Search

In recent years, we have witnessed both artificial intelligence obtaining remarkable results in clinical decision support systems (CDSSs) and explainable artificial intelligence (XAI) improving the interpretability of these models. In turn, this fosters the adoption by medical personnel and improves trustworthiness of CDSSs. Among others, counterfactual explanations prove to be one such XAI technique particularly suitable for the healthcare domain due to its ease of interpretation, even for less technically proficient staff. However, the generation of high-quality counterfactuals relies on generative models for guidance. Unfortunately, training such models requires a huge amount of data that is beyond the means of ordinary hospitals. In this paper, we therefore propose to use federated learning to allow multiple hospitals to jointly train such generative models while maintaining full data privacy. We demonstrate the superiority of our approach compared to locally generated counterfactuals. Moreover, we prove that generative models for counterfactual generation that are trained using federated learning in a suitable environment perform only marginally worse compared to centrally trained ones while offering the benefit of data privacy preservation. Finally, we integrate our method into a prototypical CDSS for treatment recommendation for sepsis patients, thus providing a proof of concept for real-world application as well as insights and sanity checks from clinical application., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2024 The Authors. Published by Elsevier B.V. All rights reserved.)

Objectives: The optimal threshold of the physician global assessment (PGA) for remission in SLE has never been evaluated systematically. The aim of this study was to assess the ideal PGA threshold associated with physician remission and to investigate its impact on remission rates in our lupus cohort., Methods: In this monocentric cross-sectional study, patients with SLE were evaluated for physician remission by asking the treating physicians whether they considered their patient to be in remission, regardless of objective remission criteria. Furthermore, two objective remission definitions were applied: (i) DORIS (Definition Of Remission In Systemic Lupus Erythematosus) remission using a PGA of <2 (0-10) (corresponding to <0.5 on a visual analogue scale 0-3 used in DORIS); and (ii) DORIS remission with omission of PGA (modDORIS). A receiver operating characteristic analysis and regression analyses were performed to assess the ideal PGA threshold and factors influencing PGA., Results: Of the 233 patients included, 126 patients (54.0%) were in physician remission, 42.5% in DORIS remission and 67.0% in modDORIS remission. A PGA of <2 [numeric rating scale (NRS) 0-10] had the highest sensitivity (79%) and specificity (81%) for physician remission and modDORIS (area under the curve 0.85 and 0.69). PGA of patients fulfilling any of the remission definitions was associated with pain and hypocomplementemia. Damage was numerically higher in patients in modDORIS only; no association between PGA and damage was found in regression analysis., Conclusion: Using a PGA threshold of <2 (0-10), corresponding to <0.6 (0-3), resulted in best prediction of physician remission. PGA levels seem to be influenced by pain and complement levels but not disease damage., (© The Author(s) 2023. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For permissions, please email: journals.permissions@oup.com.)

Background: Patient-reported outcomes (PROs) play a crucial role in assessing rheumatic diseases, offering insights into disease evaluation and treatment efficacy. This study focuses on PRO assessment in large vessel vasculitides, including Takayasu Arteritis and Giant Cell Arteritis (GCA)., Methods: We retrospectively analyzed routine data from patients treated at our rheumatology clinic over a 10-year span. Patient and physician-rated global disease activity scale (G-DAS) scores, measured on a numeric rating scale (0-10 points), were collected at each visit. Clinical variables like age, sex, body mass index (BMI), disease duration, lab values, pain perception, and questionnaire responses were recorded. Linear regression and generalized additive linear regression (GAM analysis) examined associations between PROs and these factors., Results: The study included 138 patients, primarily diagnosed with GCA (94.4%). Mean follow-up was 2.5 years (0-7.7). Patient and physician G-DAS exhibited a moderate correlation (Pearson R 0.19, CI 0.14-0.24, p < 0.001). Higher patient G-DAS correlated with younger age (CI -3.4 - -1.5, p < 0.001), increased pain (CI 3.5-4, p < 0.001), functional limitations (HAQ, CI 0.5-0.6, p < 0.001), reduced physical (CI 2.3-2.7, p ≤ 0.001) and psychological well-being (CI 2.1-2.5, p < 0.001), and higher BMI (CI 1.3-2.4, p < 0.001). Physician G-DAS correlated with Birmingham Vasculitis Activity Score (V3.0; R 0.42, p 0.046) and were significantly linked to serum CRP elevations (β = 0.04, CI 0.0-0.08, p 0.028)., Conclusions: These findings underscore the need to integrate PRO measures into vasculitis disease management strategies, enhancing the understanding of disease activity from the patient's perspective., (© 2024. The Author(s).)

Background: Biobanks are essential structures for scientific research. The RHINEVIT biobank is used to recruit biomaterials from rheumatology patients in outpatient care and to conduct clinical research studies (e.g., cohort studies) and basic research. RHINEVIT established Broad Consents (BC) to allow extensive and relevant usage of data and biospecimens without the need for specific project restrictions. For quality assurance, we compared the consent rate of individual items of the BC versions in patients with systemic lupus erythematosus (SLE) in the longitudinal study., Methods: BCs were used for biomaterial donation. Informed consent data from RHINEVIT were analyzed. Due to the content restructuring of the BC items due to changes from the templates of the working group of the Medical Ethics Commissions in the Federal Republic of Germany and GDPR requirements, content mapping of the items was performed for the analysis., Results: From September 2015 to March 2022, 291 SLE outpatients donated biomaterials. In 119 patients, the BC was renewed at least once in a subsequent biomaterial donation. Three biomaterial donations were obtained from 21 patients and four from six patients using the respective BC. However, one consent was later revoked. Consent to the BC topics showed consistently high rates of agreement (range 97.5%-100%), with only some patients disagreeing with individual topics. This remained stable over time (median 526 days [Q1 400, Q3 844]). None of the patients disagreed with a certain topic in two consecutive visits., Conclusion: Modifications to the BC did not result in any relevant changes in the approval rates for SLE patients. RHINEVIT's BC is successfully used for the quality-assured handling of excellently annotated biomaterial. The long-term use of these highly valuable biospecimens for unrestricted research, also in an international context, remains assured., Competing Interests: MSchr and RH are employed by Serrala Group GmbH. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2023 Richter, Filla, Acar, Bleck, Kernder, Düsing, Vordenbäumen, Schröder, Hansen, Distler and Schneider.)

Objective: To assess whether a healthy lifestyle is associated to beneficial effects on various systemic lupus erythematosus (SLE) health domains., Methods: In a cross-sectional study, Mediterranean Diet Adherence Score (MEDAS), physical activity energy expenditure (PAEE), and smoking status were assessed by questionnaires, along with clinical parameters and various health domains including Systemic Lupus Disease Activity Score (SLEDAI), Depression Scale (CES-D), Fatigue Severity (FSS), functional status (FFbH), physical and mental quality of life (PCS, MCS). Lifestyle choices were assessed with respect to health domains by linear regression modeling. Additionally, SLE patients with a healthy lifestyle (MEDAS ≥ 4, ≥ 1 h sport per week, no smoking) were compared to those without by Wilcoxon's signed-rank test., Results: 49 of 145 SLE patients (44.3 ± 31.7 years, 87.6% female) followed a healthy lifestyle and showed a higher physical quality of life (β = 4.5 (95%-CI 1.5-7.9) p = 0.01), lower depression (β = -5.0 (-8.2 to -0.2) p = 0.02) and lower fatigue (β = -0.8 (-1.5 to -0.2) p = 0.01) independently of SLE disease activity. Furthermore, dsDNA-antibodies were lower (146 ± 540 vs 266 ± 146 U/mL, p = 0.049). In a more detailed analysis, physical activity had the highest impact on the various health domains when compared to smoking or diet adherence, which was consistent even after adjusting for multiple potential confounders. Each 1,000 kcal of weekly PAEE was associated to a 1.8 (0.9-2.6) point increase in the PCS ( p = 0.0001), a 0.2 (0.03-0.4) point decrease in the CES-D ( p = 0.01) and a 2.8 (1.2-4.4) point increase in the FFbH ( p = 0.0006)., Conclusion: A healthy lifestyle, especially physical activity is associated with beneficial effects including quality of life, depression and fatigue in SLE.

Objective: To compare physician and patient assessments of global disease activity in anti-neutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV) and to identify associated factors., Methods: Global disease activity scores (0-10 points) were retrospectively analyzed from physicians and patients with AAV at each outpatient visit from 2010 to 2020. We compared the scores and performed a linear regression with a random effects to identify associated factors., Results: Patients ( n  = 143, 1,291 pairs, 52% female) had a mean 64 (±15) years of age and a mean disease duration of 9 (±7) years. Patients and physicians global disease activity assessments showed a moderate correlation (Pearson R 0.31, CI [0.23-0.52], p  < 0.001). Linear regression showed a strong association between the physician-documented disease activity scores and serum CRP levels (β = 0.22, CI [0.18, 0.28]), disease duration (β = -0.022, CI [-0.04,-0.01]) and patients' assessment of disease activity (β = 0.08, CI [0.04, 0.12]). By contrast, patient assessments were strongly associated with the degree of pain (β = 0.30, CI [0.25, 0.35]), functional limitations in daily living (HAQ, β = 0.49, CI [0.21, 0.78]) and the global physical well-being (NRS, β = 0.39, CI [0.32, 0.46])., Conclusion: Patients' and physicians' assessments of disease activity correlated. High CRP levels and disease duration were associated with physician-assessed disease activity scores, while subjective limitations were associated with higher patient-assessed disease activity scores. These findings highlight and support the need to develop and evaluate patient-reported outcomes to assess disease activity in patients diagnosed with AAV., Competing Interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2023 Rohde, Kernder, Acar, Düsing, Fischer-Betz, Haase, Mucke, Sander, Richter, Filla, Schneider and Chehab.)

Objective: The aim of this study was to identify factors associated with impaired work productivity and impaired daily activities in patients with systemic lupus erythematosus (SLE)., Methods: The LuLa study is a longitudinal patient-reported study. Beyond sociodemographic data, work productivity, daily activities and fatigue, several other clinical outcome parameters (e.g. mental health-related quality of life and physical functioning, disease activity, damage and pain) were surveyed with validated questionnaires. The effects of confounders on work productivity (WPAI 2) and daily activity domains (WPAI 4) were studied by multivariate regression analysis., Results: A total of 585 patients completed the questionnaire of whom 259 were employed and analysed. The median impairment in work productivity (WPAI 2) was 20% (Q1-3 0-40), and the median impairment in daily activities (WPAI 4) was 30% (Q1-3 10-50%). Multivariate regression analysis revealed that fatigue, pain, disease activity and health-related quality of life affected WPAI 2 and 4. Furthermore, we observed distinct synergistic effects of fatigue, disease activity and pain on both work productivity and daily activities: a higher impact of fatigue was associated with the reported extent of pain or disease activity., Conclusion: In employed patients with SLE, impaired work productivity and impaired daily activities were frequently reported. Fatigue, pain, disease activity and health-related quality of life demonstrated a detrimental impact, with a synergistic effect of fatigue, disease activity and pain. Hence, both optimized pain management and targeted immunomodulatory therapy are important for preserving active participation in life among patients with fatigue.

Objective: Hypomethylation of CD40-ligand (CD40L) in T-cells is associated with increased disease activity in systemic lupus erythematosus (SLE). We therefore investigated possible associations of dietary methyl donors and products with CD40L methylation status in SLE., Methods: Food frequency questionnaires were employed to calculate methyl donor micronutrients in 61 female SLE patients (age 45.7 ± 12.0 years, disease duration 16.2 ± 8.4 years) and compared to methylation levels of previously identified key DNA methylation sites (CpG17 and CpG22) within CD40L promotor of T-cells using quantitative DNA methylation analysis on the EpiTYPER mass spectrometry platform. Disease activity was assessed by SLE Disease Activity Index (SLEDAI). Linear regression modelling was used. P values were adjusted according to Benjamini & Hochberg., Results: Amongst the micronutrients assessed (g per day), methionine and cysteine were associated with methylation of CpG17 (β = 5.0 (95%CI: 0.6-9.4), p = 0.04; and β = 2.4 (0.6-4.1), p = 0.02, respectively). Methionine, choline, and cysteine were additionally associated with the mean methylation of the entire CD40L (β = 9.5 (1.0-18.0), p = 0.04; β = 1.6 (0.4-3.0), p = 0.04; and β = 4.3 (0.9-7.7), p = 0.02, respectively). Associations of the SLEDAI with hypomethylation were confirmed for CpG17 (β=-32.6 (-60.6 to -4.6), p = 0.04) and CpG22 (β=-38.3 (-61.2 to -15.4), p = 0.004), but not the mean methylation of CD40L . Dietary products with the highest impact on methylation included meat, ice cream, white bread, and cooked potatoes., Conclusions: Dietary methyl donors may influence DNA methylation levels and thereby disease activity in SLE.

Objectives: The definition of an accurate target for a treat-to-target approach in SLE has been challenging over past years, and recently the DORIS definitions of remission were presented by the international DORIS task force. It was our aim to assess the frequency of DORIS remission and LLDAS in our SLE cohort and their agreement with the treating physician's (DORIS-) independent remission judgement. Patient characteristics leading to lack of agreement and incoherence ought to be identified., Methods: In this monocentric cross-sectional study, patients with SLE were enrolled and assessed between September 2016 and December 2017. DORIS remission definitions were applied and after the clinical consultation, the treating physicians gave his/her opinion on whether the patient was in remission. Regression analyses were performed to identify parameters influencing physician remission., Results: A total of 233 patients were included (87.6% female); 99 (42.5%) patients fulfilled any of the four DORIS remission definitions, while 126 patients were in remission according to their physician's judgement. We observed discordance in the assessment of remission in 53 patients (22.7%). Physician remission was influenced by disease activity [odds ratio (OR) 0.76, 95% CI: 0.63, 0.90], disease and/or treatment-related damage (OR 0.78, 95% CI: 0.62, 0.98) and the presence of ds-DNA antibodies (OR 2.47, 95% CI: 1.06, 6.04)., Conclusions: DORIS remission proved an achievable target in our outpatient clinic. Still we found discordance regarding DORIS remission and the treating physician's judgement with a greater number of patients considered in remission by their physicians., (© The Author(s) 2021. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For permissions, please email: journals.permissions@oup.com.)

Objective: To comprehensively assess associations of site-specific CD4 + -T-cell hypomethylation of the CD40-Ligand gene ( CD40L ) with disease activity of women with systemic lupus erythematosus (SLE)., Methods: CpG-sites within the DNA of the promotor and two enhancer regions (n = 22) of CD40L were identified and numbered consecutively. The rate of methylated DNA in isolated CD4 + -T-cells of women with SLE were quantified for each methylation site by MALDI-TOF. Disease activity was assessed by SLE Disease Activity Index (SLEDAI). Associations of site-specific methylation rates with the SLEDAI scores were assessed by linear regression modelling. P values were adjusted according to Bonferroni-Holm as indicated., Results: 60 female SLE patients participated in the study (age 45.7 ± 11.1 years, disease duration 17.0 ± 8.3 years). Significant associations to the SLEDAI were noted for CpG22 hypomethylation of the promotor (β = -40.1, p = 0.017, adjusted p = 0.027), trends were noted for CpG17 hypomethylation of the promotor (β = -30.5, p = 0.032, adjusted p = 0.6), and for CpG11 hypermethylation of the second enhancer (β = 15.0, p = 0.046, adjusted p = 0.8)., Conclusion: Site-specific hypomethylation of the CD40L promotor in CD4 + -T-cells show associations with disease activity in female SLE patients.