JC: The 65 trial. Targeting MAP in sepsis. St Emlyn's at #CCR20

Editors note: This blog is based on the presentation of the 65 trial at the Critical Care Reviews conference in Belfast. We will add more data, and be able to add a more detailed critical appraisal when the full paper is published in JAMA soon. Since we have not seen the full paper yet you should add a little bit of extra scepticism to the findings described below.

What’s your target mean arterial pressure (MAP) for patients with sepsis? Do you believe in the magic number of 65? What about those patients with chronic hypertension? What about when there is no evidence of end organ failure?

Currently a lot of this is down to local practice, but international guidelines continue to endorse the magic number. It’s a little unclear why we put so much faith in this figure; 65 crept into early goal directed work as a target and several retrospective studies have previously suggested a correlation between MAP thresholds in the first 48h of sepsis and survival/organ dysfunction. However, it’s not the most robust evidence base. And it certainly feels as though we are often chasing a MAP with vasopressors just to make the numbers look good, when other parameters seem to be OK. Let us also not forget that vasopressors are not an entirely benign therapy…

The 65 trial^​1​ aims to answer the question of whether we have the MAP target right in septic patients aged over 65. You can watch the live stream of results below.

What kind of trial is this?

Design: Pragmatic, multi-centre, randomised clinical trial. Patients were randomised individually and stratified by site. As this was a trial in the critically ill there was an emergency waiver of consent.

Tell me about the patients

Patients in the 65 trial were recruited from 65 NHS adult, general, critical care units. Adults on critical care who were over 65 years, with vasodilatory hypotension and who had received vasopressors for at least one hour. They aimed to recruit 2600 patients following a fairly significant change to protocol and sample size calculation.

What was the intervention?

Patients in the intervention group had a permissive hypotension strategy (mean arterial pressure target range of 60 – 65 mmHg whilst receiving vasopressors). Those in the control group got usual care (which is variable). A wide range of vasopressors were used in this study.

What about the outcomes?

The primary outcome was 90-day mortality.

Secondary outcomes were mortality at hospital discharge, duration of survival to longest available follow-up, duration of advanced respiratory and renal support, days alive and free of advanced respiratory support and renal support, duration of critical care unit and acute hospital stay.

They also looked at cognitive function, assessed using the Informant Questionnaire on Cognitive Decline in the Elderly (short form) at 90 days and one year, health-related quality of life, assessed using the EuroQol EQ-5D-5L questionnaire, at 90 days and one year, resource use and costs at 90 days and one year, estimated lifetime incremental cost-effectiveness. 

It’s good to see functional assessment of patients in these sort of trials.

What are the main results.

They successfully managed to randomise 2600 patients. 2463 managed to make it through to final analysis. Patients were pretty similar at baseline.

In terms of the primary outcome the mortality in the intervention group was 41% vs. 43.8 % in the control group. This was not statistically signifcant

In terms of secondary outcomes the main findings were a reduction in total vasopressor dose, but did not affect time on ventilation or cognitive function.

One good question is whether they actually achieved a significant difference in blood pressure targets. The authors argue that this is not that important as the idea was to reduce exposure to vasopressors.

Anders Perner delivered a live editorial that challenged the idea that this is a negative trial. The absolute risk reduction is 3% and although not statistically significant he advocates a more open view to the traditional dichotomy of p value traditions^​2​. The confidence interval ranges from -7% to +1% which may be a better way of thinking about these results. He introduced to the concept of ‘dichotomania’ in the interpretation of clinical trials (Ed – I’m thinking of the debate around CRASH-3 here). Defo a phrase I’ll be using in journal club soon.

We also need to think about confidence intervals themselves. Within a confidence interval the probability of different results is not the same. This is elegantly illustrated below by Lars aka @load_dependent. This is really interesting stuff and a good reason to not just rely on blindly interpreting p-values.

Paul Mouncey talked about ‘how’ this trial was delivered in the UK, highlighting the difficulties of achieving recruitment rates in ICU trials. In fact the 65 trial recruited way ahead of target, but there are some really interesting data on when patients are recruited. Fewer patients get recruited at the weekend in many trials and during out of hours periods which may of course influence the results. The advantage in this trial is that recruitment and the intervention was pretty straightforward and so the overnight influence was not really seen. This is important when we consider how to mitigate these chronological effects in future trials.

Clinical Bottom Line

The 65 trial tells us that a lower MAP target in this group of patients is an acceptable strategy. Adopting it will lead to patients being on vasopressors for less time and they will receive a lower dose overall.

This fits with previous work that has suggested that a lower MAP in older patients has a mortality benefit^​3​.

Some are already going to change their practice on the basis of these results, in fact the majority of the audience said they would. A few are going to wait for the inevitable forthcoming meta-analysis.

vb

S