HACA 2002

European hypothermia study for out-of-hospital arrest

Included 275 patients with witnessed out-of-hospital cardiac arrest, VF/VT as an initial rhythm, EMS response time of 5-15 minutes, and no more than 60 minutes from collapse to ROSC
Excluded patients responding to verbal commands post-arrest, known pre-existing coagulopathy
Randomized patients to receive normothermia or hypothermia (32- 340C) with cool air blanket with or without ice packs x 24 hours, and then rewarmed over 8 hours
All patients received midazolam, fentanyl, and pancuronium
Primary endpoint (favorable neurologic outcome) occurred more often with hypothermia (55% vs. 39%, p=O.0009, NNT 6)
Hypothermia demonstrated a reduction in 6-month mortality (55% vs. 41%, p=O.02, NNT 7)
Hypothermia began a median of 105 min after ROSC, and goal temperature was achieved a median of 8 hrs after ROSC.
70% of patients required ice packs for cooling.
Hypothermia was associated with a non-significant increase in complications within the first 7 days following arrest:
- minor/major bleeding (26% vs. 19%)
- pneumonia (37% vs. 29%)
- sepsis (13% vs. 7%).
Note that these complication rates are potentially inflated due to survival bias (dead patients cannot have complications!)

TBL

Therapeutic hypothermia improved 6-month neurologic outcome and mortality among patients with out-of-hospital VT/ VF cardiac arrest. Note that hypothermia may be associated with certain complications, such as increased risk of bleeding and infection.

See the paper here

HEAT

Included 700 ICU patients with fever (>380C within 12 hours prior to enrollment) receiving antibiotics for a known or suspected infection
Randomized patients to acetaminophen 1000 mg IV Q6hr or placebo until cessation of antibiotics, discharge from ICU, resolution of fever (determined based on an algorithm), or 28 days
Primary endpoint (ICU-free days to day 28) was not different between acetaminophen and placebo (23 days vs. 22 days, p=0.07)
There was no difference in 28-day mortality, 90-day mortality, ICU length of stay, hospital length of stay, or liver dysfunction leading to discontinuation of therapy
Among patients who died in the ICU, acetaminophen was associated with a longer length of ICU stay (10.4 days vs. 4.0 days, 13<0.001) and hospital length of stay (13.9 days vs. 7.7 days, 13<0.001).
If this is a true finding (and not a type I error by chance), it would indicate that using acetaminophen to achieve normothermia may delay death.
Patients randomized to acetaminophen received a relatively short course of study drug (median 8 doses or 1.5 days of therapy), which may not have been a sufficient course of therapy to impact the primary endpoint
Open-label acetaminophen was common in the placebo arm (29.4%).
If a beneficial clinical effect does exist for acetaminophen, the effect was diluted due to this open-label usage
Those randomized to acetaminophen demonstrated a statistically significant but mediocre reduction in mean peak temperature (0.2 oc less, p<0.001) and mean daily temperature (0.30C less, p<0.001)

TBL

IV acetaminophen for fever Among ICU patients with fever due to infection, the use of IV acetaminophen produces a small reduction in temperature but does not have any impact on any clinical outcome.

See the paper here

HYPRESS

Hydrocortisone to prevent the progression from severe sepsis to septic shock

Included 353 patients with severe sepsis (defined as a suspected infection, at least two SIRS criteria, and evidence of organ dysfunction present not longer than 48 hours) who had not progressed to septic shock in 34 German intermediate or intensive care units
Randomized to either hydrocortisone (200 mg/day continuous infusion) or placebo for 5 days followed by a dose taper until day 11
Primary endpoint (14-day progression to septic shock) was not different between hydrocortisone and placebo (21.2% vs. 22.9%, p=O.7, Cl difference -10.7% to +7.2%)
Septic shock was defined as sepsis- induced hypotension (MAP < 65 mmHg or systolic BP < 90 mmHg) despite adequate volume (500 to 1000 mL of crystalloids) for longer than four hours requiring vasopressor therapy
28-day mortality was not different between hydrocortisone and placebo (8.8% vs. 8.2%, p=O.86) (Note: a mortality rate this low is uncommon for severe sepsis studies — a criticism of the study is that patients were at a relatively low risk of mortality and that hydrocortisone may have had efficacy in a “sicker” patient population)
There were no differences observed regarding time to onset of septic shock, any subgroup analyses of the primary endpoint, or in any mortality, length of stay, or organ failure endpoints
The only statistically significant adverse event of hydrocortisone was a higher risk of hyperglycemia (BG >150 mg/dL, 81.5% vs. 90.9%, p=0.009, NNH 11), although the degree of hyperglycemia was fairly mild (maximum median glucose during study medication administration 157 vs. 170 mg/dL, p=0.006).
Secondary infections, muscle weakness, impaired wound healing, and all other adverse events were not significantly different.
- Regarding the definition of sepsis, “adequate fluid” was extremely conservative. In the United States, many would consider 30 mL/kg of crystalloid an adequate fluid amount, which usually exceeds 2000 mL within the first three hours for most adult patients. This may have caused some patients in the study to be classified as having septic shock but would not have met the primary endpoint had additional fluids been given.
- This criticism is particularly relevant given the low mortality rate observed in the study, further indicating that this study may not have included a severe sepsis patient population “sick enough” to appreciably benefit from hydrocortisone.
The study was powered assuming a 40% rate in the primary endpoint, but only a rate of about 22% was observed. Although the study did rule out a 15% absolute risk difference in the primary endpoint, the 95% confidence interval still included a significant treatment effect that would be clinically meaningful to most clinicians.