Role of nutrition support in adult cardiac surgery: a consensus statement from an International Multidisciplinary Expert Group on Nutrition in Cardiac Surgery

Several scores or assessment tools have been introduced to enable the quantification of nutrition risk. These tools were neither developed for nor validated in critically ill patients [16]. Therefore, the measurement of a patient’s current nutritional status predominantly identifies those who have already reached an undernourished general state. To foresee an aggravation in the nutritional status, an assessment of nutritional risk must also identify patients at a pre-critical level of malnutrition, who will benefit (and not be harmed) by nutrition support. The Malnutrition Universal Screening Tool (MUST), the Mini Nutritional Assessment (MNA), the Short Nutritional Assessment Questionnaire (SNAQ), the Malnutrition Screening Tool (MST), and the Subjective Global Assessment (SGA) [16] are well-established assessment tools used to evaluate nutrition status in clinical practice.

Lomivorotov and colleagues demonstrated that in patients undergoing cardiac surgery, detection of malnutrition is associated with prolonged ICU stay (>2 days) and both MUST and MNA have independent predictive accuracy with regard to postoperative complications [14]. In a subsequent study, the authors further detected that the SNAQ and MUST have comparable accuracy in detecting malnutrition. Nevertheless, the authors acknowledge that whether preoperative nutritional therapy would improve the outcome in malnourished patients needs to be studied [13]. In fact, no validated pre-surgical scoring system is available to identify patients at an early enough time point for preoperative nutritional replenishment.

As the aforementioned tools consider all critically ill patients at high risk for malnutrition, the Nutrition Risk in the Critically ill (NUTRIC) score was developed to define nutrition risk in the critically ill ICU patients [17‐19]. The observation that not all ICU patients will respond the same to nutritional interventions was the critical driver for the development of the NUTRIC score [17, 20]. Yet, the NUTRIC score has not yet been validated in cardiac surgery patients.

Considering the critically ill, the Nutritional Risk Score (NRS)-2002 must be interpreted cautiously as an Acute Physiology and Chronic Health Evaluation (APACHE) score >10 in cardiac surgery patients indicates that a patient is already at high risk for malnutrition. Patients with an NRS-2002 score >3 are defined as at “risk for malnutrition”, and those with an NRS-2002 score ≥5 as “high risk for malnutrition”, which may be too broad and nonspecific for directing aggressive nutrition support. In contrast, the NUTRIC score was demonstrated in five prospective, but non-randomized studies, as a sensitive measure to identify the nutritional risk and to discriminate between patients with high risk for malnutrition and those likely to benefit from aggressive nutrition therapy [21]. Additionally, nutritional assessment by the NUTRIC score identified patients with a low nutrition risk in whom additional nutritional therapy may have neutral or negative effects. The importance of this is evident in the EDEN trial, which compared the effects of early trophic feeds to early full enteral nutrition (EN) [22]. No difference was shown between trophic vs full feeds in terms of ventilator-free days, infections, or 60-day mortality rate. However, this study was performed in a rather young, initially well-nourished, and nutritional-insensitive patient population with a low NUTRIC score [22]. Casaer et al. [23] compared those with inadequate intake (caloric restriction) to early supplemental parental nutrition (PN) with prior glucose loading. Early initiation of PN to supplement insufficient EN was associated with a higher incidence of infections, delayed recovery, and higher health care costs compared with late initiation of PN. Again, this may have been the wrong target population as the majority of study patients had a short ICU stay, suggestive of a low nutrition risk. Furthermore, tight glycemic control and high-dose glucose loading may have negatively affected the results. It is unlikely that patients having cardiac surgery and at low nutrition risk would benefit from aggressive, early PN, therefore enrolment of these patients in a large clinical trial would be nonsensical. An adequate risk assessment is mandatory to first identify high-risk patients for cardiac surgery before scientifically studying nutrition support implementation.

Ultrasound of the quadriceps muscle is an easy-to-use and readily available tool to measure muscle mass and determine changes in muscle and fat tissue [24‐26]. Computed tomography (CT) is also a well-established body composition analysis tool, though more expensive, not risk-free, and difficult to access [25, 26]. Recently, the validity of bioelectrical impedance spectroscopy (BIS) that calculates fat-free mass from measurements of body water has shown promising results in determining nutritional reserve in cardiac patients [27]. In that study a preoperative low bioelectrical impedance phase was associated with malnutrition and increased risk of adverse postoperative events [27]. Yet, the high fluid intake may significantly influence the reliability of BIS. Considering clinical practicability, both CT scan and BIS may open the opportunity for preoperative and postoperative nutrition assessment in cardiac surgery patients. In summary, biochemical inflammatory markers are unlikely to be of use. Ultrasound, CT, and BIS may represent promising future tools allowing quantitative assessment of body composition and, therefore, nutritional status.

Considering international guidelines, there is a strong consensus and most experts will report that EN is always preferred to PN. Following cardiac surgery, critically ill patients are frequently on vasopressor treatment because of an inflammatory response syndrome, vasoplegia, and/or postoperative low output syndrome due to myocardial stunning. The need for vasopressor support further results in marked changes in energy expenditure and frequent intolerance to oral feeding, leading to significant energy/protein deficits and increased risk of malnutrition. Although proven safe [28], EN is often thought to be contraindicated and considered as harmful especially in hemodynamically unstable patients on large doses of inotropes and/or vasopressors, leading to a widespread use of PN in clinical practice. Berger et al. were among the first to provide evidence from a small cohort on the feasibility and safety of early nutrition support in patients after cardiac surgery. Using a well-established paracetamol absorption test, the investigators demonstrated that hypocaloric EN was feasible in such patients with altered hemodynamic status [28]. Kahlid et al. demonstrated in a large scale, multi-center, observational study that mechanically ventilated, vasopressor-dependent patients (mixed population) had a significant survival advantage when EN feeding was started within the first 48 hours after ICU admission, compared to those receiving EN feeding later than 48 hours [29]. Furthermore, in a subgroup analysis, they demonstrated that the sickest patients (on multiple vasopressors compared to those on one vasopressor only) had a more pronounced survival advantage. In addition, nutrition support was demonstrated to be feasible in patients with extracorporeal life support systems (ECLS). All patients on ECLS (in venovenous or venoarterial mode) were fed using existing protocols that emphasize early EN in preference over PN or delayed EN [30]. Notably, the use of paralysis and sedation did not affect feeding tolerance regarding the time to reach goal rate, incidence of intolerance in the first 5 days, or time until first observed bowel motion. In contrast, actual guidelines recommend withholding EN nutrition in hemodynamically unstable patients, though this is primarily based on preclinical and observational studies [31]. The rationale behind this is that the hemodynamic failure may compromise the splanchnic blood flow, causing intestinal ischemia [32], which is associated with high mortality [33, 34].

Given the current evidence, vasopressor use per se is not a contraindication to EN. In hemodynamically unstable critically ill or cardiac surgery patients, there is at least some evidence that early EN is absorbed and metabolized without any harmful effect on systemic measurements of oxygenation and perfusion and supportive evidence from a large-scale observational study that this translates into an advantage in terms of mortality [28]. Therefore, early EN may be beneficial in patients after initial resuscitation from critical organ failure. Future well-designed studies are needed for an adequate assessment of this important question.