Abdominal symptoms during primary Sjögren’s syndrome: a prospective study

Context Abdominal symptoms are poorly documented during primary Sjögren’s syndrome (pSS). Objectives To describe abdominal symptoms among pSS patients and to assess their association with characteristics of the disease. Methods One hundred and fifty patients followed at Hospital and University Center of Limoges were prospectively included and were evaluated using a composite global symptom score (GSS) describing abdominal symptoms and their severity. Data concerning the clinical and biological characteristics of the pSS and abdominal disorders were also collected. Results Ninety-five per cent of pSS patients suffered from abdominal symptoms with a median GSS of 7.5±5.5 points out of 30. More than half of the patients experienced abdominal tension (68%), upper abdominal pain (54%), abdominal discomfort (58%) and/or constipation (54%). Regarding the pSS activity ESSDAI score items, general and central nervous system involvement was associated with a high GSS. Regarding the patients’ symptoms ESSPRI score, there was a positive correlation with the GSS (p<0.01). Multivariate analysis showed a statistical association between a high GSS and seronegative status for SSA, gastroparesis and ESSPRI score (p<0.01 for each one). Conclusion This study revealed that more than 90% of pSS patients suffered from abdominal symptoms. There is currently no therapeutic recommendation because of the lack of specific study and comprehension of the physiopathological mechanisms involved.


Introduction
Primary Sjögren's syndrome (pSS) is a chronic autoimmune disorder characterized by lymphocytic infiltration of the exocrine glands and loss of secretory function with oral and eye dryness. It affects predominantly women with a female/male ratio of 9:1 and the peak frequency of the disease is about age 50 years. The etiology is still not well understood (1).
Abdominal disorders affect approximately 25% of pSS patients but are poorly documented in the literature. Indeed, epidemiological data vary according to the studies because of small number of patients and old classification criteria. Data concerning the prevalence of abdominal symptoms and their associations with clinical and biological pSSrelated manifestations are also contradictory (2). This heterogeneity of data led us to conduct a prospective study to describe abdominal symptoms in pSS patients as defined by the new 2016 ACR-EULAR criteria (3) and to assess their relationship with the characteristics of the disease.

Population
From July 2017 to June 2018, we conducted a single-center study at the Hospital and University Center of Limoges. It was a prospective, interventional study, validated by the Committee for the Protection of Individuals and registered on clinical.trials.gov (NCT03157011). All patients (age >18 years) followed in our center for pSS in consultation or hospitalization had a proposal to participate in the study. All of them fulfilled the 2016 ACR-EULAR classification criteria for pSS (3) and tested negative for HIV and hepatitis C virus (HCV). They gave their written consent after receiving information about participating in the study. Patients' characteristics (age at disease onset, sex, duration of pSS), clinical manifestations of pSS, associated organ-specific autoimmune disorders and treatment were recorded. Activity of pSS measured by ESSDAI score (EULAR Sjögren's syndrome disease activity index) and pSS's symptoms measured by ESSPRI score (EULAR Sjogren's Syndrome Patient Reported Index) were systematically evaluated at the same time as the abdominal score (4).
Hematological, biochemical, and immunological tests included blood count, fibrinogen, C-reactive protein test, renal and hepatic function tests, β2-microglobulin, serum protein electrophoresis, anti-nuclear antibodies (ANA) (indirect immunofluorescence), precipitating antibodies to the extractable nuclear antigens Ro/SSA and La/SSB (enzymelinked immunosorbent assay), rheumatoid factor (RF) (latex fixation and Waaler-Rose tests), free light chains and complement dosage. Cryoglobulin was evaluated in case of related clinical manifestations and/or diminution of complement. All tests were performed for each patient at the time of diagnosis and during the completion of the abdominal symptoms score.

Evaluation of abdominal symptoms
All pSS patients were interviewed by a physician using a questionnaire. It concerned ten abdominal symptoms (nausea, vomiting, upper and lower abdominal pain, abdominal discomfort, bloating, diarrhea, constipation, tenesmus, dysuria). Each symptom carried a score from 0 (no symptom) to 3 (severe) and evaluated by professional judgement. A global symptomatic score (GSS), calculated as the addition of all symptom scores, was assigned to each patient (maximum score 30) (5).
Previously diagnosed digestive, pancreatic and hepatic diseases were also systematically recorded by on the medical file as well as the current symptomatic treatment taken for digestive complaints. Previously diagnosis of gastroparesis was based on international clinical guidelines (6) and irritable bowel syndrome on Rome IV classification (7). All the patients did not have a digestive exploration after the questionnaire but only those with alarm signs (endoscopy) or signs evocative of gastroparesis (gastric emptying scintigraphy) according to the French recommendations (8)(9)(10).

Statistical analysis
Quantitative variables are described by their mean ± standard deviation, and qualitative variables by number and percentage. For qualitative variables, a Chi2 or Fisher's exact test was used to compare the different groups of patients (SSA +/-SSB patients and seronegative patients). For quantitative variables in a normal distribution, the Student's t-test was used to compare groups of two or more classes. For quantitative variables that do not follow a normal distribution, the Wilcoxon signed-rank test or the Kruskal-Wallis test was used to compare groups of two or more classes. Univariate analyses between different abdominal symptoms and the rest of the variables were performed. Variables with a p value less than 0.20 were included in a multivariate logistic model, simplified by a stepwise elimination method (11), so that the final model only includes variables that are significantly associated with the digestive symptoms. Quantitative variables verifying the Logit linearity assumption were incorporated without modification, or otherwise categorized. The relevance of the model was assessed using Pearson's residual and deviance tests, and its quality with an OCR curve. All statistical analyses were done using R software (version 3.2.2). p Values less than 0.05 were considered significant.

Previous digestive disease
Forty-four patients had symptomatic gastroesophageal reflux (29%). Thirteen patients had at least one gastric emptying scintigraphy for clinical signs suggestive of gastroparesis in the cohort (9%) and among them nine had a gastroparesis confirmed. Of these nine patients, three had mild, three moderate and three severe cases of gastroparesis, depending on fixation rate at four hours (6). All patients received medical treatment, three patients were treated with botulinum toxin injections and two underwent endoscopic pylorotomy. Twenty-seven cases of gastritis (18%), mostly atrophic, were reported as well as three cases of associated pernicious anemia (2%). Thirty cases of irritable bowel syndrome were documented (20%). Sixteen patients had enteric endoscopic diverticulosis (11%), which was complicated by diverticulitis in five cases. Three cases of dolichocolons were noted. Sixty-two percent of patients (n=93) had abdominal surgery; appendicectomy (n=44; 29%), cholecystectomy (n=15; 10%) and hysterectomy (n=27; 18%) were the most common.

Characteristics influencing the abdominal symptoms score
Some characteristics of pSS were associated with a high GSS (Table 1). Regarding the ESSDAI score items, presence of general or central nervous system involvement were associated with a high GSS (p=0.0137 and 0.0365 respectively). Regarding the total ESSPRI score, there was a correlation with the GSS of 0.48 (p<0.0001). The three items of the ESSPRI score taken separately were also correlated with a high GSS (dryness p=0.0014, fatigue p<0.0001 and pain p<0.0001).
Several digestive antecedents were correlated with a high GSS including hiatal hernia (p=0.007), gastroesophageal reflux (p=0.001), gastroparesis (p=0.0032), atrophic gastritis (p=0.0097), irritable bowel syndrome (p=0.0002), existence of at least one previous abdominal surgery (p=0.0440), and associated pernicious anemia (p=0.0306). The consummation of some symptomatic gastrointestinal related treatments was correlated with a high GSS: trimebutine (p=0.0047), proton pump inhibitors (p=0.0004). In addition, a high GSS was correlated with a higher frequency of endoscopic explorations and gastric emptying scintigraphies after evaluation of the GSS (p=0.0037). Concerning pSS specific therapy, none was correlated with a high abdominal symptoms score.
There was no difference between the SSA/SSB positive and seronegative groups with regard to previous digestive diseases and treatment for abdominal symptoms (Supplemental Table 1). On the other hand, serologic status influenced the overall abdominal score as well as prevalence of some abdominal symptoms (Supplemental Table 2). Indeed, the median value for the GSS was significantly higher in seronegative patients (p=0.02).
Multivariate analyses of the total GSS (linear regression) and upper gastrointestinal symptoms (logistic regression) versus patient characteristics are summarized in Table 2.
Multivariate analysis confirmed the statistical association between the GSS total score, seronegative status and ESSPRI score (p<0.01, Table 2). There were no items significantly related with lower abdominal symptoms after multivariate analysis. Fatigue and fever of the GSS score were related with the seronegative status (OR=8.45; p=0.0152) and a high ESSPRI score (OR=1.31; p<0.0001).

Discussion
This study is the first to use the latest criteria of the 2016 ACR-EULAR and to evaluate main abdominal symptoms in order to assess their severity during pSS. Abdominal Dysfunction of the autonomic nervous system seems to have an important role.
Autonomic neuropathy is described during pSS with orthostatic hypotension, urinary retention, segmental anhidrosis and dysfunction of the digestive and urinary systems (13).
Muscarinic receptors M3 are located in vascular smooth muscle cells, particularly in gastrointestinal and genitourinary systems as well as exocrine gland cells (14,15). During pSS, it has been found that autoantibodies are directed against muscarinic receptors (14). This suggests that these antibodies may be the cause of autonomic neuropathy and could lead to bladder insufficiency or gastrointestinal disorders (15). A study revealed the involvement of these autoantibodies in the alteration of colonic contraction on an ex-vivo model (16). Multivariate analysis showed an association between a high abdominal symptom score and antecedent of gastroparesis. Gastroparesis is probably under diagnosed during pSS.
We report 69% of confirmed gastroparesis among patients performed an emptying gastric scintigraphy for evocative symptoms. Few studies evaluated gastroparesis during pSS (14,18,19). Kovacs et al. revealed 70% of gastroparesis proven on scintigraphy among 30 pSS patients (14). Another study of 28 pSS patients showed a prevalence of 43% for subjective signs of gastroparesis and 29% for gastroparesis confirmed by octanoate breath tests (18). A retrospective study of 11 patients with pSS-associating confirmed gastroparesis showed that 82% had bloating and abdominal pain (19). The link between gastroparesis and involvement of the autonomic nervous system remains unclear during pSS. The main hypothesis would come from anti-muscarinic receptor type 3 antibodies inhibiting gastric emptying (14).
Unfortunately, no study correlated the level of these antibodies and the gastric emptying time.
We acknowledge several limitations. This study was monocentric performed for patients followed in a university hospital center that could induce a selection bias. In addition, there was no control group. We did not systematically investigate for digestive disorder by endoscopy. Regarding the hypothetical link between gastrointestinal involvement and autonomic neuropathy, we did not perform physiological testing and quantitative measurement concerning the autonomic system. In the same way, the detection of antimuscarinic receptor antibodies was not made in our study.
In conclusion, abdominal symptoms, which are often ignored, are common during pSS. They may represent a source of chronic deterioration in quality of life. The neurovegetative system seems particularly involved but further prospective studies are needed. Clinicians must remain alerted to abdominal symptoms and strive to best characterize the type of underlying lesions. Moreover, there is currently no therapeutic recommendation regarding pSS abdominal involvement because of the lack of any specific study and comprehension of the physiopathological mechanisms involved.

Conflict of Interest
The authors disclose no conflicts of interest   (17) 12.0 (6.6) 124 (83) 9.2 (6.0) 0.0397 GSS global symptom score; sd standard deviation; pSS primary Sjögren's syndrome a For quantitative variables in a normal distribution, the Student's t-test was used to compare groups of two or more classes; for variables that do not follow a normal distribution, the Wilcoxon signed-rank test or the Kruskal-Wallis test was used to compare groups of two or more classes. The value p <0.05 was considered significant For quantitative variables in a normal distribution, the Student's t-test was used to compare groups of two or more classes; for variables that do not follow a normal distribution, the Wilcoxon signedrank test or the Kruskal-Wallis test was used to compare groups of two or more classes. The value p <0.05 was considered significant. 0.0208 sd standard deviation; GSS global symptoms score; n.s. not significant a For quantitative variables in a normal distribution, the Student's t-test was used to compare groups of two or more classes; for variables that do not follow a normal distribution, the Wilcoxon signed-rank test or the Kruskal-Wallis test was used to compare groups of two or more classes. The value p <0.05 was considered significant