Abtract

Poster #10: Patricia D. Maglalang [1] , James C. Cloyd [1]. The Evolution of Rescue Therapy for the Treatment of Seizure Emergencies. 

Some patients with drug resistant-epilepsy have seizure clusters, which can progress to prolonged seizures or status epilepticus. When patients or caregivers can identify such episodes, they could administer a therapy that interrupts seizure progression. In the 1980s, a team of clinical pharmacists, nurses, and epileptologists identified an unmet need for out-of-hospital management of seizure clusters. They designed an off-label therapy with rectally administered diazepam that triggered a new approach to treating seizure emergencies, now known as rescue therapy. Following reports on the off-label use of rectal diazepam, an industry-academic collaboration began development of a commercial rectal diazepam product in 1987. They obtained FDA orphan product designation in 1991 and the FDA approved diazepam rectal gel, Diastat®, in 1997. Diastat® approval launched a new approach to managing seizure emergencies. Given the objections to the use of rectal diazepam, clinicians began exploring off-label use of benzodiazepines by other routes. At the same time, industry and academic groups undertook formal development of rescue therapies. Approved intranasal and investigational intrapulmonary, intramuscular, buccal, and subcutaneous benzodiazepine products are in development. Rescue therapy has advanced the management of epilepsy, improved patient/family quality of life, and reduced health care costs. The number of rescue therapies is expanding. Non-rectal routes of administration offer convenience and may lead to faster seizure cessation resulting in fewer emergency department visits and improved quality of life. Further improvements in drug delivery coupled with seizure prediction technology hold the promise of fundamentally changing our approach to treating epilepsy.

​

Poster #11: Ashwin Karanam[1], Jacqueline A. French[2], Page B. Pennell[3], Cynthia L. Harden[4], Angela K. Birnbaum[1]. Population level simulations of LTG concentrations according to different dosing scenarios in pregnant women.

Rationale: Although therapeutic drug monitoring (TDM) for lamotrigine (LTG) is recommended during pregnancy, recommendations on dose adjustment are sparse. Studies show that LTG concentration decreases of ≥35%-of-preconception baseline increases risk of seizures. Additionally, evidence exists for the presence of two subpopulations of women based on the change in drug clearance(CL): high clearance change (HCC) group with increases of over 200% on average by the end of pregnancy and, low clearance change (LCC) group with very little CL increases in pregnancy. Our objective was to determine feasibility of developing LTG dosing recommendations during pregnancy
Methods: Literature models were used to perform simulations for both HCC and LCC groups at two levels: 1) Individual level simulations to observe the “typical individual” and; 2) population level simulations to observe population behavior for six dosage regimens at four different preconception doses. Metrics were calculated on both safety and efficacy to compare between the various scenarios
Results: Individual level simulations showed that without any dose changes in the HCC group, concentrations fell below 35%-of-baseline as early as the 8 weeks of gestation, but never fall below 35%-of-baseline for LCC group. Population level simulations show that 25 mg increases Q 2 weeks or 50 mg every month gave the best risk/benefit ratio in both HCC/LCC groups till gestational week 20. After 20 weeks the HCC group fell below 35%-of-baseline and LCC group tends to exceed toxic levels.
Conclusions: These results show that a “one-size fits all” philosophy does not work for LTG dosing in pregnant women with epilepsy and reinforces the need for TDM with individualized dosing.

​

Poster #12: Sathe AG[1,2], Ivaturi V[3], Brundage RC[2], Cloyd JC[1,2], Kapur J[4], Silbergleit R[5], Chamberlain J[6], Elm J[7], Lowenstein D[8], Shinnar S[9], Fountain N[4], Cock H[10], Coles LD[1,2]. A Pharmacokinetic Simulation Study to Assess Performance of a Sparse Blood Sampling Approach to Quantify Early Drug Exposure. 

Objectives
An ancillary study of the Established Status Epilepticus Treatment Trial to evaluate the relationship between early drug exposure of phenytoin (PHT), valproic acid (VPA) and levetiracetam (LEV), and seizure cessation used a 2-sample approach (1st collection window 20-50min (W1), 2nd window 60-120min (W2)). The objective of this simulation study is to compare the performance of an early exposure metric using non-linear mixed effects (NLME) modeling and non-compartmental analysis (NCA) approaches.
Methods
• Literature-based population pharmacokinetic models1-3 were used to simulate concentration-time profiles with and without residual unexplained variability for 100 patients. The true (without error) partial area-under-the-curve (pAUC) was obtained by integrating simulated concentrations from 0 to 2 hours.
• One timepoint each in W1 and W2 was randomly selected for each patient.
• NLME: All pharmacokinetic parameters were fixed to literature values except central volume which was estimated using population pharmacokinetic models. Individual empirical bayes’s estimates of pAUCs were obtained.
• NCA: Extrapolated concentrations at 10 min and 2 hours were determined and pAUCs were obtained.
• The percentage of subjects with a percent prediction error (PPE) within ± 20%. This was repeated 100 times to obtain a distribution of the success statistic.
Results
The median number of subjects that had their PPE within ± 20% (5th, 95th percentile) for NCA approach was found to be 72 (65, 79), 67 (58,73) and 80 (73,87) vs. NLME approach was 81 (73,88), 88 (82,93) and 92 (87,96) for PHT, VPA and LEV respectively.
Conclusions
These results support the use of 2-sample approach to quantify early drug exposure. Further, population pharmacokinetic modeling approach outperforms NCA for predicting early drug exposure.

​

Poster #13: Rui Zhong[1], Dustin Chernick[2], David Hottman[1], Ling Li[1,2]. The HDL mimetic peptide 4F mitigates vascular and cortical amyloid pathology and associated neuroinflammation in transgenic APPswDI mice. 

Cerebral amyloid angiopathy (CAA) features amyloid-β (Aβ) deposition in small arteries and capillaries of the cerebral cortex and the leptomeninges, and constitutes one of the pathological hallmarks of Alzheimer’s disease (AD). CAA is associated with cerebrovascular dysfunction and cognitive impairment. Substantial evidence has shown that high levels of high-density lipoprotein (HDL), and its main protein component, apoA-I, are associated with superior cognitive function in the elderly; in AD mice, our previous studies have shown that overexpression of human apoA-I rescues cognitive function by attenuating CAA and neuroinflammation. 4F is an 18 amino-acid HDL mimetic peptide advanced into cardiovascular clinical trials. Our preliminary data have shown that 4F inhibits Aβ aggregation. The present study was undertaken to investigate whether acute/chronic treatment with 4F mitigates CAA and associated cognitive deficits and neuropathologies in the transgenic APPswDI mouse model of CAA/AD. APPswDI mice were treated with i.p. injections of D-4F, the D-isomer that exhibits higher bioavailability and longer half-life. Two cohorts of age- and sex-matched APPswDI mice received either 1-week (acute) or 12-week (chronic) daily treatments of D-4F or PBS. After acute treatment, soluble Aβ was significantly reduced in the D-4F treated mouse brains; consistently, D-4F treatment trends toward decreased amyloid deposition and microglia recruitment in cortical/hippocampal regions. In the chronic study, D-4F was shown to rescue CAA-associated memory deficits in a sex-dependent manner in APPswDI model. Additional analyses are underway to unravel the molecular mechanisms underlying effects of D-4F. These findings suggest that HDL memetic peptides could be potentially therapeutic to mitigate CAA/AD.

​

Poster #14: Angela Jeong[1]; Shaowu Cheng[1]; David A. Bennett[2,3]; Ling Li[1]. Role of protein prenylation in the pathogenesis of Alzheimer’s disease.

Alzheimer’s disease (AD) is the leading cause of age-related dementia, but its pathogenesis
is not fully understood. Protein prenylation, a posttranslational lipid modification of proteins,
is dysregulated in aging and might contribute to AD pathogenesis. Many proteins including small GTPases undergo prenylation by farnesyltransferase (FTase) and geranylgeranyltransferases (GGTase-I, GGTase-II). Previously we showed that heterozygous deletion of FTase reduces amyloid pathology
and neuroinflammation, and rescues learning and memory in the APP/PS1 mouse model of
AD. To further investigate the role of protein prenylation in AD pathogenesis, we compared farnesylated or geranylgeranylated protein levels in 1) human postmortem brain samples from individuals with a range of clinical diagnosis from no cognitive impairment (NCI), mild cognitive impairment (MCI) to AD, 2) human amyloid precursor protein (APP)-overexpressing SH-SY5Y cells and mock-transfected controls, and finally 3) brain tissues from APP/PS1 AD mouse model and WT controls. We found that the levels of farnesylated H-Ras and its downstream protein phosphorylated ERK were higher in MCI and AD brains compared to control. Moreover, preliminary results reveal that protein farnesylation is abnormally elevated in human APP-overexpressing SH-SY5Y cells and aged APP/PS1 mouse brains compared to controls. Therefore, our study suggests that protein prenylation is dysregulated in AD and
that protein farnesylation could be a potential therapeutic target for AD.

Poster #15: Wenhui Qu[1]; Andrea Gram[2]; Ling Li[1,2]. TREM2 deficiency sustains cognitive functions and synaptic plasticity during aging. 

Human genetic studies have shown that triggering receptor expressed in myeloid cells 2 (TREM2) loss of function leads to elevated risks of developing late-onset Alzheimer’s disease (AD). TREM2 is expressed by microglia in the brain and modulates microglial phagocytosis and inflammatory responses. To recapitulate TREM2 loss of function in a mouse model, Trem2-/- (KO) mice were generated and tested for cognitive function. It has been shown that at six months of age, Trem2 deficiency does not alter cognitive performances or learning and memory function. At twelve months of age, TREM2 KO mice show a trend towards but not statistically significant deficits in learning and memory. These findings suggest that the effects of TREM2 deficiency on learning and memory may be age-dependent. We hypothesized that TREM2 deficiency and aging act synergistically to impair cognitive function. To test this hypothesis, TREM2 KO and wild-type (WT) mice at the age of six and eighteen months, respectively, are assessed for cognitive function by a battery of neurobehavioral tests, including the open field, elevated plus maze, and Morris water maze tests. In addition, electrophysiological experiments for hippocampal long-term potentiation (LTP) were conducted to evaluate synaptic plasticity in these TREM2 KO and WT mice. Results from this study are expected to shed lights on the roles of TREM2 in aging and shaping neuronal functions and provide novel insights into the pathogenesis of AD.

​

Poster #16: Lushan Wang1, Geoffry L. Curran4,5, Theresa Decklever4, Paul H. Min4, Ling Li 2,3,
Val J. Lowe4, Karunya Kandimalla1. Impact of amyloid β peptides on glucose trafficking in Alzheimer's brain

Background
Individuals with Type II diabetes mellitus (T2DM) demonstrate greater Alzheimer’s risk, whereas Alzheimer’s patients exhibit several pathophysiological attributes of T2DM, such as cerebrovascular dysfunction as well as impaired brain energy metabolism. Published studies have shown that amyloid beta (Aβ) peptides also exacerbate hepatic and peripheral insulin resistance in Alzheimer’s mouse models carrying amyloid precursor protein (APP) and human presenilin1 (PS1) transgenes (APP/PS1). These findings indicate that impairment of Aβ protein clearance, blood brain barrier (BBB) dysfunction and insulin resistance are associated with Alzheimer’s progression, even though the mechanism remains unknown.
Objective: We hypothesize that Aβ peptides cause cerebrovascular dysfunction, reduce insulin distribution to brain, impair brain glucose transport, and trigger brain insulin resistance.
Methods:
To test this hypothesis, plasma and brain Aβ kinetics were investigated in WT and AD transgenic mice (APP/PS1), which overexpress Aβ peptides. Role of Aβ peptides on the brain uptake of 18FDG was assessed in wild-type mice with/without Aβ protein infusion. Patlak plots were used to evaluate the brain influx clearance of 18FDG. Western blot/immunohistochemistry was further conducted to investigate glucose transporter (GLUT1) expression on brain microcapillaries harvested from WT mice with/without Aβ protein infusion. Next, we investigated the impact of PI3K and AKT inhibitor on 18FDG transcytosis and assessed the effect of high density lipoprotein (HDL) mimetics, D-4F, in reinstating the brain glucose metabolism in AD transgenic mice.
Results:
Patlak plots showed that the brain influx clearance of 18FDG was significantly higher in WT without Aβ infusion compared with WT without Aβ exposure. The 18FDG permeability in four groups of mice investigated in this study followed the rank order of: WT RC > WT Aβ infusion > APP/PS1 mice. Moreover, the expression of GLUT1 decreased in WT with Aβ exposure compared to WT without Aβ exposure. In vitro studies conducted in hCMEC/D3 monolayers have shown that HDL mimetic, D-4F, partially mitigated perturbation in AKT phosphorylation, which was previously shown to regulate GLUT1 activity at the BBB.
Conclusion:
Thus, Aβ exposure impaired the brain glucose metabolism and disrupted PI3K/AKT pathway that regulates glucose transport at the BBB. Furthermore, HDL mimetic peptide, D-4F, showed the potential to rectify AKT phosphorylation and reinstate glucose transport at the BBB in AD brain.

​

Poster #17: Kelly Showel, BS [1]; Amanda Wang, BS [1]; David Ferguson, PhD [1]. Variation in Cannabinoid Concentration Between Different Extraction Technique. 

The therapeutic potential of Cannabis has been explored as early as 2737 BC. Marijuana and hemp both stem from the Cannabis family, but differ in the concentrations of cannabinoids they contain. There are at least 113 known cannabinoids, the predominant four being tetrahydrocannabinol (THC), cannabidiol (CBD), cannabinol (CBN), and tetrahydrocannabivarin (THCV). Currently, there are three FDA approved CBD based drugs, whose mechanisms of action (MOA) are not fully understood. Previously it was thought that the actions of cannabinoids were limited to the endocannabinoid system. However, recent studies have demonstrated that cannabinoids – in particular CBD – have actions that extend well beyond the endocannabinoid system. Additionally, studies have shown differences in therapeutic outcome when using Cannabis flower for particular disease states. A possible explanation for these discrepancies is the variation in the drug delivery system being used. It is thought that different delivery systems lead to different concentrations of specific cannabinoids. These observations could shed light on “the entourage effect” theory – the idea that cannabinoids exert their full potential through synergistic mechanisms. The mystery behind the entourage effect, the general MOA of cannabinoids, and the variations in therapeutic outcomes, led to this study. The goal of this study was to utilize different extraction techniques on Cannabis sativa, to see if there were differences in the amount of CBD and other cannabinoids present. This was done through HPLC analysis of three different extraction techniques, with confirmation using CBD standards and additional analytical techniques – including H1NMR and mass spectroscopy.

​

Poster #18: Mélanie T. Mahoney[1], Hunter V. Brigman[1], Brian D. Johnston[2], Brian R. Raux[3,4], James R. Johnson[2,5], Elizabeth B. Hirsch[1]. Escherichia coli (EC) ST131-H30 Clonal Group is Associated with Antimicrobial Resistance, Illness Severity, Host Compromise, and Non-Cure among Patients with Bacteriuria. 

Background:
EC sequence type ST131 is the leading cause of extraintestinal EC infections, and accounts for most fluoroquinolone (FQ)-resistant and extended-spectrum β-lactamase (ESBL)-producing EC clinical isolates. The ST131-H30 subclone (H30) is responsible for most antimicrobial resistance within ST131; however, H30's impact on clinical outcomes is poorly defined.
Objective:
To compare clinical outcomes of patients with bacteriuria caused by ST131 vs non-ST131 EC, and by H30 vs non-H30 EC strains.
Methods:
Phylogroups, ST131, H30, and CTX-M-type β-lactamase genes were detected by PCR for 142 EC isolates collected prospectively from adults with monomicrobial bacteriuria at a Boston academic medical center (8/2013-1/2014). Baseline characteristics, empiric treatment, and clinical cure rates were collected retrospectively from health records and compared between ST131 vs non-ST131 cohorts.
Results:
Overall, 35% of isolates were ST131, of which 80% (39/49) were H30. Compared with other isolates, H30 isolates demonstrated a higher frequency of ESBL production (33% vs 8%; p < 0.001) and FQ resistance (90% vs 8%; p < 0.001). Patients with H30 isolates (vs. non-H30 isolates) had higher median (interquartile range [IQR]) APACHE II scores (10 [4] vs 8 [9.5]; p = 0.01), and had a lower clinical cure rate (84% vs 96%; p = 0.08). Patients with ST131 vs. non-ST131 isolates had similar median [IQR] APACHE II scores (9 [5] vs 8 [9]), frequencies of symptomatic UTI (61% vs 70%), and clinical cure rates (87% vs. 95%).
Conclusions:
ST131-H30 was associated with ESBL production, FQ resistance, illness severity, and lower clinical cure rates in symptomatic UTI.

​

Poster #19: Connor McDermott[1]; Elizabeth A. Ambrose, PhD[1,2]. Anthrax Antitoxin Lead Optimization via Bioisosteric Replacement and Other In Silico Strategies.

Bacillus anthracis, the causative agent of the deadly bacterial infection anthrax, is a well-known bioterrorism agent in need of effective countermeasures. Current treatment options include antibiotics and antibody-based therapeutics, but neither directly targets the primary cause of the lethality of these infections: the lethal factor (LF), a zinc-metalloprotease and component of the tripartite exotoxin that the bacteria secrete. LF, which interferes with cellular immune defense mechanisms and induces endothelial cell apoptosis, has therefore become a popular target for the development of new anthrax therapeutics; however, no LF inhibitors have yet been approved to treat anthrax. Recently, we performed a large-scale experimental high-throughput screen and identified two promising small molecules active against LF. Here we report the structures of these hits as well as efforts to increase their solubility for structural biology studies while retaining their inhibitory activity towards LF, primarily via targeted bioisosteric replacement and a variety of virtual screening techniques. We also employed biophysical fragment-based screening to identify functional groups that increase inhibitory activity against LF, and these results are presented herein. Finally, we report new structural biology data crucial in elucidating the binding modes of our novel compounds as well as key structural features that contribute to strong and specific LF inhibition.

​

Poster #20: Ibrahim Y. Abdelgawad, BS; Marianne K.O. Grant, BA; Christine A. Lewis; and Beshay N. Zordoky, MSc, PhD. Doxorubicin induces Senescence Phenotype in Endothelial Cells.

Background: Doxorubicin (DOX) is one of the most effective chemotherapeutic agents. Nevertheless, clinical utility is limited due to cardiotoxicity. Several mechanisms have been proposed to explain DOX-induced cardiotoxicity. Recent studies have demonstrated that DOX leads to the accumulation of senescent cells resulting in cardiovascular dysfunction. Endothelial cell senescence has been shown to be a leading cause in endothelial dysfunction which can result in a wide range of premature cardiovascular diseases. In our current work, we aim to identify the effect of DOX on the senescence of endothelial cells using EA.hy926 human endothelial-derived cell line and clinically relevant juvenile C57Bl/6 mouse model.
Methods: EA.hy926 human endothelial-derived cells were treated with different concentrations of DOX for 24 hours. Thereafter, DOX was removed and the cells were incubated in DOX free media for another 72 hours. Protein and gene expression of senescence markers were measured using Western Blot and real-time PCR, respectively. Additionally, cell cycle analysis was assessed using flow cytometry.
Results: DOX treatment induced the protein and gene expression of senescence markers in endothelial cells. Moreover, DOX induced cell cycle arrest in the G2/M phase. Additionally, DOX administration in clinically relevant mouse model upregulated the gene and protein expression of p21, a senescence marker, in different organs.
Conclusions: Our current work demonstrates DOX-induced senescence in human endothelial cells. Our future goal is to investigate multiple senolytic strategies against DOX-induced senescence in vitro. Promising senolytics will be studied in vivo using our newly developed tumor-bearing juvenile mouse model.

Poster #21: Ibrahim Y. Abdelgawad, BS[1]; Marianne K.O. Grant, BA[1]; Christine A. Lewis[1]; Flavia E. Popescu, PhD[2]; David A. Largaespada, PhD[2]; and Beshay N. Zordoky, MSc, PhD[1]. Characterizing Doxorubicin Cardiotoxicity in Young Tumor-Bearing Mice.

Background: Childhood cancer survivors have about 15-times higher risk of heart failure than their siblings mainly because of the cardiotoxicity of chemotherapy. Nearly 50% of pediatric cancer patients receive doxorubicin (DOX) which is known to be cardiotoxic. DOX-induced cardiotoxicity had been extensively studied in tumor-free mouse models. The objective of the current study is to characterize a juvenile tumor-bearing immunocompetent mouse model of DOX-induced cardiotoxicity.
Methods: Juvenile four-week old male C57Bl/6 mice were injected subcutaneously with EL4 lymphoma cells suspended in sterile PBS, while tumor-free mice were only injected with sterile PBS. One week later, both tumor-free and tumor-bearing mice were injected intraperitoneally with either a clinically relevant dose of DOX or saline for 3 weeks. One week after last DOX injection, mice were euthanized and the hearts and tumors were harvested. Gene expression of molecular markers of cardiotoxicity and inflammation was quantified using real-time PCR.
Results: DOX administration significantly reduced tumor weights in tumor-bearing mice. DOX administration caused significant cardiac atrophy in tumor-free and tumor-bearing mice. Moreover, tumor itself caused significant reduction in heart weight compared to tumor-free mice. Additionally, DOX administration induced the expression of markers of cardiotoxicity in both tumor-free and tumor-bearing mice. Saline-treated tumor-bearing mice demonstrated higher expression of TNF-alpha compared to tumor-free mice.
Conclusion: We established a juvenile tumor-bearing immunocompetent mouse model to investigate DOX-induced cardiotoxicity. Our tumor-bearing mouse model will optimize our understanding of the interplay between tumor and DOX-induced cardiotoxicity. Additionally, this model will be used for investigation of cardioprotective strategies.

​

Poster #22: Marianne K. O. Grant, BA [1], Davis Seelig, DVM, PhD [2], Ibrahim Y. Abdelgawad, BSc[1], and Beshay N. Zordoky, MSc, PhD [1]. Lack of Sex-related Differences in Isoproterenol-induced Cardiac Dysfunction in C57Bl/6 Mice. 

Background and Objective: Sex-related differences in cardiovascular diseases are complex and highly context-dependent. The objective of this work was to comprehensively determine key sex differences in the response to acute and chronic adrenergic stimulation in C57Bl/6 mice.
Methods and Results: There was no statistically significant difference in key echocardiographic parameters between male and female C57Bl/6 mice in response to acute adrenergic stimulation (a single sub-cutaneous dose of isoproterenol 10 mg/kg). After chronic adrenergic administration (sub-cutaneous injections of isoproterenol 10 mg/kg/day for 14 days), there was similar degree of cardiac dysfunction, cardiac hypertrophy, and myocardial fibrosis in male and female mice. Similarly, chronic isoproterenol administration induced hypertrophic and fibrotic genes in hearts of male and female mice to the same extent. Intriguingly, gonadectomy of male and female mice did not have a significant impact on isoproterenol-induced cardiac dysfunction as compared to sham-operated animals.
Conclusions: The current work demonstrated lack of sex-related differences in isoproterenol-induced cardiac hypertrophy, dysfunction, and fibrosis in C57Bl/6 mice. This study challenges the conventional dogma of the detrimental cardiovascular effects of male sex hormones and the beneficial effects of female sex hormones.

Poster #23: David J. Margraf, PharmD, MS, BCPS[1]; Scott Seaburg, PharmD[2]; Gregory J. Beilman, MD, FCCM[3]; Julian Wolfson, PhD[4]; Jonathan C. Gipson, MD, FACS[5]; Scott A. Chapman, PharmD[1]. Propensity Score Adjusted Comparison of 3-Factor versus 4-Factor Prothrombin Complex Concentrate for Emergent Warfarin Reversal: A Retrospective Cohort Study.

Background: Prothrombin Complex Concentrates (PCCs) are prescribed for emergent warfarin reversal (EWR). Current guidelines recommend four factor-PCC (PCC4). Few direct comparisons with three-factor PCC (PCC3) exist.
Objectives: To describe and characterize the differences in efficacy and safety between PCC4 and PCC3.
Methods: Patients receiving PCC3 or PCC4 for EWR were identified. Patient characteristics, clinical outcomes and dosing information were collected. Patients whose pre-PCC INR was > 6 hours before PCC dose or the pre-post PCC INR was > 12 hours were excluded. The primary outcome was achieving an INR ≤ 1.5 post PCC. Secondary outcomes were the change in INR over time, post INR, thromboembolic events (TE), and death during hospital stay. Logistic regression was used with and without a propensity score adjustment accounting for age, sex, actual body weight, dose, initial INR value, and time between INR measurements.
Results: Eighty patients were included (PCC3 = 57, PCC4 = 23). Baseline INR was similar (p=0.10). Dose (U/kg) was greater for PCC4 patients (29.3 (25.9-37.3) vs. 21.5 (20.4-25.9), p<0.001). However, the U/kg:INR ratio was not different (PCC4: 8.2 (7.0-10.2) vs PCC3: 7.9 (5.6-10.5), p=0.3). More PCC4 patients achieved goal INR (87.0% vs. 31.6%, odds ratio (OR) =14.4, 95% CI: 3.80-54.93, p<0.001). The result was similar after adjustment (AOR = 10.7, 95% CI: 2.17-51.24, p<0.001). Death during hospital stay (p=0.52) and TE (p=1.00) were not significantly different.
Conclusions: PCC4 was associated with goal INR achievement more than PCC3. This relationship was observed in the unadjusted and propensity score adjusted results.

Poster #24: Yiwang Guo; Changquan Calvin Sun. Effect of gastric fluid secretion rate on dissolution of erlotinib in an artificial stomach and duodenum - implication on bioavailability.

Using an artificial stomach and duodenum (ASD), we show that decreasing gastric liquid secretion rate from 2.5 to 0.5 mL/min, dissolution of erlotinib (ERL) reduced by 5-fold in duodenum chamber with AUC varying from 14,000 to 3,000 (μg‧min)/mL. This profound effect was related to a change in supersaturation ratio generated among groups that leads to the different extent of precipitation. The findings imply that, for ERL, patients with different gastric liquid secretion rates (e.g., due to differences in gender, age, disease) may suffer from large variations in clinical outcomes. Therefore, appropriate formulation strategies must be implemented to ensure a robust clinical performance. The study also showed the capability of ASD on guiding the formulation development and optimization during drug development.

​

Poster #25: Kunlin Wang, Chenguang Wang, Changquan Calvin Sun. Solid-state properties and manufacturability of two Celecoxib N-Methyl-2-Pyrrolidone solvates and insights into the structure-property relationship.

Celecoxib (CEL) is a COX-2 selective nonsteroidal anti-inflammatory drug for pain and inflammation. The marketed form III CEL has low solubility, and renders poor flow properties for satisfactory tablet manufacture. The goal of this work is to improve manufacturability of CEL, through forming two novel solvates with N-Methyl-2-Pyrrolidone (NMP).
The mono-NMP and di-NMP solvates of CEL were prepared by cooling, slow evaporation and slurry methods. The thermodynamic phase transition temperature (Tt) of the two solvates was bracketed by seeded slurry experiments at different temperatures, where equilibrium solids were characterized by Powder X-Ray Diffraction. Thermal properties were characterized by Differential Scanning Calorimetry (DSC), Thermogravimetric Analysis, and Hot-Stage Microscopy (HSM). Intermolecular interactions and mechanical properties were studied by crystal structure analysis and nanoindentation.
HSM and DSC revealed a solvent-mediated phase transformation where the di-NMP crystals converted to mono-NMP solvate upon heating. Tt was identified between 35.5 °C and 36 °C. The monosolvate exhibited superior flow properties against CEL Form III owing to its more equidimensional morphology. The Hardness (H) and Elastic modulus (E) of the disolvate were lower than those of the monosolvate, and both were lower than CEL From III. This is correlated with their intermolecular interactions and packing. These structural differences also explains the conversion of the disolvate to monosolvate under vacuum and upon heating.
We report two novel solvated phases of CEL, mono-NMP and di-NMP, their solid-state properties and improved tablet manufacturability. Structure-mechanical property relationship insights were obtained for their stability and mechanical properties to guide future product design.

​

Poster #26: Yifan Li, BS[1]; Changye Li, MS[2]; Rui Zhang, PhD[3]. Identifying and predicting risk factors of potential problem opioid use in chronic noncancer pain patients using electronic health records.

Background Identifying problem opioid use and risk factors are critical to prevent severe clinical consequences prior to and during opioid therapy. Traditional diagnostic coding system (i.e. ICD) incompletely documents opioid use problems. Relevant information can be found in clinical notes in Electronic Health Record (EHR).
Methods A total of 2,034,751 patients with chronic noncancer pain and on opioid medication, who visited Minnesota Fairview healthcare system between 2015-2017, were identified. Structured data and unstructured clinical notes in EHR were collected. A dictionary of problem opioid use related terms was developed. A natural language processing (NLP) system with the dictionary was developed to identify potential problem opioid use. Patients who were identified with problem opioid use as the case group were further compared with patients who were not as the control group to identify potential risk factors.
Results A total 11,819 patients including 3,509 from 2015, 4,071 from 2016 and 4,239 from 2017, were identified with problem opioid use using NLP system. 3,509 patients from 2015 were further compared with control group (3,722) for risk factors identification. The elder age group (P=0.028), American Indian or Alaska Native race (P=0.019), Black or African American race (P=0.004), Non-Hispanic or Latino ethnicity (P=0.015), and ICD-10 diagnosis of opioid (P<0.001), drug (non-opioids) (P<0.001), alcohol (P<0.001), nicotine (P<0.001) abuse/dependence, mental health disorder (P<0.001) and hepatitis (P<0.001) were identified as risk factors.
Conclusion NLP techniques can efficiently process free text information in clinical notes in EHR to identify potential problem opioid use and its risk factors.

Poster #27: Patricia D. Maglalang, Mahli V. Scheible, McKenzie K. Skurat, Sarah K. Jackson, Megan Fahy, Larissa Ostfeld, Tuyen Nguyen, Jazmin N. Belknap, Hoa Bui, Meg M. Little, Randall Seifert. Impact of Continuing Education Focused on Drugs of Abuse for High School Faculty and Staff.

Background: A group of student pharmacists developed an educational program on drugs of abuse (DA) for high school teachers, nurses, counselors, and other staff. The program covers the basic pharmacology of substances, current trends and statistics, signs of misuse, and safety considerations. This project aims to empower school staff to recognize drug abuse among students.
Objectives: (1) Develop and analyze the effectiveness of the educational program for high school teachers, nurses, counselors, and staff; (2) provide tools to identify drug abuse in high school students, and (3) describe methods for prevention and intervention.
Methods: Building on an established partnership, school nurses, College faculty and student pharmacists developed and pilot tested educational materials for DA. After incorporation of feedback from 6 school nurses and counselors, materials were presented at an Inservice day (n=42) at a large suburban high school. Participants completed a 10-item online survey. Questions included text entry and 5 point star rating (Likert).
Outcomes: Materials were rated as follows: would recommend to a colleague (3.9); overall quality of information and content (3.9); easy to follow (3.8); engaging (3.7); logical sequence (3.9). Phase 2 will include three school districts: two urban and one rural community bringing the total educational impact to over 200 staff members and 10,000 students. The leadership of student pharmacists in this project offers the opportunity to support additional school districts. This program opens opportunities for pharmacists to impact school health and increases access to information that educates staff, empowering them to combat substance use among students.