Chronic osteomyelitis as a risk factor for development of rheumatoid
arthritis: A nationwide, population-based, cohort study
Mao-Wang Ho1, Chun-Hung Tseng2,3, Jiunn-Horng Chen3,4, Joung-Liang Lan3,4, Chien-Chung Huang4, Chih-Hsin Muo3,5, Chung-Yi Hsu2,6, Gregory J. Tsay7
Corresponding author:
Jiunn-Horng Chen, MD, PhD Associate Professor
School of Medicine, China Medical University, Taiwan Phone: 886-4-2235-0252
Fax: 886-4-2233-1913 Email: d91842001@ntu.edu.tw Word count: Abstract-248; Text: 2,979
Keywords: rheumatoid arthritis, chronic osteomyelitis
1 Division of Infectious Disease, Department of Internal Medicine, China Medical University Hospital, Taichung, Taiwan
2 Department of Neurology, China Medical University Hospital, Taichung, Taiwan 3 School of Medicine, China Medical University, Taichung, Taiwan
4 Division of Immunology and Rheumatology, Department of Internal Medicine, China Medical University Hospital, Taichung, Taiwan
5Management Office for Health Data, China Medical University Hospital, Taichung, Taiwan
6Graduate Institute of Clinical Medical Science, China Medical University, Taichung, Taiwan
7 Institute of Microbiology & Immunology, Chung-Shen Medical University, Taichung, Taiwan
Financial Support: This work was supported by the grants from the Department of
Health, Executive Yuan, Taiwan (Grant DOH 97-HP-1101, 2008, Grant NSC97-2314-B-039-010-MY3, Grant NSC100-2314-B-039-012, Grant NSC101-2314-B-039-020, Grant MOST103-2314-B-039-020), China Medical University Hospital (Grant 1MS1, Grant DMR-97-059, Grant DMR-101-010 and Grant DMR-102-012), Taiwan Ministry of Health and Welfare Clinical Trial and Research Center of Excellence (Grant MOHW104-TDU-B-212-113002), Academia Sinica Taiwan Biobank, Stroke Biosignature Project (Grant BM104010092), NRPB Stroke Clinical Trial Consortium (MOST 103-2325-B-039 -006), Tseng-Lien Lin Foundation, Taichung, Taiwan, Taiwan Brain Disease Foundation, Taipei, Taiwan, and Katsuzo and Kiyo Aoshima Memorial Fun.
Acknowledgements: The authors thank the National Health Research Institute in
Taiwan for providing insurance claims data for analyses.
Disclosure of Competing Interest:
Disclosure of Mao-Wang Ho: None Disclosure of Chun-Hung Tseng: None Disclosure of Jiunn-Horng Chen: None Disclosure of Joung-Liang Lan: None Disclosure of Chien-Chung Huang: None Disclosure of Chih-Hsin Muo: None Disclosure of Chung-Yi Hsu: None Disclosure of Gregory J. Tsay: None ABSTRACT
An association between occult infection and the development of rheumatoid arthritis (RA) has been suggested. This study aimed to determine if patients with chronic osteomyelitis (COM) are at increased risk of developing RA.
Methods
A national insurance claim dataset of 22 million enrollees in Taiwan was used to identify 21,105 hospital inpatients with COM and 84,420 reference subjects matched by sex, age and index date of diagnosis with a mean of 5.12 years of follow-up from 2000 to 2011. The risk of RA development was analyzed using Cox proportional hazards modeling.
Results
The mean age of hospital inpatients with COM was 55.8 ± 19.4 years. The incidence of RA was 5.43 per 104 person-years in the case cohort, which was more than two-fold higher than that of 2.20 per 104 person-years in the reference cohort. After adjustment, the hazard ratio (HR) was 2.21 (95% confidence interval, 1.51–3.24). The HR was greatest in the youngest age group (<45 years, HR [95% confidence interval] = 9.08 [3.22–25.6]; 45–64 years, 1.76 [1.01–3.06]; ≥65 years, 1.68 [0.88–3.24]). In addition, HR was greatest in inpatients with more severe COM (HR [95% confidence interval] = 0.72 [0.40–1.30] and 11.2 [6.63–18.9] for patients with ≤1 or >2 hospitalization due to recurrent osteomyelitis every two follow-up years, respectively).
Conclusion
This is the first report linking COM to risk of incident RA. Patients of a younger age and with frequently recurrent COM had a greater increase in RA risk.
Background
Rheumatoid arthritis (RA) is a disease of unknown etiology. It is characterized by synovial hyperplasia that is prompted by unknown initiating events involving tissue
injury and infection. Genetic and predisposing risk factors combined with unclear random life events may influence the development of RA . Smoking is the best-described environmental risk factor for RA , and there is a strong association between smoking and increased levels of rheumatoid factor and anti-citrullinated protein antibody (ACPA) in individuals with pre-clinical RA . Exposure to microorganisms that populate the skin, oral cavity, lung and intestinal tract may contribute to the initiation of RA . Periodontitis is independently associated with established seropositive RA , and Porphyromonas gingivalis in periodontitis may initiate RA development .
Chronic osteomyelitis (COM) is a severe, persistent infection with inflammation of the bone (involving marrow, cortex, periosteum and surrounding soft tissue) that recurs for at least one month . COM is characterized by draining sinus tracts and sequestra, involucra or new bone formation around necrotic tissues in the long bone. Pathogens of COM spread through hematogenous seeding, contiguous extension from adjacent structures (soft tissue and joint) or direct inoculation at the time of trauma or surgery. Vascular insufficiency and peripheral neuropathy in the setting of diabetes predisposes to infected neuropathic ulcers, which is the most common type of contiguous spread of COM . The most common pathogens of COM include
Staphylococcus aureus, Gram-negative bacilli, anaerobic bacteria and Mycobacterium
. The incidence of COM is increasing due to the increasing prevalence of predisposing conditions such as diabetes mellitus (DM) and peripheral vascular disease . Vertebral osteomyelitis, generally involving the lower thoracic or lumbar spine, is most frequently of the hematogenous type and has an incidence of 2.4 per 105 person-years according to the French national hospital-discharge database . In developing countries, COM is a significant burden on health services.
beyond the preclinical stage . This study aimed to investigate whether there is a risk association between COM and RA.
Methods
Study population This retrospective population-based cohort study used the
beneficiary files of inpatient claims from the hospital-discharge database extracted from the National Health Insurance Research Database (NHIRD) in Taiwan from the period of January 1, 1996 to December 31, 2011. The National Health Insurance (NHI) program started in 1995 and currently covers over 98% of the population of Taiwan. The identification numbers of patients in the NHIRD were scrambled prior to this analysis. This study was approved by the ethical review board of the China Medical University in Taiwan (DMR101-IRB1-138).
Criteria for inclusion and exclusion The International Classification of
Diseases, 9th Revision, Clinical Modification (ICD-9-CM) was used to code the diseases of interest in the present study. To prevent medical fraud through overbilling for health care or inappropriate charges based on unconfirmed diagnoses, the NHI sets strict rules on disease coding and conducts regular monitoring of coding and claims submitted for reimbursement by all hospitals that provide inpatient services to NHI beneficiaries. Harsh penalties are imposed for coding errors, misdiagnoses or inappropriate treatment. Thus, the inpatient database of NHIRD provides reliable information with which to explore the risk of RA in patients with COM.
The study group consisted of 23,645 hospital inpatients admitted with the diagnosis of COM (ICD-9-CM 730.1). The date of COM diagnosis for each patient was defined as the index date. All subjects were followed up from the index date until the end date (12/31/11) or end point (development of RA). Patients diagnosed with RA (ICD-9-CM 714), systemic lupus erythematosus (SLE; ICD-9-CM 710.0), systemic sclerosis (ICD-9-CM 710.1), Sjögren’s syndrome (ICD-9-CM 710.2), dermatomyositis (ICD-9-CM 710.3), polymyositis (ICD-9-CM 710.4) or ankylosing spondylitis (ICD-9-CM 720) before the diagnosis of COM were excluded from the
study (n = 2,540). The remaining 21,105 patients were assigned to the case cohort. The reference cohort consisted of subjects in the NHIRD that had neither COM nor RA at baseline and were matched to the case cohort for age, gender and index date of diagnosis with a ratio of 4:1 (n = 84,420).
Outcome and relevant variables. The primary end point was the occurrence of
incident RA during the study period (2000–2011). Each diagnosis of RA was derived from the registry database of catastrophic illness in the NHIRD. Major or catastrophic illnesses in the NHI registry system include the following autoimmune diseases: RA, SLE, systemic sclerosis, Sjögren’s syndrome, dermatomyositis, and polymyositis. Because patients with a certified catastrophic illness are exempt from co-payment, the Bureau of NHI requires the diagnosis to be validated by at least two specialists, who carefully review all original medical records, laboratory data, image findings and pathological findings. Only those patients who meet the classification criteria for major illness are certified as having a catastrophic illness by the Bureau of NHI.
Additional variables of relevance were age, gender and comorbidities of COM, including fracture and injury (ICD-9-CM 800–829, E800–E869 and E880–E929), DM (ICD-9-CM 250), and gout (ICD-9-CM 274). All comorbidities were defined before the index date.
Statistical analysis Demographic factors and comorbidities were compared
between the case and reference cohorts using the Student’s t-test for continuous variables and the chi-square test for categorical variables. Person-years were calculated from the index date to the censored date of occurrence of RA, withdrawal from follow-up due to death, or the end of 2011. The gender- and age-specific incidence rates of RA (per 104 person-years) were compared between case and reference cohorts.
reference cohort was derived by Cox proportional hazard modeling with adjustment for age, gender and comorbidities of fracture/injury, DM, and gout, and is presented with the corresponding 95% confidence interval (CI). The risk of RA development in COM patients was also compared with respect to (1) gender, (2) age, (3) presence or absence of comorbidity, and (4) severity of COM. The age groups were <45 years, 45–64 years and ≥65 years. The severity of COM was defined by frequency of hospitalization due to recurrent osteomyelitis for antibiotic treatment and/or surgery per two follow-up years. The Kaplan-Meier survival curve was used to plot the RA-free rate of each cohort, and this was compared between cohorts using a log-rank test. A two-tailed p value <0.05 was considered significant. SAS version 9.1 (SAS Institute Inc., Carey, NC) was used for all analyses.
Results
Baseline characteristics and comorbidities We identified 21,105 hospital
inpatients with COM in the NHIRD who were included as the case cohort, and 84,420 matched subjects without COM who were included as the reference cohort. The mean duration of follow-up was 5.12 ± 3.22 years. The age and gender standardized prevalence and incidence rates of COM in the general Taiwanese population were 6.72 and 5.23 per 105 person, respectively. Demographic characteristics of both cohorts are presented in Table 1. The mean age of the case cohort was 55.8 ± 19.4 years. The gender-matched cohorts consisted of more men (64.9%) than women (35.2%). The proportion of unemployed and retired subjects was the highest in the case cohort (53.4%), which was followed with blue-collar (33.4%) and white collar occupation status (13.2%), and most of the subjects in the case cohort had monthly incomes less than TWD 22,800. The prevalence of each comorbidity examined (fracture/injury, DM and gout) was also higher in the case cohort than in the reference cohort (Table 1).
Incidence and risk of RA in patients with COM Over a total of 540,253
person-years of follow-up, 52 patients in the case cohort and 98 patients in the reference cohort developed RA, giving incidence rates of 5.43 per 104 person-years and 2.20 per 104 person-years, respectively. Figure 1 shows the Kaplan-Meier survival curve for each cohort (log rank test, p < 0.0001), which demonstrates a trend for a higher rate of RA development in the case cohort. Over the 12-year observation period, a second peak of RA development was observed in the case cohort between 6 and 8 years (Figure 2). After adjusting for age, gender, occupation, monthly income and comorbidities of fracture/injury, diabetes and gout, the risk of RA was greater in the case cohort than in the reference cohort (HR [95% CI], 2.21 [1.51–3.24]; Table 2). The increased risk was present in both genders, and was significant in women (2.47
[1.53–3.97]) and marginally significant in men (1.84 [0.97–3.50]). The risk increase was highest in the youngest age group (<45 years, 9.08 [3.22–25.6]) and dropped dramatically after age 45 years (Table 2). Finally, the risk increase was significant in subjects with (2.32 [1.20–4.47]) and without (2.62 [1.62–4.23]) comorbidity, both of which were comparable.
Risk of RA with respect to the severity of COM The incidence of RA was
assessed relative to the severity of COM, which was based on recurrent osteomyelitis with hospitalization, using a scale of 1, 2 or >2 times per two follow-up years (Table 3). Higher incidence rates were associated with more severe COM (30.46, 20.53 and 1.74 per 104 person-years for patients with severe, moderate and least severe COM, respectively). In the case cohort, the mean time from the last hospitalization to the end point of incident RA was 4.21 ± 3.23 years. Compared with the incidence of RA in the reference cohort (2.20 per 104 person-years), the risk of RA in the case cohort demonstrated a dose-response-like trend that decreased in proportion to disease severity (HRs: 11.2 [6.63–18.9], 8.24 [4.74–14.3] and 0.72 [0.40–1.30] for patients with severe, moderate and least severe COM, respectively).
Discussion
The present study took advantage of a huge Taiwanese database and confirmed that patients with COM had a more than two-fold higher risk of incident RA compared with reference non-COM subjects after a mean follow-up of 5.12 years. This risk association between COM and RA was independent of comorbidities of fracture/injury, DM and gout. These findings may suggest an independent causal effect of COM on RA, perhaps acting through the priming of a sequential autoimmune process that leads to joint destruction.
In agreement with published reports , the hospital inpatients with COM in the present study were characterized by male predominance and age >45 years. These characteristics substantiate the validity of identifying typical COM patients from the Taiwan NHIRD . The effect of episodic COM on the risk of developing RA was highest in the youngest age group and the risk declined with increasing age. Although a significant risk increase of RA was noted in women, the estimates in both genders were comparable.
In Taiwan, the most common infectious microorganism implicated in chronic refractory osteomyelitis is oxacillin-resistant S. aureus . Gram-positive pathogens such as Staphylococcus and Streptococcus account for the majority of pyogenic osteomyelitis . The usual clinical presentation is highly recurrent mixed aerobic and anaerobic osteomyelitis in open fracture . To eradicate the pathogens in patients with COM, hospitalization is provided for surgical interventions, parenteral antibiotics and hyperbaric oxygen therapy . The present study shows that patients with worse clinical conditions (determined by higher frequency of COM recurrence and hospitalization) had a greater increase risk of developing RA relative to patients with less-severe clinical conditions. The increase in the risk of RA therefore correlated with the degree of disease severity in COM patients. These findings support a contributory role of
COM in the development of RA.
Increasing evidence suggests a wide variety of genetic and environmental risk factors can trigger innate and adaptive immune responses and lead to autoimmune-mediated tissue damage in RA . Although there is little evidence to indicate how the immune response to COM results in subsequent stochastic reactions , we postulate that COM patients are predisposed to RA development through pathogen-associated and damage-associated molecular patterns that trigger Toll-like receptors (TLR) . Similar to the well-described role of P. gingivalis in periodontitis-related RA development , we suspect a potential role of TLR in COM patients, whereby TLR senses the pathogen via the cell wall (involving TLR2) of Streptococci and lipopolysaccharides (involving TLR4) in Gram-negative pathogens and initiates innate immunity, thus priming joint susceptibility to autoimmune-mediated destruction . This proposed mechanism for RA initiation requires that the production of auto-antibodies, such as rheumatoid factor and ACPA, in response to random life events may precede the development of clinical signs and symptoms of RA by several years . In the present study, there was a mean of 5.12 years between exposure to COM and development of RA, and the finding of a second peak of RA development between 6 and 8 years may warrant further exploration of whether rheumatoid factor or ACPA is present prior to the clinical onset of RA . However, even though periodontitis is known to initiate RA development, a negative association between COM and periodontitis (odds ratio, 0.60; 95% CI 0.51–0.71) was found (data not published). This may indicate the impact of periodontitis in COM patients for future RA development can be lower than that in non-COM subjects, and further strengthen an independent risk of COM for RA development.
This population-based cohort study had several limitations. First, the NHIRD does not provide certain information, including certain laboratory data (such as
inflammatory indices and bacterial culture results) or personal habits (such as cigarette and alcohol use). We were unable to determine the pathogens involved (i.e.,
Staphylococcus, Streptococcus or Mycobacterium tuberculosis) and therefore could
not determine if they were associated with the risk of incident RA. Second, although the diagnosis of RA patients with a certified catastrophic illness in this study have been validated by specialists who reviewed the medical records, laboratory and imaging finding, data related to clinical features such as serologic status, presence or absence of extra-articular disease and nature of disease severity were lacking. We were unable to compare clinical features of RA in patients with and without COM. Besides, the low difference in the incidence of RA between patients with COM and reference subjects (3.23 per 104 person-years) suggests that other factors may also contribute to the development of this complicated immune disease. Third, we were also unable to assess the impact of smoking, which is a known risk factor for RA. However, a very low smoking rate (<4.5%) was reported in Taiwanese women (26), suggesting that smoking is not likely to confound RA risk in female patients with COM in this study. We also found a greater increase in the risk of RA in younger COM patients, who are likely to have less cumulative smoking exposure than elderly patients, further suggesting that smoking did not have a large impact on RA development in this COM population. Although there may be other residual or unknown confounding effects, such as a common genetic polymorphism in COM and RA, with a huge population database of Taiwanese NHIRD and long observation period this bias may become non-differential.
On the other hand, there were several merits of this study needs to be mentioned. The NHIRD enrolled over 22 million Taiwanese citizens and employees in a national insurance program, and included more than 98% of the entire population of Taiwan during the period examined. To prevent healthcare fraud, the Bureau of
NHI rigorously monitors and audits the insurance claims made by hospitals. Thus, the reliability of diagnosis based on insurance claims is strengthened by a stringent NHI surveillance program. The demographic characteristics of COM patients in this study (majority male and >45 years of age) were comparable to those reported elsewhere . Patients with COM had a significantly higher prevalence of fracture/injury, DM and gout comorbidity than the reference subjects without COM, which could have impacted the risk of incident RA. Nevertheless, the increase in the risk of RA was comparable in groups with and without comorbidities. Although COM is not common in Taiwan, the standardized incidence of COM estimated from the NHIRD was comparable to the estimate reported by the French national hospital-discharge database . Besides, previous studies have indicated that COM is prevalent in individuals with lower socioeconomic status , which is compatible with our finding that the proportion was high in the case cohort classified as blue-collar, unemployed and retired, and with monthly incomes less than TWD 22,800. As socioeconomic status is related to lifestyle and environmental factors, these parameters were included as covariates in the Cox modeling. These evidences indicates that our findings may be generalizable to other populations. Besides, the large sample size enabled us to ascertain the impact of COM on RA risk in the different age groups. We can also exclude patients with SLE, systemic sclerosis, Sjögren’s syndrome, dermatomyositis, polymyositis, and ankylosing spondylitis at baseline because these rheumatic diseases may trigger the innate and adaptive immune system and confound the results. The long observation period in this study increased our ability to identify RA development predisposed by chronic infection. The mean follow-up period required for COM patients to develop RA was 5.12 years, and the mean duration from the last hospitalization for recurrent COM until the end point was 4.21 years, both of which are long enough to exclude a potential surveillance bias. In this case, we found a
significant dose-response-like relation between the severity of COM and the risk of RA, which further strengthens the causal role of COM in RA.
Conclusion
The present study shows that COM is an independent risk factor for RA development, regardless of the presence or absence of the comorbidities examined. COM had a greater impact on the development of RA in the youngest age group than in the older age groups, and there was a positive dose-response-like relation between COM severity and RA risk, indicating a causal relation between COM and RA development. A potential relation between episodic COM infection and development of RA may indicate a more general association between chronic infection and autoimmune diseases. Further study of COM patients focusing on the mechanism of immunological pathogenesis is warranted in the future.
References
1. Deane KD, Norris JM, Holers VM. Preclinical rheumatoid arthritis:
identification, evaluation, and future directions for investigation. Rheum Dis Clin North Am 2010;36(2):213-41.
2. Stolt P, Bengtsson C, Nordmark B, Lindblad S, Lundberg I, Klareskog L, et al. Quantification of the influence of cigarette smoking on rheumatoid arthritis: results from a population based case-control study, using incident cases. Ann Rheum Dis 2003;62(9):835-41.
3. de Rooy DP, van Nies JA, Kapetanovic MC, Kristjansdottir H, Andersson ML, Forslind K, et al. Smoking as a risk factor for the radiological severity of rheumatoid arthritis: a study on six cohorts. Ann Rheum Dis 2014;73(7):1384-7.
4. Arend WP, Firestein GS. Pre-rheumatoid arthritis: predisposition and transition to clinical synovitis. Nat Rev Rheumatol 2012;8(10):573-86.
5. Chen HH, Huang N, Chen YM, Chen TJ, Chou P, Lee YL, et al. Association between a history of periodontitis and the risk of rheumatoid arthritis: a nationwide, population-based, case-control study. Ann Rheum Dis 2013;72(7):1206-11.
6. Scher JU, Ubeda C, Equinda M, Khanin R, Buischi Y, Viale A, et al. Periodontal disease and the oral microbiota in new-onset rheumatoid arthritis. Arthritis Rheum 2012;64(10):3083-94.
7. Kaplan SL. Recent lessons for the management of bone and joint infections. J Infect 2014;68 Suppl 1:S51-6.
8. Lew DP, Waldvogel FA. Osteomyelitis. Lancet 2004;364(9431):369-79. 9. McHenry MC, Easley KA, Locker GA. Vertebral osteomyelitis: long-term outcome for 253 patients from 7 Cleveland-area hospitals. Clin Infect Dis 2002;34(10):1342-50.
Etiologic diagnosis of chronic osteomyelitis: a prospective study. Arch Intern Med 2006;166(1):95-100.
11. Zimmerli W. Clinical practice. Vertebral osteomyelitis. N Engl J Med 2010; 362(11):1022-9.
12. Grammatico L, Baron S, Rusch E, Lepage B, Surer N, Desenclos JC, et al. Epidemiology of vertebral osteomyelitis (VO) in France: analysis of hospital-discharge data 2002-2003. Epidemiol Infect 2008;136(5):653-60.
13. Deane KD, El-Gabalawy H. Pathogenesis and prevention of rheumatic disease: focus on preclinical RA and SLE. Nat Rev Rheumatol 2014;10(4):212-28.
14. Tseng CH, Chen JH, Muo CH, Chang YJ, Sung FC, Hsu CY. Increased risk of ischaemic stroke amongst patients with chronic osteomyelitis: a population-based cohort study in Taiwan. Eur J Neurol 2014 (Accpeted).
15. Chen CE, Ko JY, Wang JW, Wang CJ. Infection after intramedullary nailing of the femur. J Trauma 2003;55(2):338-44.
16. Chen LY, Chen LK, Chang CW, Kuo SC, Li LF, Chan YJ, et al. Treatment of community-onset methicillin-resistant Staphylococcus aureus (MRSA) bacteremia: a hospital-based study. Arch Gerontol Geriatr 2012;55(1):152-6.
17. Mylona E, Samarkos M, Kakalou E, Fanourgiakis P, Skoutelis A. Pyogenic vertebral osteomyelitis: a systematic review of clinical characteristics. Semin Arthritis Rheum 2009;39(1):10-7.
18. Kehrer M, Pedersen C, Jensen TG, Lassen AT. Increasing incidence of pyogenic spondylodiscitis: a 14-year population-based study. J Infect 2014;68(4):313-20. 19. Hall BB, Fitzgerald RH, Jr., Rosenblatt JE. Anaerobic osteomyelitis. J Bone Joint Surg Am 1983;65(1):30-5.
20. Chen CE, Shih ST, Fu TH, Wang JW, Wang CJ. Hyperbaric oxygen therapy in the treatment of chronic refractory osteomyelitis: a preliminary report. Chang Gung
Med J 2003;26(2):114-21.
21. Abdollahi-Roodsaz S, Joosten LA, Helsen MM, Walgreen B, van Lent PL, van den Bersselaar LA, et al. Shift from toll-like receptor 2 (TLR-2) toward TLR-4 dependency in the erosive stage of chronic streptococcal cell wall arthritis coincident with TLR-4-mediated interleukin-17 production. Arthritis Rheum 2008;58(12): 3753-64.
22. Yoshimura A, Lien E, Ingalls RR, Tuomanen E, Dziarski R, Golenbock D. Cutting edge: recognition of Gram-positive bacterial cell wall components by the innate immune system occurs via Toll-like receptor 2. J Immunol 1999;163(1):1-5. 23. Echchannaoui H, Frei K, Schnell C, Leib SL, Zimmerli W, Landmann R. Toll-like receptor 2-deficient mice are highly susceptible to Streptococcus pneumoniae meningitis because of reduced bacterial clearing and enhanced inflammation. J Infect Dis 2002;186(6):798-806.
24. Choe JY, Crain B, Wu SR, Corr M. Interleukin 1 receptor dependence of serum transferred arthritis can be circumvented by toll-like receptor 4 signaling. J Exp Med 2003;197(4):537-42.
25. Deane KD. Preclinical rheumatoid arthritis (autoantibodies): an updated review. Curr Rheumatol Rep 2014;16(5):419.
26. Bureau of Health Promotion T. Taiwan Tobacco Control Annual Report 2012. Taipei: Department of Health, Taiwan. 2012.
Figure legend
Figure 1. Rheumatoid arthritisfree rate in patients with chronic osteomyelitis and
reference subjects without chronic osteomyelitis calculated using the Kaplan-Meier
model.
Figure 2. Incidence rate of rheumatoid arthritis with respect to the time interval (every
2 years) between occurrence of chronic osteomyelitis and event onset. The number of
subjects at risk and the event number for rheumatoid arthritis are shown below the
