Patterns in Follow-Up Testing for Individuals Newly Positive for Celiac Disease Antibodies

Introduction: Rates of diagnosed celiac disease have increased over the last 25 years. Following diagnosis, North American guidelines recommend monitoring disease status through regular tissue transglutaminase antibodies (TTG) testing. However, adoption of these guidelines in a real-world, population-based setting has not been investigated. Methods: Health administrative data in Alberta Canada were used to identify incident cases and subsequent follow-up testing between 2016 and 2020. Repeated time-to-event analyses employing Prentice, Williams and Peterson gaP-time models were performed to estimate differences in follow-up testing by sex, age, geographic region, material deprivation, baseline TTG, and severity of comorbidities. Hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated overall and by event-specific tests (e.g., index to first test follow-up test, first follow-up to second). Results: A total of 1,744 children and 5,114 adults were included, with a follow-up time of 3,871 and 12,463 person-years, respectively. Within 5 years of index tTG-IgA positivity, 76% of children had at least 1 follow-up tTG-IgA test compared to 56% of adults (Figure 1A). Those living in metropolitan areas had higher probability of subsequent testing in both children (HR=1.38, 95% CI:1.23-1.54) and adults (HR=1.43, 95% CI:1.31-1.56) (Table 1). Children from the least deprived areas showed slightly higher hazards of repeated testing compared to those most deprived (HR=1.18, 95% CI:1.03-1.36) (Table 1). In adults, females had a marginally higher probability of follow-up testing (HR=1.11, 95% CI:1.04-1.19) (Table 1). Index tTG-IgA results were associated with repeated testing, but in opposite directions for children and adults (Table 1). On average, children received 2.2 follow-up tests compared to 1.3 among adults within 5 years of index positivity (Figure 1B). Event-specific associations between patient subgroups became nullified beyond the first follow-up test. Conclusion: Approximately 1 in 4 children and 1 in 2 adults did not receive a follow-up TTG test within 5 years of initial positivity. While there were notable differences in subsequent testing by geographic location and socioeconomic deprivation, these disparities may not persist after the first follow-up TTG test. Quality assurance studies are necessary to explain poor adherence to guidelines of surveillance of TTG.Figure 1.: A: Time from index TTG positivity to first follow-up test, B: Estimated mean cumulative functions for repeated tTG-IgA tests. Table 1. - Overall repeated events analyses, by age category Category Comparison HR (95% CI) P-value Children Sex Female vs male 0.96 (0.88, 1.05) 0.325 Age Continuous 0.98 (0.96, 0.99) < 0.001 Metropolitan Metropolitan vs non-metropolitan 1.38 (1.23, 1.54) < 0.001 Material deprivation (1) Least deprived vs (5) most deprived 1.18 (1.03, 1.36) 0.021 (2) vs (5) most deprived 1.19 (1.03, 1.37) 0.016 (3) vs (5) most deprived 1.07 (0.93, 1.23) 0.318 (4) vs (5) most deprived 1.01 (0.87, 1.16) 0.939 Index result* Continuous 1.05 (1.05, 1.06) < 0.001 Comorbidity1 Non-complex chronic vs non-chronic 1.05 (0.98, 1.21) 0.119 Complex chronic vs non-chronic 1.10 (0.93, 1.29) 0.267 Adults Sex Female vs male 1.11 (1.04, 1.19) 0.002 Age Continuous 1.00 (1.00, 1.00) 0.107 Metropolitan Metropolitan vs non-metropolitan 1.43 (1.31, 1.56) < 0.001 Material deprivation (1) Least deprived vs (5) most deprived 1.10 (0.99, 1.22) 0.072 (2) vs (5) most deprived 0.98 (0.88, 1.09) 0.736 (3) vs (5) most deprived 1.09 (0.98, 1.21) 0.125 (4) vs (5) most deprived 1.09 (0.98, 1.21) 0.127 Index result* Continuous 0.99 (0.98, 0.99) < 0.001 Comorbidity2 Charlson index: 1 vs 0 (no comorbidities) 0.98 (0.91, 1.06) 0.678 Charlson index: 2 vs 0 1.02 (0.91, 1.15) 0.751 Charlson index: ≥3 vs 0 1.00 (0.88, 1.13) 0.987 *Based on tTG-IgA value times the upper limit of normal. 1Pediatric Medical Complexity Algorithm (Simon et al. Pediatrics. 2014). 2Charlson comorbidity index (Quan et al. Med Care. 2005).