ABSTRACT
Background. Although SARS-CoV-2 predominantly targets the respiratory system, it has also been associated with vascular complications including stroke. Identifying COVID-19 patients at elevated risk for stroke can help inform target anticoagulation strategies. We sought to understand how symptoms and laboratory markers at presentation with COVID-19 relate to stroke risk. Methods. We enrolled a cohort of 1324 subjects who were hospitalized with COVID-19 across six PennMedicine hospitals between April and August 2020 and performed retrospective, manual chart review to measure exposures including presenting symptoms and admission inflammatory markers. Data were organized with a REDCap database, and analyses were performed using R statistical software, with Bayesian binomial regression models fit using Stan Hamiltonian Monte Carlo via the "brms" package. Results. Among 1324 subjects, 19 stroke events were observed within 30 days of COVID-19 diagnosis. Admission inflammatory markers, including C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), ferritin, and D-dimer, were poor predictors of stroke risk. Among presenting symptoms, including respiratory, gastrointestinal, dermatologic, and neurologic features of COVID-19 disease, only altered mental status documented on presentation (in 529 subjects) was significantly associated with stroke risk (odds ratio 6.06, 95% credible interval 2.16 - 18.7). Conclusion. Inflammatory markers associated with COVID-19 disease severity did not discriminate patients at high versus low risk of stroke in this cohort. Altered mental status documented on presentation was significantly associated with incident stroke during COVID-19 disease.
ABSTRACT
Background. Although CRE are a global threat, data in low- and middle-income countries are scarce. Colonization data are vital for informing antibiotic resistance strategies. We characterized the colonization prevalence of CRE in various settings in Botswana. Methods. This study was conducted in 3 districts in Botswana (1 hospital and 2 clinics per district). Adult inpatients and clinic patients were randomly selected for enrollment. Community subjects were enrolled by inviting each enrolled clinic subject to refer up to 3 adults. Each adult clinic or community subject was also asked to refer their children. All subjects had rectal swabs obtained and inoculated on selective chromogenic media for preliminary identification of CRE. Final identification and susceptibility testing were performed using MALDI-TOF MS and VITEK-2, respectively. CRE underwent genotyping for carbapenemase genes. Results. Subjects were enrolled from 1/15/20-9/4/20 with a pause from 4/2/20-5/21/20 due to a countrywide COVID lockdown. Of 5,088 subjects approached, 2,469 (49%) participated. Enrollment by subject type was: hospital - 469 (19%);clinic - 959 (39%);community adult - 477 (19%);and community child - 564 (23%). Of 2,469 subjects, the median (interquartile range) age was 32 years (19-44) and 1,783 (72%) were female. 42 (1.7%) subjects were colonized with at least one CRE;10 subjects were colonized with multiple strains. E. coli (n=17), K. pneumoniae (n=20), and E. cloacae complex (n=11) were most common. CRE colonization prevalence was 6.8% for hospital subjects, 0.7% for clinic subjects, 0.2% for adult community subjects, and 0.5% for child community subjects (p< 0.001)). CRE prevalence varied across regions (Figure 1) and was significantly higher pre- vs post-lockdown (Figure 2). VIM and NDM were the most common carbapenemase genes (Figure 3). Conclusion. CRE colonization was significantly higher in hospital vs community settings in Botswana. CRE prevalence varied by region and decreased significantly following a countrywide lockdown. With CRE prevalence still modest, elucidating risk factors for CRE colonization holds promise in developing strategies to curb further emergence of CRE. Additional investigation of the CRE isolates without identified resistance genes is warranted.
ABSTRACT
Background. Although ESCrE are a global challenge, data on ESCrE in low- and middle-income countries are limited. In particular, colonization data are critical for larger antibiotic resistance efforts. We characterized the colonization prevalence of ESCrE in various settings in Botswana. Methods. This study was conducted in 3 hospitals and 6 clinics located in 3 districts in Botswana. In each hospital, we conducted surveillance of adult patients. Adult clinic patients were also randomly selected for participation. Finally, we enrolled community subjects by inviting each enrolled clinic subject to refer up to 3 adults. Each adult clinic or community subject was also allowed to refer their children. All subjects had rectal swabs obtained which were inoculated onto chromogenic media for preliminary identification of ESCrE. Final identification and susceptibility testing were performed using MALDI-TOF MS and VITEK-2, respectively. Genotyping was done for identification of extended-spectrum beta-lactamase (ESBL) genes. Results. Enrollment occurred from 1/15/20-9/4/20 but paused from 4/2/20-5/21/20 due to a countrywide COVID lockdown. Of 5,088 subjects approached, 2,469 (49%) participated. Enrollment by subject type was: hospital - 469 (19%);clinic - 959 (39%);community adult - 477 (19%);and community child - 564 (23%). Of 2,469 subjects, the median (interquartile range) age was 32 years (19-44) and 1,783 (72%) were female. 759 (31%) subjects were colonized with at least one ESCrE;130 subjects were colonized with multiple strains. E. coli (n=663) and K. pneumoniae (n=121) were most common. ESCrE colonization prevalence was 43% for hospital subjects, 31% for clinic subjects, 24% for adult community subjects, and 26% for child community subjects (p< 0.001)). ESCrE prevalence varied significantly across regions (Figure 1) and was significantly higher pre-lockdown vs post-lockdown (Figure 2). CTX-M was the most common ESBL gene (Figure 3). Conclusion. ESCrE colonization was common in both healthcare and community settings in Botswana. Colonization prevalence varies by region and clinical setting and decreased following a countrywide lockdown. These findings provide important clues regarding potential drivers of ESCrE that might serve as targets for intervention.