Bernard
- Included 77 patients with out-of-hospital VF arrest, return of spontaneous circulation (ROSC), and persistent coma post-ROSC
- Randomized patients to therapeutic hypothermia (33°C) or normothermia x 12 hrs.
- Both patient groups received midazolam and vecuronium
- At baseline, the normothermia group was more likely to receive bystander CPR (49% vs. 71%, p=0.05)
- Primary endpoint (favorable discharge disposition to home or acute rehabilitation) was more likely with hypothermia (49% vs. 26%, p=0.046, NNT 4)
- Hypothermia was associated with a non-significant trend towards improved 30-day mortality (51% vs. 68%, p=0.145, NNT 6)
- An adequate analysis of safety (bleeding events, electrolyte disturbances) was not reported within the study
TBL
Therapeutic hypothermia in patients with out-of-hospital VF arrest improved the incidence of favorable discharge disposition and a trend towards improved mortality.
Besson
- Included 199 patients with active variceal bleeding or nonbleeding varices with evidence of GI bleed without any other source during endoscopy
- All patients received emergency sclerotherapy during the initial endoscopy
- Randomized patients to octreotide 25 mcg/hr or placebo for five total days
- Primary endpoint (survival without rebleeding in first five days) was significantly higher with octreotide than placebo (87% vs. 71%, p=0.009, NNT 6)
- The primary endpoint remained significant after adjusting for baseline characteristics (OR 3.3, 95% CI 1.5 to 7.3)
- There was no difference in 5-day mortality between octreotide and placebo (7% vs. 10%, p=0.49)
- The primary endpoint was superior with octreotide regardless of Child-Pugh score or other baseline characteristics
- Octreotide was associated with a reduced need for blood transfusion within the first 24 hours (median 1 vs. 2 units, p=0.006)
- Octreotide was associated with a higher incidence of hyperglycemia (glucose > 120 mg/dL, 23% vs. 13%, p=0.065)
- Other adverse effects were minor; there was no difference in acute renal failure or encephalopathy (two theoretical complications of octreotide)
TBL
In patients with variceal hemorrhage, sclerotherapy with octreotide is more effective in reducing rebleeding rates (but not mortality) compared to sclerotherapy alone.
Bone
- Included 382 patients with severe sepsis or septic shock at 19 centers in the United States
- Randomized patients to receive either high-dose methylprednisolone IV 30 mg/kg Q6hr x 4 doses or placebo within two hours of meeting study inclusion criteria
- One primary endpoint (14-day mortality) was not different between methylprednisolone and placebo, but did trend toward a detrimental effect (34% vs. 25%, p=0.06)
- Methylprednisolone did not reduce the incidence of shock among the severe sepsis cohort (46% MP vs. 37% placebo, p=NS) or shock reversal by 14 days (65% MP vs. 83% placebo, p=NS)
- The incidence of secondary infection was not different between methylprednisolone and placebo (19% vs. 20%); however, patients receiving methylprednisolone were more likely to die due to a secondary infection (34% vs. 7%, p<0.015, NNH 4)
- Although the data was collected, there is no mention of the glucose control among the patients receiving the very high dose of methylprednisolone
- While understandable given the publication date, there is a lack of data regarding the safety of the methylprednisolone regimen (eg, incidence of GI bleeding, wound infections, myopathy, neuropathy, etc.)
TBL
Early, high-dose methylprednisolone did not improve any clinical outcomes in severe sepsis and septic shock and may be associated with harm
Bozzette
- Included 225 patients with HIV-associated pneumocystis pneumonia
- Excluded mechanically ventilated patients or those with PaO2:FiO2 ratio < 75
- Randomized patients to prednisone (40 mg BID x 5 days, 40 mg daily x 5 days, 20 mg daily for duration of antibiotic therapy [typically 14-21 days]) or placebo
- All patients received standard antibiotic therapy (trimethoprim-sulfamethoxazole, pentamidine, or dapsone with trimethoprim)
- Primary endpoint of respiratory failure (death, mechanical ventilation, PaO2:FiO2 ratio < 75) was reduced with corticosteroids (30% vs. 14%, p=0.001, NNT 6)
- Mortality was reduced with corticosteroids (22% vs. 11%, p=0.01, NNT 9)
- Three pre-specific subgroups based on PaO2:FiO2 ratio (mild > 350, moderate 250-350, severe 75-250) suggested that corticosteroid benefit was demonstrated only in those with moderate-to-severe disease (roughly PaO2 < 70 on room air)
- At 84 days, corticosteroids were associated with a higher risk of herpes reactivation (26% vs. 15%, p=0.04, NNH 9) and oral thrush (53% vs. 41%, p<0.10)
TBL
Adjunctive corticosteroids in patients with AIDS and pneumocystis pneumonia improved mortality and respiratory failure among patients with moderate-to-severe pneumonia, although steroid therapy increased the risk of herpes reactivation and oral thrush.
Bracken I
- Included 330 patients with acute spinal cord trauma with motor or sensory loss presenting within 48 hours from injury (very long enrollment period)
- Randomized patients to low or high dose methylprednisolone x 10 days. Low dose was 100 mg IV loading dose, then 25 mg IV Q6h. High dose was 10 times the dose (1000 mg load, 250 mg IV Q6h)
- Motor, pinprick, and light sensation were tested at 6 weeks, 6 months, and 1 year.
- There were no differences in any neurologic findings at any of the specified time periods
- High dose group had higher incidence of wound infections (9.3% vs. 2.6%, NNH 15)
TBL
High dose methylprednisolone did not improve neurologic outcomes compared to low dose methylprednisolone in patients with spinal cord injury presenting within 48 hours. High dose was associated with more wound infections.
Bracken II
- Included 487 patients with spinal cord injury within 12 hours
- Excluded nerve root or cauda equina only and gunshot wounds
- Randomized patients to methylprednisolone 30 mg/kg bolus (over 15 min, then 45 min pause), then 5.4 mg/kg/hr x 23 hrs, naloxone infusion, or placebo
- Primary endpoint (motor, sensory, and pinprick sensation at 6 weeks and 6 months) was improved in patients receiving methylprednisolone vs. placebo, but only if study drug was initiated within 8 hours of injury (a pre-specified subgroup analysis)
- Naloxone showed no difference in clinical outcomes compared to placebo
- Wound infections (7.1% vs. 3.6%, NNH 29) and GI bleeding (4.5% vs. 3.0%, NNH 67) were more common with methylprednisolone than placebo
- One criticism of the study is that motor/sensory function was quantified using a numerical scale for each muscle group. This endpoint does not describe functional outcome, quality of life, or ability to perform specific activities (such as unassisted walking)
- For comparison, NASCIS I uses a lower dose of methylprednisolone (1000 mg bolus vs. 30 mg/kg [2400 mg for 80 kg patient]) and had a longer window of inclusion (48 hrs vs. 12 hrs)
TBL
In patients with spinal cord injury, methylprednisolone improved motor and sensory function if initiated within 8 hours of initial trauma. Patients receiving methylprednisolone may be at higher risk for wound infections and GI bleeding.
Bracken III
- Included 499 patients with acute spinal cord injury presenting within 6 hours of initial trauma (and able to receive study drug within 8 hrs)
- Excluded patients > 109 kg (concern for fluid overload), gunshot wounds
- Randomized patients to 24-hour methylprednisolone, 48-hour methylprednisolone, or tirilizad (lipid peroxidase inhibitor). All patients received 20-40 mg/kg methylprednisolone bolus
- Methylprednisolone infusion dose was the same as NASCIS II (5.4 mg/kg/hr)
- Primary endpoint (motor function at 6 weeks and 6 months) was equivalent for patients receiving 24 and 48 hours of methylprednisolone
- Pre-specified subgroup showed no difference if methylprednisolone was initiated within 3 hours of injury, but if injury occurred 3-8 hours prior to starting treatment, 48-hours of methylprednisolone demonstrated improvement in 6-week and 6-month motor function versus 24-hours
- Functional improvement measure (FIM), a measure of independence and quality of life, was better at 6-month (but not at 6-weeks) for patients receiving 48-hour methylprednisolone versus 24-hour. Improvement in FIM was driven by greater ability to perform self-care, sphincter tone, and higher mobility. Locomotion did not show an improvement.
- Tirilizad generally performed equally to 24-hour methylprednisolone
48-hour methylprednisolone was associated with more severe sepsis (0.6% vs. 2.6%, p=0.07) and pneumonia (2.6% vs. 5.8%, p=0.02, NNH 31) versus 24-hour
TBL
In patients with acute spinal cord injury presenting within 3 hours of injury, 24-hours of methylprednisolone was equivalent to a 48-hour infusion. For those presenting 3-8 hours post-injury, 48-hours of methylprednisolone improved motor function and quality of life measures compared to a 24-hour infusion. Patients with 48-hour infusions were more likely to develop pneumonia or severe sepsis.
Brochard I
- Included 109 patients who met objective extubation criteria but failed a two-hour T-piece trial
- Randomized within 24 hours of T-piece trail failure to intermittent T-piece trials (titrated periods of 5-120 minutes up to 8 times per day), SIMV (with a downward titration of respiratory rate), or PSV (with a downward titration of pressure support)
- The mean initial T-piece duration was 38 minutes; mean SIMV rate was 9 bpm; and the mean PSV pressure support was 18.1 cm H2O
- Primary endpoint (failure to wean within 21 days of randomization) was lowest with PSV compared to T-piece or SIMV (23% vs. 43% vs. 42%, p=0.05)
- After excluding weaning failures not due to respiratory distress (eg, laryngeal edema), PSV was still associated with the lower rate of weaning failures compared to T-piece or SIMV (8% vs. 33% vs. 39%, p<0.02)
- PSV was associated with a shorter duration of weaning versus pooled T-piece/SIMV (5.7 vs. 9.3 days, p<0.05)
- PSV was associated with a shorter ICU and hospital length of stay; however, the authors’ analysis excluded five outliers (a post-hoc exclusion decision)
- There were no difference between T-piece and SIMV with respect to time to successful extubation, mean weaning duration, or length of stay in the ICU or hospital
- Limitations of the trial include a relatively small patient population from three medical centers and the fact that clinicians were not blinded to treatment assignment.
TBL
Among mechanically ventilated patients who fail an initial two-hour T-piece trial, pressure support ventilation (PSV) was superior in facilitating ventilator weaning compared to T-piece or SIMV.
Brochard II
- Included 85 patients with COPD exacerbation and respiratory distress (respiratory rate above 30 bpm, PaO2 < 45 mmHg, pH < 7.35 on room air) from five European hospitals. Patients requiring immediate intubation were excluded.
- Randomized patients to standard therapy (oxygen up to 5 L/min with bronchodilators) or standard therapy with periods of noninvasive ventilation (NIV) via facemask for at least 6 hours per day
- Primary endpoint (need for endotracheal intubation based on objective criteria) was significantly lower with NIV (26% vs. 74%, p<0.001, NNT 2)
- Hospital mortality was much lower among patients receiving NIV (29% vs. 9%, p=0.02, NNT 5). After adjusting for endotracheal intubation, the benefit was no longer demonstrated (suggesting that avoidance of intubation was the main factor in reducing mortality)
- Patients randomized to NIV demonstrated improved encephalopathy scores, respiratory rate, PaO2, and pH within the first hour of NIV
- Of the 275 screened patients, only 85 were enrolled in the study (69% exclusion rate) which highlights that only a fraction of patients with respiratory distress may be appropriate for NIV
- Duration of mechanical ventilation was incredibly long for both those randomized to NIV and standard therapy (25±17 and 17±21 days), which limits the external validity of this study to real-world practice
- Duration of hospitalization was also very long for NIV and standard therapy (23±17 and 35±33 days, p=0.02), again, limiting the study’s external validity
- Despite the study’s threats to external validity, these data are supported by subsequent trials which have demonstrated a mortality benefit among patients with acute COPD exacerbations who receive NIV
TBL
Among a selected group of patients with acute COPD exacerbation who do not require immediate intubation, noninvasive ventilation reduced the need for endotracheal intubation, length of stay, and mortality.
