(1/x) When you trying to decongest patients in the ICU, there are lots of drugs available... not just lasix
Here are some tips for diuresing patients🧵
Furosemide - backbone of diuresis. Start at a higher than you think and back off if diuresis achieved.
Metolazone (or other thiazides). Use when Na on the higher side as helps prevent distal Na resorption. You get more naturesis.
Acetazolamide. Use to prevent metabolic alkalosis. Aids in decongestion.
Spironolactone. Use to prevent K+ wasting and to achieve more natriuresis.
3% saline. Consider in diuretic resistant heart failure with hypochloremia.
(2/x) Tip: Goal of diuresis is not just to produce lots of urine... its to produce lots of salty (high Na) urine.
This is because natriuresis (salt) > aquaresis (water) for decongesting patients.
Ways to achieve more natriuresis: 1. Multi-modal diuresis 2. Check urine Na during diuresis (ideally random urine Na > 100).. add more multi-modal diuresis if not achieving.
(3/x) If you have a patient with venous congestion (e.g. objective congestion on ultrasound) expect an increase in Cr with diuresis... ride that out and treat the patients venous congestion / overload.. it will come down once the patients high venous pressures improve.
(4/x) Hypernatremia and metabolic alkalosis due not reflect intravascular volume, but rather, your diuretic choices.
Furosemide monotherapy routinely causes high Na and high HCO3.
Add Acetazolamide and/or Metolazone from the get-go to attenuate this so you can guide diuresis clinically (e.g. overload, edema, effusions, VeXUS congestion etc.) and NOT be limited by side effects of the medications.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
(1/x) Normotensive shock is common yet often missed because there are actually multiple phenotypes to consider👇
(2/x) Shock is inadequate tissue perfusion (think end organs).
Blood pressure is not the same as flow.
Forward flow does not equal tissue perfusion.
Tissue perfusion is a complex interplay between:
Blood Pressure - minimum pressure is required
Forward Flow (cardiac output)
Back Pressure/Flow (congestion)
Tissue Level Hemodynamics (e.g. macrohemodynamics being translated into tissue level perfusion).
Any one of these alone is not sufficient for adequate organ perfusion.
(3/x) Normotensive shock can happen for a number of reasons:
1. MAP preserved but only at the expense of high vascular resistance. This is seen commonly in hemorrhagic and cardiogenic shock (including PE) where high SVR maintains MAP despite low CO.
MAP = CO x SVR
Clues: 1. Normotensive but low pulse pressure 2. Normotensive but tachycardic (compensatory) 3. Normotensive but poor perfusion (cap refill, mottling, lactate).
Easiest way to figure out --> echo/POCUS with a VTI as a surrogate of SV
2. MAP and CO preserved but significant congestion causing organ failure.
This happens commonly in heart failure patients, but also in up to 20% of patients with septic shock on ICU admission!
Forward flow AND MAP are okay, but severe congestion impairs tissue organ deliver.
Easiest way to figure out is to perform Doppler of organs like the liver, kidney. High CVP can be a cluse.
3. Normal MAP, CO, and congestion but severe microcirculatory dysfunction.
Some patients are hemodynamically incoherent, where normal systemic hemodynamics (or interventions that improve those hemodynamics) are not translated to the microcirculation.
How to assess? Ongoing poor cap refill, lactate, mottling despite optimized hemodynamic parameters.
Consider: augmenting MAP (improving capillary function) and PREVENTING. Volume overload/congestion can contribute to hemodynamic incoherence.
The intervention used Cap Refill Time as a surrogate of end-organ perfusion, and if abnormal, did physiology informed maneuvers (fluids, vasopressors, inotropes) to improve perfusion.
They found "there were 131 131 wins (48.9%) in the CRT-PHR group vs 112 787 (42.1%) in the usual care group for the hierarchical composite primary outcome, with a win ratio of 1.16 (95% CI, 1.02-1.33; P = .04)."
Most of this was driven by lower organ support in the phenotyped (intervention) arm.
Patients in the intervention arm received 1) less fluids 2) more inotropes 3) had improved SOFA/lactates 4) and less organ support (vasopressors primarily).
So is this patient important?
(3/x) It is important to contextualize this in the broader sepsis research landscape.
EGDT in 2001 --> 46.5 vs. 30.5% mortality
ARISE in 2014 --> ~19%
PROCESS in 2014 --> ~21%
PROMISE in 2015 --> ~29%
ANDROMEDA-SHOCK in 2019 ~43% vs 35% (intervention)
Why a higher mortality for Andromeda-Shock than the older ARISE/PROCESS/PROMISE?
It may be related to patient demographics, particularly recruiting patients from more low-middle income countries where sepsis mortalities are higher.
In Andromeda-Shock 2, mortality was only 26%, 10-15% lower than the previous study (and lower than expected) likely reflecting improved sepsis care globally.
The truth is, for any septic shock study, an outcome like mortality is only modifiable for a subset of patients.
Some are too sick and will die regardless of resuscitation.
Some are well enough that they will live despite what we do.
There is a window of patients in the middle whose mortality is modifiable... but as baseline mortality decreases and overall sepsis care improves, this mortality modifiable group becomes smaller.
(1/9) Here’s how to become the best doctor you can in 2025…
Some advice (e.g. learn from your pts) is timeless but some thing are different than when Osler trained.
🧵
(2/9) Learn from your patients
Learning ~= cases seen × learning extracted per case
Maximizing both is key.
Volume exposes you to varied presentations, and reflection deepens your understanding.
There’s no substitute for either. Perhaps in the coming years AI simulated presentations may assist in pattern recognition (e.g. high exposure to simulated pathology) but not quite there yet.
(3/9) Develop skills beyond knowledge
Knowledge matters, but communication, listening, problem-solving, studying, and teamwork matter more in practice.
When trainees struggle, it’s often these skills, not medical knowledge, that hold them back.
(1/x) In fellowship, I managed a peri-arrest patient in the middle of the night who changed my understanding and appreciation for hemodynamics, ultrasound, and TEE.
I've seen similar cases dozens of times now, yet this commonly gets missed, even at top institutions worldwide.
A 🧵
(2/x) When I met this pt. they had a HR of 170 and a blood pressure on arterial line of 50/30. They were mottled head to toe, ashen, and looked like they were seconds away from arresting.
They had a recent NSTEMI and my first thought was cardiogenic shock or a mechanical complication of their MI --> VSD, tamponade, free wall rupture etc.
No transthoracic views on echo.
(3/x) I started the usually therapy, phenylephrine pushes, levophed, vaso pushes, and epi infusion.
No response.
I started bolusing aliquots of 20-50mcg of epinephrine ... no response. We intubated while this was going on.
(0/x) How do I approach the first 2 minutes of meeting a new critically ill pt?
My approach has been forged from my learnings as a paramedic.
Here's my approach as an ICU doc 🧵
(1/x) I start with a scene survey.
I want to understand who's in the room and what are their roles. Is someone currently leading the resuscitation? What monitors are on the patient? What drugs are hooked up to the patient?
In EMS, one of the first things we are taught is to not rush to the patient.
This is for your own safety in case there's a hazard, but also because you can learn so much about what's going on by observing your environment as you approach the patient.
This gets missed in the hospitals.
Take some time and even by understanding what pumps they're on, you can often learn a lot about what might be going on in this very moment. You can also learn what help you might have if 💩🪭
(2/x) Next, I start perform a very quick (<15 seconds) scan of the patient
Essentially quickly checking the ABCs
This involves:
Reviewing the monitor for current vitals
Looking at the patient's skin colour
Looking for mottling
Looking at the Work of Breathing
Looking for LOC
Approaching the patient and briefly feeling their feet or hands for temperature and pulse characteristics
My goal here is to determine whether we need to go straight into resuscitation within seconds (cardiac arrest) or peri-arrest, or whether I can take a bit more detailed handover from the team that's already caring for the patient.