Identifying antinuclear antibody positive individuals at risk for developing systemic autoimmune disease: development and validation of a real-time risk model.

Objective: Positive antinuclear antibodies (ANAs) cause diagnostic dilemmas for clinicians. Currently, no tools exist to help clinicians interpret the significance of a positive ANA in individuals without diagnosed autoimmune diseases. We developed and validated a risk model to predict risk of developing autoimmune disease in positive ANA individuals. Methods: Using a de-identified electronic health record (EHR), we randomly chart reviewed 2,000 positive ANA individuals to determine if a systemic autoimmune disease was diagnosed by a rheumatologist. A priori, we considered demographics, billing codes for autoimmune disease-related symptoms, and laboratory values as variables for the risk model. We performed logistic regression and machine learning models using training and validation samples. Results: We assembled training (n = 1030) and validation (n = 449) sets. Positive ANA individuals who were younger, female, had a higher titer ANA, higher platelet count, disease-specific autoantibodies, and more billing codes related to symptoms of autoimmune diseases were all more likely to develop autoimmune diseases. The most important variables included having a disease-specific autoantibody, number of billing codes for autoimmune disease-related symptoms, and platelet count. In the logistic regression model, AUC was 0.83 (95% CI 0.79-0.86) in the training set and 0.75 (95% CI 0.68-0.81) in the validation set. Conclusion: We developed and validated a risk model that predicts risk for developing systemic autoimmune diseases and can be deployed easily within the EHR. The model can risk stratify positive ANA individuals to ensure high-risk individuals receive urgent rheumatology referrals while reassuring low-risk individuals and reducing unnecessary referrals.

Comparison of late-onset and non-late-onset systemic lupus erythematosus individuals in a real-world electronic health record cohort.

Objective: Late-onset systemic lupus erythematosus (LO-SLE) is defined as SLE diagnosed at age 50 years or later. Current studies on LO-SLE are small and have conflicting results.Methods: Using a large, electronic health record (EHR)-based cohort of SLE individuals, we compared demographics, disease characteristics, SLE-specific antibodies, and medication prescribing practices in LO (n = 123) vs. NLO-SLE (n = 402) individuals.Results: The median age (interquartile range) at SLE diagnosis was 60 (56-67) years for LO-SLE and 28 (20-38) years for NLO-SLE. Both groups were predominantly female (85% vs. 91%, p = 0.10). LO-SLE individuals were more likely to be White than NLO-SLE individuals (74% vs. 60%, p = 0.005) and less likely to have positive dsDNA (39% vs. 58%, p = 0.001) and RNP (17% vs. 32%, p = 0.02) with no differences in Smith, SSA, and SSB. Autoantibody positivity declined with increasing age at SLE diagnosis. LO-SLE individuals were less likely to develop SLE nephritis (9% vs. 29%, p < 0.001) and less likely to be prescribed multiple classes of SLE medications including antimalarials (90% vs. 95%, p = 0.04), azathioprine (17% vs. 31%, p = 0.002), mycophenolate mofetil (12% vs. 38%, p < 0.001), and belimumab (2% vs. 8%, p = 0.02).Conclusion: LO-SLE individuals may be less likely to fit an expected course for SLE with less frequent positive autoantibodies at diagnosis and lower rates of nephritis, even after adjusting for race. Understanding how age impacts SLE disease presentation could help reduce diagnostic delays in SLE.

Natural language processing to identify lupus nephritis phenotype in electronic health records.

BACKGROUND: Systemic lupus erythematosus (SLE) is a rare autoimmune disorder characterized by an unpredictable course of flares and remission with diverse manifestations. Lupus nephritis, one of the major disease manifestations of SLE for organ damage and mortality, is a key component of lupus classification criteria. Accurately identifying lupus nephritis in electronic health records (EHRs) would therefore benefit large cohort observational studies and clinical trials where characterization of the patient population is critical for recruitment, study design, and analysis. Lupus nephritis can be recognized through procedure codes and structured data, such as laboratory tests. However, other critical information documenting lupus nephritis, such as histologic reports from kidney biopsies and prior medical history narratives, require sophisticated text processing to mine information from pathology reports and clinical notes. In this study, we developed algorithms to identify lupus nephritis with and without natural language processing (NLP) using EHR data from the Northwestern Medicine Enterprise Data Warehouse (NMEDW). METHODS: We developed five algorithms: a rule-based algorithm using only structured data (baseline algorithm) and four algorithms using different NLP models. The first NLP model applied simple regular expression for keywords search combined with structured data. The other three NLP models were based on regularized logistic regression and used different sets of features including positive mention of concept unique identifiers (CUIs), number of appearances of CUIs, and a mixture of three components (i.e. a curated list of CUIs, regular expression concepts, structured data) respectively. The baseline algorithm and the best performing NLP algorithm were externally validated on a dataset from Vanderbilt University Medical Center (VUMC). RESULTS: Our best performing NLP model incorporated features from both structured data, regular expression concepts, and mapped concept unique identifiers (CUIs) and showed improved F measure in both the NMEDW (0.41 vs 0.79) and VUMC (0.52 vs 0.93) datasets compared to the baseline lupus nephritis algorithm. CONCLUSION: Our NLP MetaMap mixed model improved the F-measure greatly compared to the structured data only algorithm in both internal and external validation datasets. The NLP algorithms can serve as powerful tools to accurately identify lupus nephritis phenotype in EHR for clinical research and better targeted therapies.

Implications of the accuracy of diagnostic algorithms for systemic lupus on our understanding of racial disparities in pregnancy outcomes.

OBJECTIVE: Disparities in pregnancy outcomes among women with SLE remain understudied, with few available racially diverse datasets. We sought to identify disparities between Black and White women in pregnancy outcomes within academic institutions in the United States. METHODS: Using the Common Data Model electronic medical record (EMR)-based datasets within the Carolinas Collaborative, we identified women with pregnancy delivery data (2014-2019) and ≥1 SLE International Classification of Diseases 9 or 10 code (ICD9/10) code. From this dataset, we identified four cohorts of SLE pregnancies, three based on EMR-based algorithms and one confirmed with chart review. We compared the pregnancy outcomes identified in each of these cohorts for Black and White women. RESULTS: Of 172 pregnancies in women with ≥1 SLE ICD9/10 code, 49% had confirmed SLE. Adverse pregnancy outcomes occurred in 40% of pregnancies in women with ≥1 ICD9/10 SLE code and 52% of pregnancies with confirmed SLE. SLE was frequently over-diagnosed in women who were White, resulting in 40-75% lower rates of adverse pregnancy outcomes in EMR-derived vs confirmed SLE cohorts. Over-diagnosis was less common for Black women with pregnancy outcomes 12-20% lower in EMR-derived vs confirmed SLE cohorts. Black women had higher rates of adverse pregnancy outcomes than White women in the EMR-derived, but not the confirmed cohorts. CONCLUSION: EMR-derived cohorts of pregnancies in women who are Black, but not White, provided accurate estimations of pregnancy outcomes. The data from the confirmed SLE pregnancies suggest that all women with SLE, regardless of race, referred to academic centres remain at very high risk for adverse pregnancy outcome.

Next-generation phenotyping: introducing phecodeX for enhanced discovery research in medical phenomics.

MOTIVATION: Phecodes are widely used and easily adapted phenotypes based on International Classification of Diseases codes. The current version of phecodes (v1.2) was designed primarily to study common/complex diseases diagnosed in adults; however, there are numerous limitations in the codes and their structure. RESULTS: Here, we present phecodeX, an expanded version of phecodes with a revised structure and 1,761 new codes. PhecodeX adds granularity to phenotypes in key disease domains that are under-represented in the current phecode structure-including infectious disease, pregnancy, congenital anomalies, and neonatology-and is a more robust representation of the medical phenome for global use in discovery research. AVAILABILITY AND IMPLEMENTATION: phecodeX is available at https://github.com/PheWAS/phecodeX.

Development of American College of Rheumatology Quality Measures for Systemic Lupus Erythematosus: A Modified Delphi Process With Rheumatology Informatics System for Effectiveness (RISE) Registry Data Review.

OBJECTIVE: We aimed to develop readily measurable digital quality measure statements for clinical care in systemic lupus erythematosus (SLE) using a multistep process guided by consensus methods. METHODS: Using a modified Delphi process, an American College of Rheumatology (ACR) workgroup of SLE experts reviewed all North American and European guidelines from 2000 to 2020 on treatment, monitoring, and phenotyping of patients with lupus. Workgroup members extracted quality constructs from guidelines, rated these by importance and feasibility, and generated evidence-based quality measure statements. The ACR Rheumatology Informatics System for Effectiveness (RISE) Registry was queried for measurement data availability. In 3 consecutive Delphi sessions, a multidisciplinary Delphi panel voted on the importance and feasibility of each statement. Proposed measures with consensus on feasibility and importance were ranked to identify the top 3 measures. RESULTS: Review of guidelines and distillation of 57 quality constructs resulted in 15 quality measure statements. Among these, 5 met high consensus for importance and feasibility, including 2 on treatment and 3 on laboratory monitoring measures. The 3 highest-ranked statements were recommended for further measure specification as SLE digital quality measures: 1) hydroxychloroquine use, 2) limiting glucocorticoid use >7.5 mg/day to <6 months, and 3) end-organ monitoring of kidney function and urine protein excretion at least every 6 months. CONCLUSION: The Delphi process selected 3 quality measures for SLE care on hydroxychloroquine, glucocorticoid reduction, and kidney monitoring. Next, measures will undergo specification and validity testing in RISE and US rheumatology practices as the foundation for national implementation and use in quality improvement programs.

Phenotype Risk Score but Not Genetic Risk Score Aids in Identifying Individuals With Systemic Lupus Erythematosus in the Electronic Health Record.

OBJECTIVE: Systemic lupus erythematosus (SLE) poses diagnostic challenges. We undertook this study to evaluate the utility of a phenotype risk score (PheRS) and a genetic risk score (GRS) to identify SLE individuals in a real-world setting. METHODS: Using a de-identified electronic health record (EHR) database with an associated DNA biobank, we identified 789 SLE cases and 2,261 controls with available MEGAEX genotyping. A PheRS for SLE was developed using billing codes that captured American College of Rheumatology SLE criteria. We developed a GRS with 58 SLE risk single-nucleotide polymorphisms (SNPs). RESULTS: SLE cases had a significantly higher PheRS (mean ± SD 7.7 ± 8.0 versus 0.8 ± 2.0 in controls; P < 0.001) and GRS (mean ± SD 12.2 ± 2.3 versus 11.0 ± 2.0 in controls; P < 0.001). Black individuals with SLE had a higher PheRS compared to White individuals (mean ± SD 10.0 ± 10.1 versus 7.1 ± 7.2, respectively; P = 0.002) but a lower GRS (mean ± SD 9.0 ± 1.4 versus 12.3 ± 1.7, respectively; P < 0.001). Models predicting SLE that used only the PheRS had an area under the curve (AUC) of 0.87. Adding the GRS to the PheRS resulted in a minimal difference with an AUC of 0.89. On chart review, controls with the highest PheRS and GRS had undiagnosed SLE. CONCLUSION: We developed a SLE PheRS to identify established and undiagnosed SLE individuals. A SLE GRS using known risk SNPs did not add value beyond the PheRS and was of limited utility in Black individuals with SLE. More work is needed to understand the genetic risks of SLE in diverse populations.

Evaluation of hypertension in systemic sclerosis and systemic lupus erythematosus overlap.

Patients with systemic sclerosis and systemic lupus erythematosus serologies present a unique challenge to the clinician when hypertension is detected in the outpatient setting. Treatment choices for non-renal crisis hypertension are different for systemic sclerosis versus systemic lupus erythematosus. Urgent laboratory studies and, in the presence of certain symptoms, imaging assessment are indicated in systemic sclerosis and systemic lupus erythematosus overlap patients with systemic hypertension. Long-term assessment of systemic hypertension may be enhanced by advances in non-contrast imaging that serve as valuable biomarkers for progressive vasculopathy. In this review, the diagnostic approach to systemic sclerosis and systemic lupus erythematosus overlap patients presenting with hypertension is discussed.

Developing electronic health record algorithms that accurately identify patients with juvenile idiopathic arthritis.

BACKGROUND: The objective of this study was to develop an algorithm that accurately identifies juvenile idiopathic arthritis (JIA) patients in the electronic health record (EHR). METHODS: Algorithms were developed in a de-identified EHR by searching for a priori JIA ICD-9 (International Classification of Diseases, Ninth Revision) and ICD-10-CM (International Classification of Diseases, Tenth Revision, Clinical Modification) codes and JIA-related keywords. Exclusion criteria were selected to remove other autoimmune diseases. A training set of 200 patients was randomly selected from patients containing ≥1 occurrence of a JIA ICD-9 or ICD-10-CM code. Case status was determined by a rheumatology clinic note documenting a JIA diagnosis before age 20. For each algorithm, positive predictive value (PPV), sensitivity, and F-measure were determined using the training set. RESULTS: We developed 103 algorithms using combinations of ICD codes, keywords, and exclusion criteria. The algorithm requiring 4 or more counts of JIA ICD-9 or ICD-10-CM codes, keywords "enthesitis" and "uveitis", and exclusion of ICD-9 or ICD-10-CM codes for systemic lupus erythematosus, dermatomyositis, polymyositis, and dermatopolymyositis had the highest PPV of 97% in the training set with an F-measure of 87%. There were 1,131 JIA cases returned by this algorithm. We validated the highest performing algorithm in a separate cohort from the training set with a PPV of 92% and an F-measure of 75%. CONCLUSION: We developed and validated JIA EHR algorithms with ICD-9 and ICD-10-CM codes to accurately identify a JIA cohort. Three algorithms achieved PPVs of 97%, each with different algorithm criteria, allowing for users to select an algorithm to best fit their research needs.

Rates of Pneumocystis jirovecii pneumonia and prophylaxis prescribing patterns in a large electronic health record cohort of patients with systemic lupus erythematosus.

Objective No guidelines exist for Pneumocystis jirovecii pneumonia (PJP) prophylaxis in patients with systemic lupus erythematosus (SLE). Limited data are available on incidence of PJP infection and use of PJP prophylaxis. Using a real-world, electronic health record (EHR) cohort, we investigated the frequency of PJP infections as well as patient and provider factors that impacted use and type of PJP prophylaxis. Methods  In a large, de-identified EHR, we identified possible SLE patients using a previously validated algorithm. PJP ICD-9 or ICD-10-CM billing codes and PJP keywords were used to identify possible PJP cases within this SLE cohort. We assessed for PJP prophylaxis prescribing in all SLE patients using keywords and reviewing medication lists for prophylactic agents. Chart review was used to confirm cases of SLE, PJP, and PJP prophylaxis and to obtain data on demographics, comorbidities, and immunosuppressants. Results Of 977 SLE patients, there were only four with confirmed PJP infection. Two of these patients had concurrent Acquired Immunodeficiency Syndrome, and none were on prophylaxis. Of 977 SLE patients, 132 (14%) were prescribed PJP prophylaxis. Of 617 SLE patients ever prescribed immunosuppressants, 128 (21%) were prescribed PJP prophylaxis. Sulfonamides were the most common prophylaxis prescribed (69%), and possible adverse events were documented in 22 out of 117 instances of being placed on a sulfonamide. Patients of younger age, Black race, nephritis, and renal transplant, and on chronic glucocorticoids were all more likely to have PJP prophylaxis prescribed. Patients who were on transplant induction medications, calcineurin/mTOR inhibitors, cyclophosphamide, and mycophenolate mofetil all were more likely to be prescribed PJP prophylaxis compared to other immunosuppressants. Conclusion PJP is a rare diagnosis among SLE patients, and prior studies may even overestimate its prevalence. PJP prophylaxis was less common in our cohort than previously described. Adverse events related to sulfonamides used for PJP prophylaxis were relatively rare with lower rates than previously reported. Our study demonstrates real-world PJP prophylaxis prescribing patterns in a large cohort of SLE patients.

Systemic Lupus Erythematosus Delivery Outcomes Are Unchanged Across Three Decades.

OBJECTIVE: Using a large, de-identified electronic health record database with over 3.2 million patients, we aimed to identify trends of systemic lupus erythematosus (SLE) medication use during pregnancy and birth outcomes from 1989 to 2020. METHODS: Using a previously validated algorithm for SLE deliveries, we identified 255 pregnancies in patients with SLE and 604 pregnancies in controls with no known autoimmune diseases. We examined demographics, medications, SLE comorbidities, and maternal and fetal outcomes in SLE and control deliveries. RESULTS: Compared with control deliveries, SLE deliveries were more likely to be complicated by preterm delivery (odds ratio [OR]: 6.71; 95% confidence interval [CI]: 4.31-10.55; P < 0.001) and preeclampsia (OR: 3.22; 95% CI: 1.83-5.66; P < 0.001) after adjusting for age at delivery, race, and parity. In a longitudinal analysis, medication use during SLE pregnancies remained relatively stable, with some increased use of hydroxychloroquine over time but no increase in aspirin use. For SLE deliveries, preterm delivery and preeclampsia rates remained stable. CONCLUSION: We observed rates of preeclampsia and preterm delivery in SLE that were five times higher than the general population and higher compared with other prospective SLE cohorts. Furthermore, we did not observe improved outcomes over time with preeclampsia and preterm delivery. Despite increasing evidence for universal use of hydroxychloroquine and aspirin, we did not observe substantially higher use of these medications over time, particularly for aspirin. Our results demonstrate the continued need to prioritize educational and implementation efforts to improve adverse pregnancy outcomes in SLE.

Women with Rheumatoid Arthritis have similar rates of postpartum maternal outcomes compared to women without autoimmune disease.

OBJECTIVE: Limited data exist on the effect of rheumatoid arthritis (RA) on maternal postpartum outcomes. Using a real-world, electronic health record (EHR) cohort, we assessed maternal postpartum outcomes in RA. METHODS: In a large, de-identified EHR, we identified possible RA deliveries using ≥1 delivery ICD-9 or ICD-10-CM codes and a validated RA algorithm. RA cases were required to be diagnosed by a rheumatologist on chart review. Maternal postpartum outcomes included rates of blood transfusion, rates of infection up to 6 weeks postpartum defined by a clinician, and length of hospital stay. We also identified deliveries to women without autoimmune diseases. RESULTS: We identified 202 deliveries occurring after RA diagnosis and 596 deliveries to controls without autoimmune diseases. Postpartum infection rates were similar among RA patients and controls (8% vs. 4%, p = 0.10), as were red blood cell transfusion rates (2% vs. 2%, p = 1.00). RA case status was not significantly associated with postpartum infection (OR = 2.10, 95% CI 0.88 - 4.98, p = 0.09) but was significantly associated with preterm birth (OR = 2.11, 95% CI 1.38 - 3.23, p = 0.001). Corticosteroid use during pregnancy was common at 41%, while tumor necrosis factor inhibitor use was 13%. After adjusting for age at delivery and race, corticosteroid use at delivery was not associated with postpartum maternal infections but was associated with a significantly lower birthweight in RA cases. CONCLUSION: Women with RA have an increased risk of adverse pregnancy outcomes, particularly preterm birth. Our study highlights, however, that maternal postpartum outcomes such as postpartum infection and blood transfusion are not significantly increased in RA patients.

Developing and Validating Methods to Assemble Systemic Lupus Erythematosus Births in the Electronic Health Record.

OBJECTIVE: Electronic health records (EHRs) represent powerful tools to study rare diseases. Our objective was to develop and validate EHR algorithms to identify systemic lupus erythematosus (SLE) births across centers. METHODS: We developed algorithms in a training set using an EHR with over 3 million subjects and validated the algorithms at 2 other centers. Subjects at all 3 centers were selected using ≥1 code for SLE International Classification of Diseases, Ninth Revision (ICD-9) or SLE International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, Clinical Modification (ICD-10-CM) and ≥1 ICD-9 or ICD-10-CM delivery code. A subject was a case if diagnosed with SLE by a rheumatologist and had a birth documented. We tested algorithms using SLE ICD-9 or ICD-10-CM codes, antimalarial use, a positive antinuclear antibody ≥1:160, and ever checked double-stranded DNA or complement, using both rule-based and machine learning methods. Positive predictive values (PPVs) and sensitivities were calculated. We assessed the impact of case definition, coding provider, and subject race on algorithm performance. RESULTS: Algorithms performed similarly across all 3 centers. Increasing the number of SLE codes, adding clinical data, and having a rheumatologist use the SLE code all increased the likelihood of identifying true SLE patients. All the algorithms had higher PPVs in African American versus White SLE births. Using machine learning methods, the total number of SLE codes and an SLE code from a rheumatologist were the most important variables in the model for SLE case status. CONCLUSION: We developed and validated algorithms that use multiple types of data to identify SLE births in the EHR. Algorithms performed better in African American mothers than in White mothers.

Anti-tRNA synthetase syndrome interstitial lung disease: A single center experience.

BACKGROUND: Recognition of Anti-tRNA synthetase (ARS) related interstitial lung disease (ILD) is key to ensuring patients have prompt access to immunosuppressive therapies. The purpose of this retrospective cohort study was to identify factors that may delay recognition of ARS-ILD. METHODS: Patients seen at Vanderbilt University Medical Center between 9/17/2017-10/31/2018 were included in this observational cohort. Clinical and laboratory features were obtained via chart abstraction. Kruskal-Wallis ANOVA, Mann-Whitney U, and Fisher's exact t tests were utilized to determine statistical significance. RESULTS: Patients with ARS were found to have ILD in 51.9% of cases, which was comparable to the frequency of ILD in systemic sclerosis (59.5%). The severity of FVC reduction in ARS (53.2%) was comparable to diffuse cutaneous systemic sclerosis (56.8%, p = 0.48) and greater than dermatomyositis (66.9%, p = 0.005) or limited cutaneous systemic sclerosis (71.8%, p = 0.005). Frank honeycombing was seen with ARS antibodies but not other myositis autoantibodies. ARS patients were more likely to first present to a pulmonary provider in a tertiary care setting (53.6%), likely due to fewer extrapulmonary manifestations. Only 33% of ARS-ILD were anti-nuclear antibody, rheumatoid factor, or anti-cyclic citrullinated peptide positive. Patients with ARS-ILD had a two-fold longer median time to diagnosis compared to other myositis-ILD patients (11.0 months, IQR 8.5-43 months vs. 5.0 months, IQR 3.0-9.0 months, p = 0.003). CONCLUSIONS: ARS patients without prominent extra-pulmonary manifestations are at high risk for not being recognized as having a connective tissue disease related ILD and miscategorized as usual interstitial pneumonia/idiopathic pulmonary fibrosis without comprehensive serologies.

CD19 + CD21lo/neg cells are increased in systemic sclerosis-associated interstitial lung disease.

Interstitial lung disease (ILD) represents a significant cause of morbidity and mortality in systemic sclerosis (SSc). The purpose of this study was to examine recirculating lymphocytes from SSc patients for potential biomarkers of interstitial lung disease (ILD). Peripheral blood mononuclear cells (PBMCs) were isolated from patients with SSc and healthy controls enrolled in the Vanderbilt University Myositis and Scleroderma Treatment Initiative Center cohort between 9/2017-6/2019. Clinical phenotyping was performed by chart abstraction. Immunophenotyping was performed using both mass cytometry and fluorescence cytometry combined with t-distributed stochastic neighbor embedding analysis and traditional biaxial gating. This study included 34 patients with SSc-ILD, 14 patients without SSc-ILD, and 25 healthy controls. CD21lo/neg cells are significantly increased in SSc-ILD but not in SSc without ILD (15.4 ± 13.3% vs. 5.8 ± 0.9%, p = 0.002) or healthy controls (5.0 ± 0.5%, p < 0.0001). While CD21lo/neg B cells can be identified from a single biaxial gate, tSNE analysis reveals that the biaxial gate is comprised of multiple distinct subsets, all of which are increased in SSc-ILD. CD21lo/neg cells in both healthy controls and SSc-ILD are predominantly tBET positive and do not have intracellular CD21. Immunohistochemistry staining demonstrated that CD21lo/neg B cells diffusely infiltrate the lung parenchyma of an SSc-ILD patient. Additional work is needed to validate this biomarker in larger cohorts and longitudinal studies and to understand the role of these cells in SSc-ILD.

Human and Machine Intelligence Together Drive Drug Repurposing in Rare Diseases.

Repurposing is an increasingly attractive method within the field of drug development for its efficiency at identifying new therapeutic opportunities among approved drugs at greatly reduced cost and time of more traditional methods. Repurposing has generated significant interest in the realm of rare disease treatment as an innovative strategy for finding ways to manage these complex conditions. The selection of which agents should be tested in which conditions is currently informed by both human and machine discovery, yet the appropriate balance between these approaches, including the role of artificial intelligence (AI), remains a significant topic of discussion in drug discovery for rare diseases and other conditions. Our drug repurposing team at Vanderbilt University Medical Center synergizes machine learning techniques like phenome-wide association study-a powerful regression method for generating hypotheses about new indications for an approved drug-with the knowledge and creativity of scientific, legal, and clinical domain experts. While our computational approaches generate drug repurposing hits with a high probability of success in a clinical trial, human knowledge remains essential for the hypothesis creation, interpretation, "go-no go" decisions with which machines continue to struggle. Here, we reflect on our experience synergizing AI and human knowledge toward realizable patient outcomes, providing case studies from our portfolio that inform how we balance human knowledge and machine intelligence for drug repurposing in rare disease.

Contribution of viral infection to risk for cancer in systemic lupus erythematosus and multiple sclerosis.

Patients with autoimmune disorders (AD) have altered cancer risks compared to the general population. Systemic lupus erythematosus and multiple sclerosis lead to a heightened risk for hematological malignancies and decreased risk for breast, ovarian, and prostate malignancies. Often patients with autoimmune disease have dysregulated antiviral immune responses, including against oncogenic viruses. To uncover the relationship between viral incidence and cancer risk in the context of autoimmune disease, we extracted electronic health records (EHR) from Vanderbilt University. ICD-9/10 codes and laboratory values were collected for hematological, lung, anal-vaginal, thyroid, hepatobiliary, bladder, prostate, and breast cancers; and viruses including Epstein Barr virus (EBV), Human papilloma virus (HPV), and Hepatitis A/B/C (Hep). Only viral infections that led to a physician visit or laboratory test were entered into the EMR; therefore, only clinically relevant cases were noted and considered positive in this study. The relationship between virus infection and cancer in an SLE cohort (SLE-cases n = 2,313, and SLE-controls n = 5,702) and an MS cohort (MS-case n = 7,277, MS-control n = 7,277) was examined by multilinear logistic regression. Viral infection was strongly associated with increased risk for cancer overall. SLE and MS patients were more susceptible to all viral infections. MS patients trended toward increased risk for cancers overall, while decreased risk for hormone-based cancers in SLE patients non-significantly reduced their risk for overall cancer. Both SLE and MS patients had increased clinically relevant EBV infection, which was associated with risk for hematological cancers. Preventing viral infections by vaccination may be especially helpful in controlling risk for cancer in SLE and MS patients.

Low prevalence of bone mineral density testing in patients with systemic lupus erythematosus and glucocorticoid exposure.

SummaryPatients with systemic lupus erythematosus (SLE) have an increased risk of developing osteoporosis and fractures due to systemic inflammation and glucocorticoids (GCs). Professional organizations recommend bone mineral density (BMD) testing in SLE patients on GCs, especially within 6 months of initiation. Using a validated algorithm, we identified SLE patients in an electronic health record cohort with long-term GC exposure (≥90 days). Our primary outcome was ever BMD testing. We assessed the impact of patient and provider factors on testing. We identified 693 SLE cases with long-term GC exposure, 41% of whom had BMD testing performed. Only 18% of patients had BMD testing within 6 months of GC initiation. In a logistic regression model for BMD testing, male sex (OR = 0.49, 95% CI 0.27 - 0.87, p = 0.01) was associated with being less likely to have BMD testing after adjusting for race and ethnicity. In contrast, older age (OR = 1.04, p < 0.001) and nephritis (OR = 1.83, p = 0.003) were associated with being more likely to have BMD testing after adjusting for race and ethnicity. Bone health in SLE patients remains an area in need of improvement with attention to patients who are younger and male.

Outcomes of acute cardiovascular events in rheumatoid arthritis and systemic lupus erythematosus: a population-based study.

OBJECTIVES: Patients with RA and SLE have an excess cardiovascular risk. We aimed to evaluate outcomes of acute cardiovascular events in these patients. METHODS: Using a nationwide database of Taiwan, we identified adult patients who experienced first-time acute myocardial infarction (n = 191 008), intracranial haemorrhage (n = 169 923) and ischaemic stroke (n = 486 890) over a 13-year period. Odds ratios (ORs) of in-hospital mortality and hazard ratios (HRs) of overall mortality and adverse outcomes during long-term follow-up in relation to RA and SLE were estimated with adjustment for potential confounders. RESULTS: In each cohort, 748, 410 and 1419 patients had established RA; 256, 292 and 622 patients had SLE. Among acute myocardial infarction patients, RA and SLE were associated with in-hospital mortality (RA: OR 1.61, 95% CI 1.33, 1.95; SLE: OR 2.31, 95% CI 1.62, 3.28) and overall mortality. Additionally, RA (HR 1.28, 95% CI 1.18, 1.38) and SLE (HR 1.46, 95% CI 1.27, 1.69) increased the risk of major adverse cardiac events. After intracranial haemorrhage, patients with RA and SLE had higher risks of in-hospital mortality (RA: OR 1.61, 95% CI 1.26, 2.06; SLE: OR 3.00, 95% CI 2.33, 3.86) and overall mortality. After ischaemic stroke, RA and SLE increased in-hospital mortality (RA: OR 1.45, 95% CI 1.15, 1.83; SLE: OR 2.18, 95% CI 1.57, 3.02), overall mortality and recurrent cerebrovascular events (RA: HR 1.10, 95% CI 1.002, 1.21; SLE: HR 1.31, 95% CI 1.14, 1.51), among which ischaemic stroke (HR 1.39, 95% CI 1.19, 1.62) was more likely to recur in SLE patients. CONCLUSION: Both RA and SLE are consistently associated with adverse outcomes following acute cardiovascular events, highlighting the necessity of integrated care for affected patients.

Increased Incidence of Resistant Hypertension in Patients With Systemic Lupus Erythematosus: A Retrospective Cohort Study.

OBJECTIVE: To compare the risk of resistant hypertension (RHTN) in patients with systemic lupus erythematosus (SLE) and in controls without SLE, and to define factors associated with RHTN in patients with SLE. METHODS: We studied 1,044 patients with SLE and 5,241 control subjects using de-identified electronic health records from a tertiary care center. SLE was defined as ≥4 International Classification of Diseases, Ninth Revision codes for SLE and antinuclear antibody titer ≥1:160. RHTN was defined as uncontrolled blood pressure on 3 antihypertensive medications or requiring 4 or more antihypertensives to attain control. First, we compared the risk of RHTN between groups. Second, we examined the association between RHTN and all-cause mortality in patients with SLE. RESULTS: RHTN was nearly twice as prevalent in patients with SLE compared to control subjects (10.2% and 5.3%, respectively), with an incidence rate of 10.2 versus 6.1 cases per 1,000 person-years of observation (hazard ratio [HR] 1.72 [95% confidence interval 1.28-2.30]; P < 0.001, adjusted for age, sex, race, baseline end-stage renal disease [ESRD], creatinine, and calendar year). In patients with SLE, we found associations between RHTN and black race, lower renal function, hypercholesterolemia, and increased inflammatory markers. RHTN was associated with a significantly higher mortality risk (HR 2.91, P = 0.0005) after adjustment for age, sex, race, calendar year, creatinine, baseline ESRD, and number of visits. CONCLUSION: Patients with SLE have a higher risk of RHTN compared to frequency-matched controls, independent of multiple covariates. RHTN is an important comorbidity for clinicians to recognize in SLE, because it is associated with a higher risk of mortality.