A machine learning model identifies patients in need of autoimmune disease testing using electronic health records.

Systemic autoimmune rheumatic diseases (SARDs) can lead to irreversible damage if left untreated, yet these patients often endure long diagnostic journeys before being diagnosed and treated. Machine learning may help overcome the challenges of diagnosing SARDs and inform clinical decision-making. Here, we developed and tested a machine learning model to identify patients who should receive rheumatological evaluation for SARDs using longitudinal electronic health records of 161,584 individuals from two institutions. The model demonstrated high performance for predicting cases of autoantibody-tested individuals in a validation set, an external test set, and an independent cohort with a broader case definition. This approach identified more individuals for autoantibody testing compared with current clinical standards and a greater proportion of autoantibody carriers among those tested. Diagnoses of SARDs and other autoimmune conditions increased with higher model probabilities. The model detected a need for autoantibody testing and rheumatology encounters up to five years before the test date and assessment date, respectively. Altogether, these findings illustrate that the clinical manifestations of a diverse array of autoimmune conditions are detectable in electronic health records using machine learning, which may help systematize and accelerate autoimmune testing.

Interferon pathway lupus risk alleles modulate risk of death from acute COVID-19.

Type I interferon (IFN) is critical in our defense against viral infections. Increased type I IFN pathway activation is a genetic risk factor for systemic lupus erythematosus (SLE), and a number of common risk alleles contribute to the high IFN trait. We hypothesized that these common gain-of-function IFN pathway alleles may be associated with protection from mortality in acute COVID-19. We studied patients admitted with acute COVID-19 (756 European-American and 398 African-American ancestry). Ancestral backgrounds were analyzed separately, and mortality after acute COVID-19 was the primary outcome. In European-American ancestry, we found that a haplotype of interferon regulatory factor 5 (IRF5) and alleles of protein kinase cGMP-dependent 1 (PRKG1) were associated with mortality from COVID-19. Interestingly, these were much stronger risk factors in younger patients (OR = 29.2 for PRKG1 in ages 45-54). Variants in the IRF7 and IRF8 genes were associated with mortality from COVID-19 in African-American subjects, and these genetic effects were more pronounced in older subjects. Combining genetic information with blood biomarker data such as C-reactive protein, troponin, and D-dimer resulted in significantly improved predictive capacity, and in both ancestral backgrounds the risk genotypes were most relevant in those with positive biomarkers (OR for death between 14 and 111 in high risk genetic/biomarker groups). This study confirms the critical role of the IFN pathway in defense against COVID-19 and viral infections, and supports the idea that some common SLE risk alleles exert protective effects in antiviral immunity.

Increasing Ancestral Diversity in Systemic Lupus Erythematosus Clinical Studies.

Non-White people are more likely to develop systemic lupus erythematosus (SLE) yet are underrepresented in SLE clinical trials. The efficacy and safety of drugs may be influenced by ancestry, and ancestrally diverse study populations are necessary to optimize treatments across the full spectrum of patients. However, barriers to entry into clinical trials are amplified in non-White populations. To address these issues, a conference was held in Bethesda, Maryland, from October 15-16, 2019, entitled "Increasing Ancestral Diversity in Systemic Lupus Erythematosus Clinical Studies: Overcoming the Barriers." Conference participants included people with lupus, lupus physicians, lupus clinical trialists, treatment developers from biotechnology, social scientists, patient advocacy groups, and US government representatives (The Office of Minority Health, Centers for Disease Control and Prevention, National Institutes of Health, and the Food and Drug Administration). For all these groups, the organizers of the conference purposefully included people of non-White ancestry. Decreased participation of non-White SLE patients in clinical research was evaluated through historical, societal, experiential, and pragmatic perspectives, and several interventional programs to increase non-White patient participation in SLE and non-SLE research were described and discussed. The presentations and discussions highlighted the need for changes at the societal, institutional, research team, referring physician, and patient education levels to achieve equitable ancestral representation in SLE clinical studies.

Evaluation of Immune Response and Disease Status in Systemic Lupus Erythematosus Patients Following SARS-CoV-2 Vaccination.

OBJECTIVE: To evaluate seroreactivity and disease flares after COVID-19 vaccination in a multiethnic/multiracial cohort of patients with systemic lupus erythematosus (SLE). METHODS: Ninety SLE patients and 20 healthy controls receiving a complete COVID-19 vaccine regimen were included. IgG seroreactivity to the SARS-CoV-2 spike receptor-binding domain (RBD) and SARS-CoV-2 microneutralization were used to evaluate B cell responses; interferon-γ (IFNγ) production was measured by enzyme-linked immunospot (ELISpot) assay in order to assess T cell responses. Disease activity was measured by the hybrid SLE Disease Activity Index (SLEDAI), and flares were identified according to the Safety of Estrogens in Lupus Erythematosus National Assessment-SLEDAI flare index. RESULTS: Overall, fully vaccinated SLE patients produced significantly lower IgG antibodies against SARS-CoV-2 spike RBD compared to fully vaccinated controls. Twenty-six SLE patients (28.8%) generated an IgG response below that of the lowest control (<100 units/ml). In logistic regression analyses, the use of any immunosuppressant or prednisone and a normal anti-double-stranded DNA antibody level prior to vaccination were associated with decreased vaccine responses. IgG seroreactivity to the SARS-CoV-2 spike RBD strongly correlated with the SARS-CoV-2 microneutralization titers and correlated with antigen-specific IFNγ production determined by ELISpot. In a subset of patients with poor antibody responses, IFNγ production was similarly diminished. Pre- and postvaccination SLEDAI scores were similar in both groups. Postvaccination flares occurred in 11.4% of patients; 1.3% of these were severe. CONCLUSION: In a multiethnic/multiracial study of SLE patients, 29% had a low response to the COVID-19 vaccine which was associated with receiving immunosuppressive therapy. Reassuringly, severe disease flares were rare. While minimal protective levels remain unknown, these data suggest that protocol development is needed to assess the efficacy of booster vaccination.

Interferon pathway lupus risk alleles modulate risk of death from acute COVID-19.

Type I interferon (IFN) is critical in our defense against viral infections. Increased type I IFN pathway activation is a genetic risk factor for systemic lupus erythematosus (SLE), and a number of common risk alleles contribute to the high IFN trait. We hypothesized that these common gain-of-function IFN pathway alleles may be associated with protection from mortality in acute COVID-19. We studied patients admitted with acute COVID-19 (756 European-American and 398 African-American ancestry). Ancestral backgrounds were analyzed separately, and mortality after acute COVID-19 was the primary outcome. In European-American ancestry, we found that a haplotype of interferon regulatory factor 5 (IRF5) and alleles of protein kinase cGMP-dependent 1 (PRKG1) were associated with mortality from COVID-19. Interestingly, these were much stronger risk factors in younger patients (OR=29.2 for PRKG1 in ages 45-54). Variants in the IRF7 and IRF8 genes were associated with mortality from COVID-19 in African-American subjects, and these genetic effects were more pronounced in older subjects. Combining genetic information with blood biomarker data such as C-reactive protein, troponin, and D-dimer resulted in significantly improved predictive capacity, and in both ancestral backgrounds the risk genotypes were most relevant in those with positive biomarkers (OR for death between 14 and 111 in high risk genetic/biomarker groups). This study confirms the critical role of the IFN pathway in defense against COVID-19 and viral infections, and supports the idea that some common SLE risk alleles exert protective effects in anti-viral immunity. BACKGROUND: We find that a number of IFN pathway lupus risk alleles significantly impact mortality following COVID-19 infection. These data support the idea that type I IFN pathway risk alleles for autoimmune disease may persist in high frequency in modern human populations due to a benefit in our defense against viral infections. TRANSLATIONAL SIGNIFICANCE: We develop multivariate prediction models which combine genetics and known biomarkers of severity to result in greatly improved prediction of mortality in acute COVID-19. The specific associated alleles provide some clues about key points in our defense against COVID-19.

Evaluation of SARS-CoV-2 IgG antibody reactivity in patients with systemic lupus erythematosus: analysis of a multi-racial and multi-ethnic cohort.

BACKGROUND: Patients with systemic lupus erythematosus (SLE) are at risk of developing COVID-19 due to underlying immune abnormalities and regular use of immunosuppressant medications. We aimed to evaluate the presence of SARS-CoV-2 IgG antibodies in patients with SLE with or without previous COVID-19-related symptoms or RT-PCR-confirmed SARS-CoV-2 infection. METHODS: For this analysis, we included patients with SLE from two cohorts based in New York City: the Web-based Assessment of Autoimmune, Immune-Mediated and Rheumatic Patients during the COVID-19 pandemic (WARCOV) study; and the NYU Lupus Cohort (a prospective registry of patients at NYU Langone Health and NYC Health + Hospitals/Bellevue). Patients in both cohorts were tested for SARS-CoV-2 IgG antibodies via commercially available immunoassays, processed through hospital or outpatient laboratories. Patients recruited from the NYU Lupus Cohort, referred from affiliated providers, or admitted to hospital with COVID-19 were tested for SARS-CoV-2 IgG antibodies as part of routine surveillance during follow-up clinical visits. FINDINGS: 329 patients with SLE were included in this analysis, 146 from the WARCOV study and 183 from the NYU Lupus Cohort, and were tested for SARS-CoV-2 antibodies between April 29, 2020, and Feb 9, 2021. 309 (94%) were women and 91 (28%) were of Hispanic ethnicity. 51 (16%) of 329 patients had a positive SARS-CoV-2 IgG antibody test. Seropositive patients were more likely than seronegative patients to be Hispanic (24 [47%] of 51 vsz 67 [24%] of 278). Other demographic variables, SLE-specific factors, and immunosuppressant use were not associated with SARS-CoV-2 positivity. Of the 29 patients with COVID-19 previously confirmed by RT-PCR, 18 (62%) were on immunosuppressants; 24 (83%) of 29 patients tested positive for SARS-CoV-2 IgG antibodies. Of 17 patients who had symptoms of COVID-19 but negative concurrent RT-PCR testing, one (6%) developed an antibody response. Of 26 patients who had COVID-19-related symptoms but did not undergo RT-PCR testing, six (23%) developed an antibody response. Of 83 patients who had no symptoms of COVID-19 and no RT-PCR testing, four (5%) developed an antibody response. Among 36 patients who were initially SARS-CoV-2 IgG positive, the majority maintained reactivity serially (88% up to 10 weeks, 83% up to 20 weeks, and 80% up to 30 weeks). Seven (70%) of ten patients with confirmed COVID-19 had antibody positivity beyond 30 weeks from disease onset. INTERPRETATION: Most patients with SLE and confirmed COVID-19 were able to produce and maintain a serological response despite the use of a variety of immunosuppressants, providing reassurance about the efficacy and durability of humoral immunity and possible protection against re-infection with SARS-CoV-2. FUNDING: National Institute of Arthritis and Musculoskeletal and Skin Diseases, National Institutes of Health, and Bloomberg Philanthropies COVID-19 Response Initiative Grant.

Leveraging the United States Epicenter to Provide Insights on COVID-19 in Patients With Systemic Lupus Erythematosus.

OBJECTIVE: To characterize patients with systemic lupus erythematosus (SLE) affected by coronavirus disease 2019 (COVID-19) and to analyze associations of comorbidities and medications on infection outcomes. METHODS: Patients with SLE and reverse transcriptase-polymerase chain reaction-confirmed COVID-19 were identified through an established New York University lupus cohort, query of 2 hospital systems, and referrals from rheumatologists. Data were prospectively collected via a web-based questionnaire and review of medical records. Data on baseline characteristics were obtained for all patients with COVID-19 to analyze risk factors for hospitalization. Data were also collected on asymptomatic patients and those with COVID-19-like symptoms who tested negative or were not tested. Statistical analyses were limited to confirmed COVID-19-positive patients. RESULTS: A total of 226 SLE patients were included: 41 with confirmed COVID-19, 19 who tested negative for COVID-19, 42 with COVID-19-like symptoms who did not get tested, and 124 who remained asymptomatic without testing. Of the SLE patients with confirmed COVID-19, hospitalization was required in 24 (59%) and intensive care unit-level of care in 4, and 4 died. Hospitalized patients tended to be older, nonwhite, Hispanic, have higher body mas index (BMI), history of nephritis, and at least 1 comorbidity. An exploratory (due to limited sample size) logistic regression analysis identified race, presence of at least 1 comorbidity, and BMI as independent predictors of hospitalization. CONCLUSION: In general, the variables predictive of hospitalization in our SLE patients were similar to those identified in the general population. Further studies are needed to understand additional risk factors for poor COVID-19 outcomes in patients with SLE.

ApoL1 and the Immune Response of Patients with Systemic Lupus Erythematosus.

PURPOSE OF REVIEW: Systemic lupus erythematosus (SLE) confers up to a 50-fold increased risk of cardiovascular disease (CVD), and African Americans with SLE experience accelerated damage accrual and doubled cardiovascular risk when compared to their European American counterparts. RECENT FINDINGS: Genome-wide association studies have identified a substantial signal at 22q13, now assigned to variation at apolipoprotein L1 (APOL1), which has associated with progressive nondiabetic nephropathy, cardiovascular disease, and many immune-associated renal diseases, including lupus nephritis. We contend that alterations in crucial APOL1 intracellular pathways may underpin associated disease states based on structure-functional differences between variant and ancestral forms. While ancestral APOL1 may be a key driver of autophagy, nonconserved primary structure changes result in a toxic gain of function with attenuation of autophagy and an unsupervised pore-forming feature. Thus, the divergent intracellular biological pathways of ancestral and variant APOL1 may explain a worsened prognosis as demonstrated in SLE.