Published in the March 2018 issue of Today’s Hospitalist
AS A PULMONARY and critical care physician at the University of California, San Francisco (UCSF), Lekshmi Santhosh, MD, thinks of ICU patients as walking a very precarious tightrope.
“They’re so sick and vulnerable that they’re very prone to both diagnostic and therapeutic errors,” Dr. Santhosh pointed out during a presentation at last fall’s management of the hospitalized patient conference at UCSF. Unfortunately, she added, one estimate—which was extrapolated from autopsies and published in the July 24, 2012, BMJ Quality & Safety—holds that more than 40,000 patients may die each year in U.S. ICUs due to diagnostic errors.
That’s led her to think about “the common places where we both, intensivists and hospitalists, trip up in the ICU,” and to informally survey critical care colleagues at three hospitals in San Francisco about the common errors they see. Her Top 5 list of the most frequent pitfalls to avoid in the ICU centered on volume status, pressor choice, noninvasive ventilation, bronchoscopy and communication.
“For nearly every one,” Dr. Santhosh said, “a patient comes to mind who was affected by a near miss or an error.”
1. Volume status
THE SINGLE BIGGEST error Dr. Santhosh said she sees in the ICU is clinicians’ almost knee-jerk use of boluses for every hypotensive patient.
“We don’t stop to think: What is causing the hypotension, and why is this patient in shock? And is the patient even responding to fluids?” she said. “We need to get out of this bolus track and think critically about where the patient is at.”
“Pressors are like antibiotics. You want to target them to your patient’s physiology.”
~ Lekshmi Santhosh, MD
University of California, San Francisco
That means trying to gauge volume status, which Dr. Santhosh admitted can be tough. Central venous pressure (CVP) off a central line “was thought to be as close as we could get to a gold standard of measurement.” But that was debunked in a famous meta-analysis published in the July 2008 Chest, which found “a very poor relationship” between CVP and both blood volume and fluid responsiveness.
A study in the September 27, 2016, issue of Journal of the American Medical Association looked at different measuring modalities including CVP, IVC ultrasound, A-line pulse pressure variability and passive leg raise. Surprisingly, the research found that the passive leg raise was the most useful predictor of fluid responsiveness in hemodynamically unstable patients.
“But in real life, you have ICU patients hooked up to lines and machines, and they may have spinal stenosis,” said Dr. Santhosh. “You’re not going to be able to raise their bed 45 degrees without them hollering in pain.”
What she does with what she called “all these imperfect tests” is combine them. “I’ll glance at the CVP, although I don’t put too much weight on it, and I’ll do an IVC ultrasound and repeat that frequently,” she said. She’ll also do a straight leg raise with patients who can tolerate it and use A-line pulse pressure variability when patients aren’t in atrial fibrillation and intubated.
“I combine them all together to form a clinical picture of volume status and responsiveness,” she said.
What she doesn’t do is take a one-and-done approach. “Don’t write for a fluid bolus and then walk away,” Dr. Santhosh said. And when you’re paged about a hypotensive patient, “it’s so tempting to click that ‘reorder bolus’ button, but I urge you not to.”
Instead, go back to the bedside and reassess patients’ parameters of fluid responsiveness. How is their mental status and urine output, and are there crackles on exam? Can you repeat the straight leg raise and IVC ultrasound? “Reassess every time you bolus. That will really change your management.”
De-escalation is also key
Another common mistake around volume: not deescalating IV fluids. The FACTT trial, published in the June 15, 2006, New England Journal of Medicine, found a benefit from a conservative fluid strategy and a net negative fluid balance in terms of decreased days on a ventilator and time in the ICU.
“Multiple studies have replicated that, even in sepsis,” Dr. Santhosh noted. “After initial resuscitation with early goal-directed therapy, you want a maintenance or stabilization phase and then de-escalation.” That could mean active diuresis in patients to attain a negative fluid balance once they’re off pressors.
And while it can be a challenge to find the maintenance fluids in your EHR to discontinue them, “make sure that’s on your daily checklist,” she said. As to which fluid is best: Avoid chloride-rich fluids. “We don’t want an excess of normal saline or a chloride-rich fluid because those can lead to metabolic acidosis.” After initial resuscitation with normal saline, Dr. Santhosh said, “switch early to a buffered solution like Plasma-Lyte or Lactated Ringers.”
2. Pressor choice
WHEN CHOOSING a pressor, Dr. Santhosh said the big error doctors make is thinking that one option—norepinephrine—is always right.
It’s true that in head-to-head studies of norepinephrine and dopamine, “dopamine actually increased mortality in patients with arrhythmias, so people took home the message that it should be norepinephrine for all,” she said. “But that’s not necessarily the case.” Compounding that error is the assumption that all hypotension is sepsis when the culprit could be cardiogenic shock, acute valvular dysfunction, toxidromes or any of a host of other sepsis mimics.
As she explained, “Pressors are like antibiotics. You want to target them to your patient’s physiology.” To pick the right pressor, Dr. Santhosh reminded her audience to “go back to your MAP [mean arterial pressure] being CO [cardiac output] times SVR [systemic vascular resistance].” For a purely vascular tone problem, which typically occurs in postop patients, use phenylephrine, which targets SVR. “Spinal shock is another scenario where you often use phenylephrine.”
For sepsis, on the other hand, “norepinephrine is the pressor of choice” because it targets both SVR and cardiac output. But “you often have mixed shock scenarios, especially with cardiogenic shock.” In such cases, consider combining inotropes with a pressor and using norepinephrine and dobutamine (which affects cardiac output) together. When using an inotrope, she warned, “track a mixed venous O2 sat or a central venous sat to look at and titrate to.”
Another common error with vasopressors is doctors jumping straight to using them in hemorrhagic shock. Instead, “blood is the best vasopressor in these patients, so as you resuscitate them, their pressor requirements will decline. You want to fill up the tank before you increase the squeeze.”
She also singled out pulmonary hypertension and critical aortic stenosis as special pressor cases where, she told her audience, “you should get help.” In complex cases where you’re combining diuresis with inotropes and adding multiple pressors, turn to cardiologists, intensivists or anesthesiologists—and expect to get different answers.
In patients in shock from pulmonary hypertension, for instance, “we often use a combination of epinephrine and norepinephrine.” Meanwhile, phenylephrine is often used in patients with critical aortic stenosis, “but recent studies have found that nitroprusside has good efficacy as well.” And in such cases, she added, “be very careful with beta-blockade. You want to avoid precipitating right heart failure.”
3. Noninvasive ventilation
SOME PATIENTS are excellent candidates for noninvasive ventilation, Dr. Santhosh pointed out. Those include patients with COPD exacerbation with hypercapnic acidosis (ideally with a pH of less than 7.35), cardiogenic pulmonary edema and post-extubation respiratory failure.
But there are plenty of contraindications as well. Those include cardiac or respiratory arrest, facial or neurological surgery or trauma, patients who cannot protect their own airway or manage their own secretions, and patients at high risk of aspiration.
Dr. Santhosh also had the same caution with BiPAP as with fluids: Doctors need to continuously reassess patients started on noninvasive ventilation. “You want a repeat average blood gas in no later than an hour and to be back at the bedside every 15 minutes to see if the patient needs to be urgently intubated.” Frequent reassessment may allow you, for instance, to use BiPAP in patients who may be mildly altered, with a PCO2 in the 60s. “If you watch them closely, they may turn around quickly.”
Weaning patients from a ventilator is another great use of BiPAP, she said. “Increasingly, we are extubating to BiPAP in higher-risk patient populations” including those who are obese or are status post-abdominal surgery, or patients who may be hemodynamically tenuous, such as heart failure patients.
But noninvasive ventilation becomes complicated when considering goals of care. BiPAP won’t help palliative care patients except for those with neuromuscular disease, such as ALS. Nor is it necessarily a good choice in patients with metastatic lung or breast cancer who have increased work of breathing.
Plus, “BiPAP becomes one more item to consider withdrawing in family meetings,” she pointed out. “High-flow nasal cannula might be more appropriate.”
4. Bronchoscopy
THERE ARE MANY misconceptions about when to use bronchoscopy in the ICU to determine what’s causing hypoxemia. Considerations include: How stable is the patient, and is this the best diagnostic test for what you’re looking for?
Bronchoscopy is very safe in intubated patients with an FiO2 of 40%. “We can tolerate FiO2s up to 60% or 80%, and we’ve done some very brief ones where the FiO2 was 100%,” Dr. Santhosh said. “But in general, we want the FiO2 to be less than 60%.”
For patients undergoing bronchoscopy while awake, make sure they have the mental status to follow directions. “We use topical lidocaine, so extubated patients really have to gargle it back to undergo bronchoscopy safely.”
And “we do bronchoscopy safely all the time on patients on high-dose vasopressors,” she noted. “It really doesn’t affect it.” The same is true when doing a quick bronchoalveolar lavage, which can also safely be done in patients on anticoagulation, a heparin drip, or coagulopathy from disseminated intravascular coagulation (DIC).
One classic indication for bronchoscopy is to rule out diffuse alveolar hemorrhage. “This is the best test for that, and CT is often too nonspecific in these patients.” It’s also the gold standard to rule out pneumocystis jirovecii pneumonia (PJP).
However, “mucus plugging is not an indication in the ICU,” said Dr. Santhosh. “Numerous studies have shown that the ‘clean-out bronch’ or a ‘therapeutic bronch’ for mucus plugging is not efficacious.” Instead, rely on patients’ own cough reflex along with appropriate antibiotics and airway clearance devices.
And just because your patient is getting a bronchoscopy, “that doesn’t mean you shouldn’t get a sputum culture. You can use a sputum culture for a number of things: to rule out AFB [acid fast bacilli] or see worms, and to get a microbiological diagnosis.”
5. Communication
FINALLY, Dr. Santhosh said that doctors in the ICU commit a host of communication errors. For one, clinicians need to contact not only patients’ primary care physicians, but also any longitudinally involved outpatient specialists. Those include pulmonologists, oncologists and nephrologists—and that’s particularly true at discharge.
Communication is also important when patients are discharged from the ICU to the wards. And consider holding face-to-face meetings in the ICU with all the consultants patients have seen.
“Getting the consultants together in one room is the best way to prevent a lot of errors,” she said. In terms of end-of-life care, “be sure to loop in all consultants with the results of any family meeting.”
Phyllis Maguire is Executive Editor of Today’s Hospitalist.
Flow vs. pressure
THE FLOW VS. PRESSURE DEBATE between noninvasive ventilation and high-flow nasal cannula continues to rage, according to Lekshmi Santhosh, MD, a critical care physician at the University of California, San Francisco (UCSF).
“And there are pros and cons to each,” she said during a UCSF presentation last year. Noninvasive ventilation is great for COPD patients because it counter-balances their autoPEEP, and it reduces the work of breathing for heart failure patients.
“But the mask can be quite uncomfortable and make it hard for patients to speak with their families,” Dr. Santhosh said. “High-flow nasal cannula is much more comfortable, and it can deliver a higher FiO2.”
But while high-flow nasal cannula is great for hypoxemic respiratory failure, “it’s not as good for hypercapnic respiratory failure.” And high-flow “often falsely reassures people, so it can delay intubation and ARDS diagnoses. We forget how high an FiO2 patients are on because they’re eating, communicating and looking good.” Keep re-assessing patients’ X-rays to avoid delayed recognition of ARDS, she said, “and have a low threshold for intubating these patients.”
I really appreciated this research.