Search

IMPORTANCE: Electronic health record textual sources such as medication signeturs (sigs) contain valuable information that is not always available in structured form. Commonly processed through manual annotation, this repetitive and time-consuming task could be fully automated using large language models (LLMs). While most sigs include simple instructions, some include complex patterns. OBJECTIVES: We aimed to compare the performance of GPT-3.5 and GPT-4 with smaller fine-tuned models (ClinicalBERT, BlueBERT) in extracting the average daily dose of 2 immunomodulating medications with frequent complex sigs: hydroxychloroquine, and prednisone. METHODS: Using manually annotated sigs as the gold standard, we compared the performance of these models in 702 hydroxychloroquine and 22 104 prednisone prescriptions. RESULTS: GPT-4 vastly outperformed all other models for this task at any level of in-context learning. With 100 in-context examples, the model correctly annotates 94% of hydroxychloroquine and 95% of prednisone sigs to within 1 significant digit. Error analysis conducted by 2 additional manual annotators on annotator-model disagreements suggests that the vast majority of disagreements are model errors. Many model errors relate to ambiguous sigs on which there was also frequent annotator disagreement. DISCUSSION: Paired with minimal manual annotation, GPT-4 achieved excellent performance for language regression of complex medication sigs and vastly outperforms GPT-3.5, ClinicalBERT, and BlueBERT. However, the number of in-context examples needed to reach maximum performance was similar to GPT-3.5. CONCLUSION: LLMs show great potential to rapidly extract structured data from sigs in no-code fashion for clinical and research applications.

ObjectiveAccurate identification of lupus nephritis (LN) cases is essential for patient management, research and public health initiatives. However, LN diagnosis codes in electronic health records (EHRs) are underused, hindering efficient identification. We investigated the current performance of International Classification of Diseases (ICD) codes, 9th and 10th editions (ICD9/10), for identifying prevalent LN, and developed scoring systems to increase identification of LN that are adaptable to settings with and without LN ICD codes.MethodsTraining and test sets derived from EHR data from a large health system. An external set comprised data from the EHR of a second large health system. Adults with ICD9/10 codes for SLE were included. LN cases were ascertained through manual chart reviews conducted by rheumatologists. Two definitions of LN were used: strict (definite LN) and inclusive (definite, potential or diagnostic uncertainty). Gradient boosting models including structured EHR fields were used for predictor selection. Two logistic regression-based scoring systems were developed ('LN-Code' included LN ICD codes and 'LN-No Code' did not), calibrated and validated using standard performance metrics.ResultsA total of 4152 patients from University of California San Francisco Medical Center and 370 patients from Zuckerberg San Francisco General Hospital and Trauma Center met the eligibility criteria. Mean age was 50 years, 87% were female. LN diagnosis codes demonstrated low sensitivity (43-73%) but high specificity (92-97%). LN-Code achieved an area under the curve (AUC) of 0.93 and a sensitivity of 0.88 for identifying LN using the inclusive definition. LN-No Code reached an AUC of 0.91 and a sensitivity of 0.95 (0.97 for the strict definition). Both scoring systems had good external validity, calibration and performance across racial and ethnic groups.ConclusionsThis study quantified the underutilisation of LN diagnosis codes in EHRs and introduced two adaptable scoring systems to enhance LN identification. Further validation in diverse healthcare settings is essential to ensure their broader applicability.

This paper studies the determinants of net interest margin and of the exposure to the interest rate risk of a sample of 125 local Italian banks during the period 2006-2018. Relative to prior literature, to take advantage of the unprecedented interest rate environment determined by European Central Bank (ECB)'s unconventional monetary policy measures, we consider two sub-periods: 2006-2011 and 2012-2018. Banks' net interest margin increases with the intensity of maturity transformation, with a larger effect in the years 2012-2018, and with their exposure to interest rate risk. At this specific regard, we shed more light than previous studies by distinguishing among three sources of this risk, namely the one referred to the loans issuing and deposits collecting activity, the one stemming from the securities portfolio, and the one associated with derivatives positions. Maturity transformation is associated with an increase in interest rate risk exposure, again with an impact that is stronger over the years 2012-2018. Funding from the ECB is associated with a higher interest rate risk exposure in the years 2006-2011, while it results in a reduction in the second part of the analysed period. We argue that ECB’s (targeted) long-term refinancing operations lead to higher funding stability and strengthen banks’ capacity to withstand potential upward shocks in interest rates. The opposite occurs for the deposits held by our sample bank at the ECB

OBJECTIVE: To describe obstetric outcomes based on COVID-19 vaccination status, in women with rheumatic and musculoskeletal diseases (RMDs) who developed COVID-19 during pregnancy. METHODS: Data regarding pregnant women entered into the COVID-19 Global Rheumatology Alliance registry from 24 March 2020-25 February 2022 were analysed. Obstetric outcomes were stratified by number of COVID-19 vaccine doses received prior to COVID-19 infection in pregnancy. Descriptive differences between groups were tested using the chi-squared or Fishers exact test. RESULTS: There were 73 pregnancies in 73 women with RMD and COVID-19. Overall, 24.7% (18) of pregnancies were ongoing, while of the 55 completed pregnancies, 90.9% (50) of pregnancies resulted in livebirths. At the time of COVID-19 diagnosis, 60.3% (n = 44) of women were unvaccinated, 4.1% (n = 3) had received one vaccine dose while 35.6% (n = 26) had two or more doses. Although 83.6% (n = 61) of women required no treatment for COVID-19, 20.5% (n = 15) required hospital admission. COVID-19 resulted in delivery in 6.8% (n = 3) of unvaccinated women and 3.8% (n = 1) of fully vaccinated women. There was a greater number of preterm births (PTB) in unvaccinated women compared with fully vaccinated 29.5% (n = 13) vs 18.2% (n = 2). CONCLUSIONS: In this descriptive study, unvaccinated pregnant women with RMD and COVID-19 had a greater number of PTB compared with those fully vaccinated against COVID-19. Additionally, the need for COVID-19 pharmacological treatment was uncommon in pregnant women with RMD regardless of vaccination status. These results support active promotion of COVID-19 vaccination in women with RMD who are pregnant or planning a pregnancy.

ObjectiveNon-White populations are at higher risk of developing systemic lupus erythematosus (SLE) and have more severe outcomes, including mortality. The present study was undertaken to examine how specific causes of death vary by race and ethnicity, including Asian and Hispanic individuals.MethodsThe California Lupus Surveillance Project included SLE cases identified among residents of San Francisco County, CA during January 1, 2007 to December 31, 2009. Cases were matched to the National Death Index over a 10-year period. Logistic regression examined age-adjusted differences in causes of death by race, ethnicity, and sex. Age-standardized mortality ratios between individuals with SLE and the corresponding general population were calculated for the leading cause of death, and observed versus expected deaths were estimated.ResultsThe study included 812 individuals of White (38%), Asian (36%), Black (20%), and mixed/other/unknown (5%) race; 15% identified as Hispanic. One hundred thirty-five deaths were recorded, with a mean ± SD age at death of 62.2 ± 15.6 years. Cardiovascular disease (CVD) was the leading cause of death overall (33%), and across all racial and ethnic groups, followed by rheumatic disease (18%) and hematologic/oncologic conditions (18%). CVD as the underlying cause of death was 3.63 times higher among SLE cases than in the general population. CVD deaths for those with SLE were nearly 4 and 6 times higher for Asian and Hispanic individuals with SLE, respectively, compared to the general population.ConclusionIndividuals with SLE experience a disproportionate burden of CVD mortality compared to the general population, which is magnified for Asian and Hispanic groups.

ObjectiveTo describe people with gout who were diagnosed with coronavirus disease 2019 (COVID-19) and hospitalized and to characterize their outcomes.MethodsData on patients with gout hospitalized for COVID-19 between March 12, 2020, and October 25, 2021, were extracted from the COVID-19 Global Rheumatology Alliance registry. Descriptive statistics were used to describe the demographics, comorbidities, medication exposures, and COVID-19 outcomes including oxygenation or ventilation support and death.ResultsOne hundred sixty-three patients with gout who developed COVID-19 and were hospitalized were included. The mean age was 63 years, and 85% were male. The majority of the group lived in the Western Pacific Region (35%) and North America (18%). Nearly half (46%) had two or more comorbidities, with hypertension (56%), cardiovascular disease (28%), diabetes mellitus (26%), chronic kidney disease (25%), and obesity (23%) being the most common. Glucocorticoids and colchicine were used pre-COVID-19 in 11% and 12% of the cohort, respectively. Over two thirds (68%) of the cohort required supplemental oxygen or ventilatory support during hospitalization. COVID-19-related death was reported in 16% of the overall cohort, with 73% of deaths documented in people with two or more comorbidities.ConclusionThis cohort of people with gout and COVID-19 who were hospitalized had high frequencies of ventilatory support and death. This suggests that patients with gout who were hospitalized for COVID-19 may be at risk of poor outcomes, perhaps related to known risk factors for poor outcomes, such as age and presence of comorbidity.

ObjectiveSome patients with rheumatic diseases might be at higher risk for coronavirus disease 2019 (COVID-19) acute respiratory distress syndrome (ARDS). We aimed to develop a prediction model for COVID-19 ARDS in this population and to create a simple risk score calculator for use in clinical settings.MethodsData were derived from the COVID-19 Global Rheumatology Alliance Registry from March 24, 2020, to May 12, 2021. Seven machine learning classifiers were trained on ARDS outcomes using 83 variables obtained at COVID-19 diagnosis. Predictive performance was assessed in a US test set and was validated in patients from four countries with independent registries using area under the curve (AUC), accuracy, sensitivity, and specificity. A simple risk score calculator was developed using a regression model incorporating the most influential predictors from the best performing classifier.ResultsThe study included 8633 patients from 74 countries, of whom 523 (6%) had ARDS. Gradient boosting had the highest mean AUC (0.78; 95% confidence interval [CI]: 0.67-0.88) and was considered the top performing classifier. Ten predictors were identified as key risk factors and were included in a regression model. The regression model that predicted ARDS with 71% (95% CI: 61%-83%) sensitivity in the test set, and with sensitivities ranging from 61% to 80% in countries with independent registries, was used to develop the risk score calculator.ConclusionWe were able to predict ARDS with good sensitivity using information readily available at COVID-19 diagnosis. The proposed risk score calculator has the potential to guide risk stratification for treatments, such as monoclonal antibodies, that have potential to reduce COVID-19 disease progression.

ObjectivesTo investigate factors associated with severe COVID-19 in people with idiopathic inflammatory myopathy (IIM).MethodsDemographic data, clinical characteristics and COVID-19 outcome severity of adults with IIM were obtained from the COVID-19 Global Rheumatology Alliance physician-reported registry. A 3-point ordinal COVID-19 severity scale was defined: (1) no hospitalisation, (2) hospitalisation (and no death) and (3) death. ORs were estimated using multivariable ordinal logistic regression. Sensitivity analyses were performed using a 4-point ordinal scale: (1) no hospitalisation, (2) hospitalisation with no oxygen (and no death), (3) hospitalisation with oxygen/ventilation (and no death) and 4) death.ResultsOf 348 patients, 48% were not hospitalised, 39% were hospitalised (and did not die) and 13% died. Older age (OR=1.59/decade, 95% CI 1.31 to 1.91), high disease activity (OR=3.50, 95% CI 1.25 to 9.83; vs remission), ≥2 comorbidities (OR=2.63, 95% CI 1.39 to 4.98; vs none), prednisolone-equivalent dose >7.5 mg/day (OR=2.40, 95% CI 1.09 to 5.28; vs no intake) and exposure to rituximab (OR=2.71, 95% CI 1.28 to 5.72; vs conventional synthetic disease-modifying antirheumatic drugs only) were independently associated with severe COVID-19. In addition to these variables, in the sensitivity analyses, male sex (OR range: 1.65-1.83; vs female) was also significantly associated with severe outcomes, while COVID-19 diagnosis after 1 October 2020 (OR range: 0.51-0.59; vs on/before 15 June 2020) was significantly associated with less severe outcomes, but these associations were not significant in the main model (OR=1.57, 95% CI 0.95 to 2.59; and OR=0.61, 95% CI 0.37 to 1.00; respectively).ConclusionsThis is the first large registry data on outcomes of COVID-19 in people with IIM. Older age, male sex, higher comorbidity burden, high disease activity, prednisolone-equivalent dose >7.5 mg/day and rituximab exposure were associated with severe COVID-19. These findings will enable risk stratification and inform management decisions for patients with IIM.

BackgroundDifferences in the distribution of individual-level clinical risk factors across regions do not fully explain the observed global disparities in COVID-19 outcomes. We aimed to investigate the associations between environmental and societal factors and country-level variations in mortality attributed to COVID-19 among people with rheumatic disease globally.MethodsIn this observational study, we derived individual-level data on adults (aged 18-99 years) with rheumatic disease and a confirmed status of their highest COVID-19 severity level from the COVID-19 Global Rheumatology Alliance (GRA) registry, collected between March 12, 2020, and Aug 27, 2021. Environmental and societal factors were obtained from publicly available sources. The primary endpoint was mortality attributed to COVID-19. We used a multivariable logistic regression to evaluate independent associations between environmental and societal factors and death, after controlling for individual-level risk factors. We used a series of nested mixed-effects models to establish whether environmental and societal factors sufficiently explained country-level variations in death.Findings14 044 patients from 23 countries were included in the analyses. 10 178 (72·5%) individuals were female and 3866 (27·5%) were male, with a mean age of 54·4 years (SD 15·6). Air pollution (odds ratio 1·10 per 10 μg/m3 [95% CI 1·01-1·17]; p=0·0105), proportion of the population aged 65 years or older (1·19 per 1% increase [1·10-1·30]; p

The increasing importance of treatment of lower urinary tract symptoms (LUTS), while avoiding side effects and maintaining sexual function, has allowed for the development of minimally invasive surgical therapies (MISTs). Recently, the European Association of Urology guidelines reported a paradigm shift from the management of benign prostatic hyperplasia (BPH) to the management of nonneurogenic male LUTS. The aim of the present review was to evaluate the efficacy and safety of the most commonly used MISTs: ablative techniques such as aquablation, prostatic artery embolization, water vapor energy, and transperineal prostate laser ablation, and nonablative techniques such as prostatic urethral lift and temporarily implanted nitinol device (iTIND). MISTs are becoming a new promise, even if clinical trials with longer follow-up are still lacking. Most of them are still under investigation and, to date, only a few options have been given as a recommendation for use. They cannot be considered as standard of care and are not suitable for all patients. Advantages and disadvantages should be underlined, without forgetting our objective: treatment of LUTS and re-treatment avoidance.

Abstract. Background. The urinary retention is recognized as a promoting factor for bladder cancer, but its role as prognostic factor of therapeutic response has not yet been widely considered. To correlate bladder outlet characteristics with short-term response to treatment in non–muscle-invasive bladder cancer. Materials and methods. We carried out a case-control study on 600 consecutive patients with a first diagnosis of non–muscle-invasive bladder cancer, recruited at the first endoscopic follow-up visit after standard treatment: 200 patients were not tumor-free (cases) and 400 were tumor-free (controls). Patients were compared based on baseline and bladder-outlet characteristics ("functional parameters”: postvoid residual volume [PVR], International Prostatic Symptoms Score, perceived quality of bladder outlet). t Test, χ2 test, receiver operating characteristic curves, logistic correlations, and multivariate analysis were applied. Results. The cases had higher statistically significant PVR values compared with controls. We reported a linear correlation of no–tumor-free status with PVR (R2 = 0.087, p < 0.005); the receiver operating characteristic curves revealed an area under the curve of 0.824 (95%confidence interval, 0.783–0.865; optimal PVR cutoff, 50 mL). In the multivariate analysis, age, American Society of Anesthesiologists score ≥2, risk category ≥intermediate, and all functional parameters represented independent factors for no–tumor-free status. Conclusions. Urinary retention could represent a prognostic factor of treatment response, and its active treatment should be considered as an important therapeutic step into the clinical management of bladder cancer patients.

AimTo determine characteristics associated with more severe outcomes in a global registry of people with systemic lupus erythematosus (SLE) and COVID-19.MethodsPeople with SLE and COVID-19 reported in the COVID-19 Global Rheumatology Alliance registry from March 2020 to June 2021 were included. The ordinal outcome was defined as: (1) not hospitalised, (2) hospitalised with no oxygenation, (3) hospitalised with any ventilation or oxygenation and (4) death. A multivariable ordinal logistic regression model was constructed to assess the relationship between COVID-19 severity and demographic characteristics, comorbidities, medications and disease activity.ResultsA total of 1606 people with SLE were included. In the multivariable model, older age (OR 1.03, 95% CI 1.02 to 1.04), male sex (1.50, 1.01 to 2.23), prednisone dose (1-5 mg/day 1.86, 1.20 to 2.66, 6-9 mg/day 2.47, 1.24 to 4.86 and ≥10 mg/day 1.95, 1.27 to 2.99), no current treatment (1.80, 1.17 to 2.75), comorbidities (eg, kidney disease 3.51, 2.42 to 5.09, cardiovascular disease/hypertension 1.69, 1.25 to 2.29) and moderate or high SLE disease activity (vs remission; 1.61, 1.02 to 2.54 and 3.94, 2.11 to 7.34, respectively) were associated with more severe outcomes. In age-adjusted and sex-adjusted models, mycophenolate, rituximab and cyclophosphamide were associated with worse outcomes compared with hydroxychloroquine; outcomes were more favourable with methotrexate and belimumab.ConclusionsMore severe COVID-19 outcomes in individuals with SLE are largely driven by demographic factors, comorbidities and untreated or active SLE. Patients using glucocorticoids also experienced more severe outcomes.

IntroductionThis review aims to analyze the existing literature on local recurrence (LR) in patients undergoing partial nephrectomy (PN) for renal cell carcinoma, identifying relative risk factors, and exploring optimal clinical management strategies.MethodsA comprehensive literature search was conducted across bibliographic databases, primarily focusing on LR rates. Secondary outcomes included evaluation of positive surgical margins (PSM), nephrometry scores, pathological stage (T and grading), perioperative outcomes, time-to-LR, overall survival, and cancer-specific survival.ResultsDue to the heterogeneity, a narrative synthesis was performed. LR rates after PN varied in the literature; with PSM emerging as a significant risk factor. Other LR risk factors included pathological stage, nephrometry scores, and histological variants. However, evidence regarding optimal LR management in the absence of precise indications was lacking.ConclusionLR represents a significant clinical challenge; requiring multidisciplinary assessment and shared decision-making with patients. Given well-established risk factors, clinicians must tailor management strategies to optimize patient outcomes.

Objective Accurate identification of lupus nephritis (LN) cases is essential for patient management, research and public health initiatives. However, LN diagnosis codes in electronic health records (EHRs) are underused, hindering efficient identification. We investigated the current performance of International Classification of Diseases (ICD) codes, 9th and 10th editions (ICD9/10), for identifying prevalent LN, and developed scoring systems to increase identification of LN that are adaptable to settings with and without LN ICD codes.Methods Training and test sets derived from EHR data from a large health system. An external set comprised data from the EHR of a second large health system. Adults with ICD9/10 codes for SLE were included. LN cases were ascertained through manual chart reviews conducted by rheumatologists. Two definitions of LN were used: strict (definite LN) and inclusive (definite, potential or diagnostic uncertainty). Gradient boosting models including structured EHR fields were used for predictor selection. Two logistic regression-based scoring systems were developed (‘LN-Code’ included LN ICD codes and ‘LN-No Code’ did not), calibrated and validated using standard performance metrics.Results A total of 4152 patients from University of California San Francisco Medical Center and 370 patients from Zuckerberg San Francisco General Hospital and Trauma Center met the eligibility criteria. Mean age was 50 years, 87% were female. LN diagnosis codes demonstrated low sensitivity (43–73%) but high specificity (92–97%). LN-Code achieved an area under the curve (AUC) of 0.93 and a sensitivity of 0.88 for identifying LN using the inclusive definition. LN-No Code reached an AUC of 0.91 and a sensitivity of 0.95 (0.97 for the strict definition). Both scoring systems had good external validity, calibration and performance across racial and ethnic groups.Conclusions This study quantified the underutilisation of LN diagnosis codes in EHRs and introduced two adaptable scoring systems to enhance LN identification. Further validation in diverse healthcare settings is essential to ensure their broader applicability.

Objectives To investigate the time course of medication adherence and some of the factors involved in this process in undocumented migrants with chronic diseases.Design Retrospective cohort study.Setting A big non-governmental organisation in Milano, Italy, giving medical assistance to undocumented migrants.Participants 1918 patients, 998 females and 920 males, with at least one chronic condition (diabetes, cardiovascular diseases (CVDs), mental health disorders) seen over a period of 10 years (2011–2020). Their mean age was 49.2±13 years.Results Adherence to medications decreased over 1 year in all patients. This was more evident during the first 2 months of treatment. Patients on only one medication were less adherent than those on more than one medication; at 6 months the percentage of patients with high adherence was 33% vs 57% (p

ObjectiveUsing the American College of Rheumatology Rheumatology Informatics System for Effectiveness (RISE) registry, our objective was to examine performance on rheumatoid arthritis (RA) quality measures and to assess the association between practice characteristics and changes in performance over time among participating practices.MethodsWe analyzed data from practices enrolled in RISE between January 1, 2015 and December 31, 2017. Eight quality measures in the areas of RA disease management, cardiovascular risk reduction, and patient safety were examined. Variability in performance was evaluated at the practice level. Multivariate linear models were used to predict change in measure performance by year and to determine the effect of practice characteristics on change in performance over time.ResultsData from 59,986 patients from 54 practices were examined. The mean ± SD age was 62 ± 14 years, 77% were female, 69% were Caucasian, and most patients were seen in a single-specialty group practice (46%). The average performance on measures related to RA treatments was consistently high (>90%) across the study period. Measures related to RA functional status and disease activity assessment had the greatest improvements over time (8.4% and 13.0% increase per year, respectively; P

BackgroundParkinson's disease (PD) is a neurodegenerative movement disorder affecting 1-5% of the general population for which neither effective cure nor early diagnostic tools are available that could tackle the pathology in the early phase. Here we report a multi-stage procedure to identify candidate genes likely involved in the etiopathogenesis of PD.MethodsThe study includes a discovery stage based on the analysis of whole exome data from 26 dominant late onset PD families, a validation analysis performed on 1542 independent PD patients and 706 controls from different cohorts and the assessment of polygenic variants load in the Italian cohort (394 unrelated patients and 203 controls).ResultsFamily-based approach identified 28 disrupting variants in 26 candidate genes for PD including PARK2, PINK1, DJ-1(PARK7), LRRK2, HTRA2, FBXO7, EIF4G1, DNAJC6, DNAJC13, SNCAIP, AIMP2, CHMP1A, GIPC1, HMOX2, HSPA8, IMMT, KIF21B, KIF24, MAN2C1, RHOT2, SLC25A39, SPTBN1, TMEM175, TOMM22, TVP23A and ZSCAN21. Sixteen of them have not been associated to PD before, were expressed in mesencephalon and were involved in pathways potentially deregulated in PD. Mutation analysis in independent cohorts disclosed a significant excess of highly deleterious variants in cases (p = 0.0001), supporting their role in PD. Moreover, we demonstrated that the co-inheritance of multiple rare variants (≥ 2) in the 26 genes may predict PD occurrence in about 20% of patients, both familial and sporadic cases, with high specificity (> 93%; p = 4.4 × 10- 5). Moreover, our data highlight the fact that the genetic landmarks of late onset PD does not systematically differ between sporadic and familial forms, especially in the case of small nuclear families and underline the importance of rare variants in the genetics of sporadic PD. Furthermore, patients carrying multiple rare variants showed higher risk of manifesting dyskinesia induced by levodopa treatment.ConclusionsBesides confirming the extreme genetic heterogeneity of PD, these data provide novel insights into the genetic of the disease and may be relevant for its prediction, diagnosis and treatment.

BackgroundPatients with primary systemic vasculitis or polymyalgia rheumatica might be at a high risk for poor COVID-19 outcomes due to the treatments used, the potential organ damage cause by primary systemic vasculitis, and the demographic factors associated with these conditions. We therefore aimed to investigate factors associated with COVID-19 outcomes in patients with primary systemic vasculitis or polymyalgia rheumatica.MethodsIn this retrospective cohort study, adult patients (aged ≥18 years) diagnosed with COVID-19 between March 12, 2020, and April 12, 2021, who had a history of primary systemic vasculitis (antineutrophil cytoplasmic antibody [ANCA]-associated vasculitis, giant cell arteritis, Behçet's syndrome, or other vasculitis) or polymyalgia rheumatica, and were reported to the COVID-19 Global Rheumatology Alliance registry were included. To assess COVID-19 outcomes in patients, we used an ordinal COVID-19 severity scale, defined as: (1) no hospitalisation; (2) hospitalisation without supplemental oxygen; (3) hospitalisation with any supplemental oxygen or ventilation; or (4) death. Multivariable ordinal logistic regression analyses were used to estimate odds ratios (ORs), adjusting for age, sex, time period, number of comorbidities, smoking status, obesity, glucocorticoid use, disease activity, region, and medication category. Analyses were also stratified by type of rheumatic disease.FindingsOf 1202 eligible patients identified in the registry, 733 (61·0%) were women and 469 (39·0%) were men, and their mean age was 63·8 years (SD 17·1). A total of 374 (31·1%) patients had polymyalgia rheumatica, 353 (29·4%) had ANCA-associated vasculitis, 183 (15·2%) had giant cell arteritis, 112 (9·3%) had Behçet's syndrome, and 180 (15·0%) had other vasculitis. Of 1020 (84·9%) patients with outcome data, 512 (50·2%) were not hospitalised, 114 (11·2%) were hospitalised and did not receive supplemental oxygen, 239 (23·4%) were hospitalised and received ventilation or supplemental oxygen, and 155 (15·2%) died. A higher odds of poor COVID-19 outcomes were observed in patients who were older (per each additional decade of life OR 1·44 [95% CI 1·31-1·57]), were male compared with female (1·38 [1·05-1·80]), had more comorbidities (per each additional comorbidity 1·39 [1·23-1·58]), were taking 10 mg/day or more of prednisolone compared with none (2·14 [1·50-3·04]), or had moderate, or high or severe disease activity compared with those who had disease remission or low disease activity (2·12 [1·49-3·02]). Risk factors varied among different disease subtypes.InterpretationAmong patients with primary systemic vasculitis and polymyalgia rheumatica, severe COVID-19 outcomes were associated with variable and largely unmodifiable risk factors, such as age, sex, and number of comorbidities, as well as treatments, including high-dose glucocorticoids. Our results could be used to inform mitigation strategies for patients with these diseases.FundingAmerican College of Rheumatology and the European Alliance of Associations for Rheumatology.