Pacman
Efficacy of PA catheters in ICU patients
- Included 1014 patients in which a clinician deemed that a pulmonary artery catheter (PAC) was necessary for management.
- Most patients were enrolled to guide inotropic/vasoactive drug therapy due to multiorgan dysfunction.
- Randomized patients to either control (no PAC) or PAC.
- PAC placement occurred immediately after randomization.
- This was a very pragmatic trial, all clinical decisions regarding interpretation of PAC data and removal of the PAC was left to the discretion of the clinical team
- Primary endpoint (In-hospital mortality) was not different between control and PAC groups (66% vs. 68%, p=0.39)
- The study was adequately powered to detect a 10% absolute difference in
mortality - There was no difference in any secondary endpoint (eg, ICU/28-day mortality or ICU/hospital length of stay) or in any subgroup analyses
- Non-adherence to randomization assignment was uncommon (5-7%) even though most sites were allowed to use alternative cardiac monitoring devices
- Placement of a PAC was associated with a 10% insertion complication rate (primarily hematoma, arterial puncture, arrhythmias)
- The study’s pragmatic approach is both a strength and weakness.
- It is a strength because it reflects “typical” practice of PAC usage.
- It is a weakness because there was no assessment of whether clinicians interpreted PAC metrics appropriately to guide the decision-making process
TBL
Among a heterogeneous ICU patient population, the use of a pulmonary artery catheter (PAC) had no effect on mortality or length of stay. PAC placement was associated with a 10% complication rate.
PEITHO
Tenecteplase for intermediate-risk pulmonary embolism
- Included 1005 normotensive patients with confirmed pulmonary embolism and right ventricular dysfunction and an elevated troponin test (ie intermediate-risk PE)
- Excluded those with massive PE (defined as having hemodynamic decompensation, systolic BP < 90 mmHg), uncontrolled hypertension (BP > 180/110 mmHg), and known coagulation disorders (eg, warfarin use or platelets < 100k)
- Randomized patients to fibrinolysis (tenecteplase 30-50 mg IV push) or placebo.
- Both groups received unfractionated heparin with an initial bolus immediately after the study drug was given.
- Primary endpoint (composite of all-cause mortality or hemodynamic decompensation at 7 days) was lower with tenecteplase compared to placebo (2.6% vs. 56%, P=0-02)
- The primary composite endpoint was driven by hemodynamic
decompensation (1.6% vs. 5%, p=0.002) and not mortality (1.2% vs. 1.8%, p=0.42).- It was not powered to examine a difference in mortality.
- Tenecteplase was associated with a greater risk of major bleeding (11.5% vs. 2.4%, p<0-001) and hemorrhagic; stroke (2% vs. 0.2%, p=0-003)
- Of those with hemorrhagic stroke, 40% died within 30 days and the remaining survivors had mild-moderate disability
- Although pre-specified subgroups were not significant (p>0.05 for interaction), there is a possible signal of better efficacy and safety in those aged <75 years
- Of the post-hoc subgroup analyses, respiratory rate > 24 rpm was significantly in favor of tenecteplase for efficacy (p for interaction 0.028, OR 95% CI 0008-0489)
- The trial did not examine long-term endpoints, such as pulmonary hypertension or other sequelae of pulmonary embolism
- The tenecteplase dose used in this trial was the same as that used for acute MI.
- Some experts question whether the risk:benefit balance would have been improved with half-dose tenecteplase; a closing strategy that has been explored in other PE
trials.
TBL
Among patients with intermediate-risk pulmonary embolism, tenecteplase decreased hemodynamic decompensation but increased the risk of major bleeding.
Prodex/Midex
Dexmedetomidine vs. midazolam or propofol for mechanical ventilation
MIDEX (midazolam versus dexmedetomidine).
- Included 500 mechanically ventilated patients admitted to the ICU within 72
hours - Randomized patients to receive midazolam (0.03-0.2 mg/kg/hr) or dexmedetomidine (0.2-1.4 mcg/kg/hr) for up to 14 days. Loading doses were not allowed.
- Open-label rescue sedatives (fentanyl and propofol) were allowed for elevated RASS scores
- Primary endpoint (proportion of time within RASS goal of 0 to -3 without
rescue medications) showed non-inferiority between midazolam and dexmedetomidine (56.6% vs. 60.7%,p=0.15) - Discontinuation due to poor efficacy was more common with dexmedetomidine (4% vs. 9%, P=0-02), but use of rescue sedation was not different between groups
(45.4% vs. 43.8%, p=0.72) - Median duration of mechanical ventilation was longer with midazolam (6.8 days vs. 5.1 days, p=0.03); although there was no difference in ICU or hospital length of stay
- Patients receiving dexmedetomidine were more arousable, cooperative,
and better able to communicate pain (p<0.001) - Dexmedetomidine was associated with a higher incidence of hypotension (11.6% vs. 20.6%, p=0.07, NNH 11) and bradycardia (5.2% vs. 14.2%, p<0.001, NNH 11)
- Median midazolam dose was 0.062 mg/kg/hr (IQR 0.041 to 0.098), representing 5 mg/hr (IQR 3.3 to 7.8) for an 80 kg patient.
- Median dexmedetomidine dose was 0.45 mcg/kg/hr (IQR 0.27 to 0.76)
PRODEX (propofol versus dexmedetomidine).
- Included 498 mechanically ventilated patients admitted to the ICU within 72 hrs
- Randomized patients to receive propofol (5-66 mcq/kq/min) vs. dexmedetomidine (0.2-1.4 mcg/kg/hr) for up to 14 days. Loading doses were not allowed
- Permitted open-label fentanyl and midazolam
- Primary endpoint (proportion of time within goal RASS 0 to -3 without rescue medications) was not different between propofol and dexmedetomidine (64.7% vs. 64.6%, p=0.97)
- No difference in duration of mechanical ventilation or ICU length of stay
- Patients receiving dexmedetomidine were more arousable, cooperative, and better able to communicate pain (p<0.001)
- Discontinuation of study medication due to lack of efficacy was more common with dexmedetomidine (5% vs. 14%, p<0.001)
- Median doses of study medication:
- Propofol 1.75 mg/kg/hr (29 mcg/kg/min)
- dexmedetomidine 0.93 mcg/kg/hr
- Patients with dexmedetomidine required more midazolam rescue medication (72.5% vs. 64.4%, p=0.05)
- There was no difference in hypotension or bradycardia between groups
- Delirium, one of the most anticipated outcomes, was not collected during study treatment, and was only assessed at 48 hours post-extubation
TBL
MIDEX: In mechanically ventilated patients, dexmedetomidine was
comparable to midazolam for time within goal sedation but was associated with a shortened duration of mechanical ventilation. Compared to midazolam,
dexmedetomidine was associated with a much higher rate of hypotension and
bradycardia.
PRODEX: In mechanically ventilated patients, propofol was comparable to dexmedetomidine with respect to time within goal sedation, duration of mechanical ventilation, and ICU length of stay.
Prorata
Procalcitonin algorithm for guiding antibiotic therapy
- Included 630 ICU patients with suspected bacterial infections in seven different hospitals
- Excluded patients who were immunosuppressed or those with infections requiring a long-term duration of treatment (ie, endocarditis)
- Randomized to procalcitonin (PCT) or control groups in an open-label design
- PCT patients had a PCT level drawn at study inclusion, at each infectious episode, and every morning when receiving antibiotics. An algorithm
was used to recommend starting, stopping, or continuing antibiotic
therapy. - Control patients did not use a PCT- guided algorithm. The decision to start or stop antibiotics was based on clinical judgment.
Simplified procalcitonin algorithm:
- PCT Level
- <0.25 mcg/mL – Strongly consider stopping or not starting antibiotics
- 0.25 to 0.5 mcg/mL – Consider stopping or not starting antibiotics
- >0.5 ng/mL – Consider starting or continuing antibiotics
- >1 mcg/mL – Strongly consider starting or continuing antibiotics
- >0.5 mgg/mL and PCT increasing – Strongly consider changing antibiotics
- >80% decrease from highest level – Consider stopping antibiotics
- Primary endpoint (28-day mortality) was no different between PCT and control (21.2% vs. 20.4%, p=NS)
- 60-day mortality was also no different.
- Days without antibiotics (in first 28 days) was significantly higher (better) in the PCT group (14.3 vs. 11.6 days, 23% reduction, p<0.0001)
- Comparing PCT to control, there was no difference in relapse rates (6.5% vs. 5.1%, p=0.45), superinfections (34.5% vs. 30.9%, p=0.29), days; without mechanical ventilation (16.2 vs. 16.9 days, p=0.46), or ICU/hospital length of stay
- Adherence to the PCT algorithm’s recommendations was moderate (47% adherence), indicating that a PCT level should be evaluated in combination with other factors
TBL
Use of a procalcitonin-guided algorithm for suspected bacterial infections reduced antibiotic exposure but did not directly improve patient outcomes
Proseva
Prone positioning in severeARDS
- Included 466 patients with early (within 36 hours of ARDS criteria), severe (Pa02:Fi02 < 150 mmHg) acute respiratory distress syndrome
- After a 12-24 hour stabilization period to verify inclusion criteria, randomized patients to supine positioning (control) or prone positioning (treatment) for up to 28 days.
- Many other factors in ARDS management (mechanical ventilator adjustments,
weaning, sedation, and paralytics) were protocolized - Prone positioning was conducted for at least 16 hours per day in standard ICU patients.
- Proning was stopped once oxygenation criteria improved (Pa02:Fi02 >= 150 mmHg with PEEP <= 10 and Fi02 <= 60%)
- Primary endpoint (28-day mortality) was significantly lower with prone
positioning (16% vs 32.8%, HR 0.39, NNT 6, p<0.001). 90—day mortality was also reduced with proning (23.6% vs. 41%, HR 0.44, NNT 6,
p<0.001) - As would be expected, most patients had ARDS due to pneumonia (about 60%), and most received vasopressors (73-83%), neuromuscular blockers (82-91%), and many received glucocorticoids (40-45%)
- On average, prone patients received 4 prone sessions lasting for 17 hours. These patients were proned for 73% of the time spent on a ventilator
- Importantly, all ICUs included in this trial had at least five years of proning experience.
- Given the complexity and resources required for proning, this is a potential issue for external validity.
- Previous trials regarding proning have been conflicting. This trial may have
shown such a profoundly favourable effect due to early intervention (within 36 hours of ARDS), selecting more appropriate patients (12—24 hour stabilization period), and the significant proportion of time spent Proning (73% versus closer to 30% in other trials) - The mortality reduction of >50% in severe ARDS seems “too good to be true” — many clinicians are awaiting further studies, but for the time being, these results are difficult to ignore.
TBL
In patients with early, severe ARDS, Prone positioning for at least 16 hours per day Significantly reduced mortality.
PROTECT
Dalteparin vs. unfractionated heparin for DVT prophylaxis
- Included 3,746 patients (75% medical ICU, 90% on mechanical ventilation, 45% with vasopressors)
- Randomized to dalteparin 5000 units daily or unfractionated heparin (UFH) 5000 units Q12h (Q12h heparin may have been a weak comparator – some favor Q8h heparin)
- Patients were NOT allowed to use SCD’s unless HIT was suspected (huge problem with external validity)
- Primary endpoint: proximal leg deep-vein thrombosis (did not include other VTEs) was not different between groups
- The study was slightly underpowered for the primary endpoint (incidence rate in UFH arm 5.8%, power calculation for 8%)
- Dalteparin associated with about 50% less definite or probable PE (2.1% vs. 1.2%. p=0.02, NNT 100)
- Rate of HIT was 0.6% (UFH) and 0.3% (dalteparin) (p=NS)
- No difference in bleeding, death, duration of mechanical ventilation, or ICU/hospital length of stay
TBL
Dalteparin was comparable to unfractionated heparin in preventing proximal leg DVT. There were no differences in clinical endpoints (length of stay, mortality), but dalteparin was associated with less pulmonary embolism.
PROWESS
Xigris for severe sepsis
- Included 1690 patients with severe sepsis or septic shock within 24 hours
- Randomized to receive drotrecogin alfa (DrotAA) 24 mcg/kg/hr or placebo x 96 hrs
- Primary endpoint (28-day mortality) was lower with DrotAA (24.7% vs. 30.8%, p=0.005, NNT 16)
- Pre-specified subgroup analysis showed DrotAA was more effective with an APACHE II score of >= 25 (31% vs. 44%, NNT 8) than a lower APACHE II (19% vs. 19%)
- Serious bleeding was more common with DrotAA (3.5% vs. 2%, p=0.06, NNH 67)
- Limitations included stopping the trial early for benefit (planned enrollment of 2280 patients), study changed exclusion criteria mid-trial (excluded more chronic conditions), benefit was only seen in a subgroup of patients (APACHE II >= 25), and mortality benefit was primarily seen in sites with DrotAA experience (having enrolled at least 9 patients)
- Subsequent trials (PROWESS-SHOCK) showed a lack of benefit with DrotAA, which led to its voluntary withdrawal from the market in 2011.
- There is considerable controversy regarding whether the original PROWESS trial was flawed, or whether improvement. in sepsis treatment since 2001 blunted the efficacy of DrotAA
TBL
Drotrecogin alfa improved mortality among patients with severe sepsis or septic shock, particularly those with a high APACHE II score, but subsequent trials have failed to show a similar mortality benefit.
PROWESS SHOCK
Xigris for septic shock
- Included 1697 patients with septic shock (requiring 4+ hrs of vasopressors) and clinical evidence of hypoperfusion (metabolic acidosis, renal or hepatic dysfunction).
- Randomized to receive drotrecogin alfa (DrotAA) 24 mcg/kg/hr or placebo x 96 hrs
- Primary endpoint (28-day mortality) was no different between DrotAA and Placebo (6.4% vs. 24.2%, p=0.31).
- 90-day mortallty was also not different between treatment groups.
- All subgroup analyses did not show a group that may benefit from DrotAA (including patients with severe protein C deficiency, APACHE >= 25, number of organ failures, and use of corticosteroids)
- DrotAA was associated with more non-serious bleeding (8.6% vs. 4.8%, P=0.002, NNH 26), but there was no difference with serious bleeding events (1.2% vs. 1.0%) or intracranial hemorrhage (0.4% vs. 0.4%).
- Given the robust trial design and lack of drug efficacy, the results of PROWESS-SHOCK led to the voluntary withdrawal of DrotAA from the market in 2011
- There is considerable controversy regarding whether the original PROWESS trial was flawed, or whether improvement in sepsis treatment since 2001 blunted the efficacy of DrotAA
TBL
Drotrecogin alfa did not improve 28—day or 90—day mortality compared to placebo in patients with septic shock.
Prandharipande 2006
Lorazepam is an independent risk factor for ICU delirium
- Prospective, observational, single- center trial enrolling 198 mechanically ventilated patients in a medical or coronary ICU
- Lorazepam use was associated with the transition to delirium (OR 1.2, 95% CI 1.1-1.4).
- A dose of 20 mg/day was associated with nearly 100% incidence of delirium the following day
- Midazolam was non-significantly associated with delirium (OR 1.7, 95% CI 0.9-3.2), although sample size was very low
- The risk of delirium increased in patients > 65 years of age or those with a higher severity of illness (APACHE II > 18)
- Usage of analgesic and sedative medications was not reported (eg frequency of midazolam vs. lorazepam use, continuous vs. intermittent dosing strategies)
- Prospective, observational nature of study may not have adequately corrected for baseline characteristics, such as severity of illness, that may be present with the use of higher doses of lorazepam
TBL
Lorazepam was independently associated with the development of delirium in a dose dependent fashion. Other agents, such as midazolam, were numerically associated with higher incidence of delirium but did not reach statistical significance.
See the paper here
Pickard 1989
British aneurysm nimodipine trial
- Included 554 patients with aneurysmal (non-traumatic) subarachnoid hemorrhage within 96 hours.
- Unlike the Allen 1983 trial, patients could have neurologic deficits and still be included (better external validity)
- Randomized patients to nimodipine 60 mg PO Q4h or placebo x 21 days
- Primary endpoint (cerebral infarction) was reduced with nimodipine (22% vs. 33%, p=0.003, NNT 9)
- Poor outcome (death or severe disability) at 3 months was similarly reduced with nimodipine (20% vs. 33%, p<0.001, NNT 8)
- Safety analysis was minimal and poorly descriptive“ Nimodipine had a mean systolic BP decrease of 7.1mmHg
TBL
Nimodipine reduced cerebral infarction and 3-month functional outcomes in patients with aneurysmal subarachnoid hemorrhage.
Pontes-arruda 2006
Effect of Oxepa pharmaconutrition for ARDS
- Single center, 3-ICU study including 103 patients with severe sepsis or septic shock and ARDS (pa02;Fi02 <200)
- Randomized patients to Oxepa, an enteral nutrition formulation rich with omega-3 (EPA/GLA) and anti-oxidants, vs. a comparable isocaloric, isonitrogenous formulation.
- Study treatment was continued until extubation or at physicians’ discretion
- Primary endpoint (28-day mortality) was reduced with Oxepa (33% vs. 52%, p=0.037, NNT=5)
- Oxepa was associated with improved 28-day ventilator-free days (13.4 vs. 5.8 days, p<0.001), ICU-free days (10.8 vs. 4.6 days, p<0.001), an improvement in Pa02:Fi02 ratios at day 4 and day 7, and an improvement in new organ failures
- The study enrolled 165 patients, but excluded 62 from data analysis due to protocol violations (most commonly death or unable to meet caloric goals). By excluding these patients following randomization, the study analysis is per-protocol, not intent-to- treat, which weakens its external validity
TBL
Oxepa enteral nutrition was associated with improved mortality, days of mechanical ventilation, and ICU length of stay among septic ARDS patients. While these results are promising, the study was limited by its per-protocol analysis.
ProCESS
Protocol-Based Care for Early Septic Shock
- Included 1341 patients with septic shock from the emergency departments of 31 academic hospitals in the United States
- Randomized patients in a 1:1:1 ratio to protocolized early goal directed therapy (EGDT), protocolized standard-therapy, or usual-care
- Protocolized EGDT was the Rivers 2001 protocol, which included a central line with SCvO2 monitor, fluids, vasopressors, dobutamine, and pRBCs.
- Protocolized standard-therapy included fluids (titrated based on
Systolic BP, shock index, and clinical Symptoms of fluid overload), a central line only if peripheral access was insufficient, no dobutamine, and a lower hemoglobin threshold for pRBCs (75 g/dL). - Use of CVP and SCvO2 were discouraged.
- Usual-care did not prompt the Physician for any treatment – all decisions were left to the treating physician’s discretion
- Primary endpoint (60-day in-hospital mortality) was not different for EGDT (21.0%), standard-therapy (18.2%), or usual-care (18.9%)
- The study was powered to detect an absolute mortality reduction of about
7% - There was no difference in primary endpoint between the combined protocolized patients and usual-care (19.5% vs 18.9%, p=0.83)
- Subgroup analysis, specifically related to severity of illness, was unrevealing of any differences among groups
- There were no differences other variants of mortality outcomes, length of stay, or adverse effects
- Standard-therapy patients received a greater volume within the first 6 hours compared to EGDT (3.3 L vs 2.8 L, P<0-001), but were less likely to receive dobutamine (1.1% vs. 8.0%, p<0.001) or pRBC transfusion (8.3% vs. 14.4%, p=0.001)
- New renal replacement therapy was more common with standard-therapy than the other two groups (6.0% vs. 3.1% EGDT vs. 2.8% usual-care, p=0.04)
- This trial was designed in response to the Rivers 2001 study, which was a single center study of EGDT done more than a decade before the ProCESS trial in which septic shock mortality rates were much higher
- It is unlikely that “usual care” is equivalent among all emergency departments in the US; therefore, it is difficult to ascertain whether a protocol-less approach is appropriate for septic shock patient care based on this trial alone
- This study offers a less invasive, equally efficacious, protocolized approach to early septic shock management compared to the Rivers EGDT protocol
TBL
Traditional early goal-directed therapy, a new protocolized standard-therapy, and a no-protocol usual-therapy approach were all equally effective in treating early septic shock patients.
