Detection and Grading of Radiographic Hand Osteoarthritis Using an Automated Machine Learning Platform.

OBJECTIVE: Automated machine learning (autoML) platforms allow health care professionals to play an active role in the development of machine learning (ML) algorithms according to scientific or clinical needs. The aim of this study was to develop and evaluate such a model for automated detection and grading of distal hand osteoarthritis (OA). METHODS: A total of 13,690 hand radiographs from 2,863 patients within the Swiss Cohort of Quality Management (SCQM) and an external control data set of 346 non-SCQM patients were collected and scored for distal interphalangeal OA (DIP-OA) using the modified Kellgren/Lawrence (K/L) score. Giotto (Learn to Forecast [L2F]) was used as an autoML platform for training two convolutional neural networks for DIP joint extraction and subsequent classification according to the K/L scores. A total of 48,892 DIP joints were extracted and then used to train the classification model. Heatmaps were generated independently of the platform. User experience of a web application as a provisional user interface was investigated by rheumatologists and radiologists. RESULTS: The sensitivity and specificity of this model for detecting DIP-OA were 79% and 86%, respectively. The accuracy for grading the correct K/L score was 75%, with a κ score of 0.76. The accuracy per DIP-OA class differed, with 86% for no OA (defined as K/L scores 0 and 1), 71% for a K/L score of 2, 46% for a K/L score of 3, and 67% for a K/L score of 4. Similar values were obtained in an independent external test set. Qualitative and quantitative user experience testing of the web application revealed a moderate to high demand for automated DIP-OA scoring among rheumatologists. Conversely, radiologists expressed a low demand, except for the use of heatmaps. CONCLUSION: AutoML platforms are an opportunity to develop clinical end-to-end ML algorithms. Here, automated radiographic DIP-OA detection is both feasible and usable, whereas grading among individual K/L scores (eg, for clinical trials) remains challenging.

Phenotyping by persistent inflammation in systemic sclerosis associated interstitial lung disease: a EUSTAR database analysis.

BACKGROUND: Systemic sclerosis (SSc) is a heterogeneous disease with frequently associated interstitial lung disease (SSc-ILD). We aimed to determine the prognostic potential of phenotyping patients with SSc and SSc-ILD by inflammation and to describe disease trajectories stratified by inflammation and immunosuppressive treatment. METHODS: Patients from the European Scleroderma Trials and Research (EUSTAR) group cohort were allocated to persistent inflammatory, intermediate and non-inflammatory phenotypes if C-reactive protein (CRP) levels were ≥5 mg/L at ≥80%, at 20-80% and at <20% of visits, respectively. Cox regression models were used to analyse mortality risk and mixed effect models to describe trajectories of FVC and diffusing capacity for carbon monoxide (DLCO) %-predicted stratified by inflammation and immunosuppressive treatment. RESULTS: 2971 patients with SSc and 1171 patients with SSc-ILD had at least three CRP measurements available. Patients with SSc-ILD with a persistent inflammatory phenotype had a 6.7 times higher risk of mortality within 5 years compared with those with a persistent non-inflammatory phenotype (95% CI 3 to 15). In the inflammatory phenotype, FVC %-predicted was declining without (-1.11 (95% CI -2.14 to -0.08)/year), but stable with immunosuppressive treatment (-0.00 (95% CI -0.92 to 0.92)/year). In the non-inflammatory phenotype, patients with and without immunosuppressive treatment had a significant decline in FVC %-predicted, which was more pronounced in those with immunosuppressive treatment (-1.26 (95% CI -1.87 to -0.64) and -0.84 (95% CI -1.35 to -0.33)/year, respectively). CONCLUSIONS: Phenotyping by persistent inflammation provides valuable prognostic information, independent of demographics, disease duration, cutaneous subtype, treatment and SSc-ILD severity. The findings from this study support early immunosuppressive treatment in patients with SSc-ILD with persistent inflammation.

Systemic sclerosis and primary biliary cholangitis: Longitudinal data to determine the outcomes.

Background: Several studies described the cross-sectional characteristics of systemic sclerosis patients and coexisting primary biliary cholangitis, but longitudinal prognostic data are lacking. Aims: To describe the systemic sclerosis-primary biliary cholangitis phenotype, including baseline characteristics and outcomes. Methods: We performed a multicentre the European Scleroderma Trials and Research Group study of systemic sclerosis patients with primary biliary cholangitis or with primary biliary cholangitis-specific antibodies, matched with systemic sclerosis controls free from hepatobiliary involvement matched for disease duration and cutaneous subset. Data were recorded at baseline and at the last available visit. Results: A total of 261 patients were enrolled (115 primary biliary cholangitis-systemic sclerosis, 161 systemic sclerosis). At baseline, systemic sclerosis-primary biliary cholangitis patients had a higher prevalence of anti-centromere antibodies (p = 0.0023) and a lower prevalence of complete absence of digital ulcers. The milder vascular involvement was confirmed at follow-up when crucial differences emerged in the percentage of patients experiencing digital ulcers; a significantly higher number of patients who never experienced digital ulcers were observed among primary biliary cholangitis-systemic sclerosis patients (p = 0.0015). Moreover, a greater incidence of pulmonary arterial hypertension (p < 0.001) and of conduction blocks (p = 0.0256) was observed in systemic sclerosis patients without primary biliary cholangitis. Patients with primary biliary cholangitis had higher levels of liver enzymes at baseline than systemic sclerosis patients; a significant decrease in liver enzymes was observed at follow-up. Out of 18 patients with cholangitis, one received a liver transplant at follow-up. Conclusion: Our data show that systemic sclerosis-primary biliary cholangitis exhibit a mild systemic sclerosis and primary biliary cholangitis phenotype with outcomes being in general favourable.

Patient Assessment Chronic Illness Care (PACIC) and its associations with quality of life among Swiss patients with systemic sclerosis: a mixed methods study.

BACKGROUND: The Chronic Care Model (CCM) is a longstanding and widely adopted model guiding chronic illness management. Little is known about how CCM elements are implemented in rare disease care or how patients' care experiences relate to health-related quality of life (HRQoL). We engaged patients living with systemic sclerosis (SSc) to assess current care according to the CCM from the patient perspective and their HRQoL. METHODS: We employed an explanatory sequential mixed methods design. First, we conducted a cross-sectional quantitative survey (n = 101) using the Patient Assessment of Chronic Illness Care (PACIC) and Systemic Sclerosis Quality of Life (SScQoL) questionnaires. Next, we used data from individual patient interviews (n = 4) and one patient focus group (n = 4) to further explore care experiences of people living with SSc with a focus on the PACIC dimensions. RESULTS: The mean overall PACIC score was 3.0/5.0 (95% CI 2.8-3.2, n = 100), indicating care was 'never' to 'generally not' aligned with the CCM. Lowest PACIC subscale scores related to 'goal setting/tailoring' (mean = 2.5, 95% CI 2.2-2.7) and 'problem solving/contextual counselling' (mean = 2.9, 95% CI 2.7-3.2). No significant correlations were identified between the mean PACIC and SScQoL scores. Interviews revealed patients frequently encounter major shortcomings in care including 'experiencing organized care with limited participation', 'not knowing which strategies are effective or harmful' and 'feeling left alone with disease and psychosocial consequences'. Patients often responded to challenges by 'dealing with the illness in tailored measure', 'taking over complex coordination of care' and 'relying on an accessible and trustworthy team'. CONCLUSIONS: The low PACIC mean overall score is comparable to findings in patients with common chronic diseases. Key elements of the CCM have yet to be systematically implemented in Swiss SSc management. Identified gaps in care related to lack of shared decision-making, goal-setting and individual counselling-aspects that are essential for supporting patient self-management skills. Furthermore, there appears to be a lack of complex care coordination tailored to individual patient needs.

Mitochondrial DNA: a novel indicator of active inflammation in ANCA-associated vasculitides.

OBJECTIVES: ANCA-associated vasculitis (AAV) includes granulomatosis with polyangiitis (GPA) and microscopic polyangiitis (MPA). ANCA triggers neutrophil extracellular trap formation, which releases either mitochondrial (mt) DNA or nuclear DNA (n) DNA, contributing to inflammation. Our aim was to prospectively examine the extent and nature of circulating DNA in AAV and the clinical utility of DNA quantification. METHODS: DNA was isolated from platelet-free plasma of consecutive GPA and MPA patients and healthy controls (HCs). mtDNA and nDNA copy numbers were quantified by PCR. Clinical data, including the BVAS, were collected. RESULTS: Ninety-two HCs (median age 51 years, 58.7% female) and 101 AAV patients (80 GPA, 21 MPA, median age 64 years, 50.5% female, BVAS range: 0-30) were included. Median mtDNA copies were 13-fold higher in patients with AAV than in HCs; nDNA concentrations did not differ. Patients with active AAV (BVAS > 0) had 4-fold higher median mtDNA copies than patients in remission (P = 0.03). mtDNA, unlike nDNA, correlated with BVAS (r = 0.30, P = 0.002) and was associated with AAV activity at multivariable analysis. Receiver operating characteristic curve analysis indicated that mtDNA quantification differentiates patients with active AAV (BVAS > 0) from HCs with 96.1% sensitivity and 98.9% specificity (area under the curve 0.99). In 27 AAV patients with follow-up, mtDNA changes but not CRP or ANCA-titres correlated with BVAS changes (r = 0.56, P = 0.002). CONCLUSIONS: mtDNA, unlike nDNA, is elevated in the plasma of AAV patients and may contribute to systemic inflammation. mtDNA could be superior to established biomarkers in the laboratory monitoring of AAV activity.

Polg mtDNA mutator mice reveal limited involvement of vertebral bone loss in premature aging-related thoracolumbar hyperkyphosis.

Background: Age-related hyperkyphosis is multifactorial and involves alterations of vertebral bone, intervertebral discs (IVD) and paraspinal muscles. The relative contribution of these tissues remains unclear. Here, we compared differences in vertebral bone microarchitecture and IVD thickness between prematurely aging mice with spinal hyperkyphosis and wild type littermates. Methods: Thoracolumbar vertebral columns were dissected from homozygous Polg D257A and age-matched wild type littermates. Micro-computed tomography was performed to quantify cortical and trabecular bone parameters at anterior and posterior portions of T8-L4 vertebrae. In addition, vertebral shape, transaxial facet joint orientation and IVD thickness were quantified. Differences in anterior/posterior ratios between genotypes were compared by Student's t-test and association between vertebral bone and IVD parameters was investigated using Pearson correlation analysis. Results: Hyperkyphotic homozygous mice displayed generalized osteopenia that was more pronounced at the posterior compared with anterior portion of thoracolumbar vertebrae. An increase in the anterior/posterior ratio of trabecular bone parameters was revealed at the thoracolumbar junction (T13-L1). Polg D257A displayed diffuse loss of cortical bone thickness, yet anterior/posterior ratios were unchanged. Despite generalized and regional bone loss, vertebral shape was unaffected. PolG D257A mice showed a 10-20 % reduction of IVD thickness at both thoracic and lumbar levels, with only minimal histopathological changes. IVD thickness was negatively correlated with anterior/posterior ratios of trabecular bone parameters, as well as with more coronally oriented facet joints, but negatively correlated with the anterior/posterior ratio of cortical bone thickness. Conclusions: Aging-induced regional changes of vertebral trabecular and cortical bone did not lead to altered vertebral shape in Polg D257A mice but may indirectly cause hyperkyphosis through reduction of IVD thickness. These findings suggest a limited role for aging-induced bone loss in spinal hyperkyphosis and warrants further research on the involvement of paraspinal muscle degeneration.

Long-term persistence of antibodies after diphtheria/tetanus vaccination in immunosuppressed patients with inflammatory rheumatic diseases and healthy controls.

Many vaccines demonstrate high effectiveness for years. This prospective multicentre study was conducted in Switzerland to assess the long-term persistence of antibodies to the diphtheria/tetanus (dT)-vaccine in adult patients with rheumatic diseases (PRDs). 163 PRDs and 169 controls were included in the study. The median age of all participants was 50 years (range: 18-83 years) and 56% were female. After a median time interval of 16 years after vaccination, the median anti-vaccine antibody concentrations were lower in PRDs than in controls for tetanus (1.68 vs 2.01; p = 0.049) and diphtheria (0.05 vs 0.22; p = 0.002). Based on the currently accepted seroprotection threshold (antibody concentration ≥ 0.1 IU/ml), PRDs had lower proportions of short-term tetanus and diphtheria protection as demonstrated by crude odds ratios (OR) of 0.30 (p = 0.017) and OR: 0.52 (p = 0.004), respectively. After adjusting for 'age' and 'time since last dT vaccination', the strength of associations became weaker; for tetanus, borderline evidence remained for a true difference between PRDs and controls (OR: 0.36 [p = 0.098]), however, not for diphtheria (OR: 0.86 [p = 0.58]). We hypothesize that in the presence of rheumatic diseases and its immunosuppressive treatment, vaccine-specific long-lived plasma cells (LLPCs) may be diminished or competitively displaced by rheumatism-specific LLPCs, a process which may decrease the persistence of vaccine-specific antibodies. Novel studies should be designed by incorporating methodologies allowing to determine the attributable fraction of immunosuppressive/immunomodulatory medications and rheumatic disease itself on long-lasting vaccine-specific antibody persistence, as well as, further study the role of LLPCs.

Gout Arthritis During Admission for Decompensated Heart Failure-A Descriptive Analysis of Risk Factors, Treatment and Prognosis.

BACKGROUND: Chronic heart failure and hospital admissions are well-known risk factors for acute gouty arthritis. However, in-depth analyses of patients admitted for decompensated heart failure (DHF) who subsequently developed a gout attack are sparse. This study aims to characterize DHF patients who developed a gout attack during their inpatient treatment and describe potential risk factors, its consequences, and its management in the setting of heart failure exacerbation. METHODS: Retrospective chart review of 50 patients with an admission diagnosis of DHF who subsequently experienced a gout attack during admission at a Swiss tertiary care hospital between 2018 and 2020. Patients with a refusal of the general research consent were excluded (n = 10). RESULTS: A gout attack developed in 66/1,832 (3.6%) DHF admissions of whom 50 individual patients were analyzed. Patients were predominately male (76%), of advanced age (median 80.5 years), with several comorbidities including chronic kidney disease (74%), comorbid gout (70%, only 43% on urate lowering therapy) and hyperuricemia (median 547 µmol/l, IQR 434-667 µmol/l). Diuretics were intensified in all patients. Acute gout presented as polyarticular arthritis (62%) and was often accompanied by fever (30%). Joint aspiration was performed in 32%, and intra-articular steroid injections administered in 20% of patients. Median length of stay and 6-month mortality were 16 days (IQR 12-25) and 32%, respectively, compared to 9 days (IQR 6-14) and 16% for DHF patients without a gout attack. CONCLUSION: Our study highlights features of gout attacks in the context of DHF including the absence of comorbid gout in a significant proportion of patients, the presence of polyarticular disease during the flare, and a poor prognosis. The present study identifies the necessity to better address gout as a comorbidity in DHF patients and may assist clinicians in identifying DHF patients at risk for a gout attack.

Patient preferences for the treatment of systemic sclerosis-associated interstitial lung disease: a discrete choice experiment.

OBJECTIVES: Treatments for SSc-associated interstitial lung disease (SSc-ILD) differ in attributes, i.e. mode of administration, adverse events (AEs) and efficacy. As physicians and patients may perceive treatments differently, shared decision-making can be essential for optimal treatment provision. We therefore aimed to quantify patient preferences for different treatment attributes. METHODS: Seven SSc-ILD attributes were identified from mixed-methods research and clinician input: mode of administration, shortness of breath, skin tightness, cough, tiredness, risk of gastrointestinal AEs (GI-AEs) and risk of serious and non-serious infections. Patients with SSc-ILD completed an online discrete choice experiment (DCE) in which they were asked to repeatedly choose between two alternatives characterized by varying severity levels of the included attributes. The data were analysed using a multinomial logit model; relative attribute importance and maximum acceptable risk measures were calculated. RESULTS: Overall, 231 patients with SSc-ILD completed the DCE. Patients preferred twice-daily oral treatments and 6-12 monthly infusions. Patients' choices were mostly influenced by the risk of GI-AEs or infections. Improvement was more important in respiratory symptoms than in skin tightness. Concerning trade-offs, patients accepted different levels of increase in GI-AE risk: +21% if it reduced the infusions' frequency; +15% if changing to an oral treatment; up to +37% if it improved breathlessness; and up to +36% if it reduced the risk of infections. CONCLUSIONS: This is the first study to quantitatively elicit patients' preferences for treatment attributes in SSc-ILD. Patients showed willingness to make trade-offs, providing a firm basis for shared decision-making in clinical practice.

Representativeness of systemic sclerosis patients in interventional randomized trials: an analysis of the EUSTAR database.

OBJECTIVE: To estimate the extent of and the reasons for ineligibility in randomized controlled trials (RCTs) of SSc patients included in the EUSTAR database, and to determine the association between patient's features and generalizability of study results. METHODS: We searched Clinicaltrials.gov for all records on interventional SSc-RCTs registered from January 2013 to January 2018. Two reviewers selected studies, and information on the main trial features were retrieved. Data from 8046 patients having a visit in the EUSTAR database since 2013 were used to check patient's eligibility. The proportion of potentially eligible patients per trial, and the risk factors for ineligibility were analysed. Complete-, worst- and best-case analyses were performed. RESULTS: Of the 37 RCTs included, 43% were conducted in Europe, 35% were industry-funded, and 87% investigated pharmacological treatments. Ninety-one percent of 8046 patients included could have participated in at least one RCT. In complete-case analysis, the median [range] proportion of eligible patients having the main organ complication targeted by each study was 60% [10-100] in the overall sample of trials, ranging from 50% [32-79] for trials on skin fibrosis to 90% [34-77] for those targeting RP. Among the criteria checked, treatment- and safety-related but not demographic were the main barriers to patient's recruitment. Older age, absence of RP, and lower mRSS were independently associated with the failure to fulfill criteria for any of the included studies. CONCLUSIONS: Patient's representativeness in SSc-RCTs is highly variable and is driven more by treatment- and safety-related rather than demographic criteria.

Circulatory Neutrophils Exhibit Enhanced Neutrophil Extracellular Trap Formation in Early Puerperium: NETs at the Nexus of Thrombosis and Immunity.

Pregnancy is associated with elevated maternal levels of cell-free DNA of neutrophil extracellular trap (NET) origin, as circulatory neutrophils exhibit increased spontaneous NET formation, mainly driven by G-CSF and finely modulated by sex hormones. The postpartum period, on the other hand, involves physiological alterations consistent with the need for protection against infections and fatal haemorrhage. Our findings indicate that all relevant serum markers of neutrophil degranulation and NET release are substantially augmented postpartum. Neutrophil pro-NETotic activity in vitro is also upregulated particularly in post-delivery neutrophils. Moreover, maternal puerperal neutrophils exhibit a strong pro-NETotic phenotype, associated with increased levels of all key players in the generation of NETs, namely citH3, MPO, NE, and ROS, compared to non-pregnant and pregnant controls. Intriguingly, post-delivery NET formation is independent of G-CSF in contrast to late gestation and complemented by the presence of TF on the NETs, alterations in the platelet activity status, and activation of the coagulation cascade, triggered by circulating microparticles. Taken together, our results reveal the highly pro-NETotic and potentially procoagulant nature of postpartum neutrophils, bridging an overt immune activation with possible harmful thrombotic incidence.

Circulating mitochondrial DNA copy numbers represent a sensitive marker for diagnosis and monitoring of disease activity in systemic lupus erythematosus.

OBJECTIVES: Cell-free DNA is involved in the pathogenesis of systemic lupus erythematosus (SLE) but the clinical value of cell-free DNA measurements in SLE is unknown. Our aim was therefore to examine the utility of mitochondrial (mt) DNA and nuclear (n) DNA quantification in SLE. METHODS: EDTA plasma was drawn from 103 consecutive patients with SLE and from 56 healthy blood donors. mtDNA and nDNA copy numbers were quantified by PCR from cell-free plasma. Clinical parameters were recorded prospectively. RESULTS: Circulating mtDNA copy numbers were increased 8.8-fold in the plasma of patients with SLE (median 6.6×107 /mL) compared with controls (median 7.6×106 /mL, p<0.0001). Among all 159 individuals, a cut-off set at 1.8×107 mtDNA copies in a receiver operated curve identified patients with SLE with 87.4% sensitivity and 94.6% specificity; the area under the curve was 0.95 (p<0.0001). mtDNA levels were independent of age or gender, but correlated with Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) on multivariable analysis (p=0.004). Conversely, SLEDAI was associated with prednisone dose (p<0.001), anti-double stranded DNA-titres (p=0.003) and mtDNA levels (p=0.005), but not nDNA copy numbers. In 33 patients with SLE with available follow-up, the changes of mtDNA, but not those of nDNA concentrations, robustly correlated with the evolution of the SLEDAI (r=0.55, p=0.001). CONCLUSIONS: Circulating mtDNA unlike nDNA molecules are markedly increased in SLE plasma. Regardless of disease activity, circulating mtDNA levels distinguish patients with SLE from healthy controls with high sensitivity and represent an independent marker of SLE activity.

Patient and healthcare professional eHealth literacy and needs for systemic sclerosis support: a mixed methods study.

OBJECTIVES: We engaged patients with systemic sclerosis (SSc) and healthcare professionals to assess electronic health (eHealth) literacy and needs relating to web-based support using internet-based information and communication technologies (ICT). METHODS: We employed an explanatory sequential mixed methods design. First, we conducted a cross-sectional survey in patients (n=101) and professionals (n=47). Next, we conducted three focus groups with patients, family members and professionals (n=17). RESULTS: Of patients, 89.1% used ICT at least weekly for private communication. Patients reported relatively high comprehension of eHealth information ([Formula: see text] =6.7, 95% CI: 6.2 to 7.3, range 1-10), yet were less confident evaluating information reliability ([Formula: see text] =5.8, 95% CI: 5.1 to 6.4) and finding eHealth apps ([Formula: see text] =4.8, 95% CI: 4.2 to 5.4). Patients and professionals reported little experience with web-based self-management support. Focus groups revealed 'considering non-ICT-accessible groups' and 'fitting patients' and professionals' technology' as crucial for acceptability. In relation to understanding/appraising eHealth, participants highlighted that general SSc information is not tailored to individual's disease course. Recommendations included 'providing timely, understandable and safe information' and 'empowering end-users in ICT and health decision-making skills'. Professionals expressed concerns about lacking resources. Patients were concerned about data security and person-centredness. Key eHealth drivers included 'addressing end-user perceptions' and 'putting people at the centre of technology'. CONCLUSIONS: Patients and professionals need education/training to support uptake of eHealth resources. Key elements include guiding patients to timely/reliable information and using eHealth to optimise patient-provider communication. Design that is responsive to end-users' needs and considers individuals with limited eHealth literacy and/or ICT access appears to be critical for acceptability.

A rare disease patient-reported outcome measure: revision and validation of the German version of the Systemic Sclerosis Quality of Life Questionnaire (SScQoL) using the Rasch model.

BACKGROUND: Rare disease patient-reported outcome measures (PROMs) require linguistic adaptation to overcome the challenge of geographically dispersed patient populations. Importantly, PROMs such as health-related quality of life (HRQoL) should accurately capture responses to patient-identified concerns. The Systemic Sclerosis Quality of Life Questionnaire (SScQoL) is a 29-item tool validated in six languages. Previous evaluation of the German version revealed problems with dichotomous responses. This study aimed to revise the German SScQoL, extend the response structure, and evaluate content and construct validity, reliability and unidimensionality. METHODS: The instrument validation study involved revising the German SScQoL response structure, cognitive debriefing with patients and validation using Rasch analysis. The revised SScQoL was completed by Swiss-German-speaking patients with SSc within the Swiss MANagement Of Systemic Sclerosis (MANOSS) study. Rasch analysis was employed to test the validity, reliability and unidimensionality of the revised instrument. RESULTS: Based on cognitive debriefing with patients (n = 6) dichotomous items were extended to a polytomous 4-point response structure. A total of 78 patients completed the revised SScQoL. Initial analysis of the 29 items suggested the scale lacked fit to the model (χ2 = 51.224, df = 29, p = 0.007). Grouping items into five domains resulted in an adequate fit to the Rasch model (χ2 = 5.343, df = 5, p = 0.376) and unidimensionality (proportion of significant independent t tests: 0.045, 95% CI 0.016-0.114). Overall, the scale was well targeted, had high internal consistency (Person Separation Index, PSI = 0.931) and worked consistently in patients with different demographic and clinical characteristics. CONCLUSIONS: The revised German SScQoL has a 4-point response structure and is a valid, reliable measure. Rasch analysis is useful for validating continuous response structure of quality of life measures. Further evaluation of measurement equivalence with other German-speaking cultures is required for multinational comparisons and data pooling.

Personalized prediction of disease activity in patients with rheumatoid arthritis using an adaptive deep neural network.

BACKGROUND: Deep neural networks learn from former experiences on a large scale and can be used to predict future disease activity as potential clinical decision support. AdaptiveNet is a novel adaptive recurrent neural network optimized to deal with heterogeneous and missing clinical data. OBJECTIVE: We investigate AdaptiveNet for the prediction of individual disease activity in patients from a rheumatoid arthritis (RA) registry. METHODS: Demographic and disease characteristics from over 9500 patients and 65.000 visits from the Swiss Quality Management (SCQM) database were used to train and evaluate the network. Patient characteristics, clinical and patient reported outcomes, laboratory values and medication were used as input features. DAS28-BSR served as a target to predict active RA and future numeric individual disease activity by classification and regression. RESULTS: AdaptiveNet predicted active disease defined as DAS28-BSR >2.6 at the next visit with an overall accuracy of 75.6% (SD +- 0.7%) and a sensitivity and specificity of 84.2% (SD +- 1.6%) and 61.5% (SD +- 3.6%), respectively. Prediction performance was significantly higher in patients with a disease duration >3 years and positive rheumatoid factor. Regression allowed forecasting individual DAS28-BSR values with a mean squared error (MSE) of 0.9 (SD +- 0.05). This corresponds to a 8% deviation between estimated and real DAS28-BSR values. Compared to linear regression, random forest and support vector machines, AdaptiveNet showed an increased performance of over 7% in MSE. Medication played a minor role in the prediction of RA disease activity. CONCLUSION: AdaptiveNet has a superior capacity to predict numeric RA disease activity compared to classical machine learning architectures. All investigated models had limitations in low specificity.

Long-term safety and effectiveness of canakinumab therapy in patients with cryopyrin-associated periodic syndrome: results from the ß-Confident Registry.

OBJECTIVE: To report the long-term safety and effectiveness of canakinumab, a fully human anti-interleukin 1ß monoclonal antibody, in patients with cryopyrin-associated periodic syndromes (CAPS), including familial cold autoinflammatory syndrome (FCAS), Muckle-Wells syndrome (MWS) and neonatal-onset multisystem inflammatory disease (NOMID), in a real-world setting. METHODS: From December 2009 to December 2015, the ß-Confident Registry prospectively enrolled patients with CAPS and non-CAPS conditions who received canakinumab per routine care and were prospectively followed for up to 6 years. The registry protocol did not mandate specific visits or procedures; however, all observed adverse events (AEs) and serious adverse events (SAEs) had to be recorded. Canakinumab effectiveness was evaluated by Physician's Global Assessment (PGA). RESULTS: Of 288 patients enrolled, 3 were excluded due to missing informed consent. Among the remaining 285 patients, 243 (85.3%) were patients with CAPS and 42 (14.7%) had atypical CAPS (6.3%) or other conditions (8.4%). The median age was 26.6 years. Based on PGA, 58 of 123 (47.2%) patients with CAPS had no disease activity at 48 months, and 65 of 123 (52.8%) experienced mild/moderate disease activity at 48 months. Among CAPS phenotypes, AE incidence rates per 100 patient-years were lowest for FCAS (73.1; 95% CI 60.3 to 87.8) compared with those with MWS (105.0; 95% CI 97.2 to 113.2) or NOMID (104.6; 95% CI 86.6 to 125.2). One hundred twenty-eight SAEs were reported in 68 patients with CAPS (incidence rate/100 patient-years, 14.0; 95% CI 11.6 to 16.6). One death (metastatic rectal adenocarcinoma in a patient with MWS) was reported. CONCLUSIONS: The response to canakinumab was sustained for up to 6 years. Canakinumab demonstrated a favourable safety profile over long-term treatment in patients with CAPS. TRIAL REGISTRATION NUMBER: NCT01213641.

Impact of assessing patient-reported outcomes with mobile apps on patient-provider interaction.

OBJECTIVE: To explore the effect of apps measuring patient-reported outcomes (PROs) on patient-provider interaction in the rheumatic diseases in an observational setting. METHODS: Patients in the Swiss Clinical Quality Management in Rheumatic Diseases Registry were offered mobile apps (iDialog and COmPASS) to track disease status between rheumatology visits using validated PROs (Rheumatoid Arthritis Disease Activity Index-5 score, Bath Ankylosing Spondylitis Disease Activity Index score, Routine Assessment of Patient Index Data-3 score and Visual Analogue Scale score for pain, disease activity and skin symptoms). We assessed two aspects of patient-provider interaction: shared decision making (SDM) and physician awareness of disease fluctuations. We used logistic regressions to compare outcomes among patients who (1) used an app and discussed app data with their physician (app+discussion group), (2) used an app without discussing the data (app-only group) or (3) did not use any app (non-app users). RESULTS: 2111 patients were analysed, including 1799 non-app users, 150 app-only users and 162 app+discussion users (43% male; with 902 patients with rheumatoid arthritis, 766 patients with axial spondyloarthritis and 443 patients with psoriatic arthritis). App users were younger than non-app users (mean age of 47 vs 51 years, p<0.001). Compared with non-app users, the app+discussion group rated their rheumatologist more highly in SDM (OR 1.7, 95% CI 1.1 to 2.4) and physician awareness of disease fluctuations (OR 2.0, 95% CI 1.3 to 3.1). This improvement was absent in the app-only group. CONCLUSION: App users who discussed app data with their rheumatologist reported more favourably on patient-provider interactions than app users who did not and non-app users. Apps measuring PROs may contribute little to patient-provider interactions without integration of app data into care processes.

Discontinuation of biologic DMARDs in a real-world population of patients with rheumatoid arthritis in remission: outcome and risk factors.

OBJECTIVES: Data from randomized controlled trials have shown the feasibility of discontinuation of bDMARD therapy in patients with RA that have reached remission. Criteria for selecting patients that are likely to remain in remission are still incompletely defined. We aimed to identify predictors of successful discontinuation of bDMARD therapy in the Swiss Clinical Quality Management (SCQM) registry, a real-world cohort of RA patients. METHODS: RA patients in DAS28-ESR remission who stopped bDMARD/tsDMARD treatment were included. Loss of remission was defined as a DAS28-ESR > 2.6 or restart of a bDMARD/tsDMARD. Time to loss of remission was the main outcome. Kaplan-Meier methods were applied and Cox regression was used for multivariable analyses adjusting for confounding factors. Missing data were imputed using multiple imputation. RESULTS: A total of 318 patients in a bDMARD/tsDMARD-free remission were followed between 1997 and 2017. In total, 241 patients (76%) lost remission after a median time of 0.9 years (95% CI: 0.7, 1.0). The time to loss of remission was shorter in women, in patients with a longer disease duration >4yrs and in patients who did not meet clinical disease activity index (CDAI) remission criteria at baseline. Remission was longer in patients with csDMARD therapy during b/tsDMARD free remission [hazard ratio (HR) 0.8, P =0.05, 95% CI: 0.6, 1.0]. CONCLUSION: In a real-world patient population, the majority of patients who discontinued b/tsDMARD treatment lost remission within <1 year. Our study confirms that fulfilment of more rigorous remission criteria and csDMARD treatment increases the chance of maintaining b/tsDMARD-free remission.

A randomised, open-label trial to assess the optimal treatment strategy in early diffuse cutaneous systemic sclerosis: the UPSIDE study protocol.

INTRODUCTION: Systemic sclerosis (SSc) is a chronic, autoimmune connective tissue disease associated with high morbidity and mortality, especially in diffuse cutaneous SSc (dcSSc). Currently, there are several treatments available in early dcSSc that aim to change the disease course, including immunosuppressive agents and autologous haematopoietic stem cell transplantation (HSCT). HSCT has been adopted in international guidelines and is offered in current clinical care. However, optimal timing and patient selection for HSCT are still unclear. In particular, it is unclear whether HSCT should be positioned as upfront therapy or rescue treatment for patients refractory to immunosuppressive therapy. We hypothesise that upfront HSCT is superior and results in lower toxicity and lower long-term medical costs. Therefore, we propose this randomised trial aiming to determine the optimal treatment strategy for early dcSSc by comparing two strategies used in standard care: (1) upfront autologous HSCT versus (2) immunosuppressive therapy (intravenous cyclophosphamide pulse therapy followed by mycophenolate mofetil) with rescue HSCT in case of treatment failure. METHODS AND ANALYSIS: The UPSIDE (UPfront autologous hematopoietic Stem cell transplantation vs Immunosuppressive medication in early DiffusE cutaneous systemic sclerosis) study is a multicentre, randomised, open-label, controlled trial. In total, 120 patients with early dcSSc will be randomised. The primary outcome is event-free survival at 2 years after randomisation. Secondary outcomes include serious adverse events, functional status and health-related quality of life. We will also evaluate changes in nailfold capillaroscopy pattern, pulmonary function, cardiac MR and high-resolution CT of the chest. Follow-up visits will be scheduled 3-monthly for 2 years and annually in the following 3 years. ETHICS AND DISSEMINATION: The study was approved by the Dutch Central Committee on Research Concerning Human Subjects (NL72607.041.20). The results will be disseminated through patient associations and conventional scientific channels. TRIAL REGISTRATION NUMBERS: NCT04464434; NL 8720.

Natural variability in the disease course of SSc-ILD: implications for treatment.

Interstitial lung disease (ILD) affects approximately 50% of patients with systemic sclerosis (SSc) and is the leading cause of death in SSc. Our objective was to gain insight into the progression of SSc-associated ILD (SSc-ILD). Using data from longitudinal clinical trials and observational studies, we assessed definitions and patterns of progression, risk factors for progression, and implications for treatment. SSc-ILD progression was commonly defined as exceeding specific thresholds of lung function worsening and/or increasing radiographic involvement. One definition used in several studies is decline in forced vital capacity (FVC) of ≥10%, or ≥5-10% plus a decline in diffusing capacity of the lung for carbon monoxide ≥15%. Based on these criteria, 20-30% of patients in observational cohorts develop progressive ILD, starting early in the disease course and progressing at a highly variable rate.Risk factors such as age, FVC, extent of fibrosis and presence of anti-topoisomerase I antibodies can help predict progression of SSc-ILD, though composite risk scores may offer greater predictive power. Whilst the variability of the disease course in SSc-ILD makes risk stratification of patients challenging, the decision to initiate, change or stop treatment should be based on a combination of the current disease state and the speed of progression.