Faculty Bibliography

BACKGROUND:Sodium bicarbonate therapy (SBT) is currently indicated for the management of a variety of acute drug poisonings. However, SBT effects on serum potassium concentrations may lead to delayed QTc prolongation (DQTP), and subsequent risk of adverse cardiovascular events (ACVE), including death. Emergency department (ED)-based studies evaluating associations between SBT and ACVE are limited; thus, we aimed to investigate the association between antidotal SBT, ECG changes, and ACVE. METHODS:This was a secondary data analysis of a consecutive cohort of ED patients with acute drug overdose over 3 years. Demographic and clinical data as well as SBT bolus dosage and infusion duration were collected, and outcomes were compared with an unmatched consecutive cohort of patients with potential indications for SBT but who did not receive SBT. The primary outcome was the occurrence of ACVE, and secondary outcomes were delayed QTc (Bazett) prolongation (DQTP), and death. Propensity score and multivariable adjusted analyses were conducted to evaluate associations between adverse outcomes and SBT administration. Planned subgroup analysis was performed for salicylates, wide QRS (> 100 ms), and acidosis (pH < 7.2). RESULTS:Out of 2365 patients screened, 369 patients had potential indications for SBT, of whom 31 (8.4%) actually received SBT. In adjusted analyses, SBT was found to be a significant predictor of ACVE (aOR 9.35, CI 3.6-24.1), DQTP (aOR 126.7, CI 9.8-1646.2), and death (aOR 11.9, CI 2.4-58.9). Using a propensity score model, SBT administration was associated with ACVE (OR 5.07, CI 1.8-14.0). Associations between SBT and ACVE were maintained in subgroup analyses of specific indications for sodium channel blockade (OR 21.03, CI 7.16-61.77) and metabolic acidosis (OR: 6.42, 95% CI: 1.20, 34.19). CONCLUSION/CONCLUSIONS:In ED patients with acute drug overdose and potential indications for SBT, administration of SBT as part of routine clinical care was an independent, dose-dependent, predictor of ACVE, DQTP, and death. This study was not designed to determine whether the SBT or acute overdose itself was causative of ACVE; however, these data suggest that poisoned patients receiving antidotal SBT require close cardiovascular monitoring.

In this study, we sought to determine the burden and characteristics of orgasmic dysfunction (OD) and concomitant erectile dysfunction (ED) in men with type 1 diabetes (T1D) enrolled in the Epidemiology of Diabetes Interventions and Complications (EDIC) study. In 2010, we assessed orgasmic and erectile function using the International Index of Erectile Function (IIEF). Sociodemographic, clinical, and diabetes characteristics were compared by OD status (OD only, OD and ED, no ED or OD). Age-adjusted associations between risk factors and OD status were examined. OD and ED information was available from 563 men. Eighty-three men (14.7%) reported OD of whom 21 reported OD only and 62 reported OD and ED. Age-adjusted odds ratios demonstrated that men who reported OD only had higher odds of depression, low sexual desire, and decreased alcohol use compared with men reporting no dysfunction. Men with OD concomitant with ED had greater odds of elevated hemoglobin A1C, peripheral and autonomic neuropathy, and nephropathy. Men reporting both dysfunctions were also more likely to report smoking, lower urinary tract symptoms, and had greater odds of androgen deficiency than men with no sexual dysfunction. Men with longstanding T1D suffer from an increased burden of OD. Psychogenic factors predominate in men reporting OD only while men who present with concomitant ED report increased burden of diabetes severity, characteristics previously observed with incident ED. ED may be the central impediment to sexual function in men with OD and ED. Longitudinal studies to characterize OD and ED experience over time are warranted.

BACKGROUND:Thirty years after the Mangled Extremity Severity Score (MESS) was developed, advances in vascular, trauma and orthopedic surgery have rendered the sensitivity of this score obsolete. A significant number of patients receive amputation during subsequent admissions, which are often missed in the analysis of amputation at the index admission. We aimed to identify risk factors for and predict amputation on initial admission or within 30 days of discharge (peritraumatic amputation or PTA). STUDY DESIGN/METHODS:The Nationwide Readmission Database for 2016 and 2017 was used in our analysis. Factors associated with PTA were identified. We used XGBoost, Random Forest, and Logistic Regression methods to develop a framework for machine learning (ML) based prediction models for PTA. RESULTS:We identified 1098 adult patients with traumatic lower extremity fracture and arterial injuries, 206 underwent amputation. 176 (85.4%) underwent amputation during the index admission and 30 (14.6%) underwent amputation within a 30-day readmission period. After identifying factors associated with PTA, we constructed machine learning models based on Random Forest, XGBoost, and Logistic Regression to predict PTA. We discovered that Logistic regression had the most robust predictive ability, with an accuracy of 0.88, sensitivity of 0.47, and specificity of 0.98. We then built on the Logistic Regression by the NearMiss algorithm, increasing sensitivity to 0.71, but decreasing accuracy to 0.74 and specificity to 0.75. CONCLUSIONS:ML-based prediction models combined with sampling algorithms (such as the NearMiss algorithm in this study), can help identify patients with traumatic arterial injuries at high risk for amputation and guide targeted intervention in the modern age of vascular surgery.

We used plasma neuronal extracellular vesicles to examine how neuronal insulin signaling proteins relate cross-sectionally to brain structure in nondemented older adults with varying levels of cortical amyloid. Extracellular vesicles enriched for neuronal origin by anti-L1CAM immunoabsorption were isolated from plasma of Atherosclerosis Risk in Communities-Positron Emission Tomography study participants (n = 88; mean age: 77 years [standard deviation: 6]). Neuronal extracellular vesicle levels of phosphorylated insulin signaling cascade proteins were quantified. Brain volume and white matter hyperintensity (WMH) volume were assessed using 3T magnetic resonance imaging. After adjusting for demographic variables and extracellular vesicle marker Alix, higher levels of a neuronal insulin signaling composite measure were associated with lower WMH and greater temporal lobe volume. Secondary analyses found the levels of downstream protein kinases involved in cell survival (p70S6K) and tau phosphorylation/neuroinflammation (GSK-3β) to be most strongly associated with WMH and temporal lobe volume, respectively. Associations between neuronal insulin signaling and lower WMH volume were attenuated in participants with elevated cortical amyloid. These results suggest that enhanced neuronal proximal insulin signaling is associated with preserved brain structure in nondemented older adults.

Breast cancer screening using Mammography serves as the earliest defense against breast cancer, revealing anomalous tissue years before it can be detected through physical screening. Despite the use of high resolution radiography, the presence of densely overlapping patterns challenges the consistency of human-driven diagnosis and drives interest in leveraging state-of-art localization ability of deep convolutional neural networks (DCNN). The growing availability of digitized clinical archives enables the training of deep segmentation models, but training using the most widely available form of coarse hand-drawn annotations works against learning the precise boundary of cancerous tissue in evaluation, while producing results that are more aligned with the annotations rather than the underlying lesions. The expense of collecting high quality pixel-level data in the field of medical science makes this even more difficult. To surmount this fundamental challenge, we propose LatentCADx, a deep learning segmentation model capable of precisely annotating cancer lesions underlying hand-drawn annotations, which we procedurally obtain using joint classification training and a strict segmentation penalty. We demonstrate the capability of LatentCADx on a publicly available dataset of 2,620 Mammogram case files, where LatentCADx obtains classification ROC of 0.97, AP of 0.87, and segmentation AP of 0.75 (IOU = 0.5), giving comparable or better performance than other models. Qualitative and precision evaluation of LatentCADx annotations on validation samples reveals that LatentCADx increases the specificity of segmentations beyond that of existing models trained on hand-drawn annotations, with pixel level specificity reaching a staggering value of 0.90. It also obtains sharp boundary around lesions unlike other methods, reducing the confused pixels in the output by more than 60%.

Advances in precision oncology, including RAS testing to predict response to epidermal growth factor receptor monoclonal antibodies (EGFR mAbs) in colorectal cancer (CRC), can extend patients' lives. We evaluated uptake and clinical use of KRAS molecular testing, guideline recommended since 2010, in the Veterans Affairs Healthcare System (VA).

Maternal diet, prior to and during pregnancy, plays an important role in the immediate and long-term health of the mother and her offspring. Our objectives were to assess diet quality among a large, diverse, urban cohort of pregnant women, and examine associations with sociodemographic and health behavior characteristics. Data were from 1,325 pregnant women enrolled in New York University Children's Health and Environment Study (NYU CHES). Diet quality was assessed using the Healthy Eating Index (HEI)-2015. Mean total HEI-2015 score was 74.9 (SD = 8.5); 376 (28%), 612 (46%), 263 (20%), and 74 (6%) of women had scores that fell into the grade range of A/B, C, D, and F, respectively. Mean HEI-2015 component scores were high for fruit and whole grains and low for protein-related, sodium, and fat-related components. In multivariable linear regression models, Hispanic women scored 1.65 points higher on the total HEI-2015 (95% CI: 0.21, 3.10) compared to non-Hispanic White women, while younger age (<30 years), parity, single status, pre-pregnancy obesity, smoking, pre-existing hypertension, moderate/severe depressive symptoms, not meeting physical activity recommendations, and not taking a vitamin before pregnancy were associated with ~1.5-5-point lower mean total HEI-2015 scores. Diet is a modifiable behavior; our results suggest a continued need for pre-conceptional and prenatal nutritional counseling.

Evidence link bacterial enterotoxins to apparent crypt-cell like cells (CCLCs), and Alpha Defensin 5 (DEFA5) expansion in the colonic mucosa of Crohn's colitis disease (CC) patients. These areas of ectopic ileal metaplasia, positive for Paneth cell (PC) markers are consistent with diagnosis of CC. Retrospectively, we: 1. Identified 21 patients with indeterminate colitis (IC) between 2000-2007 and were reevaluation their final clinical diagnosis in 2014 after a followed-up for mean 8.7±3.7 (range, 4-14) years. Their initial biopsies were analyzed by DEFA5 bioassay. 2. Differentiated ulcer-associated cell lineage (UACL) analysis by immunohistochemistry (IHC) of the CC patients, stained for Mucin 6 (MUC6) and DEFA5. 3. Treated human immortalized colonic epithelial cells (NCM460) and colonoids with pure DEFA5 on the secretion of signatures after 24hr. The control colonoids were not treated. 4. Treated colonoids with/without enterotoxins for 14 days and the spent medium were collected and determined by quantitative expression of DEFA5, CCLCs and other biologic signatures. The experiments were repeated twice. Three statistical methods were used: (i) Univariate analysis; (ii) LASSO; and (iii) Elastic net. DEFA5 bioassay discriminated CC and ulcerative colitis (UC) in a cohort of IC patients with accuracy. A fit logistic model with group CC and UC as the outcome and the DEFA5 as independent variable differentiator with a positive predictive value of 96 percent. IHC staining of CC for MUC6 and DEFA5 stained in different locations indicating that DEFA5 is not co-expressed in UACL and is therefore NOT the genesis of CC, rather a secretagogue for specific signature(s) that underlie the distinct crypt pathobiology of CC. Notably, we observed expansion of signatures after DEFA5 treatment on NCM460 and colonoids cells expressed at different times, intervals, and intensity. These factors are key stem cell niche regulators leading to DEFA5 secreting CCLCs differentiation 'the colonic ectopy ileal metaplasia formation' conspicuously of pathogenic importance in CC.

Poor dietary choices are a leading cause of chronic disease, but nutrition is rarely discussed in clinical practice. Nutrition is taught in less than a third of medical schools and physicians in practice empirically report low levels of comfort and self-efficacy in discussing nutrition with patients. A two-part presentation was created and shared with second-year medical students at a college of medicine. Students were given pre-work that included a brief (15 minutes) pre-recorded presentation and an e-resource entitled "Practical Nutrition for the Primary Care Provider" and then engaged in a live virtual session with a brief lecture and question and answer period (45 minutes). A survey was administered following the live presentation to evaluate the extent to which the presentation met the stated objectives and could impact participants' future practice. One-hundred and six students participated in the live lecture. Eighty-eight students (83%) provided survey feedback. Over two-thirds of respondents indicated that the presentation completely met the objectives, 57% indicated that they would definitely talk to patients with chronic disease about nutrition, and 52% indicated they would incorporate diet assessment in visits with patients with chronic disease. Nutrition is integral to disease prevention and management. Many students provided comments on the importance of the topic and benefit of the information. Further research is necessary to determine the optimal time and place for nutrition education in medical training. This presentation and e-resource are evidence-based, brief, and provided tools for participants to access once in practice.