Impact of transient acquired hypermutability on the inter- and intra-species competitiveness of Pseudomonas aeruginosa

Once acquired, hypermutation is unrelenting, and in the long-term, leads to impaired fitness due to its cumulative impact on the genome. This raises the question of why hypermutators arise so frequently in microbial ecosystems. In this work, we explore this problem by examining how the transient acquisition of hypermutability affects inter- and intra-species competitiveness, and the response to environmental insults such as antibiotic challenge. We do this by engineering Pseudomonas aeruginosa to allow the expression of an important mismatch repair gene, mutS, to be experimentally controlled over a wide dynamic range. We show that high levels of mutS expression induce genomic stasis (hypomutation), whereas lower levels of induction lead to progressively higher rates of mutation. Whole-genome sequence analyses confirmed that the mutational spectrum of the inducible hypermutator is similar to the distinctive profile associated with mutS mutants obtained from the airways of people with cystic fibrosis (CF). The acquisition of hypermutability conferred a distinct temporal fitness advantage over the wild-type P. aeruginosa progenitor strain, in both the presence and the absence of an antibiotic selection pressure. However, over a similar time-scale, acquisition of hypermutability had little impact on the population dynamics of P. aeruginosa when grown in the presence of a competing species (Staphylococcus aureus). These data indicate that in the short term, acquired hypermutability primarily confers a competitive intra-species fitness advantage.

Therapeutic interventions alter ecological interactions among cystic fibrosis airway microbiota

The airways of people with cystic fibrosis (CF) often harbor a diverse microbiota and in recent years, much effort has been invested in cataloguing these. In spite of providing a wealth of insight, this cataloguing tells us little about how the organisms interact with one another in the CF airways. However, such relationships can be inferred using the theoretical framework of the Lotka-Volterra (LV) model. In the current work, we use a generalized Lotka-Volterra model to interrogate the nationwide data collected and curated by the UK CF Registry. This longitudinal dataset (covering the period 2008-2020) contains annual depositions that record the presence/absence of microbial taxa in each patient, their medication, and their CF genotype. Specifically, we wanted to identify trends in ecological relationships between the CF microbiota at a nationwide level, and whether these are potentially affected by medication. Our results show that some medications have a distinct influence on the microbial interactome, especially those that potentially influence the "gut-lung axis" or mucus viscosity. In particular, we found that patients treated with a combination of antimicrobial agents (targeting the airway microbiota), digestive enzymes (assisting in the assimilation of dietary fats and carbohydrates), and DNase (to reduce mucus viscosity) displayed a distinctly different airway interactome compared with patients treated separately with these medications.

Evaluation of unfractionated heparin versus low-molecular-weight heparin and fondaparinux for pharmacologic venous thromboembolic prophylaxis in critically ill patients with cancer

Essentials Critically ill cancer patients require pharmacologic prophylaxis for venous thromboembolism (VTE). Patients from 566 hospitals in the United States between 2010 and 2014 were included. Low-molecular-weight heparin (LMWH) prophylaxis was not associated in a reduction of VTE rates. LMWH prophylaxis was associated with a reduction in bleeding and heparin induced thrombocytopenia. SUMMARY: Background Critically ill patients with cancer are at increased risk of venous thromboembolism (VTE) from physical and cellular factors, requiring pharmacologic prophylaxis to reduce the risk of VTE. Objectives To assess whether low-molecular-weight heparin (LMWH) prophylaxis reduces in-hospital rates of VTE or improves clinical outcomes compared with unfractionated heparin (UFH) prophylaxis in critically ill patients with cancer. Methods We used a propensity-matched comparative-effectiveness cohort from the Premier Database. Patients aged 18 years or older with a primary diagnosis of cancer, intensive care unit admission and VTE prophylaxis within 2 days of admission between 1 January 2010 and 31 December 2014 were included. Patients were divided into LMWH or UFH prophylaxis groups. Results A total of 103 798 patients were included; 75 321 (72.6%) patients received LMWH and 28 477 (27.4%) patients received UFH. Propensity analysis matched (2 : 1) 42 343 LMWH patients and 21 218 UFH patients. Overall, LMWH was not associated with a decreased incidence of VTE (5.32% vs. 5.50%). LMWH prophylaxis was associated with a reduction in pulmonary embolism (0.70% vs. 0.99%), significant bleeding (13.3% vs. 14.8%) and heparin-induced thrombocytopenia (HIT) (0.06% vs. 0.19%). In non-metastatic solid disease, LMWH was associated with decreased VTE (4.27% vs. 4.84%) and PE (0.47% vs. 0.95%). Conclusions The use of an LMWH for VTE prophylaxis was not associated with a reduction in the incidence of in-hospital VTE as compared with UFH, but was associated with significant reductions in PE, clinically important bleeding events, and incidence of HIT in critically ill patients with cancer.

Abstract 059: Developing a Mobile Application for Cardiac Rehabilitation

Introduction: Cardiac rehabilitation improves outcomes for patients with ischemic heart disease or heart failure, but is underused. New strategies to improve delivery of cardiac rehabilitation are needed. We developed a mobile application for technology-facilitated home cardiac rehabilitation and sought to determine its usability.

Methods: We recruited patients eligible for cardiac rehabilitation who had access to a smartphone, tablet, or computer with internet access to participate in usability testing of the mobile application. The mobile application includes physical activity goal setting, logs for physical activity and health measures, health education, reminders, and feedback (Figure). Participants were introduced to the mobile application and then observed while completing pre-specified tasks with the mobile application. Participants completed the System Usability Scale (0-100), rated likelihood to use the mobile application (0-100), and participated in a semi-structured interview. Based on participant feedback, we made iterative revisions to the mobile application.

Results: We conducted usability testing in 13 participants. The first version of the mobile application was used by the first 5 participants, and revised versions were used by the final 8 participants. From the first version to revised versions, task completion success rate improved from 44% to 78% (p=0.05), System Usability Scale improved from 54% to 76% (p=0.04), and rated likelihood of using the mobile application remained high at 76% and 87% (p = 0.30). Interview responses revealed a need for introductory training ( “Initially, training with a technical person, instead of me relying on myself”) and on-demand help (“ If I had problems I’d try to find out how to fix it on this or call you” ). Additionally, many participants were interested in sharing data with providers (“ I can show my doctor what I’ve been working on ”).

Conclusions: With participant feedback and iterative revisions, we significantly improved the usability of a mobile application for cardiac rehabilitation. Patient expectations for using a mobile application for cardiac rehabilitation include introductory training, on-demand help, and sharing data with providers. Iterative mixed-method evaluation may be useful for improving the usability of health technology.

Abstract 173: Regional Variation in Cardiac Rehabilitation Participation within the Medicare and Veterans Health Administration Populations

Introduction: Cardiac rehabilitation is a guideline-recommended therapy for patients after acute myocardial infarction (MI), percutaneous coronary intervention (PCI), and coronary artery bypass surgery (CABG). However, participation in cardiac rehabilitation is notoriously poor. Evaluating variation in health care delivery can help to identify best practices that improve quality of care more broadly. Therefore, we sought to examine regional variation in the percent of patients who participate in cardiac rehabilitation across the United States.

Methods: We used ICD-9 codes from Medicare and Veterans Health Administration (VHA) data to identify patients hospitalized for MI, PCI, or CABG between 1/1/2007 and 12/31/2009 (in a 5% Medicare sample) or 10/1/2006 and 9/30/2011 (in VHA). After excluding patients who died within 30 days of hospitalization, we calculated the percent of patients who participated in one or more outpatient visits for cardiac rehabilitation (CPT code 93797 or 93798) during the 12 months after hospitalization. We then compared the percent of patients who participated in cardiac rehabilitation by state. Hawaii and Alaska were excluded from the analysis because they do not have VHA inpatient facilities.

Results: Overall participation in cardiac rehabilitation was 20% (13,435/67,115) in Medicare and 8% (3,955/47,051) in VHA. Similar regional variation was observed, with the North-Central United States having the highest participation in both health care systems (Figure). Within Medicare, participation in cardiac rehabilitation ranged from 10% (Nevada) to 52% (Nebraska). Within VHA, participation ranged from 0 (Delaware, New Jersey) to 30% (Nebraska).

Conclusion: Similar regional variation in participation in cardiac rehabilitation was observed in two separate health care systems. Nebraska had the highest rate of participation in both Medicare and VHA populations. Further study of reasons for regional variation and differences in regional variation in these populations may reveal opportunities for improving delivery of cardiac rehabilitation.

Abstract 318: Distance From Percutaneous Coronary Intervention Site is Not Associated with 30-day Readmission Length of Stay: Insights From the Veterans Affairs Clinical Assessment, Reporting and Tracking (VA CART) Program

Background: Although readmission following PCI is an emerging quality measure, it is unclear to what extent readmissions post-PCI are avoidable. Many VA patients travel long distances to receive PCI, and a prior study suggests that those living closest to their PCI site are most likely to be readmitted. Using length of stay (LOS) as a proxy for readmission complexity, we explored whether this lower readmission rate reflects a higher threshold for readmission among patients living farther away from PCI sites.

Methods: We identified all VA patients undergoing PCI from 2008-2012, excluding those with private insurance or non-VA sources of care to minimize the likelihood of readmissions outside of VA hospitals. Distance to PCI site was determined by straight-line distance (SLD) from patient residence to the site. The primary exposure was quintile SLD to PCI site. We compared risk-adjusted post-PCI readmission mean LOS by quintile, as well as the trimmed mean (excluding the top 5% LOS per quintile) to minimize outlier effects. We also grouped readmission diagnoses into 8 categories (cardiac ischemic, cardiac-nonischemic, peripheral/cerebrovascular ischemia, procedural, bleeding/anemia, respiratory, renal, and other), comparing diagnoses by SLD quintile.

Results: Among 38,937 patients undergoing PCI, the median SLD to PCI site was 37 mi., IQR 12-85. Crude readmission rates ranged from 16.5% (Q1: 0-9 miles) to 10.2% (Q5: 99-300 mi.). Compared to Q1, the relative difference in mean LOS was significantly lower in Q3 (Q3: 0.77, p = 0.03) and unchanged in all other quintiles (Q2: 0.95, p=0.75; Q4: 0.79, p=0.07; Q5: 0.92, p=0.56; Table). Comparing trimmed means, no quintile was significantly different from Q1. Diagnosis review found higher rates of cardiac ischemia diagnoses in higher SLD quintiles without other clinically significant differences in diagnoses by SLD.

Conclusions: We did not find an association between patient distance to PCI site and LOS for post-PCI readmissions, and readmission diagnoses were overall similar across quintiles. Thus, post-PCI readmissions across quintiles are likely of similar complexity and for similar diagnoses. Identifying other possible contributors to differences in readmission rates by distance may have important implications for policy and resource allocation.

Abstract 27: Association Between Cardiac Rehabilitation Participation and Health Status Outcomes After Acute Myocardial Infarction

Background: Cardiac rehabilitation (CR) is a class Ia recommendation and endorsed performance measure for the quality of care in acute myocardial infarction (AMI) survivors. While participation in CR after AMI is associated with reduced mortality, conflicting data exists on its association with health status outcomes.

Methods: Using data from 2 prospective AMI registries (TRIUMPH and PREMIER), we identified patients for whom baseline and follow-up health status scores and documentation of CR participation (attendance of 1 or more sessions within 6 months post-AMI) were available. Health status was assessed by four Seattle Angina Questionnaire (SAQ) domain scores (quality of life [QoL], angina frequency [AF], treatment satisfaction [TS], and physical limitation [PL]), as well as SF-12 physical and mental health composite scores ( PCS & MCS ). We created propensity matched cohorts of CR participants and non-participants to examine the association between CR participation with health status (6 and 12 months) and all-cause mortality (up to 7 years), using conditional repeated measures and proportional hazards models, respectively.

Results: Among 3,957 AMI patients from 31 sites, 2,015 patients (51%) participated in CR after discharge. Compared to non-participants, CR participants were more often Caucasian (83.6% vs. 65.4%), had higher rates of health insurance (90.6% vs. 79.3%), but clinically similar baseline SAQ and SF-12 scores in all domains. After propensity matching, all covariates were well-balanced (Standardized Difference <10 for all patient characteristics) between CR participants and non-participants. In the repeated measures analysis, the mean SAQ and SF-12 domain scores were clinically similar for both groups at 6 and 12 months after hospital discharge (Table). Using a conditional proportional hazards model, a decrease in all-cause mortality was noted in the CR participant group (HR 0.59, 95% CI [0.46, 0.75]).

Conclusion: In a large, multi-center AMI cohort, we found that although CR participation seemed to improve survival, CR participants had similar health status improvements after AMI as non-participants. Further investigation is required to identify how CR programs can further maximize the health status benefits to post-AMI participants.

Abstract 183: Optimal Medical Therapy for Cardiovascular Disease in the Obese Undergoing Elective Percutaneous Coronary Intervention: Insights from the VA CART Program

Background: Stigma against obese patients is well described in primary care settings and may contribute to bias in therapeutic decision-making. It is unknown if similar stigma exists in obese patients referred for elective PCI. Accordingly, we evaluated the association between body mass index (BMI) and pre-procedural guideline-recommended medication use in patients undergoing elective PCI. The presence of lower medication use in overweight and obese patients may suggest the presence of a treatment bias.

Methods: Using data from the VA Clinical Assessment, Reporting, and Tracking System (CART) Program, we identified patients undergoing elective PCI from 2007-2012. We classified patients by BMI into normal (19-25), overweight (25-30), obese (>30). Rates of guideline-indicated medication use by BMI were assessed among eligible patients: beta-blockers (BB) for HF or prior MI, statins for CAD or equivalent (DM, CVD, PAD), anticoagulation for AFib and CHADS2> 1, and ACEI/ARB for HF. We also assessed composite rates of BB and statin in eligible MI patients and BB and ACEI/ARB use in eligible HF patients, respectively. Multivariable logistic regression analyses assessed the association between BMI class and use of indicated medications.

Results: Among 9,630 patients undergoing elective PCI from 2007-2012, 13.9% of patients had normal BMI, 35.6% overweight, and 50.6% obese. Overweight and obese patients were more likely to have sleep apnea, HTN and DM, while normal BMI patients were more likely to smoke, have lung disease, and CVD. Rates of medication use ranged from 45% to 69% depending on the class of medication assessed. After adjustment for CV risk factors, overweight and obese patients were more likely to receive statins and ACE/ARBs and equally likely to receive the other classes of medications compared to normal BMI patients (Table).

Conclusions: Over 85% of patients undergoing elective PCI in the VA are overweight or obese. Rates of indicated medication use remained low across BMI categories (<70%). There was an association between overweight and obese patients with greater use of some guideline-indicated medications, suggesting that a treatment bias against obesity prior to elective PCI does not exist. Future studies should assess for any impact of BMI on treatment of patients during and following elective PCI.

Abstract 271: Anti-coagulation and Ischemic Stroke Risk in Patients with Chronic Kidney Disease and Atrial Fibrillation: Insights from the Kaiser Permanente Colorado Atrial Fibrillation Registry

Background: There is uncertainty about the benefits of anticoagulation therapy for stroke reduction in CKD patients with atrial fibrillation. Accordingly, we assessed the association between anticoagulation use and stroke risk in CKD patients with atrial fibrillation.

Methods: The Kaiser Permanente Colorado Atrial Fibrillation Registry is comprised of patients with incident atrial fibrillation between January 1, 2006 and June 30, 2012 from Kaiser Permanente Colorado. Incident atrial fibrillation was defined by ICD-9 codes 427.31 (Atrial Fibrillation) or 427.32 (Atrial Flutter) and without a diagnosis in the prior year. Patients with mitral valve replacement, renal transplant, or use of anticoagulants other than warfarin were excluded. CKD status was determined by ICD-9 codes or by two consecutive outpatient laboratory results with estimated glomerular filtration rate < 60 ml/min/1.73m2 by the CKD-EPI equation. The primary outcome was ischemic stroke identified by ICD-9 codes and validated by chart review. We assessed the association between warfarin use and ischemic stroke in patients with and without CKD using Cox proportional hazards models adjusted for CHA2DS2-VASc score and an interaction for warfarin use and CKD.

Results: Of 5,728 patients with incident atrial fibrillation, 2,070 (36.1%) had CKD. Patients with CKD were older, more likely to be female, had a higher CHA2DS2-VASc score, and were more likely to receive warfarin than those without CKD (see Table). During a mean follow up of 2.6 years (SD 1.8 years), stroke occurred in 49 (2.4%) patients with CKD and 83 (2.3%) patients without CKD. In multivariable analysis adjusting for CHA2DS2-VASc score, warfarin use was associated with lower hazard of stroke (HR 0.36; 95% CI 0.24 - 0.53). When stratified by CKD status, warfarin use remained associated with lower hazard of stroke in CKD (HR 0.35; 95% CI 0.18 - 0.66) and non CKD (HR 0.36; 95% CI 0.22 - 0.60) patients.

Conclusion: 1 in 3 patients with atrial fibrillation have CKD. There were similar reductions in the risk of stroke associated with warfarin use for CKD and non-CKD patients. These findings reinforce current clinical practice guidelines, which recommend warfarin use based on thromboembolic risk without consideration for CKD status.

Abstract 352: A Multimodal Radiation Reduction Intervention for Intra-procedural Radiation Exposure in Patients Undergoing Cardiac Catheterization in Veterans Affairs Hospitals

Background: Radiation exposure to patients from invasive cardiac procedures is substantial and contributes to a significant portion of overall radiation exposure from medical testing. Efforts to minimize intra-procedural radiation exposure are important for patient safety. This pilot study evaluated the effectiveness of a multimodal radiation intervention to reduce intra-procedural radiation exposure.

Methods: Two VA cardiac catheterization laboratories (Site 1, Site 2) were evaluated for baseline radiation dosing use over a three month period. Following this initial run-in period, the operators and cath lab staff underwent a three-tiered intervention: 1) radiation safety and minimization education, 2) an in-lab radiation monitoring protocol with verbal feedback at pre-specified radiation doses and 3) monthly site and provider-specific report cards comparing radiation dose at the site and provider level within the VA system. Radiation dosing (RD, measured as Dose-Area-Product [Gy*cm2]) was then measured following this intervention at monthly intervals over a three-month period.

Results: We examined 624 cases at Site 1 and 258 cases at Site 2 in the pre-intervention period, and 502 (site 1) and 208 (site 2) in the post-intervention period. Site 1 did not differ significantly in median RD following intervention (71.9 Gy*cm2 [IQR 48.0-114.0] pre-intervention versus 79.5 Gy*cm2 [IQR 50.0-124.8] post-intervention, p=0.34; see Fig 1). Site 2 showed a significant decrease in median radiation dose following intervention (118.72 Gy*cm2 [IQR 73.6-190.0] vs. 92.8 Gy*cm2 [IQR 56.6-158.3], p = 0.004, Fig 1). The national median radiation dose over the same time interval did not change significantly (91.53 Gy*cm2 [IQR 58.0-145.4] pre-intervention versus 90.0 Gy*cm2 [IQR 56.3-142.0] post-intervention, p=0.47, Fig 1).

Conclusion: A three-tiered, multi-modal radiation reduction intervention was associated with reduced radiation exposure in a laboratory with high baseline radiation utilization. Similar reductions were not observed in a laboratory with low baseline radiation utilization. These findings suggest that radiation reduction interventions targeted at higher radiation use centers may result in meaningful decreases in patient radiation exposure.

Abstract 5: The Association between Patient Rural Status, Distance to Cath Lab, and Likelihood to Undergo Coronary Angiography: Insights from the VA CART Program

Background. Prior studies suggest rural patients are less likely to receive coronary procedures. This finding may represent the fact that rural patients live further than urban patients from the closest cath lab, but these studies failed to incorporate a direct measure of distance to cath lab. Accordingly, we incorporated patient distance to cath lab with rural location on the likelihood to undergo elective coronary angiography as a better explanation of this relationship.

Methods. Using national data from the VA Clinical Assessment Reporting and Tracking (CART) Program, we evaluated all patients who underwent elective coronary angiography from October 2007 through September 2010. We calculated geodesic distance from patient home to cath. Patient rural status was determined using the VA Urban/Rural/Highly Rural (URH) system. Normal coronary angiography (an indirect measure of patient selection for elective coronary angiography) was defined as <20% stenosis in all vessels. Using generalized linear mixed modeling, we estimated the probability of angiographically normal coronaries as a function of distance to cath, modified by patient rural status.

Results. Among 22,523 patients undergoing elective coronary angiography, 22.8% of patients had normal coronaries. Rural patients were more likely to have high Framingham risk (highly rural 44.2% vs. rural 39.8% vs. urban 36.9%, p<0.001). Rural patients who lived further from cath labs were associated with a lower probability of normal coronaries compared to rural patients closer to cath labs and urban patients across all distances(p=0.0014). For patients living within 50 miles of a cath lab, urban/rural status did not influence the probability of normal coronary angiography. From 50 to 150 miles, rural patients were less likely to have normal coronaries at angiography. (Figure 1).

Conclusions. Among patients undergoing elective angiography in the VA, those who lived further from cath labs in rural locations were less likely to have normal coronaries. These findings suggest that patient location, rather than clinical factors, may affect the use of coronary procedures. Further study to understand the causes of this relationship, including potential for under- or over-use based on distance and rural status, may help optimize patient access for coronary procedures.

Abstract 184: Patterns of Outpatient Follow-up after Acute Myocardial Infarction: Follow-up Frequency is Associated with Utilization of Evidence-based Therapies

Background: ACC/AHA guidelines recommend that patients with acute myocardial infarction (AMI) follow-up within several weeks of hospital discharge. Recommendations regarding intensity of following-up in the year following AMI are not provided. The relationship between frequency of follow-up and use of evidence-based therapies following AMI is unknown.

Methods: 6,838 patients from 2 multicenter prospective AMI registries, PREMIER and TRIUMPH registries were studied. We divided the number of patient self-reported outpatient follow-up visits with cardiologists, primary care providers, or both into tertiles: low, medium, and high. The primary outcome was use of statins, beta blockers, aspirin, ACE/ARBs, and a composite of all four medications at 12 months among eligible patients. The association between tertiles of visits following AMI among patients who had at least one visit and primary outcome was evaluated using hierarchical multivariable modified Poisson models.

Results: Mean number of follow-up visits in the year following AMI was 6 (IQR 3 - 8) and 189 (4%) of patients had no visits. In lowest tertile, patients had 1 to < 4 visits, in the medium tertile, 4 to < 7 visits, and in highest tertile, 7 to 59 visits. Patients in medium and high intensity tertiles were older, more likely to have insurance, and had higher GRACE 6-month mortality risk scores compared to the lowest tertile. In multivariable analyses, patients in the medium tertile were more likely to use statins and ASA than those in the lowest tertile (Figure). There were no differences in use of individual medications when comparing the highest and medium tertiles although individuals in the highest tertile were less likely to use all four medications.

Conclusions: Significant variability exists in follow-up frequency following AMI and 4% of the cohort had no follow-up. Patients who had medium intensity visits were more likely to use some evidence-based medications than those with low intensity. Higher intensity visits was not associated with greater medication use. It is possible that the observed differences may be attributed to unmeasured differences among patients rather than the actual follow-up visits. Prospective studies are needed to assess key elements of outpatient visits that may lead to better utilization of evidence-based therapies.

Abstract 330: Hospital-Level Variation in One-Year Mortality and Angina Following Myocardial Infarction is Not Explained by In-Hospital Quality of Care in the TRIUMPH Registry

Background: Studies demonstrating variation in hospital quality of care using longitudinal outcomes have been limited in the amount of clinical data used to stratify patients’ risks and have not examined health status outcomes. We sought to describe hospital-level variation in risk-adjusted health status and mortality in the year following myocardial infarction (MI) and describe the extent to which hospital quality of care explains this variation.

Methods: 4,316 patients from the TRIUMPH registry, a prospective cohort study of MI patients at 24 hospitals, were included for analysis. Using hierarchical models, we described the hospital-level variation in angina (yes/no) and 1-year mortality rates. We then added hospital quality of care measures for MI applicable to the time period studied (ASA and beta blockers within 24 hours of arrival and at discharge, ACE/ARB at discharge, thrombolytics within 30 minutes, PCI within 90 minutes, and smoking cessation instructions at discharge) to these models to determine if hospital variation in one-year mortality and angina were explained by index MI quality of care.

Results: The mortality rate at one year was 6.2% and the incidence of angina at one year was 23.0%. Unadjusted hospital-level 1-year mortality ranged from 0% to 10.8% and unadjusted presence of angina ranged from 9.3% to 66.7%. Statistically significant hospital-level variation in one-year mortality and angina was observed, with risk-adjusted mortality rates ranging from 5% to 8.3% (p<0.0001) and risk-adjusted angina rates ranging from 17.6% to 31.9% (p<0.0001). In-hospital quality of care measures did not attenuate hospital-level variation in mortality or angina (Figure 1).

Conclusions: Hospital-level variation in 1-year mortality and angina was observed among the 24 hospitals participating in this MI registry. However, this variation was not explained by in-hospital MI performance measures. Future studies should assess care delivery factors that impact longitudinal outcomes following MI.

Abstract 40: Hospital Variation in Premature Clopidogrel Discontinuation following Drug Eluting Stent Placement and Adverse Cardiovascular Outcomes from the VA Clinical Assessment, Reporting, and Tracking System for Cath Labs (CART-CL)

Background: Clopidogrel is recommended for 1 year following drug eluting stent (DES) placement and premature discontinuation has been associated with adverse outcomes. The extent of variation in premature discontinuation across hospitals within an integrated healthcare system is unknown. Accordingly, we assessed variation in premature clopidogrel discontinuation across all VA PCI sites and whether there was an association between hospitals with higher rates of premature discontinuation and adverse outcomes.

Methods: We used the VA CART-CL registry which includes all PCIs with drug eluting stents performed between 10/01/08 and 09/30/09 at 55 VA cath labs that used CART. We evaluated the frequency of patients who prematurely discontinue clopidogrel at 6 and 9 months using pharmacy refill data. Multivariable regression assessed the association between premature discontinuation and all-cause mortality and/or myocardial infarction (MI). We then grouped sites into quartiles of premature discontinuation and evaluated the association between hospital level premature discontinuation and adverse outcomes.

Results: Of the 7,022 patients who received a DES, 6.3% discontinued by 6 months, and 10.2% by 9 months. After risk adjustment, patients who discontinued clopidogrel prematurely had increased risk of adverse events with HR of 5.42 at 6 months (95% CI 4.22 – 6.99), and 6.24 at 9 months (95% CI 4.98 – 7.83). There was a significant trend in the unadjusted rates within quartiles toward increased risk of adverse outcomes among hospitals with greater rates of patients who discontinue prematurely by 6 months (p < 0.01 for trend, OR 1.65 CI 1.07 – 2.62 for comparison between quartile 1 and 4).

Conclusion: Premature discontinuation of clopidogrel is associated with adverse outcomes among patients who receive drug eluting stents. Hospitals with higher rates of premature discontinuation of clopidogrel have higher rates of adverse outcomes. Hospital-level interventions to reduce early discontinuation of clopidogrel therapy have the potential to improve outcomes of patients who receive a DES.

Abstract 5: A Comparison between Antihypertensive Medication Adherence and Treatment Intensification as Potential Clinical Performance Measures

Background: Medication adherence and treatment intensification have been advocated as performance measures to assess the quality of care. While previous studies have shown that adherence and treatment intensification (TI) of antihypertensive medications is associated with blood pressure control at the patient level, less is known about whether adherence and TI is associated with blood pressure control assessed at the clinic level.

Methods: We included 162,879 patients from 89 clinics in the CVRN Hypertension Registry with incident hypertension who were started on antihypertensive medications. Medication adherence was calculated by the proportion of days covered for five different classes of antihypertensive medications. TI was calculated as the number of observed medication intensifications minus the number of expected medication intensifications divided by the number of clinic visits in the observation period. TI is expected whenever measured blood pressure during a clinic visit is higher than the blood pressure goal. A TI score of -1 indicates that no treatment intensification was made when the BP was elevated, a score of 0 indicates that treatment intensification was made only when the blood pressure was elevated, and a score of 1 indicates that treatment intensification was made when the blood pressure was normal. Multivariable logistic regression assessed the association between medication adherence and treatment intensification with blood pressure control at the patient and clinic levels.

Results: The average patient age was 56 years, 13.6% had diabetes, 1.0% had previous MI, 3.8% had chronic kidney disease, 1.7% had a previous ischemic stroke, and 0.9% had CHF. Adherence measured by proportion of days covered was 0.77 ± 0.28 at the patient level and 0.78± 0.05 at the clinic level. The average patient TI score was 0.026 ± 0.23 and the average clinic TI score was 0.01± 0.04. Both adherence and treatment intensification were associated with blood pressure control at the patient level [OR for adherence of 1.28 (1.26 - 1.29) and OR for treatment intensification of 1.55 (1.53 - 1.57)]. In contrast, treatment intensification, but not adherence, was associated with blood pressure control at the clinic level [OR for TI at the clinic level 1.21 (1.17 - 1.25) and OR for adherence at the clinic level 1.01 (0.98 - 1.04)].

Conclusion: In this study, patient adherence to antihypertensive medications was not related to blood pressure control at the clinic level and may not be suitable as a performance measure. Treatment intensification was associated with BP control, but its use as a performance measure may be constrained by challenges in measuring it and by concerns about unintended consequences of aggressive hypertension treatment in some subgroups of patients.

Predictors of cardiovascular events and associated mortality within two years of kidney transplantation

BACKGROUND

Cardiovascular disease is the most common cause of death after renal transplantation. Furthermore, acute coronary syndrome (ACS) attributable to coronary artery disease (CAD) accounts for the majority of deaths due to cardiovascular disease posttransplant. While renal transplantation is the treatment of choice for end-stage renal disease, understanding the causes of graft and patient loss is exceedingly important to improve outcomes.

METHODS

This observational case-controlled study included 780 patients who underwent a kidney transplant between 1989 and 2001 who experienced early ACS (within 2 years). Patients were compared with controls matched for gender, year of transplant, and age. The primary outcome was the occurrence of an ACS event within 2 years after renal transplantation.

RESULTS

Cardiovascular disease was the most common cause of death, with all 13 cardiovascular deaths due to CAD. An additional 15 episodes of nonfatal ACS episodes occurred. Thirty-seven percent of early ACS occurred perioperatively, the majority in the first 3 posttransplant months. On multivariate analysis, diabetes (OR [odds ratio] 5.56; P = .0007), smoking (OR 3.56; P = .034), and prior transplant (OR 2.81; P = .047) were associated with early ACS.

CONCLUSIONS

Diabetes, smoking, and prior transplant were significantly associated with early ACS. The majority of events occurred perioperatively or within 3 months of transplant, highlighting the importance of improved screening and perioperative management.

Measurement of decay parameters for Xi- --> Lambda pi- decay

Based on 1.35 x 10(6) polarized Xi(-) events, we measure the parameter phi(Xi) to be -1.61 degrees +/-2.66 degrees +/-0.37 degrees for the Xi(-)-->Lambda pi(-) decay. New results for the parameters beta(Xi) and gamma(Xi) are also presented. Assuming that the CP-violating phase difference is negligible, we deduce the strong phase difference between the P-wave and S-wave amplitudes of the Lambda pi final state to be 3.17 degrees +/-5.28 degrees +/-0.73 degrees, reducing the uncertainty in estimating the level of CP violation in Xi-hyperon decay.

Effects of water temperature and salinity on parathyroid hormone-related protein in the circulation and tissues of elasmobranchs

Parathyroid hormone-related protein (PTHrP) is a hypercalcemic factor in mammals. The PTHrP antigen has been localized in both bony and cartilaginous fish tissues. Sites of localization included gills, skin and kidney, organs involved in osmoregulation. Physiological and localization experiments were carried out in elasmobranchs to dissect PTHrP's possible role in osmoregulation. The effects of alterations in the external environment on PTHrP in sharks were examined by keeping juvenile animals under conditions of increased temperature or decreased salinity. There were no alterations in the PTHrP levels in either the circulation or tissues. Significant correlations between plasma PTHrP, electrolyte and urea levels were seen in the pretreatment samples. The localization of PTHrP by immunohistochemistry and in situ hybridization revealed conserved sites of distribution from elasmobranchs to mammals, including skin, kidney, muscle and skeleton.

Noncommutative D-brane in a nonconstant NS-NS B field background

We show that, when the field strength H of the NS-NS B field does not vanish, the coordinates x and momenta p of open string end points satisfy a set of mixed commutation relations among themselves. Identifying x and p with the coordinates and derivatives of the D-brane world volume, we find a new type of noncommutative space which is very different from those associated with a constant B field background.

Search for direct CP violation in nonleptonic decays of charged Xi and lambda hyperons

A search for direct CP violation in the nonleptonic decays of hyperons has been performed. In comparing the product of the decay parameters, alpha(Xi)alpha(Lambda), in terms of an asymmetry parameter, A(XiLambda), between hyperons and antihyperons in the charged Xi-->Lambdapi and Lambda-->ppi decay sequence, we found no evidence of direct CP violation. The parameter A(XiLambda) was measured to be 0.012+/-0.014.

Immunodetection of parathyroid hormone-related protein in plasma and tissues of an elasmobranch (Scyliorhinus canicula)

We have used antiserum to human parathyroid hormone-related protein (PTHrP) (1-16) to examine tissues and plasma of the dogfish (Scyliorhinus canicula) for the presence of immunoreactive PTHrP (irPTHrP). The plasma contained high concentrations of irPTHrP (9.34 +/- 0.37 pM), comparable to levels in humans with hypercalcaemia of malignancy. Other tissues with irPTHrP included brain neurones; epithelial cells of the saccus vasculosus, kidney, rectal gland and choroid plexus; and cells of the pituitary pars distalis. PTHrP was not detected in gut, skin, oviduct, and gill epithelia, nor in branchial cartilage. The principal source(s) of plasma PTHrP is not known.

Parathyroid hormone-related protein is a factor in normal fish pituitary

Using antibodies to the amino-terminal region of human parathyroid hormone-related protein (PTHrP) we have demonstrated PTHrP immunoreactivity in pituitaries and plasma of the sea bream (Sparus aurata). Pituitary cells at two distinct locations contained immunodetectable PTHrP; an anterior group in the rostral pars distalis which also contained immunoreactive thyroid stimulating hormone (TSH), and a posterior group lying at the border of the pars intermedia and proximal pars distalis between cells which stained with antibody to human corticotrophin-like intermediate lobe peptide. By Western blot analysis pituitary extracts contained two immunoreactive isoforms of PTHrP, one of 29 kDa and the other of 26 kDa. Media of pituitaries incubated for up to 14 days in Krebs-Ringer bicarbonate also had several isoforms of immunodetectable PTHrP, two of them corresponding to the 29- and 26-kDa molecular forms but there were in addition both larger and smaller molecules. The concentration of PTHrP in sea bream plasma was comparable with levels observed in human subjects with humoral hypercalcaemia of malignancy. There was no reaction between pituitary cells or pituitary extracts and antibody to human parathyroid hormone. Thus sea bream pituitary contains immunoreactive PTHrP, which appears to be released into medium during in vitro incubation and which may be a significant source of plasma immunoreactive PTHrP in vivo.

Parathyroid hormone-related protein: a possible endocrine function in lactation

OBJECTIVE

Parathyroid hormone-related protein (PTHrP), initially discovered as the factor responsible for the syndrome of humoral hypercalcaemia of malignancy, has also been found to be expressed in placenta, in pregnant uterus, in the fetus at many locations, and in the lactating mammary gland. This study sought to establish whether PTHrP reaches the maternal circulation when it is expressed in mammary tissue during lactation or in the maternal reproductive tract during gestation.

DESIGN

Blood samples were collected from 53 subjects: 18 pregnant women in all stages of gestation, 19 lactating mothers and 16 non-lactating post-partum controls.

MEASUREMENTS

PTHrP was measured using a specific and validated radioimmunoassay. Parathyroid hormone was measured by two-site immunoradiometric assay. Total calcium was measured by atomic absorption spectrophotometry.

RESULTS

Circulating levels of PTHrP were readily detectable in 12 of 19 nursing mothers (range 2.7-7.8 pmol/l) but in none of the mothers who were bottle feeding. PTHrP was also detected in one of 18 pregnant subjects. Parathyroid hormone concentrations were lower in lactating mothers (2.3 +/- 1.0 pmol/l), than in non-lactating mothers (3.5 +/- 1.2 pmol/l) (P < 0.01).

CONCLUSION

PTHrP reaches the maternal circulation during lactation in amounts which could produce a systemic effect.

Behavioral effects of U-78875, a quinoxalinone anxiolytic with potent benzodiazepine antagonist activity

U-78875 (3-(5-cyclopropyl-1,2,4-oxadiazol-3-yl)-5-(1-methylethyl) imidazo[1,5-a]-quinoxalin-4(5H)-one) is a chemically novel compound with a high affinity for the benzodiazepine receptors. It has anticonflict effects in both the Vogel and Cook-Davidson models of anxiety, with a potency similar to that of diazepam (1-3 mg/kg, i.p.). In unanesthetized rats implanted with cortical electrodes for EEG recording, i.p. injections of U-78875 (3-10 mg/kg) increased the EEG power density in frequencies above 12 Hz, and decreased EEG power at lower frequencies. This EEG effect is similar to that of diazepam, and was completely antagonized by pretreatment with flumazenil. In animal models measuring central nervous system depression, U-78875 is much weaker than diazepam. It produced minimal impairment of rotarod performance in rats at doses up to 30 mg/kg, but at lower doses completely reversed the impairment from 10 mg/kg of diazepam. In rats trained to avoid shocks in a shuttle box, U-78875 (3-10 mg/kg) increased avoidance responses and antagonized the suppression of avoidance from diazepam (10 mg/kg). In the mouse one-trial passive avoidance task, pretreatment with U-78875 (1-10 mg/kg) before training produced no anterograde amnesia, but completely blocked the amnesic effect from diazepam (10 mg/kg). The diazepam antagonist potency for U-78875 is 10 to 100 times that of flumazenil. This unusual profile of mixed agonist/antagonist activities suggests U-78875 to be a unique anxiolytic agent with a minimum of central nervous system depression.

Both competitive and non-competitive antagonists of N-methyl-D-aspartic acid disrupt brightness discrimination in rats

Rats were trained to avoid or escape electric shocks in a symmetrical Y-maze by choosing to enter the brighter of two arms. Pretreatment with phencyclidine-like compounds disrupted brightness discrimination with greatly increased spontaneous locomotor activity between trials. The competitive antagonists of NMDA, 2-amino-7-phosphonoheptanoate (AP7) or 3-(+/-)-2-carboxypiperazin-4-yl)propyl-1-phosphonic acid (CPP) also disrupted brightness discrimination when injected into the cerebral ventricles, with no increase in movements between trials. The results suggest that the competitive antagonists of NMDA may impair sensory and cognitive functions in a manner similar to that produced by the phencyclidine-like compounds.

A specific antagonist of vasopressin produced plasma hyperosmolality and reduced ischemic-induced cerebral edema in rats

Water diuresis was produced in rats after s.c. injections of d(CH2)5,D-Ile2,Ile4-AVP (compound 1) and d(CH2)5,Tyr(OEt)2Val4-AVP (compound 2). Compound 1 is known to be a more potent antagonist against the antidiuretic effect, while Compound 2 is a more potent antagonist against the vasopressor effect of vasopressin. Compound 1 (but not compound 2) also increased plasma osmolality significantly at the diuretic doses. In rats rendered ischemic of the forebrain by 4-hour occlusion of both common carotid arteries, the resulting increases in brain water were significantly reversed by the injection of compound 1. Compound 2 did not reduce the edema. The results suggests a novel approach to the treatment of cerebral edema.

Growing familiarity with the test situation augments the vigilance deficit of aged rats

Quantitative EEG profiles of aged and of young-adult Fisher-344 rats were studied over 10 consecutive recording (3 h) sessions. The individual recording sessions were spaced regularly over a 6-week period. In the aged rats, in recordings from both the frontal cortex and dorsal hippocampus, the EEG characteristics of a vigilance deficit state increased in prominence up through 8 recording sessions before a stable level was reached. A similar trend was not seen in the young-adult rats; instead, these rats showed a stable level of spectral voltage output in both leads over the entire period of the 10 recording sessions. Thus, the aged rat brain appears to suffer from a vigilance deficit state which is revealed most clearly by testing under the basal conditions brought about by thoroughly familiarizing the subjects with the test environment.

Arousal deficit shown in aged rat's quantitative EEG and ameliorative action of pramiracetam compared to piracetam

The basal EEG profile of the aged Fisher-344 rat was consistently different from that of the young rat, showing dominant high voltage slow-wave components. These slow waves were present in both the frontal cerebral cortex and dorsal hippocampus. Absent or greatly attenuated in the aged rat's hippocampal EEG was rhythmic theta activity, which was always dominant in the young awake rat's hippocampus. These EEG differences were clearly apparent only under basal test conditions, i.e., following habituation to the test situation. Pramiracetam sulfate acted strongly to normalize the aged rat's EEG, while the action of piracetam was weak and appeared to undergo tolerance development.

Pharmacologic therapeutic window of pramiracetam demonstrated in behavior, EEG, and single neuron firing rates

Following oral or intravenous administration, a representative cognition activator drug, pramiracetam sulfate, is shown to have a pharmacologic therapeutic window at three different levels of study: learned behavior, gross EEG activity of the frontal cortex and hippocampus, and firing rate of single hippocampal neurons.

The effect of potassium on the intestinal transport of glucose

The rate of absorption of glucose, galactose, and 3-0-methylglucose was studied in the rat's small intestine perfused in situ with isosmotic solutions containing these sugars and Na(2)SO(4) or K2SO(4). The presence of high [K(+)] in the lumen enhances absorption of glucose but not that of galactose or of 3-0-methylglucose. The potassium stimulation is apparent at higher glucose concentrations where primarily carrier-mediated diffusion is involved in the translocation. In this case potassium stimulates transport even if it is the only cation in the lumen. The potassium-stimulated intestine produces more glycogen with higher specific activity than the control gut. Lactic acid production by the intestine is markedly enhanced if the intestinal lumen is perfused with a solution containing glucose and high [K(+)]. It is concluded that potassium does not affect permeability or the specific sugar transport system of the gut, but enhances intracellular metabolic disappearance of glucose thereby creating a larger luminal intracellular concentration gradient which in turn enhances the rate of carrier-facilitated entry.