POS0656 FRAILTY AND RISK OF ADVERSE OUTCOMES IN BIOLOGIC OR TARGETED-SYNTHETIC DMARD TREATED PATIENTS WITH RHEUMATOID ARTHRITIS

Background

Recently, it has been recognized that frailty and pre-frailty are common in patients with rheumatoid arthritis (RA) [1]. Whether frailty status portends an increased risk of adverse outcomes in patients with RA on biologic or targeted synthetic disease modifying anti-rheumatic drugs (b- or tsDMARDs) remains unknown.

Objectives

To evaluate the association between frailty and adverse outcomes in patients with RA exposed to b- or tsDMARDs.

Methods

Using the IBM/Watson MarketScan Commercial Claims and Encounters Databases, we identified all patients with RA who filled new prescriptions (or received infusions) for TNFα antagonists (TNFi), non-TNFi biologics (rituximab, abatacept, tocilizumab) or Janus Kinase inhibitors (JAKi) between 2008-2019. We used a 1-year lookback period without the use of these drugs to identify new users. The date of the first prescription within these three drug categories was the index date. Patients’ frailty risk score was calculated using the Claims-Based Frailty Index (CFI) [2], which estimates a deficit-accumulation frailty index using International Classification of Diseases codes, Current Procedural Terminology codes, and Healthcare Common Procedure Coding System codes in administrative claims data in the 1-year baseline period. The index ranges from 0 (not at all frail) to 1 (severely frail). The primary outcome was time to serious infections (those requiring hospitalization); secondary outcomes: any infection (outpatient or inpatient encounters) and all-cause hospitalizations.

Patients were followed until 1) outcome occurrence; 2) disenrollment; 3) >90 days elapsed (or >180 days for rituximab) without further fills of the first drug categories; 4) they filled/received infusions of b-/tsDMARDs from a different drug category; or 5) 2 years after index. Cox proportional hazards adjusting for demographics, calendar year, serious and/or opportunistic infections in the 12-months prior to index were used to estimate the adjusted hazard ratios (aHR) and 95% confidence intervals (CIs) for each outcome. In separate model, we additionally adjusted for comorbidity burden, and health care utilization (HCU).

Results

A total of 62,246 patients with RA met our inclusion criteria of whom 50,910 (82%) started TNFi as their first biologic, 9525 (15%) non-TNFi biologics, and 1811 (3%) JAKi. Among these, 3928 (6%) were considered frail. In multivariable analyses, frail patients had higher risk of serious infections compared to non-frail patients (aHR 2.37, 95% CI 2.05-2.74) which decreased to aHR 1.34, 95% CI 1.13-1.58 (Table 1) after adjusting for comorbidity burden and the HCU. Similarly, frailty was associated with increased risk of any infection (aHR 1.18, 95% CI 1.11-1.25), and all-cause hospitalizations (aHR 1.34, 95% CI 1.21-1.49) relative to non-frail individuals.

Table 1.

Multivariable models evaluating the association between frailty status and inpatient infections as the outcome

Variable#Hazard Ratio (95% Confidence Interval)@Hazard Ratio (95% Confidence Interval)Frail2.37 (2.05, 2.74)1.34 (1.13, 1.58)

#Model adjusts for age, sex, major infection requiring inpatient admission in 12 months prior, concomitant baseline drugs such as csDMARDs, glucocorticoids, NSAIDs and opioids

@Model additionally adjusts for Chalrson comorbidy score and healthcare utilization

Conclusion

Frailty is an important predictor for the risk of adverse outcomes among patients with RA treated with b- or tsDMARDs. Our findings underscore the need for considering this parameter in patient evaluations (even among younger patients) in the clinic.

References

[1]Salaffi F et al: Prevalence of frailty and its associated factors in patients with rheumatoid arthritis: a cross-sectional analysis. Clin Rheumatol 2019

[2]Kim DH et al. Validation of a Claims-Based Frailty Index Against Physical Performance and Adverse Health Outcomes in the Health and Retirement Study. J Gerontol A Biol Sci Med Sci 2019

Acknowledgements

I have no acknowledgements to declare.

Disclosure of Interests

None declared

Vascular inflammation in patients with obstructive sleep apnoea and coronary artery disease shown on coronary computed tomography angiography attenuation

Funding Acknowledgements

Type of funding sources: None.

Introduction

Obstructive sleep apnoea (OSA) is associated with increased plaque burden in coronary artery disease (CAD), but the role of vascular inflammation in this relationship is unclear. Coronary computed tomography angiography (CTA) enables surrogate assessment of systemic inflammation via subcutaneous adipose tissue attenuation (ScAT-a), and of coronary inflammation via epicardial adipose tissue volume and attenuation (EAT-v and EAT-a) and pericoronary adipose tissue attenuation (PCAT-a).

Purpose

To investigate whether vascular inflammation is increased in patients with severe OSA and high plaque burden.

Methods

Patients with clinically indicated polysomnography and coronary CTA were included. Severe OSA was classified as apnoea/hypopnoea index (AHI) >30. High plaque burden was defined as a CT-Leaman score (CT-LeSc) >8.3. Patients with both severe OSA and high plaque burden were defined as ‘Group 1’, all other patients were classified as ‘Group 2’. ScAT-a, EAT-a, EAT-v and PCAT-a were assessed on semi-automated software.

Results

A total of 91 patients were studied (59.3 ± 11.1 years). Severe OSA was associated with high plaque burden (p = 0.02). AHI correlated with CT-LeSc (r = 0.24, p = 0.023). Group 1 had lower EAT-a and PCAT-a compared to Group 2 (EAT-a: -87.6 vs. -84.0 HU, p = 0.01; PCAT-a: -90.4 vs. -83.4 HU, p < 0.01). However, among patients without high plaque burden, EAT-a was increased in patients with severe OSA versus mild-moderate OSA (-80.3 vs. -84.0 HU, p = 0.020). On multivariable analysis, EAT-a independently associated with severe OSA and high plaque burden (p < 0.02), and PCAT-a associated with severe OSA and high plaque burden, and hypertension (all p < 0.01).

Conclusions

EAT attenuation is decreased in patients with severe OSA and high plaque burden but increased in patients with severe OSA and low plaque burden. These divergent results suggest coronary inflammation may be increased in OSA independent of CAD, but larger studies are required to validate these findings.

In vivo coronary artery thrombus imaging with 18F-GP1 PET-CT

Background

Coronary artery thrombus is typically present in type 1 myocardial infarction, but small volumes in the setting of an uncertain culprit lesion may be beyond the detection limit of current imaging modalities.

Purpose

Using a novel glycoprotein IIb/IIIa-receptor radiotracer, 18F-GP1, we investigated whether positron emission tomography-computed tomography (PET-CT) could detect thrombus formation in coronary arteries.

Methods

In a single centre cross-sectional study, patients over 40 years of age with myocardial infarction were recruited after myocardial infarction and underwent underwent CT angiography and 18F-GP1 PET-CT. Stable patients with and without coronary artery disease formed a control cohort. Coronary artery 18F-GP1 uptake was visually assessed and quantified using maximum target-to-background ratios (TBRmax).

Results

Ninety-four (44 post-myocardial infarction and 50 control patients) were included in the cross-sectional analysis. The mean age of the post-myocardial infarction group was 61±9 years, three-quarters were male and two thirds had presented with ST elevation on electrocardiography. 34 (80%) patients post-myocardial infarction, but none of the control patients, demonstrated focal 18F-GP1 uptake in the coronary arteries.

Of 42 vessels with an angiographic culprit lesion, 35 (83%) had 18F-GP1 uptake which was significantly higher than non-culprit vessels (p<0.0001) as well as control vessels (p<0.0001), while non-culprit vessel uptake was similar to control vessel uptake (p=0.567): culprit vessel median TBRmax 1.2 [interquartile range 0.96–1.44], non-culprit vessel TBRmax 0.96 [0.84–1.03] and control vessel TBRmax 0.9 [0.76 to 0.94]. Linear regression models demonstrated univariable associations between coronary 18F-GP1 TBRmax and time from myocardial infarction, male sex and presence of culprit vessel. On multivariable analysis, only culprit vessel status was associated with TBRmax (adjusted R2= 0.22, P<0.001). Based on the Youden's index of the ROC curves, the optimal cut-off of predicting the presence of a culprit vessel was 1.20 with a specificity of 97%, accuracy of 83%, sensitivity (60%) and c-statictic of 0.74.

A patient with ectatic vessel and visual thrombus demonstrated the most intense 18F-GP1 uptake (TBRmax 2.0, highest in the cohort) in the region of heaviest thrombus burden (Figure 2). Extra-coronary uptake was seen in regions of left ventricular thrombus, left atrial appendage thrombus, pulmonary thromboembolism and intramyocardial microvascular obstruction.

Conclusions

18F-GP1 PET-CT is able to detect coronary artery thrombus in culprit lesions following myocardial infarction, as well as extra-coronary thrombotic pathologies that may be important in guiding patient management. 18F-GP1 is highly specific in recognising a culprit lesion from a non-culprit lesion both visually as well as quantitatively.

Funding Acknowledgement

Type of funding sources: Foundation. Main funding source(s): British Heart Foundation Figure 1Figure 2

POS1422 CORRELATES OF TESTING POSITIVE FOR SARS-COV-2 IN PATIENTS WITH RHEUMATIC AD MUSCULOSKELETAL DISEASES

Background:

Many studies on COVID-19 outcomes in patients with RMD have either restricted to COVID positive RMD patients or compared them to the general clinic population as a comparator. Given heterogeneity in behaviors and risks, clinical characteristics associated with a positive diagnosis among patients with RMD seeking testing for Sars-CoV-2 remain less well studied.

Objectives:

Among patients with RMD receiving a Sars-CoV-2 PCR test, we aimed to identify RMD-related factors associated with a positive test result.

Methods:

Among patients seen at least once in the University of Washington (UW) rheumatology clinics between March 2018 to March 2020, we reviewed electronic medical records to identify patients undergoing Sars-CoV-2 PCR testing from March 1 through October 31, 2020. Patients with RMD were categorized into two groups: those who tested positive for Sars-CoV-2 and those who tested negative. We randomly selected patients from the negative group in a 2:1 ratio for further data abstraction. Student’s t-test and Chi-squared tests were used to compare continuous and categorical variables, respectively, between the groups. To determine the correlates of testing positive for Sars-CoV-2, specifically RMD medication use and disease activity, we constructed different multivariable logistic regression models adjusted for age, sex, race/ethnicity, presence of comorbidities, body mass index, and smoking.

Results:

A total of 2768 RMD patients underwent SARS-CoV-2 PCR testing within the UW system, of whom 43 (1.5%) were positive at least once. Three patients with incomplete information were excluded. Patients who tested positive had higher prevalence of end stage renal disease (ESRD)/chronic kidney disease (CKD) (24% versus 11%), had higher rates of active disease (24% versus 20%), were older (>55 years) (mean age 57.3 versus 54.8 years), male (63% versus 55%), non-white race/ethnicity (32% versus 26%), and higher prevalence of multiple comorbidities (42% versus 31%) (Table 1). In the multivariable models, neither RMD medication use (versus no use, Table 1) nor high disease activity (vs low disease activity/remission) were statistically significantly associated with COVID-19 positivity. Among the 41 COVID-19 positive patients, a majority recovered without specific treatments, although approximately one third of the positive patients were hospitalized and three deaths were observed.

Conclusion:

In this study, patients who tested positive did not differ in many ways from those who tested negative.

Table 1.

Baseline characteristics of the patients prior to COVID testing

VariablesAll(N=126)COVID Positive (N=41)COVID Negative(N=85)P valueAge in years – mean (SD)55.6 (15.3)57.3 (16.3)54.8 (14.9)0.40Sex0.39 Male73 (57.9)26 (63.4)47 (55.3) Female53 (42.1)15 (36.6)38 (44.7)Race0.39 White89 (71.2)26 (63.4)63 (74.1) Other race35 (28.2)13 (31.7)22 (25.9) Missing2 (1.6)2 (4.9)0 (0.0)Rheumatic disease0.64 OA/Crystal/Fibromyalgia37 (29.4)11 (26.8)26 (30.6) RA/SpA32 (25.4)9 (22.0)23 (27.1) All others57 (45.2)21 (51.2)36 (42.3)Rheumatic disease activity0.57 Active27 (21.4)10 (24.4)17 (20.0) Not active99 (78.6)31 (75.6)68 (80.0)Co-morbidities  Diabetes mellitus (%)25 (19.8)9 (22.0)16 (18.8)0.68  Hypertension48 (38.1)20 (48.8)28 (32.9)0.09  Cardiovascular disease23 (18.3)9 (22.0)14 (16.5)0.46  Lung disease25 (19.8)10 (24.4)15 (17.7)0.37  Cancer10 (7.9)3 (7.3)7 (8.2)0.86  ESRD/CKD19 (15.1)10 (24.4)9 (10.6)0.04*

BMI: Body mass index; SD: Standard deviation; OA: Osteoarthritis; Crystal: Crystalline diseases; RA: Rheumatoid arthritis; SpA: Spondyloarthritis

Acknowledgements:

The work in this study was supported by grant UL1 TR002319 to Dr Singh from the Institute of Translational Health Sciences of the University of Washington.

Disclosure of Interests:

None declared

POS0443 EXPLORING NOVEL TENOSYNOVITIS AND COMBINED INFLAMMATION IMAGING OUTCOMES: RESULTS FROM A RANDOMISED CONTROLLED TRIAL IN EARLY RHEUMATOID ARTHRITIS

Background:

Magnetic resonance imaging (MRI) trial outcomes have largely focused on synovitis, bone marrow oedema (BME) and erosions. Tenosynovitis is a common manifestation of rheumatoid arthritis (RA), but is relatively understudied; a combined inflammation score (CIS) summing synovitis, BME and tenosynovitis may be a highly responsive measure. We previously showed the responsiveness of the Outcome Measures in Rheumatology (OMERACT) RA MRI scoring system (RAMRIS) and a machine-learning derived automated tool (RAMRIQ) in a randomised controlled trial (RCT) of tofacitinib and methotrexate (MTX) in MTX-naïve patients (pts) with early RA.1

Objectives:

This post hoc analysis assessed the impact of tofacitinib ± MTX on MRI tenosynovitis and CIS in pts with early RA using semiquantitative and quantitative MRI outcomes.

Methods:

Study A3921068 (NCT01164579), a 1-year, exploratory, Phase 2, RCT, compared tofacitinib 10 mg twice daily (BID) ± MTX, and MTX monotherapy, in MTX-naïve pts with early RA.1 MRI of unilateral wrist and metacarpophalangeal joints was performed at screening/baseline (BL) and Months (M)1/3/6/12. MRI tenosynovitis and CIS were assessed using RAMRIS and RAMRIQ. Changes from BL (Δ) in RAMRIS and RAMRIQ tenosynovitis and CIS were evaluated at M1/3/6/12. Data were assessed using a mixed-effect model for repeated measures, with treatment arms as factors and BL values as covariates. Using data pooled across treatment arms, Spearman’s rank correlation coefficients were calculated for associations between BL RAMRIS and BL RAMRIQ tenosynovitis and CIS vs BL Disease Activity Score in 28 joints, C-reactive protein (DAS28-4[CRP]), and ΔRAMRIS and ΔRAMRIQ tenosynovitis and CIS at M12 vs ΔDAS28-4(CRP) at M12.

Results:

In total, 109 pts were randomised and treated. ΔRAMRIS and ΔRAMRIQ tenosynovitis and CIS were generally significantly greater at M3/6/12 in pts receiving tofacitinib ± MTX vs MTX, while ΔRAMRIQ CIS was also significantly greater at M1 (Figure 1). Compared with RAMRIS, RAMRIQ outcomes were generally more responsive to treatment with tofacitinib ± MTX. Significant correlations were seen between BL RAMRIS and BL RAMRIQ tenosynovitis and BL RAMRIS CIS vs BL DAS28-4(CRP), and between ΔRAMRIS and ΔRAMRIQ tenosynovitis and CIS at M12 vs ΔDAS28-4(CRP) at M12 (Table 1). In general, stronger correlations were seen between BL DAS28-4(CRP) and BL RAMRIS vs BL RAMRIQ parameters, while correlations were similar between ΔDAS28-4(CRP) at M12 and ΔRAMRIS and ΔRAMRIQ parameters at M12.

Table 1.

Correlations between a) BL RAMRIS and BL RAMRIQ tenosynovitis and CIS vs BL DAS28-4(CRP) (N=109) and b) ΔRAMRIS and ΔRAMRIQ tenosynovitis and CIS at M12 vs ΔDAS28-4(CRP) at M12 (N=73) across treatment armsa

a)BL DAS28-4(CRP)b)ΔDAS28-4(CRP)BL imaging featureCorrelationp valueΔImaging featureCorrelationp valueRAMRIStenosynovitis0.3660.0001RAMRIStenosynovitis0.531<0.0001RAMRIS CIS0.399<0.0001RAMRIS CIS0.554<0.0001RAMRIQtenosynovitis0.2050.037RAMRIQtenosynovitis0.543<0.0001RAMRIQ CIS0.1800.062RAMRIQ CIS0.564<0.0001

aData were pooled across the tofacitinib 10 mg BID ± MTX and MTX monotherapy arms

Spearman’s rank correlation coefficients and p values were calculated for associations between RAMRIS and RAMRIQ parameters and DAS28-4(CRP); tenosynovitis and CIS were assessed using combined data from metacarpophalangeal and wrist joints; CIS is the sum of synovitis, BME and tenosynovitis values

Δ, change from BL; N, number of pts with values at BL/timepoint of interest

Conclusion:

Responsiveness of RAMRIS and RAMRIQ tenosynovitis and CIS was demonstrated with significant improvements through M12 in pts receiving tofacitinib 10 mg BID ± MTX vs MTX. Construct validity for RAMRIS and RAMRIQ tenosynovitis and CIS was evident from correlations with DAS28-4(CRP). Further work is needed to validate these novel imaging biomarkers in terms of relative responsiveness and prediction of later structural progression.

References:

[1]Conaghan et al. Ann Rheum Dis 2016; 75: 1024-1033.

Acknowledgements:

Study sponsored by Pfizer Inc. Medical writing support was provided by Jennifer Arnold, CMC Connect, and funded by Pfizer Inc.

Disclosure of Interests:

Philip G Conaghan Speakers bureau: AbbVie, Novartis, Consultant of: AbbVie, Bristol-Myers Squibb, Eli Lilly, Flexion, Galapagos, Gilead Sciences, GlaxoSmithKline, Janssen, Novartis, Pfizer Inc, Regeneron, Mikkel Østergaard Speakers bureau: AbbVie, Boehringer Ingelheim, Bristol-Myers Squibb, Celgene, Centocor, Eli Lilly, GlaxoSmithKline, Hospira, Janssen, Merck, Mundipharma, Novartis, Novo, Orion, Pfizer Inc, Regeneron, Roche, Schering-Plough, Takeda, UCB, Wyeth, Consultant of: AbbVie, Boehringer Ingelheim, Bristol-Myers Squibb, Celgene, Centocor, Eli Lilly, GlaxoSmithKline, Hospira, Janssen, Merck, Mundipharma, Novartis, Novo, Orion, Pfizer Inc, Regeneron, Roche, Schering-Plough, Takeda, UCB, Wyeth, Grant/research support from: AbbVie, Centocor, Merck, Orrin Troum Speakers bureau: AbbVie, Amgen, Bristol-Myers Squibb, Celgene, Horizon, Novartis, Pfizer Inc, Sanofi-Genzyme, Consultant of: AbbVie, Amgen, Bristol-Myers Squibb, Eli Lilly, Horizon, Pfizer Inc, Grant/research support from: AbbVie, Amgen, Bristol-Myers Squibb, Centocor, Corrona, Novartis, Pfizer Inc, Zhiyong Xie Shareholder of: Pfizer Inc, Employee of: Pfizer Inc, Alan Brett Shareholder of: Stryker Corp, Employee of: Stryker Corp, Mark Snyder Shareholder of: Pfizer Inc, Employee of: Pfizer Inc, Abbas Ebrahim Shareholder of: Pfizer Inc, Employee of: Pfizer Inc, Douglass Chapman Shareholder of: Pfizer Inc, Employee of: Pfizer Inc, Gosford Sawyerr Consultant of: Pfizer Inc, Employee of: Syneos Health, John Andrews Shareholder of: Pfizer Inc, Employee of: Pfizer Inc

POS0442 RELATIONSHIP BETWEEN PARAOXONASE-1 GENOTYPE, ACTIVITY AND MAJOR ADVERSE CARDIOVASCULAR EVENTS IN PATIENTS WITH RHEUMATOID ARTHRITIS RECEIVING TOFACITINIB

Background:

Paraoxonase-1 (PON1) is a high-density lipoprotein (HDL)-associated enzyme with paraoxonase, lactonase and arylesterase activities.1 PON1 contributes to the antioxidant properties of HDL, and is being investigated for its atheroprotective properties.1 Patients (pts) with rheumatoid arthritis (RA) who are homozygous for the RR genotype of the Q192R gene polymorphism on PON1 (rs662) have increased paraoxonase activity, and lower risk of carotid plaques, vs those with QQ or QR genotypes.2 Tofacitinib is an oral Janus kinase inhibitor for the treatment of RA.

Objectives:

To investigate the relationship between PON1 genotype/activity and risk of major adverse cardiovascular events (MACE) in the tofacitinib RA clinical programme.

Methods:

In this post hoc analysis, data were pooled from pts enrolled in nine Phase 2/3 studies of tofacitinib in RA. Enzyme activities in pt plasma samples were measured at individual study baseline (BL) and at follow-up visits using three substrates: paraoxon (paraoxonase activity), dihydrocoumarin (lactonase activity) and phenylacetate (arylesterase activity). The effect of the PON1 Q192R genotype (QQ, QR or RR) on BL paraoxonase/lactonase/arylesterase activity was assessed using linear regression for each study, with age and sex as covariates, and then fixed-effect meta-analysis assessed effects across studies. The risk of MACE by enzyme activity was determined using Cox proportional hazards regression stratified by clinical studies. Univariate regression against BL enzyme activity and other risk factors, as well as both minimally and fully adjusted multivariable regressions against time-varying enzyme activity, are presented.

Results:

The analysis included 1969 pts with RA who received ≥1 dose of tofacitinib and had PON1 activity measures available at BL; 39 pts had ≥1 MACE event. Compared with the QQ genotype, the RR genotype had a highly significant positive association with BL paraoxonase activity, and a highly significant negative association with BL lactonase and arylesterase activity (Table 1). A univariate analysis identified several BL covariates significantly associated with risk of MACE (Figure 1a). Time-varying models found a highly significant association of increased paraoxonase activity over time with lower risk of future MACE, even after controlling for low-density lipoprotein or HDL cholesterol levels, and other traditional cardiovascular (CV) risk factors identified in univariate analysis (Figure 1b), with similar findings for lactonase and arylesterase (data not shown).

Table 1.

Effect of PON1 genotype (RR vs QQ) on BL enzyme activity

nEstimate95% CIp valueCochran’s Q testParaoxonase12291.0350.93, 1.141<0.00010.209Lactonase1188-0.375-0.505, -0.246<0.00010.025Arylesterase1231-1.016-1.382, -0.649<0.00010.251

Fixed-effects model; estimate >0 favours RR genotype and <0 favours QQ genotype

BL, baseline; CI, confidence interval

Conclusion:

Higher activity of the HDL-associated protein PON1 over time was associated with a significantly reduced risk of future MACE in pts with RA receiving tofacitinib, after controlling for traditional CV risk factors and cholesterol levels. Further investigation of PON1 as a novel functional lipid biomarker to assess CV risk in pts with RA is warranted.

References:

[1]Mackness & Mackness. Gene 2015; 567: 12-21.

[2]Charles-Schoeman et al. Arthritis Rheum 2013; 65: 2765-2772.

Acknowledgements:

Study sponsored by Pfizer Inc. Medical writing support was provided by Jennifer Higginson, CMC Connect, and funded by Pfizer Inc.

Disclosure of Interests:

Christina Charles-Schoeman Consultant of: AbbVie, Gilead, Pfizer Inc, Regeneron-Sanofi, Grant/research support from: AbbVie, Bristol-Myers Squibb, Pfizer Inc, Craig Hyde Shareholder of: Pfizer Inc, Employee of: Pfizer Inc, Shunjie Guan Shareholder of: Pfizer Inc, Employee of: Pfizer Inc, Neil Parikh: None declared, Jennifer Wang: None declared, Ani Shahbazian: None declared, Lori Stockert Shareholder of: Pfizer Inc, Employee of: Pfizer Inc, John Andrews Shareholder of: Pfizer Inc, Employee of: Pfizer Inc

POS0459 UNSUPERVISED CLUSTERING IDENTIFIES UNIQUE SUBSETS OF PATIENTS IN A RACIALLY AND ETHNICALLY DIVERSE RHEUMATOID ARTHRITIS COHORT

Background:

Single biomarkers have limited utility to date in guiding RA clinical care. Machine learning algorithms may better identify and stratify RA patients with differential outcomes.

Objectives:

To determine if unsupervised machine learning methods can be employed in a racially and ethnically diverse RA cohort to identify clusters of patients with different disease activity trajectories, as measured by DAS28ESR.

Methods:

Data are derived from the longitudinal, observational University of California, San Francisco RA Cohort. Along with routine labs, medications and disease activity assessments, a multiple biomarker of disease activity (MBDA) panel was obtained at each visit. The MBDA measures 12 serum biomarkers. Four patient clusters were identified by unsupervised K-prototype clustering after collapsing all observations into a cross sectional dataset. Plots were created to display longitudinal disease activity trajectories for each cluster. Lasso regression was applied to identify biomarkers associated with DAS28ESR.

Results:

We identified 4 distinct clusters in our cohort (Table 1) with visually different disease activity trajectories (Fig. 1). Cluster 1 (n=116) was older (63.6±9.7), had the highest proportion of Asian participants (n=73, 63%) with the most study visits and longest disease duration. Cluster 2 (n=70) had the highest mean DAS28ESR (5.5±0.7), and the highest mean dose of prednisone (8.6±4.9 mg/day). Cluster 3 had the lowest number of participants (n=14), study visits and lowest biologic use (28.6%). Cluster 4 was the largest cluster (n=173) with the shortest disease duration (4.9±3.8 years) and highest biologic use (61.3%). In the Lasso regressions, leptin was found to have significant positive associations with DAS28ESR in the whole group as well as Clusters 2 and 4. EGF had negative associations with DAS28ESR in the whole group, Cluster 1 and 4. CRP had positive associations with DAS28ESR in the whole group and Cluster 1. YKL40 and VCAM1 were found to have significant associations in Clusters 1 and 3, respectively.

Conclusion:

We identified 4 unique clusters of RA patients in a racially and ethnically diverse longitudinal cohort with different disease activity trajectories and biomarkers associated with disease activity. Although additional work is needed to explore longitudinal outcomes in each cluster, the application of machine learning methods may identify unique combinations of patient and disease characteristics influencing RA clinical outcomes.

Table 1.

Demographics and clinical characteristics of the RA cohort. Biomarkers significantly associated with DAS28ESR were determined by Lasso regression Values listed are per standard deviation of each biomarker.

Overall(N=373)Cluster 1 (N=116)Cluster 2 (N=70)Cluster 3 (N=14)Cluster 4 (N=173)Demographics:Age54.8 ± 3.663.6 ± 9.750.8 ± 14.958.2 ± 15.850.3 ± 12.1Female Sex318 (85%)101 (87%)57 (81%)11 (79%)149 (86%)Race: Hispanic/Latino181 (49%)22 (19%)47 (67%)5 (36%)107 (62%) Asian123 (33%)73 (63%)8 (11%)6 (43%)36 (21%) Black35 (9%)12 (10%)8 (11%)2 (14%)13 (8%) White & Other34 (9%)9 (7%)7 (10%)1 (7%)17 (10%)Clinical Characteristics:Rheumatoid Factor315 (85%)104 (90%)56 (80%)13 (93%)142 (82%)ACPA297 (80%)98 (85%)54 (77%)12 (86%)133 (77%)Disease Duration7.8 ± 7.613.7 ± 9.75.4 ± 4.66.7 ± 5.74.9 ± 3.8csDMARD344 (92%)108 (93%)63 (90%)13 (93%)160 (93%)Biologic DMARD185 (50%)45 (39%)30 (43%)4 (29%)106 (61%)Prednisone Dose6.7 ± 3.86.0 ± 4.08.6 ± 4.95.8 ± 1.46.3 ± 2.8Body Mass Index28.2 ± 4.526.6 ± 3.828.7 ± 3.828.0 ± 6.229.1 ± 4.7DAS28ESR4.2 ± 1.14.2 ± 1.05.5 ± 0.83.9 ± 0.83.7 ± 0.9Lasso Results:EGF-0.16*-0.41***-----0.20**Leptin0.15**--0.21*--0.21**c-reactive protein0.34**0.51***------VCAM1-------0.73*--YKL40--0.26*------

-EGF: epidermal growth factor; VCAM1: Vascular cell adhesion protein 1; YKL40: Chitinase-3-like protein 1.

-*p<0.05, **p<0.01, ***p<0.001

Figure 1.

DAS28ESR trajectory plots with 95%CIs for the whole cohort and by cluster.

Acknowledgements:

This work was supported by the Rheumatology Research Foundation Scientist Development Award.

Disclosure of Interests:

None declared

POS0537 TRENDS AND PREDICTORS OF INPATIENT MORTALITY IN PATIENTS WITH CLOSTRIDIODES DIFFICILE INFECTION AMONG THOSE WITH AND WITHOUT RHEUMATOID ARTHRITIS: A NATIONWIDE ANALYSIS

Background:

People with rheumatoid arthritis (RA) have a higher risk of serious infections. The most common nosocomial infection in the United States (US) is Clostridioides difficile infection (CDI) and yet, little is known regarding the impact of comorbid RA on outcomes in persons hospitalized with CDI.

Objectives:

To evaluate the trends over time and predictors of inpatient mortality in hospitalized patients with CDI comparing those with and without RA.

Methods:

We used the National Inpatient Sample (NIS) database, the largest available all-payer database of inpatient hospitalizations in the US. Patients aged ≥18 years admitted with CDI from 2006-2018 were included. Patients with RA were identified using the International Classification of Diseases (ICD) 9 and 10 codes. Categorical variables between CDI patients with and without RA were compared using Chi-squared test. Temporal trends of hospitalization charges, length of stay (LOS) and inpatient mortality were assessed using the Cochrane Armitage test. Predictors of inpatient mortality were assessed using multivariable logistic regression adjusting for: age ≥65, sex, race, presence of RA, and Elixhauser comorbidity index.

Results:

There were 4,396,945 hospitalizations with a discharge diagnosis of CDI in the study period, of which 111,336 (2.5%) had a coexisting diagnosis of RA. Over the study period, the proportion of CDI patients with a diagnosis of RA increased from 1.7% in 2006 to 3.3% in 2018 with a statistically significant trend of rise (p trend<0.001). There were significantly more females in the RA group (78.2%) than without RA (57.5%), p <0.001. RA group was significantly younger (68.0 vs 70.7, P<0.001). Races were similar between the groups.

The inflation-adjusted average charge for hospitalization for CDI without RA was significantly higher than for CDI with RA ($95,863 vs $73,025, p<0.001). There was an increasing trend in inflation-adjusted average charge for hospitalization for both CDI without RA ($76,966 in 2006 to $104,552 in 2018, p<0.001) and CDI with RA ($55,689 in 2006 to $77,295 in 2018, p<0.001). Patients with or without RA had the same median length of stay (LOS) of 7 days. But when trended over time, for patients without RA, the median LOS decreased from 8 days in 2006-08 to 6 days in 2016-18 whereas for patients with RA, the median LOS fluctuated between 6 and 7 days over the course of study period.

CDI patients without RA had a significantly higher mortality over the study period than those with RA (7.8% vs 6.1%, p<0.001). Among non-RA CDI patients, mortality decreased from 9.3% in 2006 to 6.3% in 2018 (p trend <0.001). Hospitalizations of people with CDI and comorbid RA showed an overall decreasing trend (p trend <0.001).

The multivariable analyses found age≥65 was associated with increased inpatient mortality (adjusted odds ratio [aOR], 1.77, 95% confidence interval [CI], 1.73-1.80), whereas a diagnosis of RA was associated with a decreased mortality (aOR 0.86, 95% CI, 0.81-0.91) (Table 1).

Table 1.

Predictors of inpatient mortality

aOR95% C.I. for aORP-valueLowerUpperAge ≥ 651.7651.7301.801<0.001Hospitalization years -Years 2006-2011Ref -Years 2011-20140.7070.6930.721<0.001 -Years 2015-20180.5820.5690.594<0.001Elixhauser Comorbidity Index1.0561.0551.056<0.001Rheumatoid Arthritis0.8590.8100.911<0.001Female0.8700.8550.885<0.001Race -WhiteRef -African American1.0711.0441.099<0.001 -Hispanic1.1441.1081.180<0.001 -Asian/Pacific Islander1.2671.2001.337<0.001 -Native American1.0220.9101.1470.716 -Other races1.2631.2011.329<0.001

aOR = Adjusted Odds Ratio; C.I.= confidence interval

Conclusion:

In this nationwide study, among hospitalized patients with CDI, those with comorbid RA had significantly decreased odds of inpatient mortality than those without RA. Further studies are needed to understand this association.

Figure 1.

Disclosure of Interests:

None declared

Thrombus formation on bioprosthetic aortic valves

Background

18F-GP1 is a novel radiotracer with a high affinity for the platelet glycoprotein IIb/IIIa receptor. Subclinical bioprosthetic valve thrombus has been postulated as a trigger for accelerated valve degeneration.

Purpose

To determine the feasibility of 18F-GP1 positron-emission tomography-computed tomography (PET-CT) for the detection of subclinical bioprosthetic aortic valve thrombus.

Methods

(i) Explanted degenerated aortic valve prostheses underwent histology and imaging. (ii) In a prospective observational study, patients with bioprosthetic aortic valve replacement (AVR) underwent echocardiography and 18F-GP1 PET-CT. Valves were assessed for hypoattenuating leaflet thickening (HALT), hypo-attenuation affection leaflet motion (HAM) and GP1 uptake.

Results

(i) GP1 correlated with thrombus on explanted valves (Figure). (ii) The first 6 patients (Table) were asymptomatic and had normally functioning surgical bioprostheses on echocardiography. At a median of 166 (range 122–189) days post-AVR, no patients had HALT or HAM on CT. There was avid focal GP1 uptake on the leaflets of all 6 patients which appeared most prominent along the leaflet edges (Figure). Only one patient had focal uptake in the valve frame, remote from the leaflets. In a separate cohort undergoing 18F-GP1 PET-CT for other conditions, there was no uptake on normal, native aortic valves (n=8).

Conclusion

For the first time, we demonstrate that 18F-GP1 PET-CT is a highly sensitive method of assessing platelet activation on bioprosthetic aortic valves. Despite the absence of CT evidence, early thrombus appeared to be a universal finding on recently implanted valve prostheses. The biological and clinical implications of subclinical bioprosthetic aortic valve thrombus have yet to be established.

GP1 uptake in AVR

Funding Acknowledgement

Type of funding source: Foundation. Main funding source(s): British Heart Foundation

Coronary artery echo-attenuated plaques in acute coronary syndromes: a serial intravascular ultrasound imaging study

Background

Echo attenuation of atherosclerotic plaque (EAP) identified with intravascular ultrasound (IVUS) has been shown to correlate with vulnerable plaque morphologies and their presence is predictive of future cardiovascular events. EAP have predominantly been assessed at a single time point and their natural history in the immediate post acute coronary syndrome (ACS) period remains unknown. We aimed to assess this and whether their presence correlated with a more modifiable plaque composition in the immediate post-ACS setting.

Methods

Serial IVUS imaging was performed in non-culprit vessels of 270 patients undergoing angiogram for ACS and at 3 month follow up. IVUS analysis of plaque burden and EAP was performed.

Results

Baseline characteristics are described in Table 1. EAP were present at baseline in 62 patients (23%) with these patients more likely to be male (89.1% vs. 76.7%, p=0.03) but no differences in other atherosclerotic risk factors. There was no difference in baseline plaque burden between patients with EAP and those without (Percent atheroma volume [PAV] 38.9% vs. 37.8%, p=0.32). At follow up IVUS change in PAV was not statistically significantly different between patients with baseline EAP and those without (ΔPAV 0.09% vs. −0.36%, p=0.43), and neither was there a difference in the frequency of plaque regressors (42.7% vs 50%, p=0.31). EAP had resolved in 25 patients (40%) within 3 months at the follow up IVUS. Despite contemporary post-ACS therapy 18 patients who had not had EAP present at baseline (9%) developed new EAP at the follow up IVUS.

Conclusion

EAP were present in a quarter of ACS patients and were not associated with baseline plaque burden or a more modifiable plaque phenotype. In the setting of contemporary ACS treatments the natural history of high risk IVUS plaque characteristics such as EAP is dynamic with significant change even over a 3 month period in the post ACS setting.

Funding Acknowledgement

Type of funding source: Foundation. Main funding source(s): Royal Adelaide Hospital Research Fund AR Clarkson Scholarship

Cardiovascular MRI evidence of reduced systolic function and reduced LV mass in rheumatoid arthritis: impact of disease phenotype

The accelerated risk of cardiovascular disease (CVD) in Rheumatoid Arthritis (RA) requires further study of the underlying pathophysiology and determination of the at-risk RA phenotype. Our objectives were to describe the cardiac structure and function and arterial stiffness, and association with disease phenotype in patients with established) RA, in comparison to healthy controls, as measured by cardiovascular magnetic resonance imaging (CMR). 76 patients with established RA and no history of CVD/diabetes mellitus were assessed for RA and cardiovascular profile and underwent a non-contrast 3T-CMR, and compared to 26 healthy controls. A univariable analysis and multivariable linear regression model determined associations between baseline variables and CMR-measures. Ten-year cardiovascular risk scores were increased in RA compared with controls. Adjusting for age, sex and traditional cardiovascular risk factors, patients with RA had reduced left ventricular ejection fraction (mean difference - 2.86% (- 5.17, - 0.55) p = 0.016), reduced absolute values of mid systolic strain rate (p < 0.001) and lower late/active diastolic strain rate (p < 0.001) compared to controls. There was evidence of reduced LV mass index (LVMI) (- 4.56 g/m2 (- 8.92, - 0.20), p = 0.041). CMR-measures predominantly associated with traditional cardiovascular risk factors; male sex and systolic blood pressure independently with increasing LVMI. Patients with established RA and no history of CVD have evidence of reduced LV systolic function and LVMI after adjustment for traditional cardiovascular risk factors; the latter suggesting cardiac pathology other than atherosclerosis in RA. Traditional cardiovascular risk factors, rather than RA disease phenotype, appear to be key determinants of subclinical CVD in RA potentially warranting more effective cardiovascular risk reduction programs.

Qualitative Ampule and Multitest for Beta-Lactam Residues in Fluid Milk Products: Collaborative Study

A collaborative study was performed on a rapid Bacillus stearothermophilus agar diffusion ampule method to detect low levels of penicillin G in 7 types of fluid milk products. A multitest technique for processing a large number of samples simultaneously was also studied. Slight modifications were made in the original method to establish more uniformity and to eliminate doubtful responses by specifying a confirmation procedure. Twenty samples spiked with penicillin G (0.000 to 0.008 IU/mL) and tetracycline hydrochloride were frozen and sent to 20 laboratories in the ampule test, and 16 laboratories in the multitest. Each analyst was asked to do a screening run and a confirmation run. Results were reported by color reaction and also as positive or negative for β-lactam inhibitors. The concentrations (penicillin G) where percent positive results equal 100 or not significantly less than 100 (α = 0.05) ranged from 0.005 to 0.007 IU/mi in the ampule test and from 0.004 to 0.007 IU/mL in the multitest. Both techniques have been adopted official first action.