2. Introduction
• The last two to three years provided several “big steps” regarding our
understanding and management of sepsis
• The increasing insight into Pathomechanism of post-
infectious defense led to some new models of host response
• Besides hyper, hypo, and anti-inflammation as the traditional
approaches to sepsis pathophysiology, tolerance and resilience were
described as natural ways that organisms react to microbes.
3. Sepsis – The Journey
• Sepsis is one of the oldest described illnesses. The term sepsis is derived from the
ancient Greek term “σῆψις” (“make rotten”) and was used by Hippocrates around 400
BCE to describe the natural process through which meat decays and swamps release
decomposing gases but also through which infected wounds become purulent.
• After this recognition, it took over 2,000 years to establish the hypothesis that it is not
the pathogen itself but rather the host response that is responsible for the symptoms
seen in sepsis4
7. Precision
Approach
• Precision Approach: It involves a more
complicated approach which applies potent
analytical platforms to assess the genome,
transcriptome, proteinome, and metabolome
of the patient.
• Together with metagenome of the
pathogen, the decision is made on the drug to
prescribe as well as its dose and timing.
https://www.embopress.org/doi/epdf/10.15252/emmm.201810128
8. Notable applications of artificial intelligence in sepsis
• Automated algorithms to identify patients at risk of having sepsis, either in real time (“sepsis
detection”) or in advance (“sepsis prediction”)
• A range of supervised learning algorithms trained on a dataset containing negative and positive
instances of sepsis
• A model based on gradient tree boosting showed good accuracy for predicting sepsis and septic shock
using only vital signs several hours before onset
• Even simpler rule-based algorithms are capable of highlighting at-risk patients, for example by
detecting end-organ damage and the non-specifc Systemic Infammatory Response Syndrome, quick
Sequential Organ Failure Assessment (qSOFA) or SOFA scores
9. AI in Sepsis
• A randomised controlled clinical trial at two medical-surgical intensive care units at the University
of California, San Francisco Medical Center,
• Enrolled patients were assigned to a trial arm by a random allocation sequence.
• In the control group, only the current severe sepsis detector was used; in the experimental
group, the machine learning algorithm (MLA) was also used.
• On receiving an alert, the care team evaluated the patient and initiated the severe sepsis bundle,
if appropriate.
• Although participants were randomly assigned to a trial arm, group assignments were
automatically revealed for any patients who received MLA alerts
Shimabukuro DW, Barton CW, Feldman MD, et al. Effect of a machine learning-based severe sepsis prediction algorithm on patient survival and hospital length of stay: a
randomised clinical trial. BMJ Open Resp Res 2017;4:e000234. doi:10.1136/ bmjresp-2017-000234
10. Effect of a machine learning-based severe sepsis
prediction algorithm on patient survival and
hospital length of stay: a randomised clinical trial
Decrease in average hospital and ICU length of stay with the
use of the machine learning algorithm. The error bars
represent one standard error above and below the mean
length of stay. ICU, intensive care uni
Reduction of in-hospital mortality rate when using the machine
learning algorithm. The error bars represent one standard error
above and below the average in-hospital mortality rate.
11. Artificial Intelligence Assisted Outcomes
• Trial demonstrated improvements in patient outcomes when using a machine
learning-based sepsis prediction algorithm.
• Found a statistically significant decrease in the hospital LOS and in-hospital
mortality when using this algorithm compared with the current rules-based
sepsis detector.
• From these results, we deduced that machine learning-based sepsis prediction
algorithms may lead to earlier clinical intervention and improved patient
outcomes
Shimabukuro DW, Barton CW, Feldman MD, et al. Effect of a machine learning-based severe sepsis prediction algorithm on patient survival and hospital length of stay: a
randomised clinical trial. BMJ Open Resp Res 2017;4:e000234. doi:10.1136/ bmjresp-2017-000234
12. Cutting head technologies in Sepsis
• Unmet need: Variety of reasons including turnaround on blood tests such as
the serum lactate that may be delayed or require expensive laboratory
equipment
• Current need: a problem-based innovation approach to allow clinicians rapid
assessment of end-organ perfusion at the bedside or emergency department
triage and be incorporated into the electronic medical record.
13. New technology for assessment of distal perfusion
in the care of sepsis
• The Flowsense finger sensor is a battery-operated
optical, temperature, and force measurement
instrument which collects highly accurate real-time
signals, provides user input and feedback through an
operation button and multicolor.
• Capillary refill in its ability to correlate with blood
lactate levels and detect sepsis earlier while monitoring
critically ill adult patients
(A) Finger Sensor; (B) Sensor on a patient; (C) Mobile application to process real-time data collection.
14. The glycocalyx: a novel diagnostic and therapeutic target
in sepsis
• The glycocalyx is a gel-like layer covering the luminal surface of vascular endothelial cells.
It is comprised of membrane-attached proteoglycans, glycosaminoglycan chains,
glycoproteins, and adherent plasma proteins.
• The glycocalyx maintains homeostasis of the vasculature, including controlling vascular
permeability and microvascular tone, preventing microvascular thrombosis, and
regulating leukocyte adhesion
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6337861/pdf/13054_2018_Article_2292.pdf
16. Sepsis and Glycocalyx
• During sepsis, the glycocalyx is degraded via inflammatory mechanisms such as
metalloproteinases, heparanase, and hyaluronidase.
• These sheddases are activated by reactive oxygen species and pro-inflammatory
cytokines such as tumor necrosis factor alpha and interleukin-1beta.
• Inflammation-mediated glycocalyx degradation leads to vascular hyper-
permeability, unregulated vasodilation, microvessel thrombosis, and augmented
leukocyte adhesion.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6337861/pdf/13054_2018_Article_2292.pdf
17. Investigational Drugs
• Several novel molecules are being investigated as possible glycocalyx-protective
therapeutics.
• S1P is a sphingolipid that may help improve glycocalyx integrity by inhibiting syndecan-1
shedding.
• S1P activates S1P1 receptor and the activation of S1P1 receptor attenuates the activity of
MMPs causing syndecan-1 ectodomain shedding.
• Sulodexide (SDX), a highly purified extraction product from porcine intestinal mucosa, has
been similarly reported to inhibit heparanase activity
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6337861/pdf/13054_2018_Article_2292.pdf
18. Advances in sepsis diagnosis and management:
a paradigm shift towards nanotechnology
• Recent advancements in nanotechnology-based solutions for sepsis diagnosis and
management
• Development of nano sensors based on electrochemical, immunological or
magnetic principals provide highly sensitive, selective and rapid detection of
sepsis biomarkers such as procalcitonin and C-reactive protein
• Nanoparticle-based drug delivery of antibiotics in sepsis models have shown
promising results in combating drug resistance
19. Advances in sepsis diagnosis and management:
a paradigm shift towards nanotechnology
• Surface functionalization with antimicrobial peptides further
enhances efficacy by targeting pathogens or specific
microenvironments.
• Various strategies in nano formulations have demonstrated the
ability to deliver antibiotics and anti-inflammatory agents
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7790597/pdf/12929_2020_Article_702.pdf