Parenteral nutrition in critical care

Kirsten Macdonald MBChB BSc (Hons) MCEM EDIC

Key points Metabolic changes in critical reserved for and initiated only after the first 7
illness days of hospitalization.5 A recent study has
Malnutrition is associated
shown faster recovery and fewer complications
with worse outcomes in During critical illness, profound metabolic
critical care patients. in patients receiving late PN (day 8) vs early PN
changes occur: protein catabolism increases,
(day 3).6 However, the study does have limita-
Parenteral nutrition (PN) basal metabolic rate (BMR) increases by up to
tions and further research is needed in this area.
should be given through a 40%, and a state of relative insulin resistance
Feeding can be given via three principal
dedicated ( preferably occurs. Malnutrition in critically ill patients is
central) i.v. line. routes: oral, enteral, or parenteral. Oral feeding is
associated with impaired immune function,
generally not possible in the critically ill patient.
Protein content is calculated muscle weakness, and results in increased
Enteral feeding, either through an oral or a nasal
in grams of nitrogen and ventilator-dependent days and intensive care
tube, or directly into the gastrointestinal (GI) tract
may include glutamine length of stay. Provision of nutrition in critically
supplementation in via a gastrostomy or jejunostomy is the preferred
ill patients decreases morbidity and mortality.
appropriate patients. method in critical care. Advantages of feeding
Indeed, the degree of cumulative energy deficit
via the enteral route as opposed to parenteral
Close monitoring of blood has been shown to be directly proportional to
include cost, maintenance of GI tract structure
glucose levels is required due the number of complications they experience.
to the risk of hyperglycaemia and function, and a decreased risk of stress ulcer-
when carbohydrate is given ation. Enteral feeding is also associated with a
parenterally. lesser incidence of hyperglycaemia, enhanced gut
General nutritional guidelines immune function, and possibly a lower risk of
PN is usually commenced at
half of calculated daily The National Institute for Health and Clinical infections. Intolerance of enteral feeding is
requirements due to the risk Excellence (NICE) guidelines recommend nu- common in critically ill patients due to impaired
of refeeding syndrome. tritional support for people who are malnour- GI motility. Patients at high risk of impaired GI
ished based on BMI (body weight accounting motility include those with head injury, burns,
Kirsten Macdonald MBChB BSc (Hons)
MCEM EDIC for height) and amount of unintentional weight sepsis, and multi-system trauma. Additionally,
Specialist Registrar in Critical Care and loss. It is also recommended in those who have drugs commonly used in critical care (e.g.
Emergency Medicine, Sheffield Teaching eaten little or nothing for more than 5 days or opioids, sedatives, adrenoceptor agonists, and
Hospitals NHS Foundation Trust
are likely to eat little or nothing for the next 5 proton-pump inhibitors and anticholinergics) can
Sheffield, UK
days and people who have poor absorptive cap- impair gastric motility. This can lead to compli-
Kevin Page MB BS MEd FRCPath
acity, high nutrient losses, or increased nutri- cations including regurgitation causing ventilator-
Associate Specialist in Chemical Pathology
Nutrition Support Team tional needs from causes such as catabolism.1 associated pneumonia and bacterial translocation
Sheffield Teaching Hospitals NHS Most patients admitted to critical care would from the gut causing sepsis.7 Success of enteral
Foundation Trust, Sheffield, UK
fall into one or all of the above categories. feeding can be maximized by the use of proki-
Lisa Brown BPharm (Hons) MSc Clin
The ESPEN (European Society of Parenteral netic agents and by post-pyloric feeding.
Pharmacy
and Enteral Nutrition) and Canadian guidelines Prokinetic agents encourage gastric motility by
Senior Critical Care Pharmacist
Sheffield Teaching Hospitals NHS state that all patients who are not expected to increasing both luminal transport and the force of
Foundation Trust, Sheffield, UK be on normal nutrition within 3 days should contraction. The most commonly used prokinetic
Daniele Bryden FRCA FFICM LLB receive nutritional support within 24 –48 h of drugs in critical care patients are erythromycin
(Hons) MML admission.2 – 4 ESPEN recommend parenteral 250 mg b.d. i.v. (motilin agonist) and metoclopra-
Consultant in Intensive Care Medicine and
nutrition (PN) is commenced within 24–48 h if mide 10 mg q.d.s. (dopamine D2 receptor antag-
Anaesthesia, Sheffield Teaching Hospitals
NHS Foundation Trust enteral nutrition (EN) is contraindicated or onist). Combination therapy of the two agents
Sheffield, UK cannot be tolerated. ASPEN (American Society may be beneficial with lower rates of tachyphyl-
Critical Care Department
Northern General Hospital of Parenteral and Enteral Nutrition) guidelines axis reported compared with monotherapy.
Herries Road recommend if early EN is not feasible or avail- Successful enteral feeding has been defined as
Sheffield S5 7AU, UK
Tel: þ44 (0)114 243 4343 able in the first 7 days no nutrition support maintenance feeding of 40 ml h21 with 4
E-mail: [email protected] therapy should be provided, that PN should be hourly NG aspirates totalling ,250 ml. It is
Fax: +44 (0)1142769342
(for correspondence)
doi:10.1093/bjaceaccp/mks056 Advance Access publication 21 November, 2012
1 Continuing Education in Anaesthesia, Critical Care & Pain | Volume 13 Number 1 2013
& The Author [2012]. Published by Oxford University Press on behalf of the British Journal of Anaesthesia.
All rights reserved. For Permissions, please email: [email protected]
Parenteral nutrition in critical care

however sometimes impossible to attain adequate nutrition enterally. through a central venous catheter (CVC) via a dedicated lumen, or
In this situation, PN can be used to supplement or entirely meet nutri- through a peripherally inserted central catheter (PICC line). If it is
tional requirements.2,4,5 anticipated that this will be administered for more than 1 month, a
tunnelled central venous line is usually inserted.
Line infection is a significant risk with administration of PN.
Assessment of nutrition This is minimized by:

Assessment of nutritional state is the first step in deciding the nutri- (i) Strict asepsis when handling the line and connecting/discon-
tional support required. In most clinical situations, the BMI is a necting feeds.
useful indicator. With critically ill patients, a number of factors limit (ii) The use of a dedicated lumen on a CVC or a PICC line for
the use of BMI, including practical difficulties in weighing patients administration of PN.
and the presence of oedema. Individual patients’ resting metabolic (iii) Use of antimicrobial-coated lines.
rate (RMR), possibly the gold standard method of assessment of
Peripheral venous lines can be used to deliver low osmolality
energy requirements in an individual patient, can be assessed using
feeds. Peripheral PN that can provide full nutritional requirements
indirect calorimetry. This however requires specialist equipment and
tends to be of larger volume and greater fat content and limited by
is impractical for routine use on the intensive care unit (ICU). As a
restrictions in electrolyte content to reduce osmolality, which
result, a number of different equations and methods for calculating
overall reduces their routine application. Although there is some
estimated energy requirements have been developed. One of the most
variation depending on the PN chosen, a typical comparison might
commonly used equations is the Schofield equation. This is not spe-
be between a volume of 1.5 litre centrally vs 2.5 litre peripherally
cific for critical care patients but estimates BMR (based on age, sex,
to give equivalent calories, nitrogen, and electrolytes.
and weight) with subsequent percentage adjustments for thermogen-
esis, activity, and additional stress factors (e.g. infection and trauma)
to produce an estimated kilocalories per day requirement. (BMR and Composition of PN
RMR are technically different, but in practical terms, there is little
difference between the two.) The composition of PN feeds can be varied according to the
A problem with using calculations in the critically ill is the requirements of individual patients. Some PN manufacturers offer
addition of activity, stress, and feeding factors to a BMR can lead a broad range of ‘off-the-shelf’ feeds that only require addition of
to an apparent indication for sicker patients to receive very large vitamins, trace elements, and additional electrolytes to patients’ in-
quantities of energy. Although metabolic requirements of these dividual requirements. Some hospitals have an Aseptic Dispensing
patients are often high, they diminish quickly and since net catab- Unit that can compound feeds from basic components that are tai-
olism cannot be stopped, they also receive considerable nutrient lored even more closely to a patient’s needs. All additions to a PN
supply from the breakdown of their own tissues. This puts critical- regimen must be carried out in a sterile environment. The precise
ly ill patients at potential risk of overfeeding, which results in composition of PN varies in its carbohydrate, lipid, and protein
worse outcomes and should be avoided. Study outcomes in inten- formulation ratios between different preparations.
sive care patients showed that survival was best among patients re-
ceiving 33 –66% of estimated nutritional needs compared with
those who received fewer than 33, or .66% of them.8 NICE and Carbohydrate
ESPEN guidelines recommend calculation of nutritional require-
Carbohydrate is supplied as glucose. Tissues such as red blood
ments based on kilocalories per kilogram per day for calories and
cells, immune cells, and renal medullary cells which have no mito-
gram per kilogram per day for nitrogen in short-term feeding, and
chondria are dependent on glucose supply for their energy through
therefore, this may be the most appropriate method in critically ill
ATP production. Brain metabolism is also partially dependent on
patients.1,2 NICE recommends 25 –35 kcal kg21 day21 and
plasma glucose concentration: it can use lactate or ketones for
ESPEN recommends 20 –25 kcal kg21 day21 in the initial phase
energy supply when blood glucose decreases gradually. Brain me-
of critical illness and 25– 30 kcal kg21 day21 in the recovery/ana-
tabolism accounts for the majority of blood glucose oxidation in the
bolic phase. It is difficult to be precise about an individual figure
body. In the absence of an exogenous supply of glucose, the body
as often a major determinant of exact calorie content is determined
can call upon hepatic glycogen stores and can synthesize glucose
by the constituent preparations/manufacturers’ formulae used,
from lactate, lipids, and amino acids through gluconeogenesis.
hence the use of a range rather than an exact figure.
The catabolic and relatively insulin-resistant state encountered
in critical illness is thought to be a potential mechanism whereby
glucose can be diverted away from less vital areas to help in
PN administration wound healing (injured tissue has been shown to lack this insulin
PN is generally supplied as an all-in-one complete balanced paren- resistance) while providing substrate for gluconeogenesis through
teral formulation and is usually administered continuously, ideally muscle breakdown. The aim of feeding is to attempt to minimize

2 Continuing Education in Anaesthesia, Critical Care & Pain j Volume 13 Number 1 2013
 Parenteral nutrition in critical care

muscle loss while providing exogenous glucose to promote tissue products contain a high proportion of the monounsaturated fatty
healing. acid, oleic acid. This may reduce oxidative damage and improve
Insulin resistance in the critically ill means that infusion of immune function and lipid profiles in the short term.11 The addition
large amounts of glucose result in an increase in blood glucose of fish oil, which contains the anti-inflammatory omega 3 fatty
levels. It has been documented that the maximum oxidation rate of acids eicosanopentanoic acid and docosahexaenoic acid, may de-
glucose in the stressed patient is 4–7 mg kg21 min21 (or 400– 700 crease synthesis of inflammatory eicosanoids and result in better
g day21 for a 70 kg patient). Therefore, in order to decrease the patient outcomes.
risk of metabolic alterations, the maximum rate of glucose infusion
should probably not exceed 5 mg kg21 min21.9 For a 70 kg
patient, this would be 2000 kcal glucose and this is not generally Protein
exceeded in standard PN regimes.
Protein, given as amino acid mixes in PN, is not given as an
Hyperglycaemia has proinflammatory effects, and indeed out-
energy substrate but in order to replace nitrogen losses and in an
comes in critically ill patients whose blood glucose levels are
attempt to prevent further skeletal muscle breakdown. Calculation
not controlled are poorer. The NICE-SUGAR study suggests that
of actual nitrogen balance in critically ill patients is difficult, so
outcomes are best when blood glucose levels are kept below 10
usually 0.13 –0.24 g kg21 day21 of nitrogen is given (again based
mmol litre21, with increased mortality secondary to hypoglycaemia
on NICE guidance and the practicalities of administration, al-
being associated with more intensive control (4.5– 6 mmol
though we would estimate and aim to administer 0.2 g kg21
litre21).10 Besides the deleterious effects of infusion of large
day21, towards the higher end of the range for the first 24–48 h,
amounts of glucose on blood sugar levels, it has also been shown
until the patient has had a nutrition/dietician review). In critical
that supplying glucose at a rate greater than its maximum oxidation
illness, glutamine, an amino acid which participates in many meta-
rate increases CO2 production in the critically ill patient, through
bolic processes, becomes an essential amino acid. This is because
lipogenesis (respiratory quotient, RQ of fat¼0.7 and RQ of
the increased utilization of glutamine in the critically ill exceeds
glucose¼1). This can be detrimental to patients with respiratory
the endogenous production rate causing plasma levels to decrease.
failure, especially those who are experiencing difficulty in weaning.
Low glutamine plasma levels have been associated with a worse
outcome in some studies, although it is controversial whether glu-
Lipid tamine supplementation in PN has been shown to improve patient
morbidity, mortality, and length of hospital stay.12
The use of lipid in PN regimens has a number of advantages. It is
a source of essential fatty acids, which have vital roles in the main-
tenance of cell membrane structure and function, and in modulat-
Trace elements, vitamins, and electrolytes
ing immune and inflammatory responses. It is a concentrated
energy source and has a lower osmolality compared with glucose. Trace elements such as copper, zinc, and selenium are not present
Lipids, however, have been associated with immune dysfunc- in commercially produced PN solutions for stability reasons, and
tion, decreases in oxygenation and pulmonary function and with these must be prescribed separately. Assessment of trace element
hepatic steatosis when given parenterally. The proportion of carbo- status on the ICU is impossible due to the effects of the acute
hydrate to lipid in any given PN formula varies as indeed do indi- phase response; however, increased losses of these can be antici-
vidual patients’ needs at any particular point in their illness. pated in particular pathological states, for example, patients with
The fatty acids linoleic acid and a-linoleic acid cannot be burns often have large losses of copper, zinc, and selenium, while
synthesized in the body and are thus essential. They are present in those on renal replacement therapy require extra selenium, zinc,
large quantities in soybean oil, and hence this is used as the lipid and thiamine. Supplementation with selenium, an important com-
source in many PN preparations. However, soybean oil is rich in ponent of the antioxidant glutathione peroxidase, has been shown
omega 6-long-chain triglycerides (LCTs) which have theoretical to improve outcome in patients with sepsis and possibly to reduce
disadvantages (e.g. platelet aggregation and impairment of immune the number of new infections on ICU.13
function) which have led to the development of several alternatives NICE guidelines state that vitamins must also be included in
which vary in the fatty acid chains attached to the glycerol moiety. every feed. Thiamine and B-vitamin replacement is important in
Medium-chain triglyerides (MCTs) such as may be found in intensive care patients who may have preexisting deficiency, for
coconut oil, for example, can be cleared from the blood stream example, alcoholics, in order to avoid possible neurological
more rapidly and this may improve tolerance and limit long-term complications.
adverse effects on the liver. It is thought that PN containing mix- Electrolytes and fluid should be given as part of PN based on
tures of LCTs and MCTs are associated with better outcomes in the typical daily requirements, clinical state, and monitored serum
critically ill, in terms of nutritional markers, physiological para- levels. Electrolytes are sometimes prescribed as acetate or phos-
meters, and disease outcome, although this does need validation in phate and chloride salts in order to minimize the risk of developing
a prospective randomized controlled trial. Olive oil-based lipid hyperchloraemic acidosis.

Continuing Education in Anaesthesia, Critical Care & Pain j Volume 13 Number 1 2013 3
Parenteral nutrition in critical care

Table 1 Example PN prescriptions. *For an average 70 kg patient without stage, by removal of the infective stimulus and specific antimicro-
significantly complicating factors such as renal failure
bial therapy.
Example starting regimen for Typical standard PN A recent NCEPOD report14 on current UK practice and stan-
patients at risk of refeeding regimen* dards observed surrounding the administration of PN showed that
Total non-N 800 1600
‘good practice’ was observed in only 19% of cases. Deficiencies in
calories (kcal) monitoring and assessment of patients on PN were observed in
Nitrogen (g) 5.7 9.2 54% of cases and metabolic complications related to PN occurred
Glutamine (g) 13.5 13.5
CHO (kcal) 300 800
in 40% of cases, in which 49% of these complications were
Fat (kcal) 500 800 deemed ‘avoidable’. Hence, there is much room for improvement.
Sodium (mmol) 50 64
Potassium (mmol) 30 48
Calcium (mmol) 5 4 Stopping PN
Magnesium 4 4.4
(mmol) There has been interest in the idea of ‘topping up’ EN with PN in
Phosphate (mmol) 20 20
Trace elements Depending on product, e.g. Depending on product, e.g. those who are not meeting daily calorie targets. Early signs are that
Decanw, Additracew Decanw, Additracew this may improve patient outcomes.
Vitamins Depending on product, e.g. Depending on product, e.g. There are no guidelines regarding when to cease PN; however,
Cernevitw, Solivitow Cernevitw, Solivitow
Volume (ml) 1250 2000 in view of what is known regarding the deleterious effects of mal-
nutrition, and the large energy deficits that can accumulate in crit-
ical care patients, perhaps the best strategy would be to stop it
when delivery of the patient’s daily energy needs are consistently
Refeeding syndrome achieved through enteral means.
All patients who have had no or very little nutritional intake for
.5 days are at risk of refeeding syndrome, which is associated Conclusion
with severe electrolyte abnormalities including hypokalaemia,
hypomagnesaemia, and hypophosphataemia, caused by movement Malnutrition in critical care patients is associated with poor out-
of phosphate, fluids, and other electrolytes intracellularly after a comes and patient harm. Initiation of adequate nutritional support
sudden carbohydrate load. A history of alcohol abuse or being on within 48 h of admission to ICU is therefore important. EN is the
drugs such as insulin, chemotherapy, antacids, or diuretics are risk first-line route, but due to many factors, there is often a need to
factors for developing refeeding syndrome. Refeeding syndrome use PN. Prescriptions for PN need to be individualized to the
can be associated with respiratory, cardiac, and neuromuscular patient based on premorbid status and current physiological needs
complications. NICE guidelines advise commencing nutritional and should involve a multidisciplinary team approach. Close daily
support at 50% of estimated energy requirements for 2 days in laboratory monitoring is required along with strict line sterility pre-
patients at risk of refeeding syndrome, thereafter increasing by cautions to prevent infection.
200– 400 kcal every day. Nutrients and fluids should not be
initiated without electrolytes and micronutrients. Close monitoring Declaration of interest
of serum potassium, magnesium, and phosphate levels is required
once feeding has been initiated. Sample regimens for PN are illu- None declared.
strated (Table 1).
References
1. National Collaborating Centre for Acute Care. NHS National Institute
Monitoring for Clinical Excellence Clinical Guideline 32: nutrition support in adults.
Available from www.nice.org.uk
Surveillance of all patients on PN should be ongoing. Metabolic
2. Singer P, Berger MM, Van den Berghe G et al. ESPEN guidelines on par-
status, in terms of regular blood glucose level assessment, and at enteral nutrition: intensive care. Clin Nutr 2009; 28: 387– 400
least daily urea, creatinine, and electrolytes including chloride, bi-
3. Kreymann KG, Berger MM, Deutz NEP et al. ESPEN guidelines on
carbonate, phosphate, and magnesium should be performed. enteral nutrition: intensive care. Clin Nutr 2006; 25: 210– 23
Plasma triglyceride levels should be monitored regularly. In critical 4. Heyland DK, Dhaliwal R, Drover JW et al. Canadian clinical practice
illness, there are no known good nutritional markers: markers guidelines for nutrition support in mechanically ventilated, critically ill
which traditionally reflect nutritional status such as albumin and adult patients. J Parenter Enteral Nutr 2003; 27: 355– 73
prealbumin do not reflect the true status of the patient. A better 5. Martindale RG, McClave SA, Vanek VW et al. Guidelines for the provi-
sion and assessment of nutrition support therapy in the adult critically ill
marker of a critically ill patient’s nutritional status may be wound
patient: Society of Critical Care Medicine and American Society for
healing and general clinical status. Signs of infection must be ac- Parenteral and Enteral Nutrition: executive summary. Crit Care Med
tively sought on a regular basis and aggressively treated at an early 2009; 37: 1757– 61

4 Continuing Education in Anaesthesia, Critical Care & Pain j Volume 13 Number 1 2013
 Parenteral nutrition in critical care

6. Casaer MP, Mesotten D, Hermans G et al. Early versus late parenteral 11. Mayer K, Fegbeutel C, Hattar K et al. v-3 vs. v-6 lipid emulsions exert
nutrition in critically ill adults. N Engl J Med 2011; 365: 506– 17 differential influence on neutrophils in septic shock patients: impact on
7. Fruhwald S, Holzer P, Metzler H. Intestinal motility disturbances in inten- plasma fatty acids and lipid mediator generation. Intensive Care Med
sive care patients pathogenesis and clinical impact. Intensive Care Med 2003; 29: 1472– 81
2007; 33: 36– 44 12. Bongers T, Griffiths RD, McArdle A. Exogenous glutamine: the clinical
8. Krishnan J, Parce P, Martinez A et al. Caloric intake in medical ICU evidence. Crit Care Med 2007; 35: S545–52
patients: consistency of care with guidelines and relationship to clinical 13. Andrews PJD, Avenell A, Noble DW et al. Randomised trial of glutam-
outcomes. Chest 2003; 124: 297–305 ine, selenium, or both, to supplement parenteral nutrition for critically
9. Wilmer A, Van den Berghe G. Parenteral nutrition. In Goldmand L, Ausiello ill patients. Br Med J 2011; 342: d1542.
D (eds). Cecil Textbook of Medicine, 23rd Edn. Philadelphia, PA: Elsevier, 14. Stewart JAD, Mason DG, Smith N et al. A mixed bag. An enquiry into
2008 the care of hospital patients receiving parenteral nutrition. Available
10. The NICE-SUGAR Study Investigators. Intensive versus conventional from www.ncepod.org
glucose control in critically ill patients. N Engl J Med 2009; 360:
1283–97 Please see multiple choice questions 1– 4.

Continuing Education in Anaesthesia, Critical Care & Pain j Volume 13 Number 1 2013 5