Episode 28: Diuresis, deescalation, and liberation with Matt Siuba

How to take the well-resuscitated critically ill patient, get fluid out of them, deescalate their antibiotics, wean their sedation, reduce vent support, extubate, and get them out of the ICU—with Dr. Matt Siuba (Twitter: @msiuba), an intensivist at the Cleveland Clinic with an abiding interest in “zentensivism,” the art of doing less.

Takeaway lessons

  1. Portal vein pulsatility is a quick and useful addition to IVC assessment when evaluating volume status, particularly in ventilated patients where quick “eyeball” assessments of IVC variability is difficult (and measuring it is a pain). Consider pulse pressure variation too, in patients with a regular cardiac rhythm and on the vent.
  2. Start a journey of diuresis with furosemide, dose 40 times the serum creatinine. If you have a lot of work to do, or you expect resistance, add a thiazide (e.g. metolazone), and if more than a couple days of diuresis is expected, add spironoloctone as well to limit potassium wasting. Aim for >1–3 liters negative fluid balance per day, and generally schedule diuretics instead of manually spot dosing. Even if scheduled, however, follow up on urine output so you can increase the next dose if the last wasn’t adequate.
  3. Alternately if kidney function is quite poor, consider a furosemide stress test of 1 mg per kg of bodyweight (or 1.5 mg per kg for those with prior loop diuretic exposure). If output is poor, e.g. <200 ml in 2 hours, consider moving towards dialysis early.
  4. Tolerate a modest pressor increase induced by diuresis, such as norepinephrine within the 0.2 micrograms/kg range. However, volume overloaded patients usually tolerate volume removal fine, even if on pressors.
  5. Furosemide has a threshold and a ceiling, and there’s not much space between them, so feel free to give go fairly big in dosing without fear of complications.
  6. An elevated creatinine due to intrinsic renal injury is no contraindication to diuresis. In fact, it may improve with volume removal, if secondary to congestive nephropathy.
  7. Labs twice a day is fine for almost everyone, if labs are needed at all.
  8. Generally, don’t tap transudative pleural effusions solely for the purpose of fluid removal, except in non-intubated patients with very large effusions and respiratory distress.
  9. In general, wean FiO2 using pulse oximetry, not the blood gas.
  10. Even more than absolute finalization of cultures, deescalation of antibiotics should be informed by identification of the source and patient risk factors. A “surprise” resistant organism not predictable using these facts is a fairly uncommon event.
  11. Have a low threshold to switch stable, intubated patients to a pressure support mode. If fatiguing, consider giving more support instead of going to a control mode. PS of up to 8–15 cmH2O is very reasonable, and usually more comfortable (and less sedation-demanding) than forcing patients to accept a low tidal volumes on VCV. If a rate is needed, consider pressure control. A very variable tidal volume on PS may also be a good reason to give a set inspiratory time or volume.
  12. Night-time or late extubations are mostly dependent on staffing. If experienced providers are on hand, the clock should be no deterrent. If staffing is limited, it might give more pause. On the other hand, if your extubation practices are aggressive, most of these candidates were probably already extubated in the morning.
  13. Consider routinely extubating very obese patients to BiPap, and elderly patients (>65) or those with chronic cardiopulmonary diseases to HFNC. The latter is reasonable in most patients whose acute hypoxic disease has not yet fully resolved.
  14. When to downgrade patients from the ICU is less dependent on how long it’s been since extubation or getting off pressors, and more dependent on whether any disease process with the potential to worsen is still present (such as an uncontrolled source of sepsis).
  15. The most common cause of error in deescalation is doing it too slowly. Being “conservative” exposes patients to iatrogenic harm from prolonged intubation, antibiotics, lines, and ICU stays; that harm of omission isn’t better or safer than the alternate harms of commission. A patient who is admitted to the ICU, aggressively resuscitated, then rapidly improves and gets turned around isn’t moving “too fast”—that’s exactly the goal. Step up fast, step down fast.

2 thoughts on “Episode 28: Diuresis, deescalation, and liberation with Matt Siuba”

  1. Great discussion, fluid status/responsiveness is a nuanced topic, but de-resuscitating on hospital day #1 in a septic patient on pressors without known source or source control is a bit fast. There is a third option (besides more fluids or diuretics) – don’t do either! Watch the patient for the next 12-24 hours then decide.

  2. That may be how I’d do it. Matt’s hardcore though. I do think as a general rule, the less you can have a any “plateau” to their arc of care (escalate, then deescalate), the better.

Leave a Reply

Your email address will not be published. Required fields are marked *