drug were mainly stomach and bowel symptoms and dizziness.
• Minocycline, 48 weeks, 1995. The trial gave 200 mg a day or placebo to 219 adults with mild to moderate disease. Joint swelling improved in 54% against 39%, and tenderness in 56% against 41%. Several other outcomes did not pass the study's strict statistical threshold. No serious toxicity occurred.
• Minocycline in early disease, 1997. The trial gave 100 mg twice a day or placebo for 6 months to 46 rheumatoid-factor-positive people with disease of under a year. A 50% improvement was held by 15 of 23 (65%) against 3 of 23 (13%).
• The same trial at four years, 1999. Remission without disease-modifying drugs or steroids was found in 8 people first given minocycline and 1 first given placebo. Ten against 16 still needed disease-modifying drugs.
• Minocycline against hydroxychloroquine, 2001. Over 2 years in 60 people with early disease, a 50% improvement on the American College of Rheumatology score was reached by 60% against 33%, and mean prednisone use at 2 years was 0.81 against 3.21 mg a day.
• Doxycycline, 50 mg twice a day, 2001. A 36-week trial in 66 people found no significant difference from placebo in disease activity or joint damage.
• Intravenous doxycycline, 2001. A 16-week exploratory trial comparing it with oral azithromycin and placebo stopped early at 31 patients. It found no significant differences in its primary measures. Infusion-related events occurred in 13 people (42%), none serious.
• Doxycycline with methotrexate, 2006. In 66 people with early disease, a 50% improvement at 2 years was reached by 41.6% on 100 mg doxycycline twice a day, 38.9% on 20 mg twice a day and 12.5% on placebo, all with methotrexate.

The 2006 result bears directly on the theory. The 20 mg dose was chosen to separate antibacterial action from the drug's effect on tissue-degrading enzymes, and the similar response at both doses led its authors to suggest that the enzyme effects mattered more than the antibacterial ones.

A 2003 pooled analysis of 10 randomized trials with 535 people reported less disease activity with tetracyclines given for 3 months or more (tender joint count, standardized mean difference −0.39, 95% confidence interval −0.74 to −0.05), and more with minocycline alone (−0.69). The effect was stronger in rheumatoid-factor-positive disease of under a year. Only 3 of the trials were rated high quality, no benefit was seen on bone erosions, and the data were too thin for a detailed look at individual side effects.

Outside trials, a US study of 15,716 patients found these drugs given to 9% of patients by 18% of rheumatologists, mainly for long-standing disease that had not responded to other drugs. Side effects were reported by 17.8% of minocycline users and 11.8% of doxycycline users. About 75% were mild or moderate. A 2025 narrative review states that current guidelines do not explicitly recommend tetracyclines, citing inconsistent evidence, small samples and antimicrobial resistance. Minocycline is also a reported cause of a lupus-like illness, described on the Drug-Induced Lupus page. Standard drug options are covered on DMARDs and Biologics for Rheumatoid Arthritis.

Sulfasalazine and Its Antibacterial Side

Sulfasalazine is an established rheumatoid arthritis drug that splits into two parts: 5-aminosalicylate and sulfapyridine, an antibacterial. That has made it part of the discussion. Its benefit against placebo is reported in pooled trials:

• A 2000 Cochrane review of 6 trials and 468 patients found a standardized difference of −0.49 for tender and swollen joints and −0.42 for pain. Withdrawals for side effects were higher on the drug (odds ratio 3.0).
• A 1999 analysis of 15 trials, averaging 2 g a day over 36 weeks, reported swollen joints improved by 51% against 26% on placebo and pain by 42% against 15%. Withdrawals for side effects were 24% against 7%.

Whether its antibacterial action explains that benefit has been studied, with mixed results:

• 1987. In 26 outpatients randomly given sulfasalazine or D-penicillamine, both groups improved, but only the sulfasalazine group had falls in bowel Clostridium perfringens and Escherichia coli. The authors suggested the antibacterial effect may be involved.
• 1993. In 31 patients, Clostridium perfringens counts fell but the change had no relationship to how well people responded. The authors concluded the benefit is independent of the antibacterial effect on bowel bacteria.
• 1994. In 17 patients over 16 weeks, the drug altered the faecal flora, but the authors found the results did not substantiate an antimicrobial explanation for the benefit, while not ruling out an undetected germ.
• Laboratory, 2021. Tested on oral bacteria, sulfasalazine was active only at concentrations far above those reached in the body.
• Laboratory, 2026. Tested on Proteus and other bacteria, sulfasalazine stopped Proteus mirabilis growth at 625 µg/mL, and its sulfapyridine part was more potent (78 to 625 µg/mL). This was growth in a dish, not a test in people.

Antibiotic Use and Infection Records in Large Populations

Large health records offer another angle on infection and antibiotics. None of these studies tested Proteus, and they point in different directions.

• Sweden, 2015. In a case-control study of 6,401 people, a urinary tract infection in the 2 years before onset went with a lower risk of the disease (odds ratio 0.78, 95% confidence interval 0.68 to 0.90), as did gastroenteritis (0.71). Respiratory infections showed no link. The abstract reports urinary infections as a group, not by germ.
• United Kingdom, 2019. Among 22,677 cases and 90,013 controls, antibiotic exposure went with 60% higher odds of the disease (odds ratio 1.60), rising with the number of prescriptions.
• United Kingdom, 2020. Among 8,482 cases and 22,661 controls, more than 10 antibiotic courses in 5 years went with an adjusted odds ratio of 2.65. The authors note that people with early undiagnosed disease may simply have been more likely to come in with infections.
• United States, 2022. Among 133,125 participants in two prospective nurses' cohorts, antibiotic use showed no association with later disease (adjusted hazard ratio 1.06 for longer use, 95% confidence interval 0.42 to 1.71).
• Flares, 2019. Among 1,192 patients with flares in a United Kingdom cohort, sulphonamide and trimethoprim antibiotics, which are widely used for urinary infections, went with more flares at 29 to 90 days (incidence rate ratio 1.71). No other antibiotic group showed a link. The authors hypothesise a route through the gut or urinary microbiomes.

A 2025 critical review of antibiotics in rheumatology limited the case for them to diseases with a clear infectious cause, naming rheumatic fever, Whipple's disease and early Lyme disease, with possible early use in reactive arthritis, until robust trials support wider use.

The Wider Microbiome and Mucosal Research

Much of the wider research on germs and rheumatoid arthritis looks at the mouth, gut and lungs rather than the urinary tract. Autoantibodies appear well before the joints are affected: anti-citrullinated protein antibodies come an average of 3 to 5 years before clinical disease, which led to the "mucosal origins hypothesis", the idea that the process begins at surfaces such as the lung, gut and gums.

• Gut. Sequencing of 114 stool samples linked Prevotella copri strongly with new-onset untreated disease, and colonized mice were more sensitive to chemically induced colitis. Among 133 relatives of patients, Prevotella was enriched in the 83 in pre-clinical stages compared with the 50 controls (p = 0.04).
• Immune response. A peptide from Prevotella copri stimulated T-cell responses in 42% of people with new-onset disease. Antibody responses to the germ were rarely found in other rheumatic diseases or in healthy controls.
• Gums. In laboratory work, Porphyromonas gingivalis was unique among the oral bacteria tested in being able to citrullinate proteins, including human fibrinogen and α-enolase, through its own enzyme. See Oral Microbiome.
• Several routes. A 2024 review argues for distinct "mucosal endotypes", in which different body surfaces and germs may drive the disease by different mechanisms in different people.

This work shares the general idea of the Proteus theory, that a microbe at a body surface may help start an autoimmune process. It differs in the germs and sites it names. The 2018 and 2024 mucosal reviews do not mention Proteus in their abstracts.

What Is Not Known, and Who Disagrees

The theory's authors regard the case as made; their 2014 title says the disease "is caused by" the infection. Others have published different readings. The questions below remain open in the studies read for this page.

• Cause or consequence. Antibodies to a germ can be raised because of the germ, or because of the general immune activity of the disease. The 1994 Newcastle authors favour the second reading; the 1995 Bermuda and England authors say their study could not separate the two.
• Whether the germ is really more common. Two urine studies found more Proteus in patients and one did not.
• Specificity. The 2010 paper describes the antibodies as disease specific. Raised levels were also reported in coeliac disease in 1988, and Escherichia coli IgM antibodies were raised alongside them in early disease in 2005.
• The DR4 link. The sequence similarity is reported, but antibody levels did not differ significantly by DR4 status in 1992 or by shared epitope in 2005.
• Treatment aimed at Proteus. The 2007 review states that prospective studies are required. The abstracts read for this page include no trial that selected patients by Proteus status or tested whether clearing the germ changes the disease.
• How the antibiotics work. Tetracycline benefit may come from effects other than killing bacteria, as the 2006 doxycycline authors suggest. Sulfasalazine studies disagree on the role of its antibacterial action.
• Diet. Starch reduction in rheumatoid arthritis rests on personal reports. The Norwegian vegetarian findings involve a small group and antibody levels, and cannot show that Proteus was the reason people improved.

Much of the wider field now studies the gut, gums and lungs. The Proteus theory has not been settled in either direction by the studies reported here.

Key Research Papers

1. Ebringer A, Ptaszynska T, Corbett M, Wilson C, Macafee Y, Avakian H, Baron P, James DC. Antibodies to proteus in rheumatoid arthritis. Lancet. 1985;2(8450):305-7. PubMed PMID: 2862470
2. Ebringer A, Rashid T. Rheumatoid arthritis is an autoimmune disease triggered by Proteus urinary tract infection. Clin Dev Immunol. 2006;13(1):41-8. PubMed PMID: 16603443
3. Rashid T, Ebringer A. Rheumatoid arthritis is linked to Proteus--the evidence. Clin Rheumatol. 2007;26(7):1036-43. PubMed PMID: 17206398
4. Ebringer A, Rashid T, Wilson C. Rheumatoid arthritis, Proteus, anti-CCP antibodies and Karl Popper. Autoimmun Rev. 2010;9(4):216-23. PubMed PMID: 19895906
5. Ebringer A, Rashid T. Rheumatoid arthritis is caused by a Proteus urinary tract infection. APMIS. 2014;122(5):363-8. PubMed PMID: 23992372
6. Rashid T, Ebringer A, Wilson C. The link between Proteus mirabilis, environmental factors and autoantibodies in rheumatoid arthritis. Clin Exp Rheumatol. 2017;35(5):865-871. PubMed PMID: 28516867
7. Gregersen PK, Silver J, Winchester RJ. The shared epitope hypothesis. An approach to understanding the molecular genetics of susceptibility to rheumatoid arthritis. Arthritis Rheum. 1987;30(11):1205-13. PubMed PMID: 2446635
8. Ebringer A, Cunningham P, Ahmadi K, Wrigglesworth J, Hosseini R, Wilson C. Sequence similarity between HLA-DR1 and DR4 subtypes associated with rheumatoid arthritis and proteus/serratia membrane haemolysins. Ann Rheum Dis. 1992;51(11):1245-6. PubMed PMID: 1466602
9. Wilson C, Ebringer A, Ahmadi K, Wrigglesworth J, Tiwana H, Fielder M, Binder A, Ettelaie C, Cunningham P, Joannou C, et al. Shared amino acid sequences between major histocompatibility complex class II glycoproteins, type XI collagen and Proteus mirabilis in rheumatoid arthritis. Ann Rheum Dis. 1995;54(3):216-20. PubMed PMID: 7748019
10. Durlik-Popińska K, Żarnowiec P, Lechowicz Ł, Gawęda J, Kaca W. Antibodies Isolated from Rheumatoid Arthritis Patients against Lysine-Containing Proteus mirabilis O3 (S1959) Lipopolysaccharide May React with Collagen Type I. Int J Mol Sci. 2020;21(24):9635. PubMed PMID: 33348817
11. Rogers P, Hassan J, Bresnihan B, Feighery C, Whelan A. Antibodies to Proteus in rheumatoid arthritis. Br J Rheumatol. 1988;27 Suppl 2:90-4. PubMed PMID: 3042079
12. Deighton CM, Gray JW, Bint AJ, Walker DJ. Anti-Proteus antibodies in rheumatoid arthritis same-sexed sibships. Br J Rheumatol. 1992;31(4):241-5. PubMed PMID: 1555039
13. McDonagh J, Gray J, Sykes H, Walker DJ, Bint AJ, Deighton CM. Anti-Proteus antibodies and Proteus organisms in rheumatoid arthritis: a clinical study. Br J Rheumatol. 1994;33(1):32-5. PubMed PMID: 8162454
14. Subair H, Tiwana H, Fielder M, Binder A, Cunningham K, Ebringer A, Wilson C, Hudson MJ. Elevation in anti-Proteus antibodies in patients with rheumatoid arthritis from Bermuda and England. J Rheumatol. 1995;22(10):1825-8. PubMed PMID: 8991977
15. Tani Y, Tiwana H, Hukuda S, Nishioka J, Fielder M, Wilson C, Bansal S, Ebringer A. Antibodies to Klebsiella, Proteus, and HLA-B27 peptides in Japanese patients with ankylosing spondylitis and rheumatoid arthritis. J Rheumatol. 1997;24(1):109-14. PubMed PMID: 9002020
16. Wanchu A, Deodhar SD, Sharma M, Gupta V, Bambery P, Sud A. Elevated levels of anti-proteus antibodies in patients with active rheumatoid arthritis. Indian J Med Res. 1997;105:39-42. PubMed PMID: 9029834
17. Newkirk MM, Goldbach-Mansky R, Senior BW, Klippel J, Schumacher HR Jr, El-Gabalawy HS. Elevated levels of IgM and IgA antibodies to Proteus mirabilis and IgM antibodies to Escherichia coli are associated with early rheumatoid factor (RF)-positive rheumatoid arthritis. Rheumatology (Oxford). 2005;44(11):1433-41. PubMed PMID: 16091399
18. Wilson C, Thakore A, Isenberg D, Ebringer A. Correlation between anti-Proteus antibodies and isolation rates of P. mirabilis in rheumatoid arthritis. Rheumatol Int. 1997;16(5):187-9. PubMed PMID: 9032817
19. Senior BW, Anderson GA, Morley KD, Kerr MA. Evidence that patients with rheumatoid arthritis have asymptomatic 'non-significant' Proteus mirabilis bacteriuria more frequently than healthy controls. J Infect. 1999;38(2):99-106. PubMed PMID: 10342649
20. Tishler M, Caspi D, Almog Y, Segal R, Yaron M. Increased incidence of urinary tract infection in patients with rheumatoid arthritis and secondary Sjögren's syndrome. Ann Rheum Dis. 1992;51(5):604-6. PubMed PMID: 1616323
21. Ebringer A, Wilson C. The use of a low starch diet in the treatment of patients suffering from ankylosing spondylitis. Clin Rheumatol. 1996;15 Suppl 1:62-66. PubMed PMID: 8835506
22. Kjeldsen-Kragh J, Rashid T, Dybwad A, Sioud M, Haugen M, Førre O, Ebringer A. Decrease in anti-Proteus mirabilis but not anti-Escherichia coli antibody levels in rheumatoid arthritis patients treated with fasting and a one year vegetarian diet. Ann Rheum Dis. 1995;54(3):221-4. PubMed PMID: 7748020
23. Kjeldsen-Kragh J, Kvaavik E, Bottolfs M, Lingaas E. Inhibition of growth of Proteus mirabilis and Escherichia coli in urine in response to fasting and vegetarian diet. APMIS. 1995;103(11):818-22. PubMed PMID: 8546847
24. Kloppenburg M, Breedveld FC, Terwiel JP, Mallee C, Dijkmans BA. Minocycline in active rheumatoid arthritis. A double-blind, placebo-controlled trial. Arthritis Rheum. 1994;37(5):629-36. PubMed PMID: 8185689
25. Tilley BC, Alarcón GS, Heyse SP, Trentham DE, Neuner R, Kaplan DA, Clegg DO, Leisen JC, Buckley L, Cooper SM, Duncan H, Pillemer SR, Tuttleman M, Fowler SE. Minocycline in rheumatoid arthritis. A 48-week, double-blind, placebo-controlled trial. MIRA Trial Group. Ann Intern Med. 1995;122(2):81-9. PubMed PMID: 7993000
26. O'Dell JR, Haire CE, Palmer W, Drymalski W, Wees S, Blakely K, Churchill M, Eckhoff PJ, Weaver A, Doud D, Erikson N, Dietz F, Olson R, Maloley P, Klassen LW, Moore GF. Treatment of early rheumatoid arthritis with minocycline or placebo: results of a randomized, double-blind, placebo-controlled trial. Arthritis Rheum. 1997;40(5):842-8. PubMed PMID: 9153544
27. O'Dell JR, Paulsen G, Haire CE, Blakely K, Palmer W, Wees S, Eckhoff PJ, Klassen LW, Churchill M, Doud D, Weaver A, Moore GF. Treatment of early seropositive rheumatoid arthritis with minocycline: four-year followup of a double-blind, placebo-controlled trial. Arthritis Rheum. 1999;42(8):1691-5. PubMed PMID: 10446869
28. O'Dell JR, Blakely KW, Mallek JA, Eckhoff PJ, Leff RD, Wees SJ, Sems KM, Fernandez AM, Palmer WR, Klassen LW, Paulsen GA, Haire CE, Moore GF. Treatment of early seropositive rheumatoid arthritis: a two-year, double-blind comparison of minocycline and hydroxychloroquine. Arthritis Rheum. 2001;44(10):2235-41. PubMed PMID: 11665963
29. van der Laan W, Molenaar E, Ronday K, Verheijen J, Breedveld F, Greenwald R, Dijkmans B, TeKoppele J. Lack of effect of doxycycline on disease activity and joint damage in patients with rheumatoid arthritis. A double blind, placebo controlled trial. J Rheumatol. 2001;28(9):1967-74. PubMed PMID: 11550961
30. St Clair EW, Wilkinson WE, Pisetsky DS, Sexton DJ, Drew R, Kraus VB, Greenwald RA. The effects of intravenous doxycycline therapy for rheumatoid arthritis: a randomized, double-blind, placebo-controlled trial. Arthritis Rheum. 2001;44(5):1043-7. PubMed PMID: 11357896
31. O'Dell JR, Elliott JR, Mallek JA, Mikuls TR, Weaver CA, Glickstein S, Blakely KM, Hausch R, Leff RD. Treatment of early seropositive rheumatoid arthritis: doxycycline plus methotrexate versus methotrexate alone. Arthritis Rheum. 2006;54(2):621-7. PubMed PMID: 16447240
32. Stone M, Fortin PR, Pacheco-Tena C, Inman RD. Should tetracycline treatment be used more extensively for rheumatoid arthritis? Metaanalysis demonstrates clinical benefit with reduction in disease activity. J Rheumatol. 2003;30(10):2112-22. PubMed PMID: 14528503
33. Smith CJ, Sayles H, Mikuls TR, Michaud K. Minocycline and doxycycline therapy in community patients with rheumatoid arthritis: prescribing patterns, patient-level determinants of use, and patient-reported side effects. Arthritis Res Ther. 2011;13(5):R168. PubMed PMID: 22008667
34. Radić M, Belančić A, Đogaš H, Vučković M, Gelemanović A, Faour A, Vlak I, Radić J. Tetracyclines in Rheumatoid Arthritis: Dual Anti-Inflammatory and Immunomodulatory Roles, Effectiveness, and Safety Insights. Antibiotics (Basel). 2025;14(1):65. PubMed PMID: 39858351
35. Suarez-Almazor ME, Belseck E, Shea B, Wells G, Tugwell P. Sulfasalazine for rheumatoid arthritis. Cochrane Database Syst Rev. 2000;1998(2):CD000958. PubMed PMID: 10796400
36. Weinblatt ME, Reda D, Henderson W, Giobbie-Hurder A, Williams D, Diani A, Docsa S. Sulfasalazine treatment for rheumatoid arthritis: a metaanalysis of 15 randomized trials. J Rheumatol. 1999;26(10):2123-30. PubMed PMID: 10529127
37. Neumann VC, Shinebaum R, Cooke EM, Wright V. Effects of sulphasalazine on faecal flora in patients with rheumatoid arthritis: a comparison with penicillamine. Br J Rheumatol. 1987;26(5):334-7. PubMed PMID: 2889501
38. Bradley SM, Neumann VC, Barr K, Troughton PR, Astbury C, Bird HA, Gooi HC, Wright V. Sequential study of bacterial antibody levels and faecal flora in rheumatoid arthritis patients taking sulphasalazine. Br J Rheumatol. 1993;32(8):683-8. PubMed PMID: 8102304
39. Kanerud L, Scheynius A, Nord CE, Hafström I. Effect of sulphasalazine on gastrointestinal microflora and on mucosal heat shock protein expression in patients with rheumatoid arthritis. Br J Rheumatol. 1994;33(11):1039-48. PubMed PMID: 7981991
40. Kussmann M, Obermueller M, Spettel K, Winkler S, Aletaha D. In vitro evaluation of disease-modifying antirheumatic drugs against rheumatoid arthritis associated pathogens of the oral microflora. RMD Open. 2021;7(3):e001737. PubMed PMID: 34588273
41. Cock IE, Whitehouse MW. Re-evaluating the antibacterial properties of DMARD and pro-drug sulphasalazine against autoimmune bacterial triggers after eighty years. Inflammopharmacology. 2026;34(3):1811-1821. PubMed PMID: 41697549
42. Sandberg ME, Bengtsson C, Klareskog L, Alfredsson L, Saevarsdottir S. Recent infections are associated with decreased risk of rheumatoid arthritis: a population-based case-control study. Ann Rheum Dis. 2015;74(5):904-7. PubMed PMID: 25656683
43. Sultan AA, Mallen C, Muller S, Hider S, Scott I, Helliwell T, Hall LJ. Antibiotic use and the risk of rheumatoid arthritis: a population-based case-control study. BMC Med. 2019;17(1):154. PubMed PMID: 31387605
44. Armstrong D, Dregan A, Ashworth M, White P, McGee C, de Lusignan S. Influence of prior antibiotic use on risk of rheumatoid arthritis: case control study in general practice. Rheumatology (Oxford). 2020;59(6):1281-1287. PubMed PMID: 31580454
45. Liu Y, Xia B, Li B, Sparks JA, He Q, Yuan J, Zhou B. No association between recent antibiotic use and risk of rheumatoid arthritis: results from two prospective cohort studies. Expert Opin Drug Saf. 2022;21(1):121-126. PubMed PMID: 34403275
46. Nagra NS, Robinson DE, Douglas I, Delmestri A, Dakin SG, Snelling SJB, Carr AJ, Prieto-Alhambra D. Antibiotic treatment and flares of rheumatoid arthritis: a self-controlled case series study analysis using CPRD GOLD. Sci Rep. 2019;9(1):8941. PubMed PMID: 31222078
47. Falagas ME, Stathopoulos P, Kontogiannis DS, Tzvetanova ID. Antibiotics for Rheumatologic Diseases: A Critical Review. Int J Mol Sci. 2025;26(21):10527. PubMed PMID: 41226564
48. Holers VM, Demoruelle MK, Kuhn KA, Buckner JH, Robinson WH, Okamoto Y, Norris JM, Deane KD. Rheumatoid arthritis and the mucosal origins hypothesis: protection turns to destruction. Nat Rev Rheumatol. 2018;14(9):542-557. PubMed PMID: 30111803
49. Holers VM, Demoruelle KM, Buckner JH, James EA, Firestein GS, Robinson WH, Steere AC, Zhang F, Norris JM, Kuhn KA, Deane KD. Distinct mucosal endotypes as initiators and drivers of rheumatoid arthritis. Nat Rev Rheumatol. 2024;20(10):601-613. PubMed PMID: 39251771
50. Scher JU, Sczesnak A, Longman RS, Segata N, Ubeda C, Bielski C, Rostron T, Cerundolo V, Pamer EG, Abramson SB, Huttenhower C, Littman DR. Expansion of intestinal Prevotella copri correlates with enhanced susceptibility to arthritis. Elife. 2013;2:e01202. PubMed PMID: 24192039
51. Alpizar-Rodriguez D, Lesker TR, Gronow A, Gilbert B, Raemy E, Lamacchia C, Gabay C, Finckh A, Strowig T. Prevotella copri in individuals at risk for rheumatoid arthritis. Ann Rheum Dis. 2019;78(5):590-593. PubMed PMID: 30760471
52. Pianta A, Arvikar S, Strle K, Drouin EE, Wang Q, Costello CE, Steere AC. Evidence of the Immune Relevance of Prevotella copri, a Gut Microbe, in Patients With Rheumatoid Arthritis. Arthritis Rheumatol. 2017;69(5):964-975. PubMed PMID: 27863183
53. Wegner N, Wait R, Sroka A, Eick S, Nguyen KA, Lundberg K, Kinloch A, Culshaw S, Potempa J, Venables PJ. Peptidylarginine deiminase from Porphyromonas gingivalis citrullinates human fibrinogen and α-enolase: implications for autoimmunity in rheumatoid arthritis. Arthritis Rheum. 2010;62(9):2662-72. PubMed PMID: 20506214

PubMed Topic Searches

1. PubMed: Proteus mirabilis and rheumatoid arthritis
2. PubMed: minocycline in rheumatoid arthritis trials
3. PubMed: sulfasalazine and bacteria in rheumatoid arthritis
4. PubMed: the microbiome and rheumatoid arthritis

Connections

• Rheumatoid Arthritis — the main page on the disease
• Proteus mirabilis — the bacterium, its infections and its treatment
• Starch-Free Diet and the Ebringer Hypothesis — the Klebsiella version of the same theory
• DMARDs and Biologics for Rheumatoid Arthritis — the standard drug treatments
• Diet and Rheumatoid Arthritis — the wider diet research
• Urinary Tract Infections — the infections themselves
• Oral Microbiome — mouth bacteria and whole-body research
• Drug-Induced Lupus — including the minocycline reports
• Autoimmunity — an interactive animation of the process

Featured Videos