Facet Joints Involvement in Rheumatoid Arthritis: A Cross-Sectional Study

Background: It has been accepted amongst rheumatologists that rheumatoid arthritis (RA) does not involve the facet joints (FJs) of the spine; nevertheless, the issue is still under debate. Objective: To compare the prevalence of FJs’ changes between patients with RA and age- and sex-matched peers. Methods: CT scans of 34 patients with RA suffering from low back pain (LBP) were compared with 70 age- and sex-matched controls (individuals without RA, suffering from LBP) in a case-control study. The degenerative changes in the FJs were evaluated (i.e., joint space narrowing, marginal osteophytes, articular process hypertrophy, subchondral sclerosis, inter-joint vacuum phenomenon, and subchondral cysts), in addition to the marginal erosions, the most characteristic feature of joint change in RA. Disease activity characteristics (CRP, ESR, DAS-28, SDAI, and CDAI), duration of RA, age, and sex were obtained from patients’ clinical charts. Results: The prevalence of FJs’ changes in patients with RA and age- and sex-matched controls were not significantly different at any spinal level or in a total L1-S1 score. Marginal erosions, a characteristic feature of joint changes in RA, were not found in any of our RA subjects. No difference was found in disease parameters and markers in individuals with RA with affected FJs and those without. The occurrence of FJs’ changes amongst subjects with RA demonstrated no correlation with disease duration and activity. Conclusions: FJs of the lumbar spine are not involved in the inflammatory process of RA, and their LBP is not due to inflammation in FJs of the spine.


Introduction
It has commonly been accepted amongst rheumatologists that rheumatoid arthritis (RA) does not involve the facet joints (FJs) of the spine [1]. The FJs are synovial joints of the spine with hyaline cartilage overlying subchondral bone, a synovial membrane, and a joint capsule. Due to their high level of mobility and the considerable forces influencing FJs, especially in the lumbar area, significant degenerative changes can develop and constantly be a potential source of pain and disability [2]. There have been a few studies describing the prevalence of FJs' pathological changes due to RA [3] [4] [5], however, these studies are scarce, i.e. Stoddard and Chiverton's case report [4] and Kawaguchi et al. [3] usage of plane x-rays, which did not allow a detailed description of pathological findings. Only Yamada et al. [5] studied a large sample (n = 201) and used an MRI for evaluating pathological changes. However, there is still a need for a study of FJs in RA in different populations and samples.
Due to a precise demonstration of osseous details [6] [7] [8] and a relatively low cost, computed tomography (CT) was found to be the preferred method for imaging lumbar facet joint osteoarthritis. Abnormalities of the FJs that can be observed and categorized by a CT include osteophyte formation, hypertrophy of articular processes, joint space narrowing, vacuum joint phenomenon, synovial and subchondral cysts, and calcification of the joint capsule [9] [10] [11].
Aim: To compare the prevalence of FJs' changes between patients with RA and age-and sex-matched peers and to evaluate the prevalence of changes in FJs that are characteristic of the RA inflammatory process. Our hypothesis was that changes in FJs are of similar prevalence to changes in an age-and sex-matched sample.

Design
Cross-sectional analytical case-control study of matched samples.

Reliability of CT Readings
Initially, TR and LL read one batch of CTs and developed a reading protocol for evaluating FJs' radiographic changes. Adhering to this protocol, the reader (LL) read and re-read 15 CTs performed two weeks apart, blinded to the identity of the patient in order to assess the intra-rater reliability of the FJs' evaluations. Intra-observer reliability tests (kappa statistics) were performed.

FJs Evaluation
One assessor, blinded to the patient's health status, read CTs of the lumbar spine (L1-S1) and assessed the FJs. To evaluate the changes in the FJs, the assessor evaluated the features proposed by Kalichman et al. [11] [13] such as joint space narrowing, marginal osteophytes, articular process hypertrophy, subchondral sclerosis, inter-joint vacuum phenomenon, and subchondral cysts. The most distinguishing features of joint change in RA, the marginal erosions without subchondral sclerosis or osteophytes [5] [14] were also evaluated. Changes in each joint were scored according to Kalichman et al.'s protocol [11]. The joint that received a score ≥ 3 was considered affected. In addition, we created a total L1-S1 score, where the individual who exhibited at least one affected FJ was considered Open Journal of Rheumatology and Autoimmune Diseases positive and those without affected joints, negative. We also computed the sum scores of changes in the intervertebral discs, FJs' osteophytes, and the total score of FJs' changes.

Additional Data Collection
Age, sex and disease characteristics and markers (years of RA, presence of a rheumatoid factor, C-reactive protein (CRP); erythrocyte sedimentation rate (ESR); Disease Activity Score C-reactive protein (DAS-CRP); Disease Activity Score erythrocyte sedimentation rate (DAS-ESR); Clinical Disease Activity Index for RA (CDAI); Simple Disease Activity Index for RA (SDAI) were collected from the patients' clinical charts, close to the date of performing the CT scan.

Statistical Analysis
Statistical analysis was performed by the SPSS statistical package (Version 23).
Significance levels were set at P < 0.05. Descriptive statistics characterized the study sample. Prevalence of FJs' changes was calculated at each spinal level and in the total lumbar spine. Subsequently, the χ 2 -test compared individuals with and without RA. A one-way ANOVA evaluated the association between these sum scores in RA patients vs. the controls.

Results
Descriptive statistics of the studied sample are shown in Table 1. No differences were found between subjects with RA and those without as to age and sex. The mean length of disease in subjects with RA was 7.15 ± 4.17 years; 76.5% were rheumatoid factor positive. The prevalence of FJs' changes in patients with RA and the age and sex-matched controls is shown in Table 2. No significant difference in the prevalence of affected FJs was found at any spinal level or in the total L1-S1 score. Virtually no subjects in our sample were found with characteristic features for RA erosion in FJs. Osteophytosis of FJs and disc degeneration were distributed equally between patients with RA and without. In addition, RA individuals with affected FJs and those without exhibited no difference in any of the disease parameters and/or markers (Table 3).

Discussion
Low back pain is a frequent health problem, especially amongst the adult population. Its prevalence is estimated at approximately 20% for ages 30 -60 [15].   vs. degenerative), may assist in delivering a better differential diagnosis of FJs' changes, and encourage medical practitioners to introduce appropriate treatment for patients suffering from low back pain.
The earliest study found in the literature was performed in 1964 [16]. The authors found that disk narrowing without vertebral osteophytes, erosion in FJs, Open Journal of Rheumatology and Autoimmune Diseases and osteoporosis, were more common in their series of lateral lumbar spine x-rays of patients with RA. However, the same changes, which in their opinion could be attributed to RA, were found in a series of x-rays of the lumbar spine from a random population of males and females. In a more recent study, Yamada et al. [5]  Sakai et al. [17] examined 104 patients with RA together with their lumbar spine x-rays and MRI, regardless of low back pain. In 45.2%, disc lesions were found related to RA, thus, emphasizing the correlation between frequency of disc involvement and the number of affected RA peripheral joints. One of the fundamental views in the pathogenesis of RA is that inflammation occurs in the synovial membrane; hence, we cannot attribute the disc changes found in the studies to the inflammation process of RA. We found no difference between individuals with and without RA in relation to the prevalence of disk changes in the lumbar spine. These results are comparable with findings in the aforementioned studies. Moreover, we found that changes in synovial joints observed in the lumbar spine were similar in prevalence with the random population. Low back pain amongst the studied patients appeared while they were in the stage of low RA disease activity or in remission. We found no correlation between their disease activity parameters and FJs' changes. On the basis of our findings, we

Limitations
Our study had several limitations. Firstly, the number of subjects was limited. Secondly, RA subjects commenced their treatment nearing the date of diagnosis; hence, it is possible that treatment was effective in preventing changes in the lumbar spine intervertebral discs and the FJs.

Conclusion
There were no differences found in FJs' changes between the RA subjects and controls. The occurrence of FJs' changes amongst the RA subjects revealed no correlation with disease duration and activity. According to our findings, we may assume that FJs of the lumbar spine are not involved in the inflammatory process of RA and low back pain amongst patients with RA is not due to inflammation in this region of the spine.

Compliance with Ethical Standards
The study was approved by the Ethical (Helsinki) Committee of the Barzilai Medical Center (#0101-16 BRZ).