Screen Shot 2018-08-06 at 09.58.23.png

Late rescue steroids for ARDS

  • Included 180 patients with persistent ARDS for 7—28 days
  • Main reasons for exclusion included prior corticosteroid use and history of COPD
  • Randomized patients to methylprednisolone or placebo Methylprednisolone dosing schedule:
    • 2 mg/kg (ideal body weight), then 0.5
      mg/kg Q6hr x 14 days, then 0.5
      mg/kg Q12hr x 7 days, and then a 2-
      4 day taper depending on clinical
      status *Note that this taper duration is
      quicker than the Meduri protocols
  • Primary endpoint (60-day mortality) was not different between methylprednisolone and placebo (29.2% vs. 28.6%, p=1.0)
  • Importantly, patients randomized after 14 days of persistent ARDS had higher mortality rates with methylprednisolone (35% vs. 8%, p=0.02, NNH 4). This effect was not seen in those randomized 7-13 days after onset of ARDS (27% vs. 36%, p=0.26)
  • Methylprednisolone improved 28—day vent-free days (11.2 vs. 6.8 days, p<0.001) and ICU-free days (8.9 vs. 6.2 days, p=0.02)
  • Methylprednisolone was associated with a higher incidence of myopathy or neuropathy (9 vs. 0 events, p=0.001)
  • In contrast to previous, high-dose studies, methylprednisolone was not associated with a higher rate of secondary infections
  • Main criticisms of this study are the rapid final taper duration, the small screening enrollment ratio (5% enrolled), and the 7-year study duration over a period of significant advances in ICU and ARDS management (low tidal volume ventilation, tighter glucose control, protocolized sedation with daily awakening)


Methylprednisolone in persistent ARDS (7-28 days after onset) did not improve 60-day mortality, but did improve ventilator-free and ICU-free days. Patients randomized after 14 days of ARDS had higher mortality rates with methylprednisolone compared to placebo.

See the paper here

Lau 2000

Screen Shot 2018-08-06 at 10.03.11.png

PPI drip for upper GI bleeding ulcer

  • Included 240 patients with upper GI bleed and successful endoscopic treatment of an ulcer with epinephrine and thermocoagulation
  • Randomized patients to placebo or omeprazole (80 mg bolus, then 8 mg/hr x 72 hours) following endoscopy.
  • All patients received omeprazole 20 mg PO daily x 8 weeks following the 72 hour study period
  • Primary endpoint (30-day recurrent bleeding) was less common with omeprazole than placebo (6.7% vs. 22.5%, NNT 6, P<0.001).
  • Recurrent bleeding was most common during the first 72 hours.
  • Omeprazole was associated with fewer units of blood transfused after endoscopy (1.7 vs. 2.4 units, p=0.03)
  • There was no difference in 30-day mortality between omeprazole and placebo (4.2% vs. 10%, P=0-13)
  • No side effect related to the infusion was reported in either group, although the authors do not describe a robust evaluation of adverse events;
  • The inclusion criteria in this study are important to clinical practice. Study drug was only started after successful endoscopic treatment of an upper GI ulcer
  • Although a placebo comparator seems strange, there was significant controversy at the time regarding the role of acid suppression therapy following successful endoscopy


In patients with an upper GI bleeding ulcer, a high-dose omeprazole infusion reduced recurrent bleeding compared to placebo following successful endoscopic treatment.

See the paper here

Leuven I

Screen Shot 2018-08-06 at 10.09.43

Intensive insulin therapy in the SICU

  • Included 1548 patients admitted to a surgical ICU (primarily cardiac) receiving mechanical ventilation
  • Randomized patients to intensive insulin (BG goal 80-110 mg/dL) or conventional insulin (BG < 215 mg/dL)
  • Study protocol was conducted by a team that was not involved in the daily management of each patient (limits external validity)
  • Primary endpoint (ICU mortality) was higher with conventional therapy (8% VS_ 45%, p<0.04, NNT 29).
  • Hospital mortality was also higher (10.9% vs. 7.2%, p=0.01, NNT 27)
  • Average morning blood glucose was lower (as expected) with intensive control (153 vs. 103 mg/dL)
  • Intensive therapy was associated with a reduction in renal impairment (both 50 > 25 mg/dL and need for renal replacement therapy), bloodstream infections, and the number of RBC transfusions per patient
  • There was no difference in ICU length of stay or duration of mechanical ventilation Hypoglycemia (BG < 40 mg/dL) occurred more frequently with intensive insulin (5% vs. 0.8%, NNH 24).
  • Note that subsequent studies have suggested hypoglycemic rates that are much higher.


In surgical ICU patients (primarily cardiac), intensive insulin therapy reduced ICU mortality, renal impairment, and bloodstream infections. The rate of severe hypoglycemia was higher with intensive insulin.

See the paper here

Leuven II

Screen Shot 2018-08-06 at 10.17.18

Intensive insulin therapy in the MICU

  • Included 1200 medical ICU patients with an anticipated ICU stay of > 3 days
  • Excluded all surgical patients and those able to receive oral nutrition
  • Randomized patients to intensive insulin (BG goal 80-110 mg/dL) or conventional insulin (BG < 215 mg/dL)
  • Primary endpoint (hospital mortality) was not different between intensive or conventional therapy (37.3% vs. 40.0%, p=0.33). ICU, 28-day, and 90-day mortality were also not different between groups
  • Intensive insulin reduced newly acquired kidney injury (doubling of SCr or peak SCr > 2.5_mg/dL, 8.9% vs. 5.9%, p=0-04, NNT 33), earlier weaning from mechanical ventilation (adjusted HR 1.21, p=0.03), and earlier ICU discharge (adjusted HR 1.15, p=0.04)
  • A reduction in hospital mortality was seen with intensive insulin among patients who stayed in the ICU for 3 or more days (43% vs. 52.5%, p=0.009, NNT 11), although these patients were not prospectively identified
  • Hypoglyemia (BG < 40 mg/dL) was more common with intensive insulin (18.7% vs. 3.1%, p<0.001, NNH 6)


In patients in the medical ICU, intensive insulin therapy did not improve mortality. While intensive therapy may have had a positive effect on duration of mechanical ventilation and length of ICU stay, it was associated with a higher incidence of severe hypoglycemia.

See the paper here


Powered by WordPress.com.

Up ↑

%d bloggers like this: