Staying woke I: An introduction to ICU liberation

A case

A 33-year-old male, previously well, is admitted after a single gunshot wound to the left abdomen. He is brought directly to the OR, where a splenectomy is performed as well as anastomosis of a large bowel perforation. Perioperatively, he is unstable and inflammatory, requiring a modest trauma resuscitation and a brief course of pressors; however, he does well and is transferred to ICU still on the ventilator.

Over the next few days, he is agitated and dyssychronous, requiring midazolam and fentanyl infusions. Vent weaning is unsuccessful. On ICU day 6 he develops a ventilator-associated pneumonia and is started on antibiotics. By day 15, with his course still indolent, he undergoes placement of a tracheostomy and percutaneous feeding tube.

On day 20, DVTs are diagnosed and anticoagulation is initiated. On day 22, he again develops pneumonia, this time resulting in septic shock requiring pressors. After stabilization, he finally begins to wean from his sedation, and his lungs begin to dry out.

By day 35, he is awake and tolerating trach collar trials off the ventilator. He is downgraded from the ICU and discharged to rehab.

The problem

This man arrived critically ill, and was discharged home alive. A save — right? Well, maybe not.

Here’s one reason to hold off on the celebration: he was in the ICU for over a month, and mechanically ventilated for most of it. While recovery from a major injury like this assuredly takes time, by day 25 or 30 it seems like a reach to suggest that respiratory failure, DVTs, and pneumonia were a direct consequence of the initial gunshot wound. No, these were complications of critical illness and intensive care itself… which suggests that they might have been preventable by better management.

The other reason is that the story doesn’t end there.

He is discharged to acute inpatient rehabilitation, where he undergoes three weeks of rigorous physical therapy, and is finally sent home.

One year later, he is still out of the hospital — but he is not normal. He was unable to return to work at his previous skilled position, lacking the capacity for the physical and mental workload. He can ambulate at home and perform most of his activities of daily living, but only with assistance from his family. He cannot climb more than one flight of stairs without stopping to rest, and his memory and ability to focus have become seriously compromised. He is plagued by persistent symptoms of depression, anxiety, and PTSD; personality changes and emotional lability have begun to tax his family and personal relationships.

Great save. Right?

What happens when you don’t get better?

We need to start by understanding what the normal course of critical illness looks like. In our post on this topic, we discussed how acute illness develops suddenly, and how intensive care strives to halt the freefall before it becomes irreversible. If death is prevented, the patient enters the long tail of recovery, a process that takes much longer than the initial descent into sickness, and is punctuated by setbacks and complications that then recapitulate the recovery process in miniature.

This is the normal, expected pattern. Much of the energy in critical care is focused upon the first phase: rescuing patients from the freefall. But if we turn our attention to the second, post-acute period, we find that we can have an impact there as well. Here success is not manifested by creating more survivors (although some lives can probably be saved), but by bending the curve of recovery, so that it returns toward normality sooner, and plateaus at a higher level.

You see, once a patient has survived, they can get better in different ways. They can recover quickly, and end up very close to their previous functional baseline. Or they can recover slowly — so slowly that they don’t seem to be getting better at all.

This latter phenomenon has become common enough to earn a name: chronic critical illness. (Other names, such as “persistent critical illness,” have also been used.) While definitions vary, the phenotype is familiar to anyone who has rounded in the ICU.

… prolonged dependence on mechanical ventilation… usually measured in weeks… profound weakness… alterations of body composition… increased vulnerability to infection, often with multiresistant microbial organisms… brain dysfunction manifesting as coma or delirium… [and] skin breakdown associated with nutritional deficiencies, edema, incontinence, and prolonged immobility. [Nelson et al.]

These patients make up a tremendous part of the resource drain and mortality burden in critical care, and represent a vexing problem for the patient, their families, and the healthcare team. Their ICU and hospital course is burdened by long stays, prolonged ventilator courses, numerous hospital-acquired conditions, and heavy costs for the entire system.

The story doesn’t end at discharge, either. As our case study demonstrated, patients who have been weighed down by sluggish recovery don’t suddenly bounce back when they leave the hospital. Their obstacles persist, recovery remains poor, and in some cases, they plateau at a level far below their previous norm.

We’ve come to call this the post-intensive care syndrome, and it’s characterized by physical weakness, cognitive deficits, and emotional changes that persist after discharge, sometimes for years — or forever. See this evocative description from an ICU survivor, courtesy of the Johns Hopkins OACIS group.

While the post-intensive care syndrome remains an active area of research — why it happens, how to prevent it, and how best to manage and support these patients — we do know about some of the risk factors. They include:

  • Sepsis, ARDS, multisystem organ failure
  • Age and prior cognitive impairment
  • Prolonged mechanical ventilation (> 7 days)
  • Prolonged duration of deep sedation
  • Incidence and duration of ICU delirium

You’ll notice that some of these factors, such as age and disease severity, are not modifiable. Yet others may be.

The terrible trio

For some time, the Society of Critical Care Medicine has been promoting an “ABCDEF” approach to this topic, which they call ICU liberation. Each letter of the acronym stands for a good practice, but I think that much of it can be broken down into three key pillars: oversedation, immobility, and delirium.

All three of these factors contribute to the syndrome we’ve described: the ICU patient who has survived the acute phase of illness, but has now been sucked into a downward spiral of complications and poor recovery. Day after day, even with the passage of weeks or months, you visit them and find that little that has changed. “Why is he still here?” you ask. “Did something happen?” People shrug. Nothing much happened. They just aren’t getting any better.

These factors are also deeply synergistic and intertwined. Oversedation leads to delirium, which dramatically increases morbidity and even mortality. Unfortunately, delirium results in additional sedation, because agitated patients often need drugs to be safe. This leads to a positive feedback loop from which it’s hard to escape.

Indeed, sedation and delirium exist on a see-saw: a delirious patient can be sedated, at which point delirium is no longer apparent (they’re simply comatose), or they can be allowed to rouse and manifest their delirium, but either way they’re not awake and normal.

At the same time, physical mobility has the potential to reduce delirium and improve outcomes. Yet patients who are either sedated or delirious can’t cooperate with mobility activities.

Around and around it goes. To successfully escape from this syndrome — or better yet, prevent it — an intelligent, evidence-based approach is needed.

Come back next time, where we’ll look at the data and make some recommendations regarding practices for ICU sedation.

This series is adapted from a Grand Rounds lecture given at the University of Maryland Prince George’s Hospital.

4 Replies to “Staying woke I: An introduction to ICU liberation”

Leave a Reply

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