Citations

Abstract Introduction Distinct temperature patterns have long been recognized to correlate with fevers of differing etiologies. While the use of wearable sensors for high-frequency temperature monitoring (HFTM) on a near minute-by-minute basis has been shown to detect fevers earlier than standard-of-care nursing vital sign assessments in hospitalized patients, leveraging these high-resolution datasets to computationally identify unique digital signatures for real-time diagnosis of underlying fever etiology has not been widely explored. Diagnostic uncertainty is common in patients undergoing hematopoietic stem cell transplantation (HCT), with only 20–30% of febrile neutropenic episodes being microbiologically documented. We hypothesized that unique temperature patterns extracted from HFTM data collected during episodes of febrile neutropenia could be used to develop a supervised machine learning classifier capable of accurately predicting underlying fever etiology in HCT patients. Methods We analyzed 68 clinically independent fever episodes recorded in HCT patients (n=90) outfitted with an FDA-cleared wireless temperature sensor (TempTraq®, BlueSpark Technologies) that measured axillary temperature every 2 minutes throughout hospitalization. Time-series features were extracted from temperature traces spanning 1 hour before to 3 hours after fever onset and used to train a suite of machine-learning models to distinguish engraftment fevers from other fever etiologies. Model training and evaluation were performed using repeated stratified 5-fold patient-level cross-validation, yielding 100 train-test evaluations. Results Among all classification models, the logistic regression classifier provided the best overall performance and interpretability, achieving 94% specificity (95% CI, 0.84–1.0) for identifying engraftment fevers with a mean AUROC of 0.88 ± 0.10. Feature importance analysis demonstrated that both clinical variables and HFTM-derived temperature dynamics contributed to model performance, with a strong reliance on time-series features captured within the first 4 hours of fever onset. Conclusion Our study provides a demonstration that continuous temperature data collected from patients outfitted with wearable sensors can be leveraged not only for early fever detection but also for machine learning–based diagnosis of fever etiology. These findings suggest that dynamic temperature patterns contain clinically meaningful physiologic information that with further studies could support real-time diagnostic decision-making and guide safe de-escalation of empiric antibiotics during febrile neutropenia in patients undergoing intensive cancer therapy.

592 Background: While not all fevers occurring during hematopoietic stem cell transplantation (HCT) are dangerous, episodes of febrile neutropenia tend to be managed aggressively given the significant mortality and morbidity associated with infectious complications of transplantation. This can potentially lead to unnecessary diagnostic testing, antibiotic treatment, and prolonged hospital stays for non-infectious fever etiologies such as engraftment. Our group has previously demonstrated that high-frequency body temperature monitoring (HFTM) using a wearable device detects fevers earlier in transplant recipients (Flora et al, Cancer Cell, 2021). We hypothesized that applying supervised machine learning to HFTM data collected from inpatient allogeneic HCT patients would allow for accurately distinguishing between non-infectious and infectious etiologies of fever and could potentially guide safe de-escalation of antibiotics. Methods: We analyzed 74 independent fever events detected in HCT patients (n=42) outfitted with an FDA-cleared, wireless sensor measuring axillary temperatures every 2 minutes for the duration of their respective hospitalizations. After data filtering, independent fever events were then retrospectively grouped by fever etiology using physician-determined annotation as previously described (Flora et al., 2021). A split ratio of 80:20 was used to create training and test sets from the parent data set, consisting of 59 and 15 fever events for each split, respectively. Training sets were used to create a classification algorithm identifying bacterial infections requiring antibiotics using the machine learning model XGBoost. The performance of this model was then evaluated using test sets of the remaining fevers (n=15). Results: Averaging over 1000 test sets of 15 independent fevers comprising 5 bacterial infection-related fevers and 10 non-infectious fevers, our model correctly identified 82% (mean recall) of cases requiring antibiotics with a mean precision of 0.41. The model also correctly identified 42% of non-infectious fevers (mean true negative rate, 0.42) that would otherwise have been treated with unnecessary antibiotics. The mean AUCPR was 0.58, with the best performing individual model achieving an AUCPR of 0.97. Conclusions: The preliminary results of this study demonstrate that our classification model shows promise for differentiating fevers requiring antibiotics due to bacterial infections from other non-infectious fever etiologies. Prospective studies further developing supervised machine learning models for subcategorization of febrile neutropenia could pave the way for safely de-escalating antibiotics in HCT and other immunocompromised patients undergoing cancer treatment monitored using HFTM.

AbstractThe characterization of single extracellular vesicle (EV) has been an emerging tool for the early detection of various diseases despite there being challenges regarding how to interpret data with different protocols or instruments. In this work, standard EV particles were characterized for single CD9+, single CD81+ or double CD9+/CD81+ tetraspanin molecule positivity with two single EV analytic technologies in order to optimize their EV sample preparation after antibody labelling and analysis methods: NanoImager for direct stochastic optical reconstruction microscopy (dSTORM)‐based EV imaging and characterization, and Flow NanoAnalyzer for flow‐based EV quantification and characterization. False positives from antibody aggregates were found during dSTORM‐based NanoImager imaging. Analysis of particle radius with lognormal fittings of probability density histogram enabled the removal of antibody aggregates and corrected EV quantification. Furthermore, different machine learning models were trained to differentiate antibody aggregates from EV particles and correct EV quantification with increased double CD9+/CD81+ population. With Flow NanoAnalyzer, EV samples were prepared with different dilution or fractionation methods, which increased the detection rate of CD9+/CD81+ EV population. Comparing the EV phenotype percentages measured by two instruments, differences in double positive and single positive particles existed after percentage correction, which might be due to the different detection limit of each instrument. Our study reveals that the characterization of individual EVs for tetraspanin positivity varies between two platforms—the NanoImager and the Flow NanoAnalyzer—depending on the EV sample preparation methods used after antibody labelling. Additionally, we applied machine learning models to correct for false positive particles identified in imaging‐based results by fitting size distribution data.

Background. Danlou tablet (DLT), the traditional Chinese medicine has been commonly used for dyslipidemia, atherosclerosis, and coronary heart disease. Whether it was effective against vascular injury caused by CIH has remained unknown. The aim of the current study was to observe the effects of DLT on chronic intermittent hypoxia (CIH)-induced vascular injury via regulation of blood lipids and to explore potential mechanisms. Methods. Sixteen 12-week-old male ApoE−/− mice were randomly divided into four groups. The sham group was exposed to normal room air, whereas the other three groups were exposed to CIH. Mice in the CIH + normal saline (NS) group were gavaged with NS. Mice in the CIH + Angptl4-ab group were intraperitoneally injected with Angptl4-antibody. Mice in the CIH + DLT group were gavaged with DLT. After four weeks of intervention, serum lipid concentrations, and serum lipoprotein lipase (LPL) activity were detected. The changes in atherosclerosis in vascular tissue were detected by hematoxylin and eosin (H&E) staining. Quantitative real-time polymerase chain reaction (qRT-PCR) and Western blot analysis were applied to detect the expression levels of hypoxia-induciblefactor-1 (HIF-1), factor-inhibiting HIF-1 (FIH-1), angiopoietin-like 4 (Angptl4), and LPL in different tissues. Results. CIH exposure increases serum lipid levels, decreases serum LPL activity, and exacerbates atherosclerosis. Both Angptl4-ab and DLT treatment reversed the changes in lipid concentration, LPL activity, and atherosclerosis caused by CIH. In the epididymal fat pad, CIH exposure decreased the expression of FIH-1 and increased the expression of HIF-1, whereas DLT treatment increased the expression of FIH-1 and LPL and inhibited the expression of HIF-1 and Angptl4. In heart tissue, the expression levels of LPL and Angptl4 were not affected by modeling or treatment. Conclusions. DLT improved vascular damage by improving the increase in blood lipids induced by CIH, potentially by upregulating FIH-1 and downregulating HIF-1 and Angptl4 in adipose tissue. Therefore, DLT may be a promising agent for the prevention and treatment of CIH-induced vascular injury.

Background Obstructive sleep apnea ( OSA ) is an important risk factor for the development of cardiovascular diseases including myocardial infarction ( MI ). The aim of this study was to investigate the effects of OSA on prognosis after MI , and to determine which specific measures of OSA severity best predicted outcomes. Methods and Results We performed a prospective study, in which 112 patients without a prior diagnosis of sleep apnea underwent comprehensive polysomnography within a median of 7 days after MI . Patients were followed up at 6‐monthly intervals (±2 weeks) for a total of 48 months. Patients classified with central apnea (n=6) or those using continuous positive airway pressure (n=8) after polysomnography were excluded from analyses. The primary end point was major adverse cardiac events, including death from any cause, recurrent MI , unstable angina, heart failure, stroke, and significant arrhythmic events. Forty of 98 patients (41%) had OSA (apnea‐hypopnea index ≥15 events/h). OSA patients had higher major adverse cardiac event rates when compared to those without OSA (47.5% versus 24.1%; χ 2 =5.41, P =0.020). In a multivariate model that adjusted for clinically relevant variables including age, left ventricular ejection fraction, diabetes mellitus, oxygen desaturation index, and arousal index, significant hypoxemia, as defined by nocturnal nadir oxygen saturation ≤85%, was an independent risk factor for major adverse cardiac events (hazard ratio=6.05, P =0.004) in follow‐up 15 months after baseline. Conclusions Nocturnal hypoxemia in OSA is an important predictor of poor prognosis for patients after MI . These findings suggest that routine use of low‐cost nocturnal oximetry may be an economical and practical approach to stratify risk in post‐ MI patients.

Complementary and alternative medicine (CAM) has gained acceptance throughout the industrialized world. The present study was performed to provide information about the use of CAM at Mayo Clinic, an academic medical center in Northern Midwest of the US. We retrospectively reviewed the electronic medical records of 2680 patients visiting the CAM program at Mayo Clinic, Rochester, between 1 July 2006 and 31 March 2011. Services provided included acupuncture, massage, integrative medical consultations and executive stress management training. Data including age, gender, race, diagnosis and the number of treatment/consultation sessions were collected to describe the use of CAM in our institute over the last several years. It was found that the mean (standard deviation) age of patient was 52.6 (15.5) years. Of those, 73.1% were female and 26.9% were male. Most patients were white. The number of patients referred to CAM increased significantly from 2007 to 2010. The three most common diagnostic categories were back pain (12.9%), psychological disorders (11.8%), and joint pain (9.6%). Back pain was the most common diagnosis for patients receiving acupuncture, and fibromyalgia was the most common for patients receiving massage therapy. Psychological disorders (i.e., stress) were the major diagnosis referred to both integrative medical consults and executive stress management training. These results suggest that the diseases related to pain and psychological disorders are the main fields of CAM use. It also shows the increasing trend of the use of CAM at an academic medical center in the US.

Obstructive and central sleep apnea have been associated with increased risk of adverse cardiovascular events and mortality. Sympathetic dysregulation occurring as a result of the respiratory disturbance is thought to play a role in this increased risk. Sleep apnea increases the risk of arrhythmias, myocardial ischemia/infarction, stroke, and heart failure, all of which may increase mortality risk. A higher incidence of nocturnal arrhythmias, cardiac ischemia, and sudden death has been noted in subjects with sleep-disordered breathing (SDB). In this review, the association between SDB and each of these conditions is discussed, as well as the potential mechanisms underlying these risks and the effects of treatment of SDB. Particular emphasis is placed on the relationship between SDB and nocturnal atrial and ventricular arrhythmias, myocardial ischemia/infarction and sudden death.

Background Epidemiologic evidence suggests a link between short sleep duration and cardiovascular risk, although the nature of any relationship and mechanisms remain unclear. Short sleep duration has also been linked to an increase in cardiovascular events. Endothelial dysfunction has itself been implicated as a mediator of heightened cardiovascular risk. We sought to determine the effect of 8 days/8 nights of partial sleep restriction on endothelial function in healthy humans. Methods and Results Sixteen healthy volunteers underwent a randomized study of usual sleep versus sleep restriction of two‐thirds normal sleep time for 8 days/8 nights in a hospital‐based clinical research unit. The main outcome was endothelial function measured by flow‐mediated brachial artery vasodilatation ( FMD ). Those randomized to sleep restriction slept 5.1 hours/night during the experimental period compared with 6.9 hours/night in the control group. Sleep restriction was associated with significant impairment in FMD (8.6±4.6% during the initial pre‐randomization acclimation phase versus 5.2±3.4% during the randomized experimental phase, P =0.01) whereas no change was seen in the control group (5.0±3.0 during the acclimation phase versus 6.73±2.9% during the experimental phase, P =0.10) for a between‐groups difference of −4.40% (95% CI −7.00 to −1.81%, P =0.003). No change was seen in non‐flow mediated vasodilatation ( NFMD ) in either group. Conclusion In healthy individuals, moderate sleep restriction causes endothelial dysfunction. Clinical Trial Registration URL : ClinicalTrials.gov. Unique identifier: NCT 01334788.

New Findings What is the topic of this review? This review summarizes chemoreflex physiology in health and disease, with specific focus on chemoreflex‐mediated pathophysiology in obstructive and central sleep apnoea. What advances does it highlight? Chemoreflex mechanisms are thought to contribute significantly to the pathophysiology and adverse outcomes seen in sleep apnoea. Clinical implications of altered chemoreflex function in sleep apnoea from recent studies in humans, including cardiac arrhythmias, coronary artery disease, systolic/diastolic heart failure and sudden cardiac death are highlighted. Activation of the chemoreflex in response to hypoxaemia results in an increase in sympathetic neural outflow. This process is predominantly mediated by the peripheral chemoreceptors in the carotid bodies and is potentiated by the absence of the sympatho‐inhibitory influence of ventilation during apnoea, as is seen in patients with sleep apnoea. In these patients, repetitive nocturnal hypoxaemia and apnoea elicit sympathetic activation, which may persist into wakefulness and is thought to contribute to the development of systemic hypertension and cardiac and vascular dysfunction. Chemoreflex activation could possibly lead to adverse cardiovascular outcomes, such as nocturnal myocardial infarction, systolic and/or diastolic heart failure, cardiac arrhythmias and sudden death in patients with sleep apnoea. This review summarizes chemoreflex physiology in health and disease, with specific focus on chemoreflex‐mediated pathophysiology in obstructive and central sleep apnoea. Measurement of the chemoreflex response may serve as a potential avenue for individualized screening for cardiovascular disease. Whether modulation of this response in sleep apnoea may aid in the prevention and treatment of adverse cardiovascular consequences will require further study.