The Latest in Critical Care, 2/5/24 (Issue #28)
New recommendations for fever in the ICU: what to know
Fever in the ICU: Guideline Update
The Society of Critical Care Medicine (SCCM) and the Infectious Diseases Society of America (IDSA) issued an interim update to their 2008 recommendations for the management of fever in the ICU. PulmCCM is not affiliated with SCCM or IDSA.
Virtually all the recommendations were based on weak evidence and represent a consensus of expert opinion. They shouldn’t be considered automatically applicable to immunosuppressed patients.
What’s a Fever? Depends On the Patient, and Who You Ask
The Centers for Disease Control and Prevention defines a fever (for general hospitalized patients) as >38°C.
For ICU patients, the SCCM and IDSA have agreed fever is ≥38.3°C (100.9°F).
For patients >65 years old living in long-term care facilities, IDSA feels lower thresholds should be used (e.g., >37.2°C oral measured multiple times).
Also, many patients with infection won’t have fevers, and those that don’t may be at the highest risk for poor outcomes. And so the recommendations “may generally apply to ICU patients with suspected infection regardless of the presence of temperature elevation.”
(Isn’t that most of the ICU, usually? Let’s keep moving…)
You’re Already Doing It Right
Most of the panel’s suggestions will feel like second nature to the experienced clinician. (Everything that follows are PulmCCM’s interpretations of the recommendations. Read the original document here.)
Fever in early postoperative patients is likely noninfectious and may not require workup.
Collect at least two sets of peripheral blood cultures, near-simultaneously from different venipuncture sites.
Check chest radiographs in most febrile ICU patients.
Collect viral PCR panels on patients with suspected pneumonia.
For patients developing fevers several days after abdominal/pelvic/thoracic surgery, check a CT scan in collaboration with the surgeon.
Measure core temperature with internally placed thermistors if compatible invasive devices are already in place (e.g., bladder catheter), or use oral/rectal temperatures if not. Avoid less-accurate temperature measurement methods like axillary, tympanic membrane, temporal, infrared, etc.
Don’t give acetaminophen or other antipyretics just to reduce a fever. If a patient needs it for comfort, do provide antipyretics.
PET Scans for Fevers?
Something new: 18F-FDG PET/CT was suggested (weakly) if other testing fails to identify a persistent fever’s source. The evidence for this comes from small retrospective case series:
In 33 ICU patients receiving PET scans for suspected infection over 3 years at a Dutch center, 21 of 35 scans were true positives, 3 false positives, and 11 true negatives (as adjudicated clinically), with a high accuracy rate overall.
In 18 severe sepsis patients in 3 years at one center who underwent PET scanning (apparently done ad hoc), 14 of the 18 had positive tests, and 11 were true positives, leading to surgery in 2 and pacemaker removal in 2 (and longer / new antibiotics in 2).
In 42 patients getting PET scans in a neuro-ICU over 10 years, four of the six with sepsis had infectious foci identified.
These are tiny numbers (less than one PET per month at these centers, which presumably were doing it more than most).
Pretty thin stuff. But this is more evidence of utility than is available for tagged white blood cell nuclear medicine scans, which have been used for decades to look for occult infections. The panel did not find enough evidence to recommend using (or not using) tagged white blood cell scans.
PET scanning looks like a niche test to be considered on selected ICU patients with negative CT scans and a high suspicion of infection, probably with infectious diseases already consulted.
Ultrasound Is Upvoted
The panel weakly suggested bedside ultrasound for patients with fever and an abnormal chest radiograph, but only if an expert is available to interpret the often complex or borderline results. As all radiologists are ostensible experts, this reads more like an endorsement of POCUS in the ICU for intensivists well-trained and experienced in the technique.
(The panel notes thoracic ultrasound’s ability to detect parenchymal and interstitial abnormalities: absent specialized thoracic training, most practicing radiologists should not be expected to identify and describe such lung findings with sufficient specificity to be clinically helpful.)
Complicated (infected parapneumonic) pleural effusions should be promptly identified and drained. But thoracic ultrasounds on every febrile ICU patient with a mildly abnormal chest film would likely result in many unhelpful thoracenteses of harmless effusions, and some avoidable pneumothoraces. There is no prospective data to suggest this practice is beneficial. There were enough caveats in this section to signal the panel’s emphasis on clinical judgment and expertise as guiding determinants.
Moving down to the abdomen:
“Diagnostic abdominal ultrasound has not been studied in the evaluation of fever in critically ill patients,” the panel notes, but still anoints bedside abdominal ultrasound as a new “best practice” for any patient with fever with signs/symptoms of an abdominal source (postoperative or not).
Recall the panel also recommended abdominal CT earlier in the guideline for postoperative patients, and acknowledges CT might be the better first test (rather than ultrasound) in at least some post-op and non-operative critically ill patients.
Ultrasound can quickly and more accurately identify significant biliary disease (e.g., acute cholecystitis and choledocholithiasis) than CT, and so it makes sense as a first test for patients with lab abnormalities consistent with biliary disease. But should it be for all febrile patients with abdominal signs or symptoms?
They acknowledge ultrasound provides “a more limited abdominal evaluation” than CT, and they warn that an inexperienced ultrasound technician could miss important findings. Radiologists are nevertheless empowered in this guideline with consistently and accurately identifying a wide range of pathologies beyond the usual biliary system abnormalities, including abscesses and bowel perforation. No prospective data is offered to show this to be the case at sufficient accuracy compared to CT.
In slower-moving care settings, delaying a CT (by hours or a day) to first obtain an ultrasound, or overly relying on an inadequately performed or interpreted ultrasound, could easily result in delayed identification (or non-identification) of critical problems.
Ultrasound before CT for patients with suspected abdominal source of fever would be ideal, all else being equal, but it seems like a “best practice” only in settings without the above concerns.
Blood Cultures for Patients With Central Lines
Blood cultures should ideally not be drawn solely through an intravascular catheter, as the rate of misleading false positives is high.
Febrile patients with central venous catheters should have blood cultures drawn peripherally (most important) and from the central line, near-simultaneously, the panel suggests. If all the cultures grow the same organism (especially if the centrally-drawn blood cultures result positive first), it increases the likelihood of true bacteremia.
Also: nurses should remove or replace the needleless connector on the central line, prior to drawing cultures.
For “hard stick” patients with central lines from whom peripheral blood cultures prove impossible to obtain, at least two lumens of the catheter should be sampled, they suggest.
No Molecular Testing Without Blood Cultures
Proprietary tests can identify genetic material in blood associated with fungal species, along with E. faecium, S. aureus, Klebsiella, Pseudomonas, and E. coli. More tests, covering more organisms, will likely become available. The FDA-approved tests identified bloodstream infections within a few hours with reasonably high accuracy in pre-approval testing, but their real-world clinical utility is unproven, they don’t identify all organisms, and provide no information on antibiotic sensitivity. As the inexpensive gold standard, blood cultures are therefore recommended to be obtained along with any molecular tests collected.
Avoid False CAUTIs: Replace Urinary Catheters Before Culture
Many electronic orders include a combined urinalysis with “reflex” urine cultures performed if the UA is abnormal. In ICU patients with urinary catheters and asymptomatic bacteriuria, this leads to overuse of antibiotics and overdiagnosis of catheter-associated urinary tract infections (CAUTI), with implications for “quality” reporting and hospital payment.
One way to avoid this, the panel suggests, is to obtain urinalysis and if pyuria is present (increased WBC on UA), replace the urinary catheter before collecting another sample for urine culture.
Suspect Nosocomial Covid
ICU patients with new fevers should receive Covid testing, says the panel, “based on levels of community transmission” (which appear to be virtually indeterminable in the current era, unless you count wastewater metrics).
They seemed to limit this suggestion to “critically ill patients … with fever and pneumonia,” perhaps sensing the potential for unintended consequences of overdiagnosis.
Procalcitonin and C-Reactive Protein
The utility of procalcitonin (PCT) and CRP vary widely across studies. When the data from their use in sepsis patients are aggregated together into meta-analyses, use of PCT to guide antibiotic decisions appears to reduce antibiotic use and even mortality. (CRP performed adequately, but less accurately than PCT.)
However, both PCT and CRP simply return too many false negatives, in clinically high-stakes situations (i.e., serious infections). Aggregating the data from many studies (with a high risk of bias) does not eliminate that fact, it obscures it.
This is a complex area and the suggestions were appropriately tempered and nuanced: “PCT and CRP provide only supportive and complementary information to clinical assessment,” the panel warned. Clinical judgment (with bedside evaluation) is paramount. Adding either PCT or CRP seemed advisable for febrile ICU patients with no obvious source of infection, but the tests have no utility (or negative utility) in febrile ICU patients in whom bacterial infection is already strongly suspected.
Using “Best Practice” Is Not the Best Practice
The panel prescribed a lot of “best practices”: at least nine.
“Best practice” is an overused term that is being nudged toward retirement by the GRADE working group, which instead uses the phrase “good practice” in its publications.
The term “good practice” is reserved for “motherhood” situations that are
… usually something that is already commonly accepted as beneficial or practical advice. It could even be seen as irrefutably “good” as motherhood and apple pie (hence the term). The nature of these types of statements is such that the action is seen as so obviously beneficial that it would be unduly onerous to conduct a review to demonstrate its efficacy.
This guideline panel states they followed GRADE’s guidance when issuing its best practice statements, but that’s debatable:
The GRADE Working Group recommends that good practice statements be used sparingly, if at all. Because good practice guidance is typically based on several linked sources of indirect evidence, there is no way to tell whether the benefits of the proposed recommended action are as truly obvious or incontestable as they seem. (Grade Working Group)
Other than ultrasound before CT for fevers of suspected abdominal source, none of the nine here stand out as problematic. So why does it matter?
“Best practice” is an often-invoked, almost magical phrase in healthcare settings that tends to shut down inquiry, debate, and critical thinking. In fact, it seems to often be deployed precisely for this purpose (e.g., to enforce compliance with an administrative policy). It’s implied that the science is settled by incontrovertible evidence, but the irony is that “best practice” statements are usually made in the absence of evidence, as a substitute for it. So how do we know they’re the “best?”
There’s even an unavoidable judgy moral and professional component to it: Oh, you didn’t follow the best practice? I guess you’re someone who settles for SECOND-BEST FOR YOUR PATIENTS?!
I think that’s what the working group was getting at. The SCCM/IDSA panel made valuable, strong suggestions based on experts’ consensus opinion. Let’s call them good practices, and save the “best” for Mom’s apple pie.
Work up fevers ≥38.3°C (100.9°F) in the ICU with chest films, blood cultures (taking care to avoid contamination and false positives, especially in patients with central lines), urinalysis (changing urinary catheters before collecting urine for culture), and viral PCR (including Covid) for pneumonia.
Obtain additional imaging according to your local conditions—with ultrasound a recommended first option for abdominal and thoracic examinations, if appropriate resources and expertise are available. Ultrasound is well-suited for identifying most serious biliary system abnormalities. CT scanning remains the most widely-used care standard for undetermined abdominal, pelvic or thoracic pathology, especially for post-operative patients.
For patients with persistent fevers despite thorough investigation, consider PET scanning. Use procalcitonin and CRP as ancillary tests only.
Please comment below if you have any suggested additions or edits to this document.