Post on 24-Jan-2017
Evidence-based critical care – Update 2006
Joel Peerless MD3 January 2006
Intensivist shortage
Experts predict that as the US population ages, the shortage of intensivists will become increasingly acute
By 2020, the supply of intensivists will meet only 22% of the demand for their services
The old way…
Intern/junior resident (Dr. X) was taughta concept by his/her:Senior residentChief residentFellowAttending
Dr. X practiced what he was taught…
The old way…
Dr. X went on to become:A senior residentA chief residentA fellowAn attending
And taught the same concept to his/her junior resident, and so on and so on….
The old way…”Why do you do it that way??” “Well, I learned this from Dr. X” “We’ve always done it this way” “We have good outcomes” “I have an article to prove it” (more on
this later…)
The old way… dopamine
Dopamine in low doses activates dopamine receptors in the kidney
Renal blood flow is increased Urine output is increased (sometimes) The assumption, and teaching,
became…
The old way… dopamine
Dopamine is indicated for:Preventing renal failureTreating renal failureReversing renal vasoconstriction when vasopressors are usedPreventing renal failure during aortic and renal cross-clamping
Dopamine flowed like water (often better than urine) in ICUs worldwide
Evidence-based critical care
THERE NOW EXISTS A MORE “SCIENTIFIC” METHOD TO PROVIDE CARE…
The new way…
Try to adapt practice based on the quality of clinical studies that support the test or intervention – “Evidence-based”
A careful evaluation of existing studies Type of trial Number of centers Number of patients Quality of the study
Evidence-based concepts
The best study is:ProspectiveRandomizedDouble Blinded – not always possibleMulticenter
Meta-analyses evaluate a number of similar studies
Evidence-based concepts
Surviving Sepsis Campaign Guidelines for the Management of Severe Sepsis and Septic Shock
11 critical care societies - international Initial document: CCM March, 2004 Evidence-based review: CCM
supplement November, 2004
Grading system
Grade A: At least 2 large, randomized trials with clearcut results
Grade B: At least 1 large, randomized trial with clearcut results
Grade C: Small, randomized trials, uncertain results
Grading system
Grade D: Nonrandomized, contemporaneous controls
Grade E: Nonrandomized, historical controls; case series; uncontrolled studies; expert opinion
Evidence-based critical care
VENTILATOR MANAGEMENT OF ARDSVt settingsLevel of PEEPUse of steroids in “late” ARDS
Low tidal volume strategy
Barotrauma High pressure generated with flow of air
into lungs Volume trauma (volutrauma)
Overdistention of alveoli can lead to further lung damage
Low tidal volume strategy
Classic practice: Vt 10-15 cc/kg Hypothesis: lower Vt is protective of
alveolar damage and outcome in ARDS ARDS-net trial, NEJM, 2000 Comparison of 6 ml/kg vs 12 ml/kg 861/1000 patients Mortality 31% vs 39.8% Reduction in mortality of 22%
Low tidal volume strategy
When ARDS diagnosed, adjust Vt for ideal body weight 60 kg x .6 cc/kg = 360 cc
pCO2 will likely increase – permissive hypercapnea
GRADE B
PEEP levels in ARDS
ARDS-net, NEJM, 2004 Comparison of low and high PEEP in
ARDS 549 patients Low versus high PEEP algorithm All patients had low Vt No difference in ventilator free days or
mortality
Steroids for “late” ARDS
Fibroproliferation can occur after 5-7 days of ARDS
This inflammatory process may present with worsening oxygenation, fever and leukocytosis (infection may present with same signs)
Case series of steroids to minimize the late inflammatory fibroproliferation
Steroids for “late” ARDS
ARDS-net randomized placebo-controlled trial
Goal 200 patients over 6-8 years No value in use of steroids in “late”
ARDS
Other ARDS modalities
We no longer consider:Prone positionInhaled nitric oxideSurfactantLiquid ventilation
Evidence-based critical care
Glucose control in critically ill patients
Intensive insulin therapy
NEJM, 2001, single center study 1,548 surgical patients Randomized to tight vs non tight control 80-120 mg/dl vs 150 mg/dl ICU mortality 4.6% vs 8.0% 43% risk reduction
Intensive insulin therapy
50% decrease in BSI, ARF requiring RRT, critical illness polyneuropathy and transfusion requirement
Shorter duration of mechanical ventilation and ICU care
Less multiple organ failure
Intensive insulin therapy
Mayo Clin Proc, 2004 800 critically ill medical-surgical patients
in a community ICU Hospital mortality decreased Decreased ICU LOS, ARF, transfusion
requirement Most benefit in septic patients
Intensive insulin therapy
GRADE D based on original study, (randomized but single center, mainly surgical patients), likely upgrade to C
German study (600 patients) – negative Two large scale randomized trials
underway European study – 3500 medical/surgical
patients Australia/NZ – 4500 medical/surgical
patients
Evidence-based critical care
Activated protein-C in sepsis (Xigris)
RH Activated Protein C
Prowess Study Recombinant human activated protein C
worldwide evaluation in severe sepsis Randomized, double blinded, placebo
controlled 96 hour infusion of activated protein C
(APC) Endpoint: death at 28 days
RH Activated Protein C
Prowess study 1690 patients (840 placebo; 850 APC) Mortality 30% vs 24.7%, p=0.005 6.1% reduction of death Serious bleeding 2% vs 3.5%, p=0.06 Similar number of blood transfusions
RH Activated Protein C
Prowess study The difference in outcome was greatest in
patients with an APACHE score >25 Address study
PRCT of RHAPC in patients with APACHE <25
No improvement in outcome May be worse outcome in SURGICAL
patients with single organ dysfunction
RH Activated Protein C
Contraindications Internal bleeding, hemorrhagic CVA, CHI,
neurosurgical procedures, surgery within 12 hours
Many patients with abdominal sepsis require repeat explorations
The drug is expensive ($100/kg)
RH Activated Protein C
GRADE B for patients with severe sepsis, multiple organ dysfunction, with APACHE >25
“Dear doctor” letter from Eli Lilly re: surgical patients with single organ dysfunction
Evidence-based critical care
CORTICOSTEROIDS AND SEPTIC SHOCK
Corticosteroids and sepsis
Old teaching: don’t give septic patients high dose steroids
New thought: patients with septic shock who are unresponsive to pressor agents may have adrenal insufficiency
Corticosteroids and sepsis
Incidence of adrenal insufficiency 30-50% in critically ill patients 50-60% in patients with septic shock
Clinical presentation resistant hypotension hyponatremia, hyperkalemia
Low or “normal” serum cortisol level
Corticosteroids and sepsis
JAMA, 2002; multicenter, RCT, 300 patients Septic shock, vasopressor therapy A baseline cortisol level of <25 mcg/dl Corticotropin stimulation test (cortrosyn) An increase in serum cortisol of 9 mcg/dl after
corticotropin – “responders” Unable to increase serum cortisol 9 mcg/dl –
“nonresponders”
Corticosteroids and sepsis
Nonresponders randomized to placebo vs hydrocortisone 50 mg IV every 6 hours + fludrocortisone
Mortality improved in patients with adrenal insufficiency
Faster time to vasopressor withdrawal No difference in patients with adequate
corticotropin response
Corticosteroids and sepsis
Should low doses of corticosteroids be used in the treatment of septic shock?
GRADE C SSC guidelines suggest steroids in
pressor-dependent patients with low serum cortisol (don’t even bother with the cortrosyn stimulation)
Evidence-based critical care
PULMONARY ARTERY CATHETER USE IN THE ICU
PA catheter use in the ICU
Old practice: any unstable patient who didn’t respond to fluids or who had possible cardiac dysfunction; “we didn’t know where we were…”
Preoperative “optimization” of high risk patients
Old practices questioned: Connors, et al, JAMA, 1996
Editorial: “abandon PACs until PRCT”
PA catheter use in the ICU
JAMA, 2005 Meta-analysis, 13 RCTs, 5051 patients Use of PA catheter neither increased
overall mortality or hospital days, nor conferred benefit
Use of the PA catheter was associated with a higher use of inotropes and vasodilators
PA catheter use in the ICU
RCT in progress ARDS-net, 1000 patient goal FACTT Study: PRMCT of PAC vs CVP
for management of ALI and ARDS and PRMCT of “fluid conservative” vs “fluid
liberal” management of ALI and ARDS
Evidence-based critical care
EARLY GOAL-DIRECTED THERAPY IN SEPSIS
Early goal-directed therapy
“Standard” goals for sepsis (resuscitation) MAP >65 CVP 8-12 Urine output >0.5 cc/hr
Early goal-directed therapy
DO2 = Hgb (SaO2) x 1.34 x CI x 10 “Supranormal” goals for PAC (1980s)
CI > 4.5 L/min/m2
DO2I > 600 ml O2/min/m2
VO2I > 170 ml 02/min/m2
Fluids, blood, inotropes, pressors… We could measure our progress with a
calculator!!
Early goal-directed therapy
Supranormal goals disproven in established sepsis (NEJM x2)
Will this theory work if we get to the patient earlier?
Early goal-directed therapy
NEJM, 2001, Henry Ford Hospital 263 ED patients Control group: MAP, CVP, UO Protocol group: MAP, CVP, UO,
SCVO2 >70 Mortality 46.5 vs 30.5%, p=.009
Early goal-directed therapy
There IS value to optimizing cardiac output, oxygen delivery and SVO2 if you intervene early enough
(Is it really the SVO2 that makes the difference, or is it resuscitation??)
Evidence-based critical care
ANTIBIOTICS and INFECTION CONTROLWHO NEEDS TREATMENT AND FOR HOW LONGVENTILATOR ASSOCIATED PNEUMONIACATHETER-RELATED BACTERIAL STREAM INFECTIONC-DIFFICILE COLITIS
Who needs antibiotics?
Systemic inflammatory response syndrome (SIRS)
Fever means inflammation, not necessarily infection
Emergence of multiresistant bacteria Treating fever alone or fever + WBC
without a source of infection is usually not in the best interest of the patient
Duration of antibiotic coverage
Minimal data over duration of antibiotic usage for most infections Bacteremia – 2 weeks Endocarditis – 6 weeks
What about “double coverage” for pneumonia SPACE bugs: Serratia, Pseudomonas,
Acinetobacter, Citrobacter, Enterobacter VAP coverage was always 2 weeks
SICU antibiotic usage
Perioperative coverage – single dose or 24 hours of coverage
UTI – 5 days Sinusitis – 5-7 days Bacteremia – 10 days
Evidence-based critical care
VENTILATOR-ASSOCIATED PNEUMONIADIAGNOSISTREATMENTPREVENTION
Ventilator-associated pneumonia - diagnosis Classic teaching: Fever, WBC, purulent
sputum, + culture, infiltrate 75% of intubated patients are colonized
with GNB after 3 days in the ICU You could have bronchitis or sinusitis +
atelectasis, or pulmonary edema Endotracheal aspirate vs BAL
BAL samples give quantitative cultures
Ventilator-associated pneumonia - diagnosis Fever/WBC, + culture without purulent
sputum or CXR – NO ANTIBIOTICS Fever/WBC, purulent sputum or
changing CXR, + sputum culture → empiric therapy
Fever/WBC, purulent sputum or changing CXR, + sputum culture → BAL + empiric therapy
Ventilator-associated pneumonia treatment JAMA, 2003 PRCDBT, 51 French ICUs, 401 patients VAP treatment for 8 vs 15 days No excess mortality, recurrent infections,
ventilator days, ICU LOS with 8 day treatment
But, recurrence of VAP was higher with PSD
Ventilator-associated pneumonia prevention Subglottic suctioning of secretions
Additional suction port above the ETT cuff Some studies suggest decreased VAP Who will get them, will you risk changing
the tube?? Role of gastric contents
Gastric residuals: not shown to correlate with VAP (a major cause of poor nutrition)
Ventilator-associated pneumonia What is known
More likely from mouth contents than from regurgitation of stomach contents
Head of bed 30 degrees up is the only proven preventive technique!!
Early effective antibiotic therapy is essential !!
Start with broad-spectrum coverage, then taper antibiotics based on EA/BAL results
Catheter-related BSI
Prevention – what doesn’t work Changing lines every three days Performing wire changes Putting antibiotic solutions (goop) on the
site Antibiotic lock solutions
Catheter-related BSI
Prevention – what DOES work Subclavian site Prep: Chlorhexidine (not povidone) Barrier precautions: Hat, mask, gown,
gloves, FULL body drape Antiseptic coated catheters ($$$; beware
of hospital administrators) Management of stopcocks: alcohol prior to
injections, keep ports closed
C-difficile colitis
Related to indiscriminate antibiotic usage May occur with only a single dose of
antibiotics (even perioperative coverage) May require colectomy May be fatal A huge topic in the lay press and
certainly will be seized on by the legal community
Infection control techniques
We all have the power to limit the occurrence of hospital acquired infections
Hand washing is the most effective method of limiting hospital acquired infections
Infection control techniques
Soap and water Hand disinfectant solutions : Steris
product, Purell Remember: C-difficile toxin spores are
not killed by hand solutions; use soap and water
Wear protective gowns and mask when examining c-diff patients
EBM in perspective
EBM in perspective
We have very good data (A) DVT prophylaxis Stress ulcer prophylaxis Spontaneous breathing trials Preventive practices for CR-BSI
EBM in perspective
We have good data (B), but these studies can likely never be repeated Early goal-directed therapy in septic shock Activated protein-C for severe sepsis Low Vt in ARDS
EBM in perspective
We have good data (B) that will probably never get better despite further studies Crystalloids vs colloids, albumin Bicarbonate in acidosis
We are able to eliminate unproven or potentially unsafe practices NO for ARDS Prone positioning for ARDS
EBM in perspective
We practice with lots of E’s – these probably can’t ethically be studied Use of antibiotics early in sepsis Draining abscesses Giving fluids for hypovolemia Vasopressors in septic shock Mechanical ventilation in respiratory failure
EBM in perspective – the good
We know the quality of the data Development of consensus statements
ASA, SCCM, ACCP, EAST www.guideline.gov
Guidelines and protocols Keep us current and consistent in our care Allows ongoing monitoring of practices Educate our trainees
EBM in perspective – the bad Extending data to non-studied or non-
proven conditions:Low tidal volumes for all: There is no data that patients who do not have ARDS should be ventilated with low tidal volumesDaily breaks in sedation: It is not reasonable to break sedation in some patients: ARDS, elevated ICP, etc
EBM in perspective – the bad
Blanket referrals to EBM:PAC usage in critically ill patients: A PAC may still be useful in managing your patientThe anti-dopamine conspiracy: Lancet, 2000: 328 patients; 23 ICUs; no effect on renal dysfunction, LOS or mortality
EBM in perspective: the bad
There are patients with low urine output or impending renal failure secondary to low cardiac output who may benefit from the use of low dose dopamine
EBM in perspective – the scary
OUTCOMES RELATED TO EVIDENCE-BASED CARE WILL BE MONITORED AND WILL BECOME A MEASURE OF QUALITY AND POTENTIALLY A SOURCE OF MEDICOLEGAL LIABILITY
EBM in perspective – keep reading The data may change !! JAMA, July 13, 2005 Contradicted findings (16%)
Hormone therapy and CAD risk Vitamin E and CAD risk Vit A and breast cancer Monoclonal AB to endotoxin (HA1A) Nitric oxide in ARDS
EBM in perspective – keep reading Studies with initially stronger effects
(16%) Zidovudine in HIV Angioplasty vs TPA in AMI TPA in CVA CEA and risk reduction of CVA or death Acetylcysteine preventing contrast
nephropathy
EBM in perspective – reading the literature Cautious use of “the literature”,
“landmark” articles It’s easy to quote the last line of the
abstract LOOK FOR THE EDITORIAL
Comments on the quality and limitations of the study
“This study raises more questions than it answers”
“Large, randomized trials are needed”
Summary
We should use EBM consensus statements to keep our knowledge current
Reaching the highest rankings (grade A) may be limited by ethics, resource restrictions and marketing strategies
Be watchful for “new” data or negation of previously accepted data
Summary
Realize that the public is watching our adherence to practice guidelines
Accept the fact that external monitoring of evidence-based practices and outcomes will be standard in the future
Be proactive in maintaining high standards of care