Cohort profile of a US military population for evaluating pre-disease and disease serological biomarkers in rheumatoid and reactive arthritis: Rationale, organization, design, and baseline characteristics

Abstract

Purpose: The etiology of several autoimmune disorders, including rheumatoid arthritis, remains unknown. While there are clear phases of disease progression, the mechanisms of transition between these phases are poorly understood. Additionally, treatment focuses on an alteration of the biological processes to prevent joint damage and functional decline. A goal is to potentially treat the disease during the preclinical phase to mitigate the disease process. Reactive arthritis is another rheumatologic condition known to be secondary to a distal infection. While prevention of infection would mitigate risk, serologic profiling patients with the disease may assist in the elucidation of potential disease risk factors. This study was initiated to enable an assessment of pre-disease biomarkers in patients newly diagnosed with rheumatoid arthritis and reactive arthritis.

Participants: A retrospective cohort of 500 rheumatoid and 500 reactive arthritis cases with 500 matched controls was drawn from a population of active component US military personnel. Appropriate inclusion criteria limited subject selection. Additionally, 4 serum samples (3 pre-disease and 1 disease-associated) were obtained for each case and control.

Findings to date: The established cohort provides the framework for novel exploration of the host response through serum profiling and seroepidemiology prior to disease onset.

Future plans: This study establishes the framework for the evaluation of novel serum biomarkers enabling the identification of signals prior to clinical disease that may enable disease prediction, elucidate disease pathogenesis and identify novel exposures leading to increased disease risk and/or disease severity.

Keywords: Cohort; Reactive arthritis; Rheumatoid arthritis; Seroepidemiology.

Fig. 1

Planned serum sample selection concurrent with and prior to rheumatoid or reactive arthritis diagnosis Fig. 1 Legend: For RA subjects, sample A was the first sample prior to the first RA diagnosis unless a sample was available within 30 days of the first diagnosis. In this case, the sample 30 days subsequent to diagnosis was obtained. Sample D for RA cases was the first sample available in the repository, and samples B and C were interim samples approximately evenly distributed across the service member's duty time. For ReA, samples A and B represent the first available sample after and before an initial ReA diagnosis, respectively. Sample D was the first sample available in the repository, and sample C was an interim sample approximately evenly distributed between samples B and D. For control subjects, sample A was matched by year ( ±1 year) to ReA and RA cases (50% of controls matched to RA cases and 50% matched to ReA cases) such that an equal proportion of Sample A from the controls were collected at approximately the same time ( ±1 year) as Sample A from the cases. Samples B and C for control subjects represented the two preceding serum samples available from the repository and sample D was the earliest serum sample available.

Fig. 2

Year of incident diagnosis (or matching) Fig. 2 legend: The cumulative percent of subjects by the first year in which the subject had their incident medical encounter for the outcome of interest (based on ICD9-CM code) or the year in which a control was identified for either the matched RA case or ReA case.

Fig. 3

Box and Whisker Plots of Serum Samples Timing Stratified by RA, ReA, and Healthy ControlsReACase, reactive arthritis cases; RACase, rheumatoid arthritis cases; ReACtrl, reactive arthritis healthy controls; RACtrl, rheumatoid arthritis healthy controls. The box and whisker plots represents the median (mid-line) 25% and 75% quartiles (boxes) and the 1st and 3rd quartiles + 1.5 * the interquartile range (whiskers).