Fever in the ICU Infectious Diseases in Clinical Practice February - - PDF document

1

Fever in the ICU

Jennifer Babik, MD, PhD Associate Clinical Professor Division of Infectious Diseases, UCSF

Infectious Diseases in Clinical Practice February 2020

Disclosures

• I have no disclosures.

2

Learning Objectives

By the end of this talk, you will be able to:

• 1. Construct a framework for the differential diagnosis of

fever in a patient in the ICU

• 2. Describe the common clinical presentation, diagnosis,

and management of common infections in the ICU

• 3. Recognize the common non‐infectious etiologies for

fever in the ICU

Roadmap

• Introduction/Framework
• Case‐based approach to common infectious and

non‐infectious etiologies for fever in the ICU

• CLABSI
• CA‐UTI
• VAP
• Non‐infectious etiologies
• Short takes (nosocomial sinusitis, acalculous

cholecystitis)

• “Double covering” GNRs

3

Definition of Fever

• Definition of fever is arbitrary
• ≥38.3°C (101°F) commonly used (IDSA/ACCCM)
• Use a lower threshold in immunocompromised patients
• T < 36.0°C should also prompt work‐up for infection
• Note that patients on CRRT or ECMO may not mount

a fever even when infected

O’Grady et al, Crit Care Med 2008, 35:1330.

Measurement of Fever

• Central thermometers (bladder, rectal, esoph) ≈

pulmonary artery temperatures

• Peripheral thermometers have:
• Poor correlation with central temperatures (± 0.5‐2˚C)
• High specificity (~95%) but poor sensitivity for detecting fever
• Oral or tympanic: 75% sensitive
• Temporal 63% sensitive
• Axillary 42% sensitive

Niven et al, Ann Intern Med 2015, 163:768.

4

Does the Height of the Fever Help With Etiology?

• Hyperpyrexia = Fever >41.1˚C (or >106˚F)
• Classic teaching is that infections are a rare cause of

hyperpyrexia

• But the data shows that infections are a frequent

cause of hyperpyrexia in adults (at least 50% solely due to infection)

• Causes:
• Most common: Staph, Strep, GNRs
• TB, candida, malaria, viruses

Sioson and Brown, South Med J 1993, 86:773. Simon,JAMA 1976; 236:240.

Fever in the ICU: Epidemiology

• Fever is common (25‐70%
• f ICU patients)
• Non‐infectious etiologies
• ccur frequently
• Most common causes:
• Infections: PNA,

bloodstream, abdominal

• Non‐infectious: post‐op,

central fever, drug fever

Niven et al, J Intensive Care Med 2012, 27:290. von Vught et al, JAMA 2016, 315:1469.

Infectious 35‐55% Non‐infectious 45‐65%

Etiology for Fever in the ICU

5

Framework for Building the DDx

• 1. Is this a complication of the underlying reason for admission?
• Untreated, relapsed, or metastatic focus of infection
• Post‐surgical infection (surgical site infection, abdominal abscess)
• 2. Is this a separate nosocomial process?
• Hospital‐acquired PNA (HAP, VAP)
• CA‐UTI
• Central Line‐Associated Blood Stream Infection (CLABSI)
• Clostridium difficile
• 3. Is this non‐infectious?
• Drug fever
• Central fever
• Post‐op fever

“big 4”

Initial Evaluation

• History:
• Any change in secretions or

respiratory status?

• Any diarrhea?
• Exam to include:
• Careful neuro exam
• Sinus exam
• Back and joint exam
• Skin exam:
• Line sites
• Decubitus ulcers
• Rashes
• Remove bandages
• Labs:
• CBC with diff (look for eos)
• LFTs (drug reaction,

acalculous cholecystitis)

• Micro:
• Blood cultures
• UA +/‐ Ucx
• Respiratory cultures?
• Cdiff testing?
• Imaging:
• CXR
• Chest or abdominal CT?

6

Approach to Management

• Do you need to treat empirically or can you wait for

cultures/diagnostics?

• Is there a source control procedure needed?
• For empiric therapy:
• How sick is the patient?
• Where do you think the patient is infected?
• Prior positive cultures?
• Prior antibiotics?
• Is the patient at risk for MDR organisms?

Case #1

A 36 year old man with AML is in the ICU for leukopheresis and induction therapy and clinically

• improves. He then spikes a fever

but remains stable.

• He is bacteremic with Staph

epidermidis from both his line and peripheral blood cultures

• He improves with vancomycin.

Can we leave the tunneled line in?

7

Would You Change the Line?

• 1. Yes
• 2. No

Central Line Infections

Exit site infection (<2cm from exit site)

• With or without BSI
• If blood cultures neg, can

try to salvage the line.

• Tunnel infection (>2cm)
• Port pocket infection
• With or without BSI
• Remove the line, even

if blood cultures neg. Bacteremia without

• verlying skin changes
• BSI by definition
• Line removal depends on
• rganism, clinical situation

8

Central‐Line Associated BSI (CLABSI): Diagnosis

• Clinical findings at exit site in <3%
• Catheter tip culture:
• (+) peripheral bcx and > 15 cfu/plate

from catheter tip

• 80% sensitive, 90% specific
• But >80% of catheters removed

unnecessarily

Mermel et al, Clin Infect Dis 2009, 49:1. Safdar and Maki, Crit Care Med 2002, 30:2632.

CLABSI: Differential Time to Positivity

• Allows for diagnosis without removing the line
• Culture from line + peripheral blood at the same time
• CLABSI = blood culture drawn from central line turns

positive at least 2 hrs before the peripheral culture

• Test characteristics
• 85‐95% sensitive
• 85‐90% specific

Liñares, Clin Infect Dis 2007, 44:827. Bouza et al, Clin Infect Dis 2007, 44:820. Bouza et al, Clin Microbiol Infect 2013, 19: E129. Safdar et al, Ann Intern Med 2005, 142:251.

9

DTTP: Possible Scenarios

Line (+) and peripheral (+) Line (+) and peripheral (+)

DTTP ≥ 2 hrs DTTP ≥ 2 hrs CLABSI CLABSI DTTP < 2 hrs DTTP < 2 hrs Look for another source Look for another source

Line (+) and peripheral (−) Line (+) and peripheral (−)

Possibilities

• Line colonization
• Contaminant
• Bacteremia from other source

with 1/2 positive cultures Possibilities

• Line colonization
• Contaminant
• Bacteremia from other source

with 1/2 positive cultures

DTTP for Candida?  Not as good

• DTTP cut‐off of 2h is 85% sensitive, 82% specific
• The special case of C. glabrata:
• Most slow growing Candida with median TTP of 37h

(other species <30h)

• Using 2hr cut‐off DTTP: sensitivity 77%, specificity 50%
• Best DTTP cut‐off = 6h  sensitivity 63%, specificity 75%

Park et al, J Clin Microbiol 2014, 52:2566.

10

When to Remove the Line

1. Severe sepsis 2. Persistent bacteremia (>72h of appropriate ABx) 3. Septic thrombophlebitis 4. Exit site or tunnel infection 5. Metastatic infection: endocarditis, osteomyelitis

• 1. Staphylococcus aureus
• 2. Pseudomonas
• 3. Candida

Mermel et al, Clin Infect Dis 2009, 49:1

Virulent Organisms Complicated Infections

Line Management for Other Organisms

Organism Coag‐negative staphylococci Enterococcus Other GNRs (not Pseudomonas)

Mermel et al, Clin Infect Dis 2009, 49:1

Less aggressive with line removal

HD Catheter Remove, retain, or guidewire exchange Remove, retain or guidewire exchange Remove, retain or guidewire exchange Tunneled Cath/Port Remove or retain Remove or retain Remove or retain PICC/Short‐term CVC Remove or retain Remove Remove

Use clinical judgment based on:

• Severity of infection
• Access options (talk to renal or onc)
• Risk of removal/replacement

11

Line Salvage: General Principles

• Which patients?
• Not for complicated infections, exit site infections, or virulent
• rganisms
• Only studied in long‐term catheters
• How to treat?
• Give systemic ABx + antibiotic lock therapy for 7‐14 d
• Get surveillance blood cultures (1 wk after Abx stop)

Mermel et al, Clin Infect Dis 2009, 49:1

Antibiotic Lock Therapy

• Goal is to get supra‐therapeutic ABx

concentrations to penetrate biofilms

• Logistics
• Work with pharmacy and nursing
• Mix with heparin, dwell times are variable but usually <48h
• Common Abx:
• Gram positives: linezolid, vancomycin, cefazolin
• Gram negatives: ceftazidime, ciprofloxacin, gentamicin

12

Line Salvage with Antibiotic Lock Therapy

Mermel et al, CID 2009, 49:1 Aslam et al. JASN 2014;25:2927. Fernandez‐Hidalgo and Almirante, Expert Rev Anti‐Infect Ther 2014, 12:117. Ashby et al, Clin J Am Soc Nephrol 2009, 4:1601. Beathard, JASN 1999, 10:1045.

10 20 30 40 50 60 70 80 90 100 30‐45% 60‐75% >90%

Systemic Abx Systemic Abx + Lock

10 20 30 40 50 60 70 80 90 80‐90%

CoNS GNRs S.aureus

40‐55%

Abx Lock Efficacy by Organism (%) Overall Success Rate (%)

Line removal

80‐90%

What About Guidewire Exchange?

• Goal is to eliminate biofilm entirely
• How good is it?
• Limited data, mostly HD catheters
• At least equal to ABx lock (~70% cure), maybe better
• Likely better than ABx lock for S. aureus
• When to consider using?
• If HD catheter removal is clearly indicated but not feasible

(especially for S. aureus)

Robinson et al, Kidney Int 1998, 53:1792. Shaffer, Am J Kid Dis 1995, 25:593. Mokrzycki et al, Dial Transpl 2006, 21:1024. Aslam et al. JASN 2014;25:2927

13

Line Management: Take‐Home Points

• Differential time to positivity (line positive ≥ 2 hours

before peripheral) allows for diagnosis of CLABSI without line removal

• All lines should be removed for:
• Any complicated infection
• S. aureus, Pseudomonas, or Candida
• Line management for other organisms depends on line

type (lower barrier to remove line for short term catheter > long‐term catheter > HD catheter)

• Use antibiotic lock when possible for line salvage

Case #2

55 y/o woman in the ICU after a complicated spinal surgery. She remains intubated, spikes a fever

• n POD#3 and is pan‐cultured.
• She has thick secretions and a

new CXR infiltrate.

• mBAL is growing MRSA.
• UA (catheter): 25‐50 WBC, Ucx

positive for VRE.

14

Do You Need to Treat the VRE?

• 1. Yes
• 2. No
• 3. Not sure

Asymptomatic Bacteriuria

ASB = (+) urine culture AND no signs/symptoms of UTI

15

• Seen in up to:
• 25% of elderly, diabetic, HD patients, short‐term catheters
• 50% of patients in long term care facilities
• ~100% of patients with long‐term catheters
• Of positive urine cultures obtained on the wards after

hospital admission  ~90% are ASB

• Do not treat EXCEPT in pregnant women, GU

procedures, neutropenia/renal transplant

Asymptomatic Bacteriuria is COMMON!

Nicolle et al, Clin Infect Dis 2005, 40:643. Leis et al, Clin Infect Dis 2014, 58:980

The Heart of the Problem

• It’s Hard to Ignore a Positive Culture
• Proof of concept study:
• At Mount Sinai, 90% of their inpatient urine cultures were

ASB, and 50% were treated with ABx

• They stopped reporting these (+) urine cultures in the EMR
• Results:
• The % of ASB that was treated dropped by 80%
• No untreated UTIs and no sepsis

Leis et al, Clin Infect Dis 2014, 58:980.

16

How To Distinguish ASB vs. CA‐UTI?

• Does the UA help?  Yes, but only if negative
• Pyuria is seen in >50% of catheterized patients with ASB
• But the absence of pyuria suggests an alternative dx
• Does the organism help?  NO
• The same organisms cause ASB and UTI
• Use clinical context – does the patient have

signs/symptoms of UTI?

Nicolle et al, Clin Infect Dis 2005, 40:643. Tambyah et al, Arch Intern Med 2000, 160:678. Lin et al, Arch Int Med 2012, 172:33.

How to define UTI in patients with a catheter or AMS?

Nicolle et al, Clin Infect Dis 2005, 40:643.

Signs/symptoms consistent w/ UTI

• Fever, rigors, AMS, malaise
• Flank pain, CVAT, pelvic pain
• Acute hematuria
• Spinal cord injury: spasticity,

autonomic dysreflexia, unease No other source of infection (i.e., diagnosis of exclusion)

What if I Can’t Assess Symptoms?

AND

17

Alternate Diagnosis Likely? (Signs/ sx of other illness present) Yes Do not order U/A, urine cx No Send U/A, urine cx U/A, urine cx (‐) Do not treat for UTI U/A (‐), urine cx (+) Asymptomatic bacteriuria U/A (+), urine cx (+) Treat for UTI (If no alternate dx identified)

Slide courtesy of Catherine Liu.

U/A (+), urine cx (‐) Do not treat

How to Interpret Urine Studies in a Patient With a Foley or AMS

CA‐UTI: Treatment

• Antibiotics
• Empiric Rx: ceftriaxone, ertapenem, pip/tazo, cefepime
• Duration:
• 7 days if there is prompt resolution of symptoms
• 10‐14 days if response is delayed
• Catheter change?
• IDSA guidelines recommend yes as one study showed

 CA‐UTI at 28d and  time to resolution of sx

• But a 2018 study called this in to question – it showed no

difference in outcomes with catheter removal

Hooton et al, Clin Infect Dis 2010, 50:625. Babich et al, J Am Geriatr Soc 2018, 66:1779.

18

• Candiduria is very common in cathetrized patients
• Candiduria is usually asymptomatic
• In general, don’t treat! (exceptions: same for ASB)
• Change the foley: can eliminate candiduria in 20‐40%
• Symptomatic candiduria (uncommon)
• Look for same symptoms as bacterial UTI
• Treat if you are convinced

Candiduria: Who Needs Treatment?

Pappas et al, Clin Infect Dis 2009, 48:503.

Candida UTI: Treatment Options

• Fluconazole is the drug of choice
• Excellent urine levels
• 10‐fold higher than in serum
• Can get levels > MIC for fluc‐resistant species like C glabrata
• What about a fluconazole‐resistant organism?
• Try fluconazole and re‐check a urine culture
• Other options: flucytosine, conventional amphotericin,

ampho bladder washes

• Other azoles, echinocandins have poor urinary penetration

Fisher et al, Clin Infect Dis 2011, 52:S457.

19

ASB vs. CA‐UTI: Take‐Home Points

• ASB is very common and rarely needs Rx in the ICU
• Pyuria ≠ UTI, but the absence of pyuria  alternative dx
• UTI diagnosis in a patient with a catheter requires:
• Signs and symptoms compatible with UTI
• No other source for infection (i.e., diagnosis of exclusion)
• CA‐UTI can be treated with 7 days of antibiotics if

symptoms resolve quickly

• Fluconazole is the drug of choice for C. albicans (and
• ften non‐albicans)

Case #3

57 y/o man admitted with SAH s/p coiling, c/b vasospasm and stroke, now with persistent fevers. He has a PICC line, foley, and is intubated. He has been on vanc + pip/tazo for 5 days and continues to spike. No diarrhea, secretions unchanged.

• F 39.1, HR 65, other vitals normal.
• Exam is unremarkable.
• WBC is 11 (eos 0.63), Cr and LFTs normal.
• Blood cultures, UA, CXR, LP all negative

20

What Would You Do With His Antibiotics?

• 1. Escalate pip/tazo to meropenem
• 2. Add tobramycin
• 3. Add ganciclovir
• 4. Stop antibiotics

Non‐infectious Etiologies for Fever

• Drug Fever
• Central fever
• DVT/PE
• Malignancy
• Post‐op fever
• Rheumatologic
• Transfusion reaction
• Transplant rejection
• Adrenal insufficiency

21

Drug Fever

• Diagnosis of exclusion
• Clinical features:
• May appear well and be unaware of fevers
• No typical fever pattern
• Pulse‐temperature dissociation (11%)
• Rash (5‐10%)
• Eosinophilia (~20%)
• Timing:
• 7‐10 d after starting a drug (with re‐challenge, can be hours)
• Usually defervesce within 1‐2 days of stopping the drug
• Labor. Drug Intell Clin Pharm 1986; 20:413‐20. Mackowiak, Ann Int Med 1987; 106:728‐33. Foster, et al. Med Clin North Am

1966;42:523‐39

Drug Fever is Usually High‐Grade

Mackowiak, et al, Ann Intern Med 1987, 106:728.

40 39.6 41 39.9 39.8

38.5 39 39.5 40 40.5 41 41.5 Cardiac ABx Chemo CNS Other

Temp (˚C)

22

Drug Fever: Treatment

• Discontinue or change to another drug class if possible
• If benefit > risk to continue, can try to pre‐treat:
• Corticosteroids and/or antihistamines
• But watch for signs/sx of progression of hypersensitivity
• If fever with severe adverse effects, avoid rechallenge
• Important to document potential allergy with as much

detail as possible

Patel, et al, Pharmacotherapy 2010, 30:57. Joint Task Force on Practice Parameters. Ann Allergy Asthma Immunol 1999; 83:665‐700.

Central Fever

• Accounts for ~50% of fever in the NICU
• Which patients?
• Brain tumors, SAH, intraventricular bleed
• Associated with vasospasm
• Clinical characteristics:
• Appears within 72 h of admission
• Persists for longer than infectious causes of fever
• No difference in height of fever c/w infectious fever

Hocker et al, JAMA Neurol 2013, 70:1499.

23

8 5 1 2 4 6 8 10

Distribution of Fever in PE

VTE and Fever

• 5‐15% with PE/DVT
• Characteristics:
• Usually <38.9
• Peaks on day of PE
• Gradually subsides

within 1 week

Stein et al, Chest 2000, 117:39. Nucifora et al, Circulation 2007, 115:e173. Barba et al, J Thromb Thrombolysis 2011, 32:288.

% patients

VTE and Leukocytosis?

• Patients presenting to the hospital with acute PE and no
• ther cause for leukocytosis (n=266)

80 16 3 <10K 10‐15K 15‐20K >20K 20 40 60 80 100

WBC count % patients

Afzal et al, Chest 1999, 115:1329.

24

Tumor Fever

• Which cancers?
• Most common: lymphoma, leukemia, renal cell
• But any cancer can do it
• Pathophys: cytokines, tumor necrosis
• Clinical features:
• Tmax usually between 38‐39˚C
• Usually intermittent fevers, spiking once (most common) or twice daily
• +/‐ Leukocytosis
• Data is conflicting on use of naprosyn test, but some studies

show that defervesence with naprosyn predicts tumor fever

Zell and Chang, Support Care Cancer 2005, 13:870. Liaw et al, J Pain Symptom Manage 2010, 40:2.

Non‐infectious Fever: Take Home Points

• Always consider it, but it’s a diagnosis of exclusion
• Drug fever is usually high grade (>39˚C) ‐ look for

eosinophils, temp‐pulse dissociation, and rash although these are only seen in <20%

• Central fever is associated with SAH, vasospasm
• Fevers due to VTE or malignancy are usually <39˚C

25

Case #4

65 y/o man with cirrhosis is intubated for severe influenza and ARDS. He had been slowly improving but then over the last 2 days has starting having fevers to 38.4 with new production of thick secretions. He has trouble following commands when sedation is lifted.

• Blood and urine cultures neg
• CXR unchanged
• Head CT: pansinusitis

Your Next Diagnostic Step is:

• 1. Sinus puncture
• 2. Lumbar puncture
• 3. Mini‐BAL or endotracheal aspirate
• 4. BAL

26

Pneumonia in the ICU

• Hospital‐Acquired PNA (HAP) = PNA acquired after 48h in

the hospital and not incubating at admission

• Ventilator‐Associated PNA (VAP) = PNA acquired after

48h of intubation (subset of HAP)

• Microbiology overall is similar:
• Gram (+): S. aureus, particularly MRSA
• Gram (‐): Pseudomonas, E. coli, Klebsiella
• Pseudomonas, Stenotrophomonas, Acinetobacter more

common in VAP

IDSA/ATS Guidelines, Am J Resp Crit Care Med 2005. Weber et al, ICHE 2007 Kalil et al, IDSA/ATS guidelines, CID 2016.

HAP/VAP IDSA Guidelines 2016: What Changed?

• 1. HCAP no longer included (not at high risk for MDR)
• 2. Recommendation for semi‐quantitative endotrachaeal

aspirate over invasive methods (BAL, mini‐BAL)

• 3. Slightly less emphasis on using 2 antibiotics against

Pseudomonas for empiric coverage

• 4. Duration of therapy = 7 days for all pathogens

Kalil et al, IDSA/ATS Guidelines, CID 2016

27

VAP: Microbiologic Diagnostics

• Get blood cultures (~15% are positive)
• 2016 guidelines recommend semi‐

quantitative endotracheal aspirate over invasive sampling (mini‐BAL, BAL) (weak recommendation, low quality evidence)

Kalil et al, IDSA/ATS Guidelines, CID 2016

• Why?
• No difference in outcomes (mortality, ICU days, ventilation)
• Requires less resources
• Both ~75% sensitive but mini‐BAL/BAL more specific (80% vs 50%)

VAP: Clinical Diagnosis

• Also look at change in oxygenation over time
• In ARDS, consider PNA if have only ≥ 1 clinical criteria

because may not see CXR change

IDSA/ATS Guidelines, Am J Resp Crit Care Med 2005

New or progressive CXR infiltrate 2 clinical criteria

• Fever
• Leukocytosis/leukopenia
• Purulent secretions

70% sensitive 75% specific

+

28

VAP/HAP: Empiric ABx

• Cover for S. aureus, Pseudomonas,

GNRs

• Do you need MRSA coverage?
• Yes if MDR risk, >20% local S. aureus

isolates are MRSA, high risk of mortality

• Do you need 2 drugs for Pseudomonas?
• Yes if MDR risk, >10% local GNRs resistant

to monotherapy Abx, high risk mortality

• Use clinical judgment

Kalil et al, IDSA/ATS Guidelines, CID 2016

Risk of MDR VAP

• Prior IV Abx in 90 d
• Septic shock
• ARDS
• ≥5 d in hospital
• Acute HD/CRRT

Risk of MDR HAP

• Prior IV Abx in 90 d

VAP/HAP: ABx Menu

MRSA Vancomycin Linezolid Anti‐pseudomonal (β‐lactam) Piperacillin/tazobactam Cefepime/ceftazidime Meropenem/imipenem Aztreonam HAP only: levo/ciprofloxacin 2nd Anti‐pseudomonal Levo/ciprofloxacin Aminoglycosides

+ +/‐

Kalil et al, IDSA/ATS Guidelines, CID 2016

*Use local resistance patterns to help guide therapy

29

• RTC of 400 patients with VAP randomized to 8 vs. 15 days
• f ABx
• 8‐day group had:
• No difference in mortality, recurrent infections, ICU LOS
• More ABx‐free days and less MDR organisms if recurrent
• But…higher pulmonary reinfection rate (41 vs 25%) if had a

glucose nonfermenter (Pseudomonas, Acinetobacter, or Stenotrophomonas)

• This led to the recommendation for 15 days for glucose

nonfermenters and 8 days for everyone else

Duration of Antibiotics in VAP (8 vs 15 days)

Chastre et al, JAMA 2003, 290:2588.

New IDSA Guidelines: Duration of ABx in VAP

• Systematic reviews of 6 RCTs comparing short (7‐8 days)

vs long (10‐15 days) course therapy:

• Confirmed benefit of short course Rx (more Abx free days, less

recurrences with MDRO) and no difference in cure, mortality

• Glucose‐nonfermenter subgroup: no difference in recurrence,

mortality

• Bottom line:
• 7d treatment course, even for glucose non‐fermenters
• Extrapolate data to HAP
• Note MRSA IDSA guidelines recommend 7‐21d for MRSA PNA

Kalil et al, IDSA/ATS Guidelines, CID 2016.

30

HAP/VAP: When to Stop Empiric Vanco?

• Clinical factors which make MRSA less likely:
• Low clinical suspicion based on disease severity
• Negative respiratory cultures (before antibiotics)
• Note: negative blood cultures alone are not

sufficient as these are positive in only 5‐10% of MRSA PNA

Wunderink et al, Chest 2003. Wunderink et al, Clin Infect Dis 2012; 54: 621. Kalil et al, IDSA/ATS Guidelines, CID

• 2016. Parente et al, Clin Infect Dis 2018, 67:1.

The MRSA nasal swab:

• A negative MRSA nasal swab with a low prevalence of MRSA PNA has a

NPV of 95% for VAP, 98% for CAP

• Can also avoid starting vanco in the first place in stable patients if you

have a negative nasal swab within the last 7 days

HAP/VAP: Take Home Points

• Diagnosis is based on a combination of clinical and

microbiologic parameters

• Think about risk factors for MDR pathogens and local

resistance patters to guide empiric therapy

• Duration of therapy = 7 days in most cases
• MRSA nasal swab can be helpful to avoid starting or

for stopping vancomycin

31

Short Take: Nosocomial Sinusitis

• Epidemiology:
• Radiographic sinusitis in up to 75% of intubated pts but

clinical sinusitis in only 0.14%

• Etiology of fever in the ICU in <5%
• Radiographic sinusitis ≠ infectious sinusitis
• Micro/Treatment:
• Micro: Pseudomonas, S. aureus, can be polymicrobial
• Treatment: 7 days

Borman et al, Am J Surg 1992, 164:412. Talmor et al, Clin Infect Dis 1997, 25:1441. George et al, CID 1998, 27:463.. O’Grady et al, Crit Care Med 2008, 35:1330. Huyett et al, Laryngoscope 2016, 126:2433. Metheny et al, Am J Crit Care 2018, 27:24.

True nosocomial sinusitis is a rare cause of fever in the ICU True nosocomial sinusitis is a rare cause of fever in the ICU

Short Take: Acalculous Cholecystitis

• Rare (~1%) of all ICU patients
• Diagnosis:
• Symptoms/signs often not helpful
• LFT abnormalities in >60% but nonspecific
• US > CT
• GB wall thickness ≥ 3.5 mm (80% sensitive, 98% specific)
• Sludge, pericholecystic fluid, GB distention > 5cm, sonographic Murphy’s
• HIDA: sensitivity only 70‐80%
• High risk death (30%), GB gangrene (50%), perforation (10%)
• Treatment
• Cholecystectomy often not possible  percutaneous chole tube
• Antibiotics  target GNRs, Enterococcus, anaerobes +/‐ Candida

Barie and Eachempati, Gastroenterol Clin N Am 2010. Laurila et al, Acta Anaesthesiol Scand 2004. Zeissman, J Nucl Med Technol 2014.

32

Case #5

A 57 year old woman with metastatic breast cancer undergoing chemo and extensive prior antibiotic treatment is admitted to the ICU with septic shock.

• She is febrile to 39.6˚C, tachy to 120s, rapidly

uptitrated to max doses on 3 pressors.

• WBC is 1.4 (ANC 800), Cr 1.8, other labs normal.
• Blood and urine cultures are drawn and she is

started on vancomycin plus meropenem.

What Would You Do With Her ABx?

• 1. No changes (this is a source control issue)
• 2. No changes (ABx have not had time to work yet)
• 3. Add an aminoglycoside
• 4. Add a fluoroquinolone

33

Case #5 Continued

• Blood cultures: Pseudomonas susceptible to all Abx.
• Pressor requirement is downtrending.

What Would You Do With Her ABx Now?

• 1. Continue “double coverage”
• 2. Change to beta‐lactam monotherapy

34

“Double‐Covering” GNRs

• Also known as “combination therapy”
• Usually refers to a beta‐lactam + (aminoglycoside or

fluoroquinolone)

• Caveats to Combination Therapy Data :
• Often observational, non‐blinded studies
• Empiric vs definitive therapy not always defined
• Different beta‐lactams, different combinations, old ABx

3 Reasons To Consider Combination Rx

• 1. Increase the probability of appropriate empiric

coverage by expanding the spectrum of activity

• 2. Synergy between 2 active antibiotics
• 3. Prevent the development of resistance

Empiric combination therapy Definitive combination therapy

35

Reason #1: Empiric Combination Therapy

• mortality if inappropriate empiric Abx for GNR bacteremia
• Using empiric combination therapy will increase the

likelihood of having at least one active antibiotic

Paul and Leibovici, Clin Infect Dis 2013; 57:217.

When to empirically “double cover” for GNRs?

• Patient is critically ill
• Patient is at high risk for MDR pathogens
• How to choose between fluorquinolone(FQ) and

aminoglycoside (AG)?

• Know your local antibiogram: how good is the beta

lactam? What is the benefit of adding a FQ vs AG?

• Balance risk of nephrotoxicity from AG with risk of

inappropriate coverage

• Has the patient been on recent FQ?

Example: UCSF ICU VAP Coverage

VVanc/mero covers 80% Vanc/mero/cipro covers 85% Vanc/mero/tobra covers 87% So tobra does not add much

• ver cipro for VAP coverage

36

Reason #2: Combination Rx for Synergy?

• In vitro and animal studies
• Best data is for beta‐lactam plus aminoglycoside
• Data for beta‐lactam plus fluoroquinolone more sporadic
• Does this translate into clinical benefit?
• NO mortality benefit based on recent observational data

and meta‐analyses

Tamma et al, Clin Microbiol Rev 2012; 25:450. Paul and Leibovici, Clin Infect Dis 2013; 57:217.

What About in Certain Subgroups?

• Older studies from the 1980s/1990s showed benefit of

combination Rx in septic shock, neutropenia, Pseudomonas

• Issues with older studies:
• Monotherapy arm was often with an aminoglycoside
• Older beta‐lactams used, some without anti‐Pseudomonal activity
• Newer observational data/meta‐analyses show no benefit of

definitive combination therapy for:

• Septic shock
• Pseudomonas
• Neutropenia (although less data)

Tamma et al, Clin Microbiol Rev 2012; 25:450.

37

Reason #3: Combination Rx to Prevent Resistance?

• Combination therapy may prevent development of

resistance in vitro

• But in clinical practice, no evidence that combination

therapy prevents the development of resistance

Paul and Leibovici, Clin Infect Dis 2013; 57:217. Bliziotis et al, Clin Infect Dis 2005; 41:149. Paul et al, Cochrane Database Syst Rev 2014, Tamma et al, Clin Microbiol Rev 2012; 25:450.

Combination Rx for GNRs: Take Home Points

• Consider empiric combination therapy in critically ill

patients who are at risk of having MDR organisms

• The goal of “double‐covering” for GNRs is to ensure an

appropriate Abx is included in the initial empiric regimen

• Once susceptibilities are known, narrow to monotherapy
• There is no evidence that definitive combination therapy

is “synergistic” in vivo or prevents the development of resistance

38

Thank you!

• Questions?