Hypothermia in sepsis. The CASS trial.



Induced hypothermia in patients with septic shock and respiratory failure (CASS): a randomised, controlled, open-label trial, by Theis Skovsgaard Itenov et al. [1]

I know, past weeks we faced some lag in our state-of-the-art servers due to unexpected end-of-the-year laziness. But now we’re back! We’re comin’ in hot! And talking about hot, how about the out of the oven CASS (Induced hypothermia in patients with septic shock and respiratory failure: a randomised, controlled,open-label trial) trial.

To summarize: hypothermia is a treatment looking for a disease. More about that later.

Sepsis and especially septic shock (no matter what definition you use) is a recipe for multi-organ dysfunction and poor prognosis. In the past few years, we saw a lot of failed attempts trying to find something that could decrease sepsis mortality. From immunoglobulin to recombinant activated protein C, and of course Marik’s elixir (hydrocortisone, vitamin C, and thiamine). Well, the later is an unproven intervention, while the first didn’t work, and the second was withdrawal from the market.

There is animal data showing hypothermia decreases sepsis mortality, maybe it protects the organs against further damage. It’s interesting that hypothermic (non-induced) septic patients have a poorer prognosis than non-hypothermic ones. Also, the HEAT [2] trial showed that fever control has no impact in ICU-free days. So, what to do regarding temperature control? We know that hypothermia (non-induced) is bad and fever control doesn’t have any influence on outcomes, yet good data was still missing on induced hypothermia. I won’t mention all the other hypothermia trials here. Maybe in the near future, we write some thoughts about them. But I’ll give you a spoiler: go for normothermia.

Therefore, our friends from the North designed a multicenter, randomized, controlled trial. From 2011 to 2016, they recruited ICU patients with severe sepsis or septic shock (remember the new sepsis definitions only came to life two years ago), ≥50y/o (including older patients would increase the study power) who were mechanically ventilated. The exclusion criteria were:  bleeding disorders, recent open surgery, pregnancy and others. It’s interesting to see that everybody is scared about hypothermia induced bleeding disorders, yet nobody really knows how it works. Patients were randomized on a 1:1 ratio to either Intervention or Control. The variable block-sizes randomization was stratified by age, APACHE II, and site center. I love allocation concealment. Unfortunately, the healthcare professionals were not blinded (it’s an impossible blinding), but the assessors and investigators were.

The intervention

All patients were treated accordingly to the SSC guidelines (is that a good thing?). The Intervention group was cooled to a target of 32-34°C within 2h of randomization. They used either a external pad-based hypothermia device or a IV catheter to induce hypothermia. The target temperature was maintained for 24h. Thereafter, the patients were rewarmed at a rate of 0.5°C/h until 37°C. The target for the next 48h was 36-38°C (normothermia). In the control group, no physical or pharmacological thermal interventions were permitted during the initial 24h. Antipyretic drugs were allowed after 24h.

The primary outcome was 30d mortality. Based on a mortality rate of 40-56%, with 80% power and 5% alpha, a sample size of 560 patients was needed. Three safety interim analyses were made. The investigators were particularly worried about hypothermia induced coagulopathy. An O’Brien and Fleming approach was used (a conservative approach in the early stages, which I like), and there was no change in the final alpha (since it was a safety interim analysis only). A futility analysis was planned at the third interim analysis. Although the industry donated all the equipment, they had no role in study design, data collection, data interpretation, or writing of the report.

What they found

From 2011 to 2016, 5695 patients were screened, and after 436 patients were included, the trial was stopped for futility. Fun fact: 1383 patients were “ineligible” because the primary investigator was not on duty! Want it or not is a huge source of bias. C’mon! This was completely unnecessary.

Both groups were fairly balanced, except for small unbalances, for now: the Intervention group had more patients with RASS≤4 (80% vs. 60%) and renal failure (32% vs. 18%). In most of the patients (93%) an external cooling method was used. The median time to target temperature achievement was 3.2h (graph below). It’s interesting that even without any temperature control method, patients in the Control group had a median temperature <38°C.

Itenov TS, Johansen ME, Bestle M et al. Induced hypothermia in patients with septic shock and respiratory failure (CASS): a randomised, controlled, open-label trial The Lancet Respiratory Medicine. 2018.

There was no difference in the primary outcome, with mortality of 44.2% in the Intervention groups vs. 35.8% in the Control (difference 8·4% [95% CI –0·8 to 17·6]; relative risk 1·2 [1·0–1·6]; p=0·07). The Kaplan-Meier is shown below:

Itenov TS, Johansen ME, Bestle M et al. Induced hypothermia in patients with septic shock and respiratory failure (CASS): a randomised, controlled, open-label trial The Lancet Respiratory Medicine. 2018.

The secondary outcomes showed a difference in days alive without RRT and vasoactive drugs in the Control group. Also, the Control group had lower C-reactive protein levels. Two interesting points here. First: hypothermia trials showed a reduction in vasoactive drugs in the hypothermia group (higher MAP), which went the other way around here. Second: while one might think about the protective effect of hypothermia acting on inflammatory pathways, the CRP (if we consider it a useful inflammatory marker) showed otherwise. Note that CRP was already hight at baseline in the Intervention group.

One big concern was about the coagulation function. There was no difference in INR between groups, while the Intervention group had lower platelets (which was already lower than Control at baseline). There is lab data about platelet dysfunction in extreme temperatures (~0°C), but it’s hard to tell if mild hypothermia (the case in this study) would have an impact in coagulation, despite low platelet levels in this group. It would have been interesting to see data about thromboelastography, but unfortunately, the investigators didn’t go that way.

To wrap it up

At the end of the day, is interesting to see a good trial about hypothermia in sepsis. I’m not a big fan of hypothermia myself and that bullshit about the investigator be on duty for recruitment drove me crazy! But the data doesn’t lie. There’s something there, and although one might think this was a negative trial, I’m not so sure. Throughout all the analyses, we can see a trend favoring the Control group. And if you look close enough, you can see that given enough sample size, hypothermia would be shown to be harmful. Ok, I’m biased because I do not love hypothermia as I do ketamine. But if anyone saw things from other light, please, feel free to create your own blog and write something there. Case closed.



1- Itenov TS, Johansen ME, Bestle M et al. Induced hypothermia in patients with septic shock and respiratory failure (CASS): a randomised, controlled, open-label trial. The Lancet Respiratory Medicine. 2018;

2- Young P, Saxena M, Bellomo R et al. Acetaminophen for Fever in Critically Ill Patients with Suspected Infection N Engl J Med. 2015; 373(23):2215-2224.



Photo Credit

What's Your Reaction?

Confused Confused
Damn Damn
Geeky Geeky
Like Like
Scary Scary
Win Win
Angry Angry
Happy Happy

Hypothermia in sepsis. The CASS trial.

log in

Become a part of our community!

reset password

Back to
log in