Raising the D-dimer threshold for patients with low or moderate clinical pretest probabilities of pulmonary embolism (PE) effectively rules out PE
and reduces the need for chest imaging by one-third (N Engl J Med 2019; 381:2125–34).
The primary goal of diagnostic testing for suspected PE is to determine which patients should be treated with anticoagulants and which should not, according to investigators in the Pulmonary Embolism Graduated D-Dimer study. Current practices generally rule out PE in patients with low clinical pretest probabilities when they have D-dimer results <500 ng/mL. However, only about one-third of patients being evaluated for PE fit this picture, and chest imaging with computed tomography (CT) is expensive, exposes patients to radiation and contrast reactions, and can be time-consuming.
The authors evaluated a strategy that prospectively ruled out further testing in patients who had a low clinical pretest probability and D-dimer results <1,000 ng/mL or moderate clinical pretest probability and D-dimer results <500 ng/mL. In all, the study enrolled and evaluated 2,017 emergency or outpatients at university-based clinical centers in Canada and followed them for 90 days to determine whether they developed venous thromboembolism (VTE). Physicians used the Wells clinical prediction rule to assign pretest probabilities; patients with low or moderate pretest probabilities had their D-dimer results measured via locally available assays, including STA-Liatest, HemosIL HS 500, Innovance, Triage, Hemosil HS, and Roche Cardiac Reader.
In all, 1,325 patients had either a low clinical pretest probability of PE and D-dimer results <1,000 ng/mL (1,285) or a moderate clinical pretest probability and D-dimer results <500 ng/mL (40). None of these individuals developed a VTE, and just one person had VTE out of 1,863 who received neither an initial diagnosis of PE nor anticoagulant therapy. This diagnostic strategy led to CT imaging in 34.4% of patients. In comparison, a strategy that would have ruled out PE in patients with a low clinical pretest probability and D-dimer results <500 ng/mL would have resulted in CT imaging in 51.9%.
Kratom, Marijuana Drive Increased Calls to Poison Control Centers About Natural Psychoactive Substances
Exposures to marijuana, nutmeg, and kratom led a statistically significant 74% increase in rate of exposure to natural psychoactive substances that occurred between 2000 and 2017, rising from 17.6 to 30.7 per million population (Clin Toxicol 2019; doi.org/10.1080/15563650.2019.1688341). While the rate of exposure to marijuana increased over this period by 150% and nutmeg by 64%, the rate of exposure to kratom increased by a remarkable 4,948.9%.
The investigators, who analyzed 67,369 calls to poison control centers in the U.S. related to natural psychoactive substances, also found kratom had the highest proportions of healthcare facility admissions and serious medical outcomes (25.7% and 47.1% of single-substance cases, respectively), and was associated with seven deaths from 2016 to 2017.
Kratom, the authors note, is currently classified as a dietary supplement, but because its main active alkaloid, mitragyrine, interacts with opioid receptors, concerns have been raised about its addictive potential. “The continued rise in kratom usage coupled with the serious medical outcomes identified in our study support the need for federal regulation of kratom along with further research on this public health problem,” they wrote.
Exposure rates to other substances, including hallucinogenic mushrooms, anticholinergic plants, morning glory plants, peyote, kava kava, salvia species, absinthe, and khat, all declined.
Sepsis Digital Alert System Improves 3 Key Outcomes
Implementation of a digital sepsis alert system in a multisite U.K. hospital network was associated with improved outcomes—including lowered odds of 30-day mortality across all patients (0.76 odds ratio)—and shorter hospital length of stay and more timely start of antibiotics for patients who alerted in emergency departments (0.93 and 1.71 odds ratios, respectively) (J Am Med Inform Assoc 2019; doi:10.1093/jamia/ocz186).
This natural experiment involving phased implementation of the alert system without randomization across 27,000 hospital stays is the largest undertaken to date and the first to show an association between use of an alert system and better care outcomes, according to the authors.
Imperial College Healthcare NHS Trust incrementally rolled out the alert system, which is based on Cerner’s St. John Sepsis Algorithm. Initially this integrated part of Imperial’s electronic health record (EHR) ran in silent mode, during which alerts were not visible to clinical staff, then was switched to live mode. The implementation started in inpatient units, before spreading to emergency departments, hematology units, and finally all other inpatient units.
Coupled with the alert system, Imperial implemented a novel multidisciplinary care pathway that would launch when a clinician confirmed suspicion of sepsis.
The alert system has two levels of alerts, suspicion of sepsis and suspicion of severe sepsis. The five parameters for the former include white blood cell count >12,000 cells/mm3 or <4,000 cells/mm3 or 10% immature (band) forms, glucose >141 mg/dL or <200 mg/dL, temperature >38.3°C or <36°C, and heart rate >95. When at least three of the criteria were met, an alert would fire.
Similarly the system would fire an alert for suspicion of severe sepsis when two or more of these criteria were met along with at least one criteria indicative of organ dysfunction, including: serum creatinine increase over baseline ≥0.5 mg/dL (72 hr lookback); total bilirubin ≥2.0 mg/dL or <10.0 mg/dL (30 hr lookback); serum lactate >2.0 mmol/L (12 hr lookback); and systolic blood pressure <90 mmHg or mean arterial pressure <65 mmHg.
“This study has clearly shown that the introduction of a network-wide digital screening tool embedded in EHRs is associated with improvement in patient outcomes, demonstrating that digital-based interventions can be successfully introduced and readily evaluated,” the authors concluded.