Abstract
Objective To assess whether patient and physician global assessment of gastrointestinal tract (GIT) disease in systemic sclerosis (SSc) are associated with a meaningful change in disease status.
Methods One hundred forty-three participants from the Australian Scleroderma Cohort Study were recruited to this study. Using logistic regression analysis, we evaluated the relationship between patient-reported and physician-assessed GIT disease status and symptoms, measures of health-related quality of life (36-item Short Form Health Survey [SF-36]) and GIT disease severity, measured by the Scleroderma Clinical Trials Consortium UCLA Gastrointestinal Tract 2.0 (GIT 2.0) score.
Results Patient-reported worsening of GIT symptoms in the month preceding assessment was significantly associated with more severe GIT disease (odds ratio [OR] 6.14, P < 0.01) and progressive worsening GIT disease severity as measured by the GIT 2.0 score (OR 45.98, P < 0.01). The new onset of reflux was the only specific symptom associated with patient-reported GIT disease activity (OR 2.98, P = 0.04). Physician-assessed GIT disease activity was not significantly associated with higher GIT 2.0 scores or increasing severity of disease. Patient-reported and physician-assessed GIT activity was not associated with SF-36 scores.
Conclusion In the absence of objective measures of GIT disease activity in SSc, patient-reported symptoms of GIT disease could be used to indicate disease activity and should merit consideration for inclusion in a multisystem disease activity index.
Systemic sclerosis (SSc) is a multisystem disease characterized by a triad of inflammation, vasculopathy, and fibrosis.1 Gastrointestinal tract (GIT) involvement is near-universal with involvement of all parts of the GIT from the mouth to anorectum.2 GIT disease has commonly been assessed using the University of California Los Angeles Scleroderma Clinical Trials Consortium (SCTC) Gastrointestinal Tract 2.0 (GIT 2.0) instrument, a 34-item patient-reported scale that measures the severity of GIT involvement.3 Gastrointestinal (GI) symptoms can also be measured using the Patient-Reported Outcomes Measurement Information System (PROMIS) GIT instrument.4 The Medsger Severity Scale (MSS)5 includes a GIT score, and the SSc-specific Health Assessment Questionnaire includes a visual analog scale assessing the impact of GIT disease on overall function.6 These outcome measures evaluate GIT disease severity, meaning they capture both disease activity, which refers to aspects of disease that are considered reversible, as well as damage, which is considered irreversible. There is currently no activity-specific instrument to evaluate only reversible aspects of GIT disease.
The SCTC has convened a working group (WG) to develop an activity index (AI). Measurement of GIT disease activity has been a significant hurdle in the development of the AI as many common manifestations of GIT disease are irreversible, representing damage rather than activity. Also, there is the recognized discordance between certain patient-reported symptoms and objective measures of GIT involvement.7 There are no validated clinical measures or biomarkers to differentiate GIT damage from activity.8
A potential solution to the absence of symptoms or biomarkers of disease activity is to use patient- or physician-eported assessments. Global assessments are included in other multisystem outcome measures in an effort to capture involvement of organ systems that are otherwise challenging to measure.9 We have previously demonstrated that patient-reported symptoms are significantly associated with meaningful progression of disease in specific organ systems.10 The use of a GIT-specific assessment by either the physician or patient has not previously been investigated. Therefore, we sought to evaluate whether a patient- and physician-reported GIT disease assessment could be used to assess clinically meaningful changes in disease status. We hypothesized that patient-reported worsening (PRW) of GIT symptoms and a physician-reported global assessment (PGA) of GIT disease activity would be associated with a clinically meaningful progression of GIT disease. Second, we explored whether a PRW and PGA were associated with the development of new symptoms of GIT involvement.
METHODS
Participants. All patients were enrolled in the Australian Scleroderma Cohort Study (ASCS). ASCS data are prospectively collected at annual review. Consecutive patients, aged ≥ 18 years, who fulfilled the 2013 American College of Rheumatology/ European Alliance of Associations for Rheumatology Classification Criteria for SSc,11 who had data available to define disease subtype and were reviewed by face-to-face consultation between January 2020 and November 2021 were included in this study. The ASCS is carried out in accordance with the National Statement on Ethical Conduct in Research Involving Humans (May 2015). The Human Research Ethics Committees at St. Vincent’s Hospital Melbourne and Royal Adelaide Hospital approved the study (HREC-A 020/07). Written informed consent was provided before any data were collected.
Clinical data. Demographic data collected included age, sex, and disease duration. Disease duration was recorded from onset of the first non-Raynaud manifestation. Annual data collected from all participants included examination findings, presence of SSc disease manifestations (recorded yes/no), medications, and patient-reported outcomes including the GIT 2.0, 36-item Short Form Health Survey (SF-36), and 29-item PROMIS (PROMIS-29). All participants had baseline autoantibody testing. Any routinely collected data that were not collected were recorded as missing. GIT investigations were performed at the discretion of the treating clinician.
Outcomes measured. Physicians (n = 6) were informed of the definition of disease activity, as defined through consensus by the SCTC AI WG: “Disease activity in SSc refers to aspects of disease, attributable to SSc, that are potentially reversible, or can be arrested, with time and/or effective therapy. Disease activity may be associated with morbidity and uncontrolled activity may lead to organ dysfunction and mortality.” Physicians were asked: “Do you think your patient currently has active, progressive GIT disease?” (PGA). Potential responses were as follows: (1) cannot assess, (2) no, (3) low activity, (4) moderate activity, or (5) high activity. Patients were asked, “Do you think any of your GIT symptoms have worsened in the past month?” (PRW), with the following possible responses: (1) no, (2) a little worse, (3) mild worsening, (4) severe worsening, (5) very severe worsening (need for hospitalization). This ASCS review for the purposes of this study was termed the study entry visit. Any PRW response of any of a little, mild, severe, or very severe worsening was considered a positive response and in study analyses, considered to indicate the presence of patient-reported worsening of GIT symptoms. Any physician rating of low, moderate, or high disease activity was considered to indicate the presence of GIT disease activity. We assessed the relationship between the PRW and PGA with the GIT 2.0 score measured at the study entry visit, as well as the change in GIT 2.0 score from the preceding study visit. Progression of GIT disease severity was defined as increase in GIT 2.0 scores of > 0.12, in accordance with the previously defined minimal important difference (MID).12 In the absence of an established standard for screening investigations for GIT involvement and the infrequent nature of invasive GIT investigations performed in this cohort, the PRW and PGA were compared to the GIT organ score of the MSS5 and the GIT component of the SCTC–Damage Index (DI) score.13 The MSS score rates GIT severity on a numeric scale: 0 (normal); 1 (mild) – distal esophageal hypoperistalsis, abnormal small bowel series; 2 (moderate) – antibiotics required for small intestinal bacterial overgrowth (SIBO); 3 (severe) – malabsorption syndrome, episodes of pseudo-obstruction; and 4 (endstage) – hyperalimentation.5 In the absence of a defined MID of the MSS, an MSS increase of ≥ 1 was considered significant. The SCTC-DI is a weighted score of organ damage that includes esophageal dysmotility (1 point), esophageal stricture (1 point), refractory gastro-esophageal reflux disease (1 point), gastric antral vascular ectasia (GAVE; 2 points), pseudo-obstruction (3 points), and low BMI < 18.5 kg/m2 or weight loss of > 10% over 12 months (2 points).13 A significant increase in SCTC-DI score was considered present if an increase of GIT DI score ≥ 1 was recorded, consistent with the SCTC-DI definition of worsening damage.13
Any associations between GIT disease assessment ratings and new-onset symptoms were evaluated. GIT symptoms collected as part of the ASCS protocol were considered new-onset if they were recorded as present at the study entry visit and had been absent at the preceding review. GIT symptoms of interest were those that had been nominated by the AI WG as potential AI items in a Delphi exercise. Potential GIT activity items collected by the ASCS were reflux, dysphagia, bloating, anemia, diarrhea, constipation, fecal incontinence, number of bowel actions per day, weight loss, esophageal strictures, GAVE, SIBO (defined as the concurrent presence of diarrhea and prescription of antibiotics for bacterial overgrowth), and episodes of pseudo-obstruction.
Statistical analysis. Data are presented as number (percentage) for categorical variables and mean (SD) for normally distributed or median (IQR) for nonnormally distributed continuous variables. Differences in frequency were tested using the chi-square test. The agreement between the PRW and PGA was assessed by Cohen 
coefficient. Univariable logistic regression analysis was performed to determine the association between the PRW and PGA with GIT 2.0, SF-36, MSS, and DI scores (analyzed as continuous variables) and GIT symptoms of interest (analyzed as dichotomous variables). The potential effects of confounding factors such as age, weight, medications, smoking and alcohol use, and depression and anxiety (rated on the PROMIS-29 Depression and Anxiety scales) on GIT symptoms were evaluated using multivariable logistic regression analysis. Study data were managed using REDCap tools hosted at the University of Melbourne. All statistical analyses were performed using Stata 14.2 (StataCorp).
RESULTS
Participant characteristics. This study included 143 participants. Patients had a median (IQR) disease duration of 13.67 (6.77-20.12) yrs and a median interval of 378 days (364-574) between study visits. All patients had GIT symptom data available for analysis and 118 (82.52%) patients completed the GIT 2.0 score at the study entry visit with a median total score of 0.32 (0.09-0.66). One hundred ten patients (76.92%) had SF-36 scores at study entry available for analysis, with a median physical component summary score of 39.20 (29.86-51.56) and median mental component summary score of 48.21 (38.67-56.14). Thirty-one patients (21.68%) were found to have a significant change in GIT 2.0 score at their study entry visit compared to the previous ASCS review (Supplementary Figure S1, available with the online version of this article). The mean MSS GIT organ score was 0.42 (SD 0.88) and mean SCTC-DI GIT score was 1.24 (1.66) at the study entry visit. Seven (4.90%) patients had ≥ 1 point increase of MSS compared to their previous review. Baseline characteristics of the study population are summarized in Table 1.
Baseline population characteristics (N = 143).
Assessment of GIT disease. Thirty-three (23.08%) patients had PGA and 29 (20.28%) had PRW. There was moderate agreement between the PRW and PGA (
0.51, P < 0.01). No patient had new-onset esophageal stricture, GAVE, or pseudo-obstruction during the study. One patient had a new, initial diagnosis of esophageal dysmotility by barium swallow study and a further 3 patients had new reflux esophagitis confirmed by endoscopy for the first time. Only the patient with new dysmotility recorded a positive PRW. No patient who had new-onset SIBO reported any recent change in GIT symptoms.
Associations of patient- and physician-reported GIT disease assessments. There was a significant association between the PRW and higher GIT 2.0 scores as well as a strong association with increased GIT 2.0 scores between study visits (Table 2). There was a nonsignificant trend toward an association between the PGA and higher GIT 2.0 scores (Table 3). The PRW was most closely associated with the onset of symptoms of the upper GIT (reflux, P = 0.04; dysphagia, P = 0.08; bloating, P = 0.08) as well as constipation (P = 0.06). However, only new-onset reflux reached statistical significance. Controlling for potential confounders strengthened the association between new-onset GIT symptoms and the PRW, except for upper GIT symptoms and comorbid depression and anxiety (Table 2). There was no significant association between the PGA and symptoms (Table 3). Neither patient nor physician GIT assessment was associated with change in SF-36 scores. The PRW was associated with higher MSS GIT scores (P = 0.01). There was a trend toward a positive PRW and higher burden of GIT disease damage measured by the SCTC-DI (P = 0.07).
Logistic regression analysis of patient-reported GIT disease status and clinical variables.
Logistic regression analysis of PGA of GIT disease and clinical symptoms.
DISCUSSION
This study has shown that a PRW is associated with more severe GIT disease as measured by the GIT 2.0 and MSS. The PRW was most likely to be associated with the new onset of upper GIT system symptoms and development of new constipation. A PGA was not associated with any other measure of GIT disease status or the onset of new GIT symptoms.
GIT disease is highly burdensome to patients2 and the measurement of GIT disease continues to prove challenging. There are no recommendations for the regular investigation of the GIT and there is ongoing reliance on clinical symptoms for the longitudinal assessment of the GIT.14 Existing measures such as the GIT 2.0 score can be used to measure overall GIT disease severity.3 However, the inclusion of aspects of both activity and damage mean that longitudinal assessment of progressive GIT disease is limited when using these instruments and has compelled the AI WG to consider a novel method of capturing GIT disease activity. We have shown that PRW is significantly associated with more severe GIT involvement, as measured by the GIT 2.0 score and MSS. These results suggest that a recall of change in symptoms over the preceding month is associated with the onset of new, clinical important symptoms of GIT involvement, supporting the inclusion of patient-reported GIT symptoms as a measure of disease in the AI.
Whether GIT symptoms reflect an active, potentially reversible disease process or are evidence of irreversible damage remains controversial. Our results indicate that the new onset of individual symptoms correlate poorly with both the patient and physician assessment of GIT disease. The exception to this may be symptoms of the upper GIT as we did observe trends toward an association between PRW and symptoms of reflux, dysphagia, and bloating. Conceivably, each of these symptoms may reflect a component of activity; reflux esophagitis can be reversible with aggressive gastric acid suppression and SIBO may be improved with antibiotics.2 The potential reversibility of these symptoms with treatment is consistent with the construct of disease activity defined by the AI WG. However, the validity of individual symptoms to measure disease activity requires careful evaluation in future studies, given the often-nonspecific nature of symptoms and recognized discordance between symptoms and invasive measures of SSc GIT manifestations.7
These results need to be considered within the limits of our study design. The sample size was small, from a single cohort with long-standing SSc, and results may not be applicable to SSc populations with differing demographics. Each patient was assessed at one individual timepoint by a single assessor. It was not possible to assess the inter- or intrarater reliability of these measures. There were 6 independent physicians who contributed data to this study. Although all physicians were informed of the definition of disease activity at the time of assessment, it is possible that variability of physician interpretation of this definition influenced the PGA. The ASCS collects data on an annual basis. Therefore, a comparison between patient-recalled changes in symptoms over 1 month could be compared only to the onset of new symptoms and change in other measures of disease over a 12-month period. It may be the case that a comparison between the PRW, recalled over 1 month, and individual symptom changes over the same period may yield different results.
Further, the ASCS does not include protocolized routine GIT investigation; thus, it is not possible to correlate the PRW with investigation abnormalities. Consequently, the frequency of accrual of new GIT manifestations may be underreported in this study. The ASCS does not record GIT investigations that have been performed with no abnormality detected, further limiting the accuracy of the estimates of accrual of new GIT disease manifestations. The relatively short duration of follow-up means that rarer events such as new esophageal strictures, GAVE, and pseudo-obstruction were not recorded in this study. The ASCS does not collect data on the dosage of medications, so it was not possible to account for change in medication dose as a contributing factor to PRW.
In conclusion, GIT disease is highly prevalent in SSc and of great clinical importance to patients. For these reasons, the SCTC AI WG has nominated to include measures of GIT disease in the AI. In the absence of consensus as to which individual symptoms or investigation findings could be used to measure activity, we have shown that measurement of PRW has partial face, construct, and criterion validity and is a feasible method of assessing change in disease status. Therefore, it may be an appropriate item to measure GIT disease activity pending the development of more robust outcome measures.
Footnotes
The development of the Scleroderma Clinical Trials Consortium (SCTC) Activity Index has been supported by a Betty Z. Benedict Award, SCTC Working Group grants, and private philanthropic donations. LR is supported by a Musculoskeletal Australia PhD Scholarship and an Australian Government Research and Training Scholarship. MN is supported by a National Health and Medical Research Council of Australia Investigator Grant (GNT 1176538). KM holds a National Health and Medical Research Council of Australia Investigator Grant (APP 1197169). The Australian Scleroderma Cohort Study is supported by Actelion Australia, Scleroderma Australia, Scleroderma Victoria, Arthritis Australia, Musculoskeletal Australia, SCTC, St. Vincent’s Hospital Research Endowment Fund, The Australian Rheumatology Association, philanthropic donations, GSK, Roche, Pfizer, Bayer, CSL Biotherapies, and BMS.
The authors declare no conflicts of interest relevant to this article.
- Accepted for publication October 31, 2022.
- Copyright © 2023 by the Journal of Rheumatology
