Let Them Eat Cake – Sanctioned Hunger on ICU

Permissive Underfeeding or Standard Enteral Feeding in Critically Ill Adults

By Dr Tamara Tajsic and Emma Service

Nutritional support for critically ill patients is a persistently exasperating problem. The various debates about timing, route, and composition seem to never settle on a conclusion. There have been four recent important trials on ICU patients, see Roger Harris’ podcast below, but a clear signal has yet to emerge and surprisingly we find ourselves in the slightly perplexing situation that near-starving people for six days is thought to be harmless. How have we got here?

One of the problems is that, in common with nearly all of the studies of critically ill patients which we discuss, we collect every patient admitted to ICU into one group as if they as subject to the same pathology and physiology. How can the metabolic and nutritional demands of an 18 year old poly-trauma patient be the same as a 62 year obese diabetic with sepsis? Intuitively they can’t be, but the studies largely treat them equally. Happily there are some good aspects to this, we are now moving towards a concept of ‘nutritional support’ rather than feeding and we are starting to recognise the role of nutrition in modulating disease and, perhaps, attenuating the metabolic response to stress, to prevent oxidative cellular injury, and to enhance the immune response.

Studies looking at different doses of enteral feeds have not yet provided consistent evidence but certainly raised a question of what ‘optimal caloric targets’ truly are. Standard caloric targets are based on the assumption that attenuating malnutrition and protein catabolism, which are associated with increased morbidity and mortality, will, in turn, improve outcomes. However, over-feeding does not reduce mortality, but may risk complications whereas underfeeding is perhaps beneficial when done deliberately but we actually always do it by accident.

The study we will discuss this week sought to tease out the importance of protein intake vs calorie intake (recent meta-analysis). It is an unblinded, randomized controlled trial involving 7 centres in Saudi Arabia and Canada over a 5 year period, which compared permissive underfeeding (40-60% of caloric requirements) and standard feeding regime (70-100% of caloric requirements) for up to 14 days. Critically, the study is interesting because protein requirements (1.2-1.5 g/kg/day) did not differ between the two groups, rather it was the amount of non-protein calories which varied. They did not detect any differences across a range of primary, secondary or tertiary outcomes which lead the authors to conclude that standard feeding goals do not improve outcomes and higher caloric intake does not attenuate protein catabolism.

How disappointing but has it told us anything useful?

This study brings into sharp focus the persistent conundrum within research into nutrition in critically ill patients, why can we not establish the amount of calories and protein which benefit patients?

The usual questions which we need to think about before discussing the study in detail are:

Is the study generalisable to our patients?

Was the methodology robust?

Should we change our practice based on this methodology?

To think more specifically:

This study is part of a larger group of studies which broadly tell us that over feeding (just calories?) is bad and that under feeding may have some advantages but it is a very muddy area. We already underfeed though, by accident and we should try to catch up, but how do we do this without avoiding daily overfeeding?

What are the easily modifiable factors which influence protein and calorie intake in critically ill patients?

Does fluid balance influence ICU stay?

This is an area where, typically, clinicians have a relatively poor baseline level of knowledge. Happily our excellent specialist dietician, Emma Service, has identified some interesting aspects to this study and some questions which we also will need to think about when we discuss the study.

Current protein intake recommendation currently 1.2-1.5 g/kg body weight/day (ESPEN guidelines). However, a recent systematic review suggested that a higher protein content of 2.0-2.5 g/kg of body weight/day may be actually be optimal whilst recognising that well designed trials are needed to identify answers to these questions. To add to the confusion, in the context of traumatic brain injury, a single centre study from Tennessee suggested 2g/kg/day. This serves to illustrate the persistent debate about protein vs calorie requirements which has been developed into the concept of permissive underfeeding (<25% of target)  – why would you do this?

How good is the evidence for optimal calories – usually aiming to provide >80%?

Equally, what about the evidence for protein requirements of 1.2-2g/kg/day?

The study this week only achieved average of ~0.7g/kg/day protein (68-69% of requirements) which is very low – what would be the consequences?

If protein was fed to 1.2g/kg would this have changed the outcomes?

How can we get better at working out what patients need, indirect calorimetry is perhaps the gold standard for measuring energy expenditure while predictive equations are used more commonly in practice on NCCU?

Lots of comment in the ICM, surgical, and nutrition literature:

Annals of Translational Medicine 1 – excellent – essential reading.

Annals of Translational Medicine 2

Critical Care Nutrition

Journal of Enteral and Parenteral Nutrition

Journal of Thoracic Disease 1

Journal of Thoracic Disease 2 – excellent.

JICS Journal Club – essential reading.

Reflecting the lack of good quality data there is an abundance of guidelines, two of the best are below.

Scottish ICS


PlumCCM Blog.

Great podcast by Roger Harris on ICN – essential material.