You are on page 1of 11

Metabolism,

Metabolomics, and
N u t r i t i o n a l Su p p o r t
of Patients with Sepsis
Joshua A. Englert, MDa, Angela J. Rogers, MD, MPHb,*

KEYWORDS
 Sepsis  Metabolism  Metabolomics  Nutrition  ICU outcomes  Biomarker

KEY POINTS
 Sepsis is characterized by profound metabolic derangements.
 Metabolic changes can serve as diagnostic and prognostic biomarkers for patients with sepsis.
Lactate is already in widespread clinical use, but technological advances make broader metabolic
profiling possible.
 Many large-scale clinical trials have been conducted to optimize nutrition in septic patients. Most
failed to show a benefit to early full feeding or supplementation with specific nutrients or metabo-
lites in patients with normal nutritional status at presentation.

INTRODUCTION (because of anorexia) or unable (because of


encephalopathy, respiratory failure requiring
Metabolism, derived from the Greek word to mechanical ventilation, and so forth) to eat, which
change, refers to all chemical reactions required can lead to a large energy deficit and worse out-
by cells. In the healthy state, human metabolism is comes in critically ill patients.2 This deficit, in
characterized by synchronized catabolic and turn, leads to profound skeletal muscle wasting
anabolic processes that not only allow cells to main- and prolonged recovery.3
tain homeostasis but also respond to their microen- In this review, the authors highlight metabolic
vironment. The main source of cellular energy is ATP changes that occur in sepsis, including both the
from aerobic metabolism and nutritional needs are systemic and cellular alterations that lead to the
largely met through nutrient intake, not catabolism dysregulation of normal human metabolism. The
of endogenous lipid and protein stores. authors then review the use of metabolic changes
This state of metabolic homeostasis is as biomarkers for disease severity, with a focus
massively disrupted in sepsis. Sepsis is a syn- first on lactate, the most widely used intensive
drome characterized by a dysregulated inflamma- care unit (ICU) biomarker, and then a broader dis-
tory response leading to organ damage in cussion of metabolomics in general. Finally, given
following a microbial infection. Sepsis is associ- the marked changes in metabolism in sepsis and
ated with an overall catabolic state leading to the the association of worse short- and long-term
breakdown of carbohydrates, lipid, and protein prognosis in patients with severe metabolic
stores.1 Despite increased nutritional require- derangements, the authors review the seminal
ments, patients with sepsis are often unwilling trials conducted to optimize nutrition in the ICU.
chestmed.theclinics.com

a
Division of Pulmonary, Allergy, Critical Care, and Sleep Medicine, The Ohio State University Wexner Medical
Center, 473 West 12th Avenue, Columbus, OH 43210, USA; b Division of Pulmonary and Critical Care Medicine,
Stanford University, 300 Pasteur Drive, H3143, Stanford, CA 94305-5236, USA
* Corresponding author.
E-mail address: ajrogers@stanford.edu

Clin Chest Med - (2016) --


http://dx.doi.org/10.1016/j.ccm.2016.01.011
0272-5231/16/$ see front matter 2016 Elsevier Inc. All rights reserved.
2 Englert & Rogers

METABOLIC CHANGES IN SEPSIS insufficiency in patients with septic shock. A


Mediators of Altered Metabolism in Sepsis randomized, multicenter, placebo-controlled trial
of hydrocortisone in patients with septic shock
The metabolic changes associated with sepsis are
did not decrease 28-day mortality regardless of
complex, with many of the key features highlighted
whether an adrenocorticotropin hormone stimula-
in Table 1. Many of these metabolic derange-
tion test was positive.7 Patients treated with
ments are mediated by changes in the endocrine
hydrocortisone did have a more rapid resolution
and autonomic nervous systems. The activation
of shock but also had an increased incidence of
of these two systems occurs simultaneously and,
new infection.7 In light of these results, debate
in general, increases energy consumption. Of
persists among experts regarding the use of cor-
note, neuroendocrine activation in sepsis is
ticosteroids in sepsis and septic shock. In addition
dynamic and can change frequently throughout a
to the changes in the HPA axis with sepsis, altered
patients course.
function of other endocrine organs, such as the
thyroid, can also lead to hormonal changes that
Altered endocrine physiology
alter metabolism.8
Acute illness, including sepsis, typically leads to
activation of the hypothalamic-pituitary-adrenal
Activation of the adrenergic nervous system
(HPA) axis and increased cortisol release.4
Activation of the adrenergic nervous system in
Increased circulating corticosteroid levels act to
septic patients leads to the release of endoge-
preserve vascular tone and reactivity in order to
nous catecholamines; in addition, patients with
maintain perfusion of vital organs.5 Although the
septic shock frequently require the administra-
normal response to stress is to increase adrenal
tion of exogenous catecholamines for blood
corticosteroid secretion, there are many different
pressure support. The release or administration
factors that can lead to the impairment of adrenal
of epinephrine, norepinephrine, and dopamine
function in septic patients. High levels of circu-
can have profound effects on metabolism that in-
lating cytokines can directly impair adrenal
crease catabolism of most macronutrients.9 One
corticosteroid production,6 and the use of medica-
of the main effects of catecholamine release is to
tions that can impair adrenal function is common in
increase the production of glucose by increasing
septic patients.5
hepatic glycogenolysis and gluconeogenesis.10
Although absolute adrenal insufficiency is
Furthermore, the insulin resistance that occurs
rare in patients with sepsis, there has been sub-
in sepsis is mediated, at least in part, by activa-
stantial controversy surrounding the use of
tion of the adrenergic system.11 In addition to
adjunctive corticosteroids to treat relative adrenal
the effects of catecholamines on metabolic
regulation, they also can affect immune function.
Table 1 Immune cells express adrenergic receptors9 and
Summary of major metabolic changes in sepsis catecholamines are known to affect cytokine
production12 and cell migration.13 These effects
Physiologic Change have implications for the ability of patients to
in Sepsis Metabolic Impact clear the inciting infection and return to a state
[ Gluconeogenesis, Hyperglycemia of metabolic homeostasis.
glycolysis
[ Protein Altered circulating amino Metabolic effects of cytokine release
catabolism acids The inflammatory cytokines that mediate that
[ Lipolysis [ Triglycerides, pathogenesis of sepsis play a key role in the acti-
Y lipoproteins vation of the neuroendocrine system described
Y Micronutrients [ Oxidative stress earlier. In addition, these cytokines can also
[ Neuroendocrine [ Catecholamines, directly alter metabolism in septic patients. The
activation [ counter-regulatory role of cytokines in the pathogenesis of sepsis
hormones has been reviewed in detail by others.14,15 Here
[ Cortisol Hyperglycemia the authors focus on the metabolic effects of
some of the classic proinflammatory cytokines.
[ Catecholamine [ Gluconeogenesis,
release [ glycolysis Many years after its initial discovery, tumor necro-
sis factor alpha (TNFa) was reported to be the
[ Cytokine release Hyperglycemia, insulin
resistance
same substance as the hormone cachectin.
Cachectin was initially described for its role in
Impaired oxygen [ Reactive oxygen species
increasing catabolism by upregulating lipolysis in
utilization
the setting of malignancy and chronic infection
Metabolism and Nutrition in Patients with Sepsis 3

before its description in the pathogenesis of shock impaired oxidative phosphorylation in multiple
and organ failure.16 In experimental models, infu- organs. This defect in the activity of the
sion of TNFa was also found to recapitulate the mitochondrial respiratory chain enzymes leads
hyperglycemia and insulin resistance found in to a shift toward anaerobic ATP production and
sepsis.1 is accompanied by increased free radical
TNFa is also a potent inducer of other cytokines generation.26 In one small study, septic patients
that play a role in metabolism, such as interleukin- were found to have decreased tissue ATP and
1 beta (IL-1b) and IL-6. Similar to TNFa, infusion of glutathione levels compared with nonseptic con-
recombinant IL-1b is also able to induce a hyper- trol subjects.27 Interestingly, decreased tissue
metabolic state in preclinical models17; IL-6 ATP levels were associated with worse out-
seems to work synergistically with these other cy- comes in this group of patients.27 Although it is
tokines. Although there are extensive data known that cellular metabolism is altered in
regarding the key roles of these cytokines in the sepsis, a clear understanding of the exact meta-
pathophysiology of sepsis, there have been bolic derangements has been elusive because
more than 100 clinical trials in patients with sepsis until recently the technology to simultaneously
targeted at these pathways that have failed to assess multiple metabolic pathways was not
lead to the approval of novel therapies.18 This available.
failure is likely multifactorial because of the com-
plex pathogenesis of sepsis, issues with some
Macronutrients
preclinical models, and substantial heterogeneity
among septic patients.18 It is also possible that The metabolism of all major types of macronutri-
an incomplete understanding of how cellular and ents (carbohydrate, protein, and lipid) is dysre-
systemic metabolism is altered in sepsis may be gulated in sepsis. Hyperglycemia is one of the
limiting progress in the development of novel most common metabolic derangements in pa-
therapeutics. tients presenting with sepsis and results from
altered glycogen metabolism and profound insu-
lin resistance.1,28 In late stages, sepsis can also
Mechanisms of Altered Cellular Metabolism in
be characterized by hypoglycemia due to multi-
Sepsis
system organ failure. The molecular events lead-
It is well known that some patients with sepsis ing to sepsis-induced hyperglycemia are
develop end organ failure despite appropriate complex and include the effects of inflammatory
therapy. Organ failure frequently develops in the cytokines and alterations in the regulatory hor-
later phases of sepsis following initial resuscita- mones that maintain normal glucose homeosta-
tion, and the molecular mechanisms of multior- sis, as discussed earlier.28,29 Hyperglycemia
gan failure remain incompletely understood.19 impairs the function of innate immune system
Given that sepsis is often accompanied by shock that further impairs the ability of the host to com-
and lactic acidosis, it was initially thought that bat infection.30 Given these effects, it is not sur-
organ failure in sepsis was primarily due to prising that hyperglycemia is an independent
tissue hypoxia from impaired oxygen delivery predictor of adverse outcomes in critically ill
in the setting of increased microvascular patients.29,31
permeability.20 In addition to abnormalities in carbohydrate
Although this may be the case early in sepsis, metabolism, sepsis is also characterized by
in the later phases (that are characterized by or- altered protein and lipid metabolism. Accelerated
gan failure), tissue oxygen delivery has been protein breakdown leads to a net negative nitrogen
shown to be normal in animal models21 and sep- balance1 that, in turn, leads to skeletal muscle
tic patients.22,23 This finding led to the realization wasting, deconditioning, and prolonged recovery
that sepsis is characterized by altered cellular for critically ill patients. In addition to generalized
metabolism and impaired oxygen utilization protein breakdown, sepsis is associated
despite adequate oxygen delivery. This impair- with altered concentrations of circulating amino
ment in cellular respiration, termed cytopathic acids.32,33 In general, amino acids from the break-
hypoxia, may be one of the mechanisms respon- down of peripheral tissues are shunted to the liver
sible for multiorgan failure in the setting of to support the synthesis of acute phase reac-
sepsis.24 The concept of energetic dysfunction tants.1 One of the goals of supplemental nutrition
in sepsis was first proposed when abnormal in sepsis is to try to mitigate protein catabolism
swollen mitochondria were visualized in animal by providing adequate amino acids for protein
models of sepsis.25 Since that time it has synthesis, although controversy remains regarding
become clear that sepsis is characterized by the ideal strategy to prevent protein catabolism.
4 Englert & Rogers

In addition to accelerated protein breakdown, dysfunction can outstrip the capacity of the cell
sepsis is also characterized by increased lipolysis, for aerobic glycolysis, and the cell switches to
as lipids are the primary source of energy in less efficient anaerobic metabolism in which
patients with infections.1 Patients with sepsis pyruvate is processed into 2 molecules of lactic
have altered lipid metabolism characterized by acid and 2 ATP. This increase in lactate locally is
increases in serum triglycerides and decreased a major reason for elevated plasma lactate levels,
levels circulating lipoproteins.34 Furthermore, though additional mechanisms, including impaired
administration of certain antiinflammatory classes lactate clearance in sepsis, may contribute.44,45
of lipid mediators has been shown to improve Lactic acid level is the most widely used
outcomes in patients with sepsis.35 biomarker used by clinicians caring for patients
with severe sepsis today. The importance of lactic
Micronutrients acid elevation has long been recognized.46
In addition to changes in the metabolism of mac- Elevated lactate levels can occur because of
ronutrients, sepsis is also associated with impaired organ perfusion (type A lactic acidosis)
changes in various micronutrients, including or in the absence of tissue hypoperfusion (type
trace minerals and vitamins. Micronutrients play B) due to malignancy, liver disease, or mitochon-
key roles in metabolism and cellular homeosta- drial disorders. Despite this lack of specificity, an
sis, and evidence suggests that lower levels of elevated lactate level has been validated in both
micronutrients in critically ill patients are associ- sepsis diagnosis and prognosis; failure to
ated with a higher risk of death and multisystem normalize lactate during resuscitation is similarly
organ failure.36 Two of the most well-studied associated with a poor prognosis. This evidence
micronutrients in sepsis are selenium and zinc. is discussed in further detail later.
Selenium is a trace mineral with antioxidant and
anti-inflammatory properties that is deficient in Lactate as a biomarker in the diagnosis of
patients with sepsis.37 Low selenium levels are sepsis
associated with poor outcomes in critically ill Severe sepsis is part of the inflammatory cascade
patients, and supplementation with selenium in that occurs in the setting of infection, from
critically ill patients has been shown to decrease systemic inflammatory response syndrome
mortality in some studies.38,39 However, given (SIRS) to sepsis, sepsis with organ hypoperfusion,
methodological concerns with some of these and finally septic shock. Elevated lactate (most
studies, experts suggest additional studies are commonly 4 mmol/L, though upper limits of
needed before the routine use of selenium can normal may vary for a given laboratory) is a marker
be recommended for septic patients.40 Similarly, of sepsis-induced hypoperfusion, even in the
zinc is another essential micronutrient that plays absence of frank septic shock.47 This cutoff was
a key role in cellular homeostasis, immune func- used for entry into pivotal human sepsis trials,
tion, and response to stress.41 Zinc deficiency including the early goal-directed therapy (EGDT)
increases mortality in preclinical models of protocol by Rivers and colleagues,48 whose entry
sepsis,42 and patients with sepsis have lower criteria included suspected infection with at least
circulating zinc levels compared to nonseptic 2 SIRS criteria and either systolic blood pressure
controls.43 Although many agree that treatment of 90 mm Hg or less or lactate of 4 mmol/L or
with micronutrients may be beneficial in sepsis, greater. The same threshold was incorporated
controversy remains regarding patient selection, into the Surviving Sepsis Campaign guidelines
choice of specific micronutrients, and optimal for early identification of sepsis in 2008 with grade
dosing.3,36 1C evidence26 and used as an entry criterion for
the follow-up early goal-directed therapy Pro-
CESS trial.49
METABOLIC CHANGES AS BIOMARKERS IN
SEPSIS
Lactate as a prognostic biomarker in sepsis
Lactate: the Prototypical Biomarker for Sepsis
Not only does lactate serve as an important
Diagnosis and Prognosis
biomarker for the diagnosis of sepsis but it is
Human cells use ATP for energy. In the resting also very clearly associated with increased risk
state, most glucose is metabolized through the for mortality in patients with sepsis, in
aerobic pathway, with mitochondria processing both the ICU and emergency department
pyruvate into carbon dioxide, water, and 38 ATP setting.50,51 Even a moderate elevation of lactate
via the citric acid cycle. In times of stress, such (range 2.14.0 mmol/L) has been associated
as exercise or sepsis; however, high cellular with an increased risk of death in normotensive
ATP requirements coupled with mitochondrial patients who present with sepsis.52,53
Metabolism and Nutrition in Patients with Sepsis 5

In addition to the importance of baseline lactate exploration of the metabolic changes in sepsis
levels, the clearance of lactate during the first beyond single-metabolite studies in small numbers
6 hours of sepsis treatment is highly associated of individuals. However, recent technological
with mortality. For example, patients who fail to advances in liquid and gas chromatography and
decrease their lactate level by at least 10% within mass spectroscopy now allow high-resolution
this timeframe are twice as likely to die as those screening of the human metabolome.58
who reach that threshold.54 Two randomized A full discussion of technical aspects of meta-
controlled trials (RCTs) have evaluated targeting bolic profiling and analysis is beyond the scope
lactate clearance as an end point in early of this review, but can be found elsewhere.59,60
sepsis resuscitation. A trial conducted by the Briefly, metabolic profiling is usually performed in
EMShockNet Investigators showed a 17% a 2-step fashion: (1) a combination of gas and/or
mortality rate in subjects randomized to lactate lipid chromatography is used to separate metabo-
clearance-guided resuscitation versus 23% in lites followed by (2) mass spectroscopy or nuclear
those randomized to central venous oxygen satura- magnetic resonance (NMR) mass spectroscopy
tionguided resuscitation, with the conclusion that for quantification. Untargeted metabolic profiling
this resuscitation goal was noninferior and led to is designed to measure all metabolites in a given
very similar early fluid resuscitation.55 Another sample, whereas targeted profiling is designed to
RCT randomized patients to routine EGDT or to identify a fixed subset of metabolites of interest.
lactate measurement every 2 hours for the first Both methods have their strengths. The major
8 hours; although lactate clearance was not appre- advantage of untargeted profiling is the lack of
ciably different between the 2 groups, the lactate bias and broader ability to discover novel, unantic-
group did receive more fluids, inotropes, and had ipated metabolites; but limitations include the sub-
a lower hospital mortality after adjustment for stantial time and cost required to definitively
severity of illness.56 Measurement of 6-hour lactate characterize newly identified metabolites and the
clearance was included in the 2012 Surviving lack of absolute quantification of metabolites.
Sepsis guidelines with grade 2C evidence.47 Even in untargeted profiling, resolution of metabo-
lites based on hydrophilicity and size may vary
depending on the strategy used, with greatest res-
Broader Metabolic Profiling in Sepsis
olution across a part of the metabolite spectrum.
Although lactate is a highly useful biomarker for Conversely, targeted metabolomic profiling iden-
sepsis diagnosis and prognosis, it is highly tifies only a subset of metabolites (often on the
nonspecific and often elevated in patients without order of 100500); but use of internal calibration
sepsis (type B lactic acidosis). Given the myriad standards allows both absolute quantification
metabolic changes induced by sepsis, it makes and high confidence in the individual metabolites
sense that changes in many other metabolites in attained.
addition to lactate would occur. To date, several groups including the authors
Metabolomics, the study of chemical products own have performed broader metabolic profiling in
(metabolites) used and produced by cellular meta- sepsis to test whether incorporation of multiple me-
bolism, measures small molecules, including tabolites could serve as prognostic indicators in
lipids, nucleotides, amino acids, carbohydrates, sepsis. Plasma metabolomic changes associated
and even drug metabolites. Metabolomics is a with sepsis mortality published to date are summa-
rapidly growing field of study in genomics, in part rized in Table 2. The studies vary dramatically in (1)
because it represents the end of the genomic sample size, (2) metabolites measured, and (3)
cascade, from single nucleotide polymorphism modeling method used to differentiate survivors
and methylation changes (many of which are pre- versus nonsurvivors. Given these varying methodol-
sent at birth) through gene expression, protein ogies in study design, coupled with high correlation
translation, and finally to metabolic changes. among many metabolites, it is perhaps not surpris-
These last genomic markers (gene expression, ing that the metabolites identified vary greatly
protein translation, and metabolites) are dynamic across studies. Although settling on one particular
and can change in response to environmental per- metabolic network is, premature at this time, it is
turbations, making them particularly compelling as worth noting that all studies reveal profound meta-
potential biomarkers in sepsis. The human plasma bolic derangements in sepsis, most (but by no
metabolome includes greater than 4000 metabo- means all) metabolites are upregulated; these
lites identified to date; however, because the changes extend far beyond the anaerobic meta-
human metabolome is not complete, this is likely bolism/lactate cycle. Identifying the optimal network
an underestimate.57 Metabolic profiling was previ- of metabolites, establishing the importance of these
ously extremely technically difficult, limiting networks across varied populations with sepsis,
6
Englert & Rogers
Table 2
Published metabolomic studies in sepsis

No. of Cases/ Metabolites Analytical


Reference Controlsa Control Population Profiled Strategy Final Metabolites
Mickiewicz 10/10 Pediatric septic 58 PCA & PLS 11 Metabolites (not identified)
et al,75 2013 shock survivors
Langley 31/119 Surviving subjects >300 Support vector Cis-4-decenoylcarnitine, 2-methylbutyroylcarnitine, butyroylcarnitine,
et al,76 2013 33/67 with SIRS or sepsis machine hexanoylcarnitine, and lactate
25/65
Rogers 30/60 Surviving subjects 167 Bayesian Sucrose, mannose, b-hydroxyisovalerate, methionine, and arginine
et al,77 2014 115/34 with SIRS or sepsis Network
Mickiewicz 4/4 ICU survivors with 60 PLS Network of 20 (not identified)
et al,78 2014 sepsis

Langley and colleagues76 and Rogers and colleagues77 analyses use highly overlapping datasets.
Abbreviations: PCA, principal components analysis; PLS, partial least squares.
a
When multiple case/control populations are shown, these represent the total number for testing and replication populations in that work.
Metabolism and Nutrition in Patients with Sepsis 7

and addressing whether incorporation of these net- to have both increased ICU length of stay and
works into critical illness severity models improves increased ICU mortality.63
their performance in prospective cohorts are all Given these substantial metabolic changes in
needed to determine the clinical utility of metabolo- sepsis, the potential to reduce these effects
mics in sepsis. through nutritional therapy has been studied
Metabolomic profiling in sepsis is still in its extensively. Major topics include timing, route,
infancy, and its diagnostic promise is not yet clear. rate of nutrition, and the nutrient composition to
Even such basic issues as type of body fluid to optimize sepsis survival. The authors highlight
sample are still evolving. Although the studies the strongest evidence for each of these issues
summarized in Table 2 have all focused on serum later and in Table 3. Although many of these
and plasma, Stringer and colleagues61 recently studies are not restricted to only patients with
suggested that whole-blood sampling may be a sepsis, most of them include critically ill patients
better target, as it also reflects endothelial cell who have sepsis as their primary or secondary
metabolism and free hemoglobin level. Additional ICU risk factor. For further reading on this
options include studying samples that reflect a evidence, several excellent reviews of ICU nutri-
particular organ function in sepsis, for example, tion have been written previously.64,65
urine in acute kidney injury or bronchoalveolar
lavage fluid in septic patients with acute respira- Timing of Nutrition
tory distress syndrome (ARDS).
Most critically ill patients are unable to take in
NUTRITION IN PATIENTS WITH SEPSIS adequate oral nutrition, particularly early in the
course of critical illness. Although several meta-
As noted throughout this review, patients with analyses have suggested a mortality benefit to
severe sepsis and septic shock have profoundly early feeding within the first 48 hours in critically
altered metabolism, with a catabolic state leading ill patients,66 methodological concerns about
to breakdown of both protein and lipids coupled high potential for bias in the small trials included
with decreased production of new muscle mass. in these meta-analyses limit their usefulness.
Critically ill patients have been shown to have a Several recent large, high-quality RCTs are
profound loss of muscle mass, a mean of 17% highly relevant. The EDEN trial focused on patients
loss of femoral mass by day 10, worse among pa- with the ARDS; most of these patients had under-
tients with increased severity of illness.62 Patients lying sepsis (>70% with either sepsis or pneu-
who develop critical illness neuropathy are known monia as their ARDS risk factor). Patients were

Table 3
Highlighted articles in intensive care unit nutrition

Trial (Ref) Treatment Groups Population (% Sepsis) Primary Outcome


EDEN68 Trophic vs full ARDS (>70%) No change in vent time, infections,
feeds in 60-d mortality, 1-y physical
first 6 d function
Early PN70 Early PN vs ICU patients with short-term No difference in mortality, early PN
standard care contraindication to EN (6%) had fewer vent days
EPaNIC69 Early PN vs ICU patients at nutritional risk Late PN: 6% more likely to be
delayed (22%) discharged alive from ICU &
to day 8 hospital, fewer infections
SPN71 PN or EN if not ICU patients not meeting No change in ICU stay or
meeting caloric nutritional needs by day 3 60-d mortality, but Y vent days
need by day 3 (w45%)
REDOXS72 Glutamine or ICU patients with 2 or more organ [ In-hospital & 6-mo mortality in
antioxidants failures (w30%) glutamine group; no effect with
antioxidants
SIGNET73 Glutamine or ICU needing at least 50% of calories No change in mortality or infection
selenium via PN (w60%) rate in all patients
OMEGA74 Omega-3 fatty ARDS (75% sepsis or pneumonia) Y Vent-free and ICU days, trend
acids toward [ 60-d mortality

enteral nutrition; PN, parental nutrition.


8 Englert & Rogers

randomized within 48 hours of mechanical ventila- oxygen species and a depletion of antioxidant nutri-
tion to either trophic or full enteral feeds for the first ents associated with increased mortality, as
6 days. Despite a marked difference in calories discussed earlier.27 Not surprisingly, numerous
(400 vs 1300), there was no difference in duration trials of therapeutic administration of macronutri-
of mechanical ventilation, infectious complica- ents and antioxidants have been performed in an
tions, or 60-day mortality in the two groups. attempt to improve sepsis mortality. Although
Patients followed for up to 1 year after discharge many small RCTs and subsequent meta-analyses
also demonstrated no difference in physical or have shown encouraging trends toward improved
cognitive function based on nutrition strategy.67 sepsis mortality,39 this has not borne out in larger
These data suggest that delay of full nutrition up RCTs to date, which are summarized in Table 3.
to 6 days is likely safe in patients without baseline Low glutamine levels are associated with worse
malnutrition (who were excluded from these prognosis in critical illness; it has, thus, been the
trials).68 subject of several large RCTs. The REDOXS trial
randomized 1223 critically ill patients to receive
Enteral Versus Parenteral Nutrition glutamine, antioxidants (including selenium, zinc,
vitamins C, E, and b carotene), both, or placebo.
Enteral feeding is the accepted first choice for
The groups that received glutamine had higher
nutrition in critically ill patients who can tolerate
in-hospital and 6-month mortality, whereas antiox-
it, given consistent evidence of improved
idants had no effect.72 The SIGNET trial random-
outcomes, including fewer infections and
ized 500 subjects to a lower dose of glutamine
improved gut integrity.65 However, a large propor-
(approximately one-third the dose in REDOXS)
tion of critically ill patients have relative contraindi-
and identified no difference in mortality or new in-
cations to enteral feeding and fail to meet their
fections in all patients randomized.73
caloric needs in the early days of ICU care. Thus,
The ARDS Network OMEGA trial randomized
several large RCTs have been conducted to
272 patients with ARDS (75% with sepsis or pneu-
address the timing of initiation of parental nutrition
monia as their ARDS risk factor) to receive omega-
(PN) in high-risk ICU patients.
3 fatty acids, which favor production of less active
The EPaNIC trial randomized 4640 critically ill
prostaglandins and leukotrienes. The trial was
subjects at high risk for malnutrition to receive
stopped early for futility, with patients receiving
either early PN (glucose  48 hours then full-
supplements having fewer ventilator-free and
caloric PN) or a late-initiation group that did not
ICU free days and a trend toward higher 60-day
receive PN until day 8. The late-initiation strategy
hospital mortality.74 Although multiple studies
was associated with faster recovery (6% increased
have identified deficiencies of specific nutrients/
likelihood of discharge alive from ICU and hospital)
metabolites in septic patients, there are currently
and fewer ICU infections.69 The Early PN Trial ran-
no data to support the use of replacement therapy
domized 1372 with short-term contraindications to
in sepsis.
enteral feeding to receive either early PN versus
pragmatic standard care (of the latter group, 29%
commenced enteral nutrition, 27% parenteral, SUMMARY
and 40% unfed). The early PN strategy led to no dif-
In summary, sepsis is characterized by profound
ference in 60-day mortality but fewer days of me-
metabolic changes. Some of these metabolic
chanical ventilation.70 Finally, the SPN trial
changes contribute to sepsis pathophysiology,
randomized 305 critically ill patients who were
and others (eg, high lactate, hyperglycemia, low
reaching less than 60% of nutrition targets by day
selenium and zinc) are recognized markers of
3 to receive either PN or enteral nutrition. They
ICU outcomes. Given these profound metabolic
found substantially lower rates of infection in the
changes, numerous large-scale trials have been
PN group (the primary outcome).71 Cumulatively,
designed to optimize nutrition in the ICU, though
these trials suggest that PN and full nutrition can
most have failed to show that nutrition strategies
be delayed for at least 7 days in most critically ill pa-
improve outcomes for septic patients, at least
tients with normal nutritional status at presentation
early in the ICU course of patients with adequate
and that any advantages to early PN initiation are
baseline nutrition.
likely modest.
Finally, technical advances in metabolomic
profiling have enabled cheaper, high-resolution
Role of Macronutrient and Micronutrient
testing than was previously available. The option
Replacement
to simultaneously examine multiple metabolic
Sepsis is characterized by marked systemic inflam- pathways in large populations is now more
mation, with increased production of reactive feasible and will likely allow a more nuanced
Metabolism and Nutrition in Patients with Sepsis 9

picture of the metabolic changes that occur in 16. Tracey KJ, Beutler B, Lowry SF, et al. Shock and tis-
sepsis. As these data emerge in the coming years, sue injury induced by recombinant human cachec-
developing a more individualized approach to tin. Science 1986;234(4775):4704.
metabolism and nutrition in septic patients may 17. Molloy RG, Mannick JA, Rodrick ML. Cytokines,
become possible. sepsis and immunomodulation. Br J Surg 1993;
80(3):28997.
REFERENCES 18. Marshall JC. Why have clinical trials in sepsis failed?
Trends Mol Med 2014;20(4):195203.
1. Michie HR. Metabolism of sepsis and multiple organ 19. Englert JA, Fink MP. The multiple organ dysfunction
failure. World J Surg 1996;20(4):4604. syndrome and late-phase mortality in sepsis.
2. Alberda C, Gramlich L, Jones N, et al. The relation- Curr Infect Dis Rep 2005;7(5):33541.
ship between nutritional intake and clinical out- 20. Brealey D, Singer M. Mitochondrial dysfunction in
comes in critically ill patients: results of an sepsis. Curr Infect Dis Rep 2003;5(5):36571.
international multicenter observational study. Inten- 21. VanderMeer TJ, Wang H, Fink MP. Endotoxemia
sive Care Med 2009;35(10):172837. causes ileal mucosal acidosis in the absence of
3. Casaer MP, Van den Berghe G. Nutrition in the acute mucosal hypoxia in a normodynamic porcine model
phase of critical illness. N Engl J Med 2014;370(25): of septic shock. Crit Care Med 1995;23(7):121726.
24501. 22. Boekstegers P, Weidenhofer S, Pilz G, et al. Periph-
4. Khardori R, Castillo D. Endocrine and metabolic eral oxygen availability within skeletal muscle in
changes during sepsis: an update. Med Clin North sepsis and septic shock: comparison to limited
Am 2012;96(6):1095105. infection and cardiogenic shock. Infection 1991;
5. Cooper MS, Stewart PM. Corticosteroid insufficiency 19(5):31723.
in acutely ill patients. N Engl J Med 2003;348(8): 23. Sair M, Etherington PJ, Peter Winlove C, et al. Tissue
72734. oxygenation and perfusion in patients with systemic
6. Catalano RD, Parameswaran V, Ramachandran J, sepsis. Crit Care Med 2001;29(7):13439.
et al. Mechanisms of adrenocortical depression dur- 24. Fink MP. Bench-to-bedside review: cytopathic
ing Escherichia coli shock. Arch Surg 1984;119(2): hypoxia. Crit Care 2002;6(6):4919.
14550. 25. Levy E, Slusser RJ, Ruebner BH. Hepatic changes
7. Sprung CL, Annane D, Keh D, et al. Hydrocortisone produced by a single dose of endotoxin in the
therapy for patients with septic shock. N Engl J Med mouse. Electron microscopy. Am J Pathol 1968;
2008;358(2):11124. 52(2):477502.
8. Meyer S, Schuetz P, Wieland M, et al. Low triiodothy- 26. Trager K, DeBacker D, Radermacher P. Metabolic
ronine syndrome: a prognostic marker for outcome alterations in sepsis and vasoactive drug-related
in sepsis? Endocrine 2011;39(2):16774. metabolic effects. Curr Opin Crit Care 2003;9(4):
9. Norbury WB, Jeschke MG, Herndon DN. 2718.
Metabolism modulators in sepsis: propranolol. Crit 27. Brealey D, Brand M, Hargreaves I, et al. Association
Care Med 2007;35(9 Suppl):S61620. between mitochondrial dysfunction and severity and
10. Chu CA, Sindelar DK, Igawa K, et al. The direct effects outcome of septic shock. Lancet 2002;360(9328):
of catecholamines on hepatic glucose production 21923.
occur via alpha(1)- and beta(2)-receptors in the dog. 28. Marik PE, Raghavan M. Stress-hyperglycemia, insu-
Am J Physiol Endocrinol Metab 2000;279(2):E46373. lin and immunomodulation in sepsis. Intensive Care
11. Lang CH. Sepsis-induced insulin resistance in rats is Med 2004;30(5):74856.
mediated by a beta-adrenergic mechanism. Am J 29. Hirasawa H, Oda S, Nakamura M. Blood glucose
Physiol 1992;263(4 Pt 1):E70311. control in patients with severe sepsis and septic
12. Elenkov IJ, Wilder RL, Chrousos GP, et al. The sym- shock. World J Gastroenterol 2009;15(33):41326.
pathetic nervean integrative interface between two 30. Turina M, Fry DE, Polk HC Jr. Acute hyperglycemia and
supersystems: the brain and the immune system. the innate immune system: clinical, cellular, and mo-
Pharmacol Rev 2000;52(4):595638. lecular aspects. Crit Care Med 2005;33(7):162433.
13. Oberbeck R. Therapeutic implications of immune- 31. Taylor JH, Beilman GJ. Hyperglycemia in the inten-
endocrine interactions in the critically ill patients. sive care unit: no longer just a marker of illness
Curr Drug Targets Immune Endocr Metabol Disord severity. Surg Infect (Larchmt) 2005;6(2):23345.
2004;4(2):12939. 32. Druml W, Heinzel G, Kleinberger G. Amino acid
14. Blackwell TS, Christman JW. Sepsis and cytokines: kinetics in patients with sepsis. Am J Clin Nutr
current status. Br J Anaesth 1996;77(1):1107. 2001;73(5):90813.
15. Rittirsch D, Flierl MA, Ward PA. Harmful molecular 33. Su L, Li H, Xie A, et al. Dynamic changes in amino
mechanisms in sepsis. Nat Rev Immunol 2008; acid concentration profiles in patients with sepsis.
8(10):77687. PLoS One 2015;10(4):e0121933.
10 Englert & Rogers

34. Wendel M, Paul R, Heller AR. Lipoproteins in inflam- and septic shock. N Engl J Med 2001;345(19):
mation and sepsis. II. Clinical aspects. Intensive 136877.
Care Med 2007;33(1):2535. 49. Angus DC, Yealy DM, Kellum JA, et al. Protocol-
35. Pontes-Arruda A, Martins LF, de Lima SM, et al. Enteral based care for early septic shock. N Engl J Med
nutrition with eicosapentaenoic acid, gamma-linolenic 2014;371(4):386.
acid and antioxidants in the early treatment of sepsis: 50. Aduen J, Bernstein WK, Khastgir T, et al. The use
results from a multicenter, prospective, randomized, and clinical importance of a substrate-specific
double-blinded, controlled study: the INTERSEPT electrode for rapid determination of blood lactate
study. Crit Care 2011;15(3):R144. concentrations. JAMA 1994;272(21):167885.
36. Manzanares W, Langlois PL, Hardy G. Update on 51. Shapiro NI, Howell MD, Talmor D, et al. Serum
antioxidant micronutrients in the critically ill. lactate as a predictor of mortality in emergency
Curr Opin Clin Nutr Metab Care 2013;16(6):71925. department patients with infection. Ann Emerg
37. Forceville X, Vitoux D, Gauzit R, et al. Selenium, sys- Med 2005;45(5):5248.
temic immune response syndrome, sepsis, and 52. Mikkelsen ME, Miltiades AN, Gaieski DF, et al.
outcome in critically ill patients. Crit Care Med Serum lactate is associated with mortality in severe
1998;26(9):153644. sepsis independent of organ failure and shock.
38. Angstwurm MW, Engelmann L, Zimmermann T, et al. Crit Care Med 2009;37(5):16707.
Selenium in intensive care (SIC): results of a 53. Howell MD, Donnino M, Clardy P, et al. Occult hypo-
prospective randomized, placebo-controlled, multi- perfusion and mortality in patients with suspected
ple-center study in patients with severe systemic infection. Intensive Care Med 2007;33(11):18929.
inflammatory response syndrome, sepsis, and sep- 54. Nguyen HB, Rivers EP, Knoblich BP, et al. Early
tic shock. Crit Care Med 2007;35(1):11826. lactate clearance is associated with improved
39. Manzanares W, Dhaliwal R, Jiang X, et al. Antioxi- outcome in severe sepsis and septic shock.
dant micronutrients in the critically ill: a systematic Crit Care Med 2004;32(8):163742.
review and meta-analysis. Crit Care 2012;16(2):R66. 55. Jones AE, Shapiro NI, Trzeciak S, et al. Lactate
40. Allingstrup M, Afshari A. Selenium supplementation clearance vs central venous oxygen saturation as
for critically ill adults. Cochrane Database Syst Rev goals of early sepsis therapy: a randomized clinical
2015;(7):CD003703. trial. JAMA 2010;303(8):73946.
41. Haase H, Rink L. Functional significance of zinc- 56. Jansen TC, van Bommel J, Schoonderbeek FJ, et al.
related signaling pathways in immune cells. Annu Early lactate-guided therapy in intensive care unit
Rev Nutr 2009;29:13352. patients: a multicenter, open-label, randomized
42. Knoell DL, Julian MW, Bao S, et al. Zinc deficiency controlled trial. Am J Respir Crit Care Med 2010;
increases organ damage and mortality in a murine 182(6):75261.
model of polymicrobial sepsis. Crit Care Med 57. Wishart DS, Jewison T, Guo AC, et al. HMDB 3.0the
2009;37(4):13808. human metabolome database in 2013. Nucleic
43. Besecker BY, Exline MC, Hollyfield J, et al. Acids Res 2013;41(Database issue):D8017.
A comparison of zinc metabolism, inflammation, 58. Dettmer K, Aronov PA, Hammock BD. Mass
and disease severity in critically ill infected and spectrometry-based metabolomics. Mass Spectrom
noninfected adults early after intensive care unit Rev 2007;26(1):5178.
admission. Am J Clin Nutr 2011;93(6):135664. 59. Dunn WB, Ellis DI. Metabolomics: current analytical
44. Jansen TC, van Bommel J, Bakker J. Blood lactate platforms and methodologies. Trends Analyt Chem
monitoring in critically ill patients: a systematic 2005;24(4):28594.
health technology assessment. Crit Care Med 60. Korman A, Oh A, Raskind A, et al. Statistical
2009;37(10):282739. methods in metabolomics. In: Anisimova M, editor.
45. Levraut J, Ciebiera JP, Chave S, et al. Mild hyperlac- Evolutionary genomics, 856. New York: Humana
tatemia in stable septic patients is due to impaired Press; 2012. p. 381413.
lactate clearance rather than overproduction. Am J 61. Stringer KA, Younger JG, McHugh C, et al. Whole
Respir Crit Care Med 1998;157(4 Pt 1):10216. blood reveals more metabolic detail of the human
46. Cohen RD, Woods HF. Clinical and biochemical metabolome than serum as measured by 1H-NMR
aspects of lactic acidosis. Oxford: Blackwell Scientific spectroscopy: implications for sepsis metabolo-
Publications; 1976. mics. Shock 2015;44(3):2008.
47. Dellinger RP, Levy MM, Rhodes A, et al. Surviving 62. Puthucheary ZA, Rawal J, McPhail M, et al. Acute
sepsis campaign: international guidelines for skeletal muscle wasting in critical illness. JAMA
management of severe sepsis and septic shock: 2013;310(15):1591600.
2012. Crit Care Med 2013;41(2):580637. 63. Garnacho-Montero J, Madrazo-Osuna J, Garcia-
48. Rivers E, Nguyen B, Havstad S, et al. Early goal- Garmendia JL, et al. Critical illness polyneuropathy:
directed therapy in the treatment of severe sepsis risk factors and clinical consequences. A cohort
Metabolism and Nutrition in Patients with Sepsis 11

study in septic patients. Intensive Care Med 2001; nutrition: a randomized controlled trial. JAMA 2013;
27(8):128896. 309(20):21308.
64. Casaer MP, van den Berghe G. Nutrition in the acute 71. Heidegger CP, Berger MM, Graf S, et al. Optimisation
phase of critical illness. N Engl J Med 2014;370(13): of energy provision with supplemental parenteral nutri-
122736. tion in critically ill patients: a randomised controlled
65. Martindale RG, McClave SA, Vanek VW, et al. Guide- clinical trial. Lancet 2013;381(9864):38593.
lines for the provision and assessment of nutrition 72. Heyland D, Muscedere J, Wischmeyer PE, et al.
support therapy in the adult critically ill patient: A randomized trial of glutamine and antioxidants in
Society of Critical Care Medicine and American critically ill patients. N Engl J Med 2013;368(16):
Society for Parenteral and Enteral Nutrition: execu- 148997.
tive summary. Crit Care Med 2009;37(5):175761. 73. Andrews PJ, Avenell A, Noble DW, et al. Rando-
66. Doig GS, Heighes PT, Simpson F, et al. Early enteral mised trial of glutamine, selenium, or both, to sup-
nutrition, provided within 24 h of injury or intensive plement parenteral nutrition for critically ill patients.
care unit admission, significantly reduces mortality BMJ 2011;342:d1542.
in critically ill patients: a meta-analysis of rando- 74. Rice TW, Wheeler AP, Thompson BT, et al. Enteral
mised controlled trials. Intensive Care Med 2009; omega-3 fatty acid, gamma-linolenic acid, and anti-
35(12):201827. oxidant supplementation in acute lung injury. JAMA
67. Needham DM, Dinglas VD, Bienvenu OJ, et al. 2011;306(14):157481.
One year outcomes in patients with acute lung injury 75. Mickiewicz B, Vogel HJ, Wong HR, et al.
randomised to initial trophic or full enteral feeding: Metabolomics as a novel approach for early
prospective follow-up of EDEN randomised trial. diagnosis of pediatric septic shock and its mortality.
BMJ 2013;346:f1532. Am J Respir Crit Care Med 2013;187(9):96776.
68. National Heart Lung and Blood Institute Acute 76. Langley RJ, Tsalik EL, Velkinburgh JC, et al. An inte-
Respiratory Distress Syndrome Clinical Trials grated clinico-metabolomic model improves predic-
Network, Rice TW, Wheeler AP, et al. Initial trophic tion of death in sepsis. Sci Transl Med 2013;5(195):
vs full enteral feeding in patients with acute lung 195ra95.
injury: the EDEN randomized trial. JAMA 2012; 77. Rogers AJ, McGeachie M, Baron RM, et al. Metabolo-
307(8):795803. mic derangements are associated with mortality in crit-
69. Casaer MP, Mesotten D, Hermans G, et al. Early ically ill adult patients. PLoS One 2014;9(1):e87538.
versus late parenteral nutrition in critically ill adults. 78. Mickiewicz B, Duggan GE, Winston BW, et al. Meta-
N Engl J Med 2011;365(6):50617. bolic profiling of serum samples by 1H nuclear mag-
70. Doig GS, Simpson F, Sweetman EA, et al. Early netic resonance spectroscopy as a potential
parenteral nutrition in critically ill patients with diagnostic approach for septic shock. Crit Care
short-term relative contraindications to early enteral Med 2014;42(5):11409.

You might also like