Patients

A total of 29 patients with celiac disease were evaluated in rheumatology clinic for joint complaints. Thirteen (44.8%) were diagnosed with arthritis and 16 (55.2%) had arthralgia but no evidence of active arthritis. The median time from the first to the last recorded rheumatology visit for patients with arthritis was 46 months (IQR 16.5–78.5).

Patient demographics and clinical characteristics at presentation are shown in Table 1. The average age at the index visit was younger in patients with arthritis versus arthralgia only (8.9 vs. 12.6 years). Both groups were predominantly female, white, and not Hispanic or Latino, consistent with prior observations of increased CD prevalence in non-Hispanic white patients [13]. No patients self-identified as American Indian or Alaska Native, Asian, Black or African American, or Native Hawaiian or Other Pacific Islander. About 80% of all patients reported a family history of autoimmune disease, with the most frequent including psoriasis (31.0%), rheumatoid arthritis (27.6%), inflammatory bowel disease (17.2%), systemic lupus erythematosus (13.8%), and Hashimoto thyroiditis (10.3%). Approximately 1/3 also reported a family history of celiac disease.

Table 1 Demographics and clinical characteristics

Of the patients with arthritis, 30.8% and 7.7% were positive for ANA and RF, respectively. Of the 11 tested for HLA-B27, 1 (9.1%) was positive. Whereas ESR was normal at presentation in most patients (76.9%), CRP was elevated in 8 (61.5%) patients (> 2x normal in 5 patients and 1-2x normal in 3 patients). One patient was already on a gluten-free diet at the time of rheumatology presentation. Only one patient had concomitant uveitis, and that patient was ANA negative.

Characteristics of joint involvement

On average the diagnosis of arthritis occurred 4.7 months (SD 3.6) after the onset of joint symptoms. The presentation was most commonly oligoarticular (76.9%) and asymmetric (84.6%). The median number of active joints at presentation was 2 (IQR 1–5). The most frequently involved joint was the knee (n = 9, 69.2%), and knee involvement was most often unilateral. Other commonly affected joints are listed in Table 2 and included ankles (n = 4, 30.8%), finger PIPs (n = 3, 23.1%), and wrists (n = 3, 23.1%).

Table 2 Joints with arthritis at index visit

Of the patients who had arthralgia without arthritis, the mean number of tender joints was 4.6 (SD 4.9). Patients most often reported large joint arthralgia. The knee was the most common tender joint (62.5%), followed by the ankle (50.0%), shoulder (25.0%), and hip (25.0%). Patients also frequently reported non-articular pain, including generalized body pain (n = 3), shin pain (n = 2), arm pain (n = 1), and pain involving feet, hands and/or digits (n = 5).

Celiac disease features among patients with arthritis

Celiac disease features are summarized in Table 3. The celiac disease diagnosis was made prior to the arthritis diagnosis in only 2 cases (15.4%). The timing of celiac diagnosis ranged from 51 months prior to 59 months after the arthritis diagnosis, with a median of 3 months following the arthritis diagnosis (IQR 0–19). Initial celiac testing that led to the diagnosis was obtained by the patient’s rheumatologist in 6 (46.2%) cases, by a gastroenterologist in 4 (30.8%) cases, by the PCP in 2 (15.4%) cases, and by an endocrinologist in 1 (7.7%) case. Of the 10 patients who were not IgA deficient, all had an elevated TTG IgA. Of the three that were IgA deficient, two had TTG IgG testing, and one was positive. All but one of the cases were biopsy-confirmed. The single case that was not biopsy-confirmed was due to family preference to adopt a gluten-free diet based on laboratory testing without endoscopy. Eight patients (61.5%) reported concomitant GI symptoms. Of these, only 3 patients had a BMI z-score less than − 1.64, and only one patient had impaired linear growth.

Table 3 Celiac Disease Features Treatment

Treatments received over the course of therapy for celiac-associated arthritis are demonstrated in Table 4.

Table 4 Treatments received over course of therapy for celiac-associated arthritis

Four of the twelve patients (33.3%) who were seen at a 6-month follow-up visit reported starting a gluten-free diet. However, of those, none had cleared celiac antibodies. The remaining patients did not yet have a confirmed diagnosis of celiac disease. All patients were started on a gluten-free diet once celiac disease diagnosis was confirmed, and all but one reported compliance with the diet at the last documented rheumatology visit. Nevertheless, only 3 patients (23.1%) had cleared celiac antibodies by the last documented visit (median time from celiac diagnosis 64 months, IQR 42–164); clearance was documented a median of 14 months (IQR 3–35) after initiation of the gluten-free diet.

Ten children (76.9%) underwent intra-articular joint injections, eight of which occurred in the first 6 months of treatment. Eleven (84.6%) were started on systemic therapy with a DMARD (n = 4), a biologic (n = 1) or both (n = 6). Seven of these started systemic medications in the first 6 months after diagnosis. Two patients (16.7%) were never on systemic medications and achieved remission with intra-articular corticosteroid injections and gluten-free diet alone. Both patients had mild joint disease from onset, with only one joint involved at diagnosis. One was a 15-year-old male with negative ANA and right wrist arthritis, and the other was a 3-year-old female with positive ANA and right knee arthritis.

Of the 10 patients who were started on systemic medications and reported compliance with the gluten-free diet, 3 (30%) successfully stopped systemic medications. Two of the patients who stopped medications had cleared celiac antibodies. Seven patients (70%) remained on systemic medications at the final visit, despite reported adherence to the gluten-free diet. Of the 7 patients who remained on systemic medications, only 1 had cleared celiac antibodies. Of those, 2 were on methotrexate, 1 was on adalimumab, 3 were on a combination of methotrexate and adalimumab, and one was on methotrexate and had received a rituximab infusion about 1 month prior. The latter was the only patient who had active arthritis at the final visit.

Disease course

Physician and patient-reported measures of disease activity at the baseline visit, at the 6-month follow-up, and at the final recorded visit are demonstrated in Fig. 1.

Fig. 1

Disease activity and patient-reported outcomes

Legend. *p < 0.05

a5 documented.

b9 documented.

c9 documented.

d8 documented.

e10 documented.

f12 documented.

g12 documented. One patient seen intelemedicine visit and had no joint exam documented.

h6 documented.

i9 documented.

j10 documented.

k6 documented.

l9 documented.

m10 documented.

There was a statistically significant improvement in number of joints with active arthritis (p = 0.02) between the baseline and final visit. Only one patient had active arthritis at the final visit, with 13 joints involved despite adherence to systemic medications. Her underlying diagnosis was in question given her severe presentation despite RF negativity. There was also a statistically significant improvement in physician assessment of disease activity (p = 0.03) between the baseline and final visit. Patient reported outcomes trended towards improvement, but did not achieve statistical significance.