The study budget was not sufficient to support the collection of complete information on all patients screened for eligibility into the trial at each site. The total number of screened patients, and reasons they were not enrolled, are not available. ICU indicates intensive care unit.
Day 1 is the day of study enrollment. Energy received was calculated from enteral nutrition, parenteral nutrition, and intravenous infusions with ≥10% glucose. Error bars indicate standard error; ICU, intensive care unit.
Doig GS, Simpson F, Sweetman EA, et al. Early parenteral nutrition in critically ill patients with short-term relative contraindications to early enteral nutrition: a randomized controlled trial. JAMA. doi:10.1001/jama.2013.5124.
Table of Contents
Early PN Trial Conduct and Management
eTable 1: Complete Eligibility Criteria
Inclusion Criteria (All YES answers for enrolment)
Exclusion Criteria (All NO answers for enrolment, YES to any for exclusion)
Contraindications to Kabiven G19% based on TGA Licensing Indications
eTable 2: PN Protocols
eTable 2 continued: PN Protocols
Study PN Protocol B: MALNOURISHED Early PN PATIENTS (Ex. BMI = 17)
eTable 3: Harris-Benedict equations and adjustment factors used by study web site
eTable 4: Nutritional therapy received during ICU stay, Standard Care and Early PN patients
eTable 5: Clinically Significant Organ Failure and Concomitant Interventions, crude event rates, not adjusted for time at risk (ICU stay)
eFigure 1. Energy intake from enteral nutrition and/or parenteral nutrition during ICU stay
eFigure 2: Subjective Global Assessment of muscle wasting during ICU stay
eFigure 3: Subjective Global Assessment of fat loss during ICU stay
Doig GS, Simpson F, Sweetman EA, Finfer SR, Cooper DJ, Heighes PT, Davies AR, O’Leary M, Solano T, Peake S, Early PN Investigators of the ANZICS Clinical Trials Group FT. Early Parenteral Nutrition in Critically Ill Patients With Short-term Relative Contraindications to Early Enteral NutritionA Randomized Controlled Trial. JAMA. 2013;309(20):2130-2138. doi:10.1001/jama.2013.5124
Caring for the Critically Ill Patient Section Editor: Derek C. Angus, MD, MPH, Contributing Editor, JAMA (email@example.com).
Author Affiliations: Northern Clinical School Intensive Care Research Unit (Dr Doig and Mss Simpson, Sweetman, and Heighes), The George Institute for Global Health (Dr Finfer), University of Sydney, Sydney, Australia; Alfred Hospital, Melbourne, Australia (Drs Cooper and Davies); Royal Prince Alfred Hospital, Sydney (Dr O’Leary); Westmead Hospital, Sydney (Dr Solano); and Queen Elizabeth Hospital, Adelaide, Australia (Dr Peake).
Importance Systematic reviews suggest adult patients in intensive care units (ICUs) with relative contraindications to early enteral nutrition (EN) may benefit from parenteral nutrition (PN) provided within 24 hours of ICU admission.
Objective To determine whether providing early PN to critically ill adults with relative contraindications to early EN alters outcomes.
Design, Setting, and Participants Multicenter, randomized, single-blind clinical trial conducted between October 2006 and June 2011 in ICUs of 31 community and tertiary hospitals in Australia and New Zealand. Participants were critically ill adults with relative contraindications to early EN who were expected to remain in the ICU longer than 2 days.
Interventions Random allocation to pragmatic standard care or early PN.
Main Outcomes and Measures Day-60 mortality; quality of life, infections, and body composition.
Results A total of 1372 patients were randomized (686 to standard care, 686 to early PN). Of 682 patients receiving standard care, 199 patients (29.2%) initially commenced EN, 186 patients (27.3%) initially commenced PN, and 278 patients (40.8%) remained unfed. Time to EN or PN in patients receiving standard care was 2.8 days (95% CI, 2.3 to 3.4). Patients receiving early PN commenced PN a mean of 44 minutes after enrollment (95% CI, 36 to 55). Day-60 mortality did not differ significantly (22.8% for standard care vs 21.5% for early PN; risk difference, −1.26%; 95% CI, −6.6 to 4.1; P = .60). Early PN patients rated day-60 quality of life (RAND-36 General Health Status) statistically, but not clinically meaningfully, higher (45.5 for standard care vs 49.8 for early PN; mean difference, 4.3; 95% CI, 0.95 to 7.58; P = .01). Early PN patients required fewer days of invasive ventilation (7.73 vs 7.26 days per 10 patient × ICU days, risk difference, −0.47; 95% CI, −0.82 to −0.11; P = .01) and, based on Subjective Global Assessment, experienced less muscle wasting (0.43 vs 0.27 score increase per week; mean difference, −0.16; 95% CI, −0.28 to −0.038; P = .01) and fat loss (0.44 vs 0.31 score increase per week; mean difference, −0.13; 95% CI, −0.25 to −0.01; P = .04).
Conclusions and Relevance The provision of early PN to critically ill adults with relative contraindications to early EN, compared with standard care, did not result in a difference in day-60 mortality. The early PN strategy resulted in significantly fewer days of invasive ventilation but not significantly shorter ICU or hospital stays.
Trial Registration anzctr.org.au Identifier: ACTRN012605000704695
Parenteral nutrition has been in common use since the 1960s1 and is accepted as the standard of care for patients with chronic nonfunctioning gastrointestinal tracts.2 In critical illness, controversy surrounds the appropriate use of parenteral nutrition,3 but large-scale trials have begun to answer important questions.
Published in 2011, EPaNIC (Early Parenteral Nutrition Completing Enteral Nutrition in Adult Critically Ill Patients)4 enrolled 4640 critically ill patients to investigate the effects of using parenteral nutrition when enteral nutrition failed to reach a caloric target. EPaNIC did not find any benefits from using additional parenteral nutrition in patients who could receive enteral nutrition5; however, many other important questions regarding parenteral nutrition remain.
For example, key questions were raised by a systematic review that reported benefits associated with parenteral nutrition use in patients with relative contraindications to early enteral nutrition.6 In contrast to EPaNIC, which targeted a wide spectrum of patients who were eligible to receive enteral nutrition, this systematic review focused on clinical trials that enrolled a narrow subset of critically ill patients, those with short-term relative contraindications to enteral nutrition. When these patients were randomized to receive standard care (unfed for 2-5 days) or parenteral nutrition within the first 24 hours of intensive care unit (ICU) admission, mortality was significantly reduced in the group receiving parenteral nutrition. However, infection rates increased.
Because the small size of the clinical trials included in this systematic review led to uncertainty surrounding the clinical importance of increased infections, some authors expressed reluctance to use parenteral nutrition in patients with short-term relative contraindications to enteral nutrition.7 This uncertainty is mirrored by major practice guidelines, some of which recommend parenteral nutrition within 24 to 48 hours of ICU admission if enteral nutrition is contraindicated,8,9 and other guidelines recommending parenteral nutrition only after 7 to 14 days of starvation.10
We conducted a multicenter clinical trial (the Early PN Trial) to assess the effects of providing parenteral nutrition within 24 hours of ICU admission to adult critically ill patients who would not otherwise receive nutrition therapy because of short-term relative contraindications to enteral nutrition.
Adult patients were eligible for enrollment within 24 hours of ICU admission if they were expected to remain in the ICU on the calendar day after enrollment, were considered ineligible for enteral nutrition by the attending clinician due to a short-term relative contraindication and were not expected to receive parenteral or oral nutrition on the day of enrollment or the day after enrollment, and had a central venous line through which parenteral nutrition could be delivered. Patients who were moribund and not expected to survive 24 hours, who were to receive palliative care only, or who had a licensing/labeling contraindication to the study parenteral nutrition (eg, known pregnancy, documented allergy, etc) were excluded. (See eTable 1 for complete eligibility criteria.) Approval was obtained from each participating site's human research ethics committee. Written consent was documented in accordance with local and national laws.
Allocation concealment was maintained by use of a central randomization web server. Randomization was stratified within site by age and body mass index (BMI) using permuted blocks of variable sizes and random seeds. Use of a small number of odd-sized permuted blocks (size = 1), referred to as random seeds, increases the difficulty of counting into the allocation sequence and helps maintain concealment of upcoming group assignments.11 After investigators attended a small-group start-up meeting, a run-in phase allowed them to become familiar with the application of eligibility criteria before they recruited their first patients.12
Intervention patients received standard parenteral nutrition from a ready-to-mix 3-chamber bag containing amino acids, glucose, lipids, and electrolytes (Kabiven G19%, Fresenius Kabi Australia), with starting rates and rate increases defined by study protocol and trace elements, minerals, and vitamins added as clinically appropriate.13 Targets were to be achieved by study day 3. The protocol reminded clinicians to consider additional vitamins and minerals on each study day and to consider enteral or oral nutrition on study day 3. The protocol also defined timing and rates for parenteral nutrition discontinuation.
Energy targets were calculated using the Harris-Benedict equation, allowing for stress factors.14 Targets were capped at 35 kcal/kg/d, and obese patients (BMI ≥30) were fed to their ideal body weight (BMI = 21). (BMI is calculated as weight in kilograms divided by height in meters squared.) Total energy content (including amino acids) of the study parenteral nutrition was used in all calculations. Protein targets were not set independently of calculated energy targets. Malnourished patients (BMI <17 or clinical diagnosis) were assigned to a parenteral nutrition protocol designed to reduce the risk of refeeding syndrome. eTable 2 and eTable 3 present complete study parenteral nutrition protocols and Harris-Benedict equations, respectively.
Standard care was defined pragmatically and was not via protocol. The attending clinician selected the route, starting rate, metabolic targets, and composition of nutrition to be provided to patients receiving standard care based on current practice in their ICU. In Australia and New Zealand, it is not an accepted routine practice to infuse isolated glucose solutions to provide caloric intake prior to initiating enteral, parenteral, or oral nutrition.
For all enrolled patients, regardless of group, an insulin infusion was recommended if blood glucose levels exceeded 180 mg/dL to achieve peak glucose levels less than 180 mg/dL. (To convert glucose to mmol/L, multiply by 0.0555.)
The primary outcome was death by study day 60. Secondary outcomes, including quality of life and physical function measures (RAND-36 general health status and physical function15 plus Eastern Collaborative Oncology Group performance status16), were ascertained by unblinded site investigators using scripted telephone or face-to-face interviews at study day 60. Tertiary outcomes included days of clinically significant organ failure,17 infection rates,18,19 ICU and hospital stay, vital status at ICU and hospital discharge, days of invasive mechanical ventilation, days of renal replacement therapy, days of treatment for pressure ulcers, days of antibiotic usage, plus others.13
Twice each week while the patient was in the study ICU, trained site investigators assessed mid-arm muscle circumference, the Subjective Global Assessment (SGA)20 item scoring muscle wasting, and the SGA item scoring fat loss.21 These SGA items assess muscle wasting and fat loss on scales ranging from 1 to 4, with higher scores indicating greater muscle loss or fat loss (1, no obvious loss; 2, mild loss; 3, moderate loss; and 4, severe loss).20
Hospital radiologists' written reports were used to identify new or worsening infiltrates for use in the Clinical Pulmonary Infection Score.18 Hospital radiologists reviewed chest x-rays unaware of group assignment. The independent safety and data monitoring committee conducted 1 planned blinded interim analysis using Haybittle-Peto stopping rules at mid-recruitment.22,23
Data from the control group of a 27-hospital cluster randomized trial conducted in target participating sites was used to identify potentially eligible patients and estimate expected baseline mortality.8 Because meta-analyses of small randomized trials may overestimate treatment effects observed in subsequent larger confirmatory trials,24 the magnitude of the treatment effect reported in a previous meta-analysis on this topic6 was conservatively deflated by 45%.
Assuming an estimated baseline mortality of 29.7%, we calculated 1470 patients would need to be enrolled to provide 90% power to detect a 7.7% risk difference (RD) between groups.25 Complete details regarding the assumptions underlying the sample size calculation are provided in the statistical analysis plan.13
A detailed intention-to-treat analysis plan is published elsewhere.13 Crude (unadjusted) analysis of the effect of treatment on the primary outcome, and all other dichotomous outcomes, was conducted using an exact Pearson χ2 test, with unconditional exact 95% confidence intervals calculated around the RD and the odds ratio metrics.
Outcomes based on count data (eg, length of stay, days of clinically significant organ failure, etc) were analyzed using Poisson regression. If the scaled deviance exceeded 1.4 units per degree of freedom, a conservative negative-binomial model was used instead. An offset term (ICU length of stay) was used to account for time at risk where appropriate. Changes in body composition over time (mid-arm muscle circumference, SGA muscle, and SGA fat stores) were assessed using fully factorial nested analysis of variance.
A prespecified algorithm was used to identify baseline characteristics for inclusion in a covariate-adjusted logistic regression model to control confounding on the primary outcome.13 A covariate-adjusted average risk difference and appropriate 95% CI were also calculated.26 Four a priori defined subgroup analyses were conducted (BMI <18.5, BMI ≥30, age >70 years, and chronic insulin-treated diabetes) based on logistic regression tests of interaction.13 Missing data were accepted to be missing at random unless prespecified thresholds were exceeded. If thresholds were exceeded, missing values were imputed as study group–specific pooled mean values.13 Two-sided 5% significance levels were used to identify statistically significant results. All analyses were conducted in SAS version 9.2 (SAS Institute).
From October 19, 2006, to June 30, 2011, 1376 submissions were received by the study website from 31 participating ICUs throughout Australia and New Zealand (Figure 1). Three submissions represented duplicate patient data (1 internet failure, 2 attempts to correct data at time of enrollment), and 1 submission represented a patient enrolled on consecutive days by different research team members. All randomized patients were analyzed in the groups to which they were initially allocated except if consent was withdrawn.
Of the 1372 unique patients enrolled, 686 were randomized to receive pragmatic standard care and 686 were randomized to receive early parenteral nutrition. Nine patients withdrew consent (4 standard care and 5 early parenteral nutrition), and 5 could not be contacted at day 60 (2 standard care and 3 early parenteral nutrition). The mean time from ICU admission to enrollment was 13.8 hours (95% CI, 13.6-14.5, standard care, vs 95% CI, 13.2-14.1, early parenteral nutrition).
Table 1 presents baseline characteristics. Covariate-adjusted day-60 mortality was controlled for age, gender, BMI, Acute Physiology and Chronic Health Evaluation (APACHE) II score, chronic liver disease, chronic lung disease, and source of admission to ICU.
At least 1 component of the baseline APACHE II score was missing in 57 of 1363 patients (4.18%). No other baseline variables required imputation. Outcome variables were not imputed.
A mean of 1.98 days (95% CI, 1.43-2.78) after enrollment, 199 of 682 patients receiving standard care (29.2%) were given enteral nutrition. Of these 199 patients, 48 (24%) eventually received supplemental parenteral nutrition, which was commenced a mean of 5.58 days (95% CI, 3.9-7.96) after enteral nutrition. An additional 27.3% of patients receiving standard care (186/682) started parenteral nutrition first, with a mean time to start of 1.99 days (95% CI, 1.45-2.70) after enrollment. Enteral and parenteral nutrition was started at the same time in 19 of 682 standard care patients (2.8%), with a mean time from enrollment to start of 5.58 days (95% CI, 3.90-7.96). The remaining 278 of 682 standard care patients (40.8%) never received enteral or parenteral nutrition during their 3.72-day (95% CI, 2.65-5.20) ICU stay. Of all patients allocated to standard care, 298 of 682 patients (43.7%) received enteral nutrition at some time during their ICU stay. Thirteen standard care patients received enteral nutrition, parenteral nutrition, or oral intake within 24 hours of ICU admission. Overall, standard care patients remained unfed for a mean of 2.8 days after randomization (95% CI, 2.3-3.4). Standard care patients received 2.9 days of parenteral nutrition (95% CI, 2.7-3.1) and 4.0 days of enteral nutrition (95% CI, 3.6-4.6) while in the study ICU.
Two early parenteral nutrition patients never received any parenteral nutrition and 10 had parenteral nutrition started later than 24 hours after ICU admission. Parenteral nutrition was begun a mean of 44 minutes after randomization (95% CI, 36-55), with additional vitamin supplementation commencing 2.8 days after study parenteral nutrition initiation (95% CI, 2.7-3.0) and additional mineral/trace element supplementation starting 2.2 days after study parenteral nutrition initiation (95% CI, 2.1-2.3). Of all patients allocated to early parenteral nutrition, 274 of 681 patients (40.2%) received enteral nutrition at some time during their ICU stay. Early parenteral nutrition patients received a mean of 6.0 days of parenteral nutrition (95% CI, 5.6-6.4) and 3.1 days of enteral nutrition (95% CI, 2.8-3.5) while in the study ICU. There were no study intervention–related serious adverse events reported.
Figure 2 presents nutrition delivery over the first 7 ICU days. Additional information regarding nutrition therapy provided to patients in both groups of the trial is reported in eTable 4 and eFigure 1.
Crude day-60 mortality (22.8% [155/680] for standard care vs 21.5% [146/678] for early parenteral nutrition; RD, −1.3; 95% CI, −6.6% to 4.1%; P = .60) and covariate-adjusted day-60 mortality, controlled for age, gender, BMI, APACHE II score, chronic liver disease, chronic lung disease, and source of ICU admission, were not found to be significantly different between groups (adjusted RD, 0.0%; 95% CI, −4.2% to 4.3%; P = .99).
A statistically significant improvement in quality of life (RAND-36 General Health Status)15 (45.5 for standard care vs 49.8 for early parenteral nutrition; mean difference, 4.3; 95% CI, 0.95 to 7.58; P = .01) was detected in favor of patients receiving early parenteral nutrition; however, the magnitude of the difference did not exceed one-half the standard deviation,27 which was defined a priori as clinically meaningful to patients.13
Additional details regarding physical function are reported in Table 2.
Details regarding vital status at ICU discharge and hospital discharge, along with lengths of stay, are reported in Table 2.
Adjusted for time at risk (ICU stay), patients receiving early parenteral nutrition required significantly fewer days of invasive mechanical ventilation (−0.47 days per 10 patient × ICU days; 95% CI, −0.82 to −0.11; P = .01) and experienced fewer days with clinically significant coagulation failure17 (−0.34 days per 10 patient × ICU days; 95% CI, −0.57 to −0.08; P = .01). Expressed as crude day counts, the magnitude of these differences were 1.07 fewer days of invasive mechanical ventilation (95% CI, −1.77 to −0.29) and 0.43 fewer days of coagulation failure (95% CI, −0.62 to −0.21) attributable to early parenteral nutrition (eTable 5). There were no other significant differences (Table 3).
There were no significant differences between groups in rates of new infection (Table 4).
Standard care patients experienced significantly greater muscle wasting (0.43 vs 0.27 increase in SGA score per week; mean difference, 0.16; 95% CI, 0.038 to 0.28; P = .01) and significantly greater fat loss (0.44 vs 0.31 increase in SGA score per week; mean difference, 0.13; 95% CI, 0.01 to 0.25; P = .04) over the duration of their ICU stay. Changes in mid-arm muscle circumference were evident between groups by day 2 (0.2-cm loss for standard care vs 0.0-cm loss for early parenteral nutrition over 2 days; mean difference, −0.2; 95% CI, −0.39 to −0.01; P = .04); however, these early mid-arm muscle circumference differences did not remain significant over the entire ICU stay (0.8-cm loss vs 0.4-cm loss per week; mean difference, −0.4; 95% CI, −1.2 to 0.3; P = .28).
There were no significant differences in day-60 mortality between a priori defined subgroups (BMI <18.5, P = .75 for interaction; BMI ≥30, P = .66 for interaction; age >70 years, P = .15 for interaction; and chronic insulin-treated diabetes, P = .80 for interaction).
The provision of parenteral nutrition within 24 hours of ICU admission to critically ill patients with short-term relative contraindications to enteral nutrition did not result in significant differences in day-60 all-cause landmark mortality or ICU infection rates. Patients who received early parenteral nutrition required significantly fewer days of invasive mechanical ventilation, but this did not result in a statistically significant shortening of ICU or hospital length of stay. No harm was attributable to the use of early parenteral nutrition in this trial.
Patients who received early parenteral nutrition had significantly higher energy and amino acid/protein intakes on each of the first 6 days of ICU stay after study enrollment. Considering the magnitude of nutritional differences reported in previous trials8,28 and the intended target population,6 the lack of mortality effect cannot be attributed to unsuccessful implementation of the study intervention.
It is accepted that meta-analyses of small trials often overestimate treatment effects obtained from subsequent large-scale clinical trials.24 The most unbiased estimate of treatment effect, obtained from the covariate-adjusted model, suggests early parenteral nutrition may have zero effect on mortality. Use of early parenteral nutrition also had no significant effects on infection rates.
The Early PN Trial provides reliable and precise estimates of infection rates, including catheter-related infections.5 Previous clinical trials each used different definitions of infections, with some trials reporting only the total number of positive cultures by study group.6 The Early PN Trial used robust definitions of infections appropriate for a critically ill population,18 with subjective information (chest x-ray interpretation) obtained from blinded sources.29 Furthermore, the clinical importance of each infection was graded according to site and organism, with infections having an attributable excess case mortality greater than 15% pooled for analysis as major infections.19 Despite extensive reporting and analysis, there were no differences in any type of infection between groups.
Previous studies have shown critically ill patients lose 15.5% of their total body protein over the first 3 weeks of their ICU stay, with the majority of total body protein loss coming from skeletal muscle early during the ICU stay (up to 1.2% per day).30
We found early administration of parenteral nutrition may be protective against both muscle wasting and fat loss, with significant early benefits appearing to persist over the patient's entire ICU stay. Although preservation of muscle mass might be expected to translate into improved recovery of physical function, measures of physical function obtained 60 days after enrollment did not differ between groups. It is possible the RAND-36 is not sensitive to differences in early recovery, and we strongly recommend longer-term follow-up in future trials in this field.31 However, liberation from invasive mechanical ventilation may reflect early differences in recovery attributable to maintenance of muscle mass.
Diaphragmatic function is a major determinant of the ability to successfully wean from invasive mechanical ventilation.32 It is well known that mechanical ventilation itself has adverse effects on the structure and function of the diaphragm, with biopsies demonstrating marked atrophy of myofibers after only 18 hours of mechanical ventilation,33 characterized by increased catabolic activity.34 Furthermore, the specific catabolic changes observed in the diaphragm are known to be down-regulated by amino acids.35 It is possible that the overall preservation of muscle mass attributable to early parenteral nutrition resulted in some degree of preservation of diaphragmatic structure and function, leading to improved respiratory mechanics at time of weaning.36 We strongly recommend future studies investigate the veracity of this mechanistic hypothesis.
Objective and repeatable eligibility criteria were developed to identify patients with short-term relative contraindications to enteral nutrition within the first 24 hours of ICU admission. It was anticipated this trial would recruit a large proportion of critically ill patients who had multiple surgical procedures scheduled within the first 2 days of ICU admission and patients who were ordered by surgeons not to take oral food or fluids for 48 hours after major surgery. In both types of patients, initial short-term relative contraindications to early enteral nutrition could resolve over time, and under usual care conditions, it would be appropriate to commence enteral nutrition. Indeed, the finding that approximately 40% of patients in both groups eventually received delayed enteral nutrition suggests our pragmatic approach to standard care resulted in a comparison group that reflects real-world usual-care conditions: as appropriate, some standard care patients received delayed enteral nutrition, others received delayed parenteral nutrition, and some never received enteral or parenteral nutrition during their entire ICU stay.
The Early PN Trial differs from the other major trial in this field with regards to 2 key factors: the patient population targeted and the intervention studied. Whereas EPaNIC targeted critically ill patients regardless of adequacy of planned enteral intake,6 the Early PN Trial maintained a strict focus on a narrow subset of patients unlikely to receive any early enteral nutrition. Furthermore, it is important to understand that complete nutrition therapy did not commence in either group of the EPaNIC trial until ICU day 3, while the Early PN Trial commenced early parenteral nutrition on the first day of ICU care, relative to a standard care group commencing nutrition therapy on day 3 or later. Although the results of the Early PN Trial apply to a much narrower subset of patients than EPaNIC, the Early PN Trial does address an important question in this narrow subset concerning the timing of the provision of nutrition therapy: if my patient cannot receive early enteral nutrition, should I start feeding my patient immediately with parenteral nutrition?
The Early PN Trial was terminated when 100 patients short of its target of 1470 patients. The decision to terminate was based on financial considerations due to slower-than-expected recruitment and was made without knowledge of treatment effects.37 Because the 95% CIs for both the crude and covariate-adjusted estimates of treatment effect rule out the presence of a 7.7% RD mortality effect, we conclude it is unlikely that a loss of post hoc power due to early termination, or a lower-than-expected baseline mortality rate, explains our negative findings.38
Many aspects of trial conduct were not blinded (study intervention, outcome assessment), so it is possible that preexisting biases influenced concomitant treatment decisions and outcome ascertainment; however, a large body of literature expresses both positive6,9 and negative3,10 views on parenteral nutrition use. Based on numerous visits to, and discussions with, clinicians at all 31 study sites, we believe participating clinicians represented a broad spectrum of all possible views and opinions with no evidence of overall net positive or negative personal views on parenteral nutrition use.
A prespecified regression algorithm identified age, gender, BMI, APACHE II score, chronic liver disease, chronic lung disease, and source of ICU admission as potential confounders for inclusion in a covariate-adjusted mortality model. However, we do not recommend this subset of variables for use in prespecified covariate-adjusted models in future studies. There is small likelihood that this subset will completely capture confounding in future studies, and each individual study is likely to require adjustment with a unique subset of alternate confounders.
The major strengths of this trial lie in its rigorous design and conduct. A meaningful difference in nutrition therapy was achieved between groups, with minimal treatment crossovers and minimal loss to follow-up. Conduct at 31 sites throughout 2 countries enhances generalizability to similar health care settings.
The use of early parenteral nutrition did not result in significant differences in day-60 mortality or infection rates. Early parenteral nutrition resulted in a significant reduction in days of invasive mechanical ventilation, but this did not result in a significant shortening of ICU or hospital length of stay. No harm was associated with the use of early parenteral nutrition in this trial.
Corresponding Author: Gordon S. Doig, PhD, Royal North Shore Hospital, Intensive Care Unit, Pacific Hwy, St Leonards, NSW 2065, Australia (firstname.lastname@example.org).
Published Online: May 20, 2013. doi:10.1001/jama.2013.5124
Author Contributions: Dr Doig had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Doig, Simpson.
Acquisition of data: Doig, Simpson, Sweetman, Finfer, Heighes, Davies.
Analysis and interpretation of data: Doig, Simpson, Sweetman, Finfer, Cooper, Davies, Solano, Peake.
Drafting of the manuscript: Doig.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Doig.
Obtained funding: Doig, Simpson, Finfer, Cooper.
Administrative, technical, or material support: Doig, Simpson, Sweetman, Heighes.
Study supervision: Doig, Simpson, Sweetman, Heighes.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Doig reported receiving academic research grants from Fresenius Kabi Deutschland and Baxter Healthcare; consulting fees from Nestle Nutrition, Fresenius Kabi Deutschland, and Baxter Healthcare; speakers' honoraria from Nestle Nutrition, Baxter Healthcare, and Fresenius Kabi Deutschland; and travel support from Fresenius Kabi Deutschland, Baxter Healthcare, and Nestle Nutrition. Ms Simpson reported receiving academic research grants from Fresenius Kabi Deutschland and Baxter Healthcare, speakers' honoraria from Fresenius Kabi, and travel support from Fresenius Kabi Deutschland and Baxter Healthcare. Dr Solano reported receiving consulting fees from Pfizer and Astra and conference attendance support from Pfizer. Dr Finfer reported receiving academic research grants from Fresenius Kabi Deutschland and Fresenius Kabi Australia and consultancy fees from Esai and Eli Lilly. Ms Heighes reported receiving academic research grants and travel support from Fresenius Kabi Deutschland and Baxter Healthcare. Dr O’Leary reported receiving speakers' honoraria from Baxter Healthcare. Ms Sweetman reported receiving academic research grants from Fresenius Kabi Deutschland and Baxter Healthcare. No other disclosures were reported.
Funding/Support: This work was supported by a peer-reviewed academic grant from the Australian National Health and Medical Research Council (NH&MRC) and unrestricted academic grants from Fresenius Kabi Deutschland and Baxter Healthcare. Fresenius Kabi Australia supplied the study parenteral nutrition.
Role of the Sponsor: As a peer review funding body, the NH&MRC provided constructive comments on the study design. Fresenius Kabi Deutschland, Fresenius Kabi Australia, and Baxter Healthcare had no role in the design and conduct of the study; in the collection, analysis, and interpretation of the data; or in the preparation, review, or approval of the manuscript. Although participating sites were compensated for the costs of conducting the trial, site investigators did not receive financial compensation for their contributions.
Early PN Trial Contributing Sites and Site Investigators, alphabetical by site:Auckland City Hospital, New Zealand: Jodi Brown, Heidi Buhr, Vicki Cochrane, Michelle Eccleston, Eileen Gilder, Shay McGuiness, Rachael Parke, Anna Whitley. Austin Hospital, Victoria, Australia: Rinaldo Bellomo, Glenn Eastwood, Donna Goldsmith, Inga Mercer, Kim O’Sullivan, Leah Peck, Helen Young. Bendigo Hospital, Victoria: Catherine Boschert, John Edington, Jason Fletcher, Gary Koch, Mainak Majumdar, Tracey Shard, Julie Smith. Blacktown Hospital, New South Wales, Australia: Kalpesh Gandhi, Kiran Nand, Treena Sara. Box Hill Hospital, Victoria: David Charlesworth, Suzanne Eliott, David Ernest, Angela Hamilton (deceased), Belinda Howe, Inga Mercer, Sam Radford, Jaspreet Sidhu. Cabrini Hospital, Victoria: Jonathon Barrett, Felicity Hawker, MariaGrazia de Luca. Calvary Mater Hospital Newcastle, New South Wales: Irene Bailey, Jorge Brieva, Katrina Ellem. Campbelltown Hospital, New South Wales: Gillian Bishop, Olivia Mulligan, Ray Eckhardt. Concord Hospital, New South Wales: David Milliss, Helen Wong. Dandenong Hospital, Victoria: Subhash Arora, Michael Buist, Bridget O’Bree, Kate Shepherd, Susan Van Dyk. Frankston Hospital, Victoria: Sharon Allsop, Subhash Arora, John Botha, Himangsu Gangopadhyay, David Lewis, Naomi Pratt, Fiona Turnbull, Jodi Vuat. Geelong Hospital, Victoria: Allison Bone, Claire Cattigan, Tania Elderkin, Melissa Fraser, Anne Kilmonth, Neil Orford, Tania Salerno. Gold Coast Hospital, Queensland, Australia: Alan Spencer, Mandy Tallott, Rosemary Whitbread. Gosford Hospital, New South Wales: Rob Cameron, Sheridan Hatter, Jackie Hyslop, Peter Rye. John Flynn Private Hospital, Queensland: Robin Holland, Roslyn van der Vooren. John Hunter Hospital, New South Wales: Elise Crowfoot, Miranda Hardie, Peter Harrigan, Sam Jenkins. Liverpool Hospital, New South Wales: Deepak Bhonagiri, Sharon Micallef, Michael Parr. Lyell McEwin Hospital, South Australia, Australia: Rajaram Ramadoss, Josette Wood, Julie Zuppa. Middlemore Hospital, New Zealand: Marilyn Beggs, Peter Dzendrowskyj, Chantal Hogan, Judy Tai, Anna Tilsley, Tony Williams. Monash Medical Centre, Victoria: Jonathon Barrett, Sue Burton, Tim Crozier, Pauline Galt, Ainsley Gillies, Rebecca Ioannidis, Marnie Reilly, Carly Thornhill. Nepean Hospital, New South Wales: Cheryl Cuzner, Rebecca Gresham, Larissa Hoyling, Tony Maclean, Maria Nikas, Phoebe Palejs, Ian Seppelt, Leonie Weisbrodt, Sarah Whereat. Royal North Shore Hospital, New South Wales: Anthony Delaney, Gwen Hickey. Royal Hobart Hospital, Tasmania, Australia: David Cooper, Kathryn Marsden, Rick McAllister, Ram Sistla, Andrew Turner. St George Hospital, New South Wales: Vanessa Dhiacou, Deb Inskip, Theresa Jacques, Alina Jovanovska, Michael O’Leary, Rebecca Sidoli. St Vincent's Hospital Melbourne, Victoria: Nicole Groves, Jenny Holmes, John Santamaria, Roger Smith, Antony Tobin. St Vincent's Hospital Sydney, New South Wales, Australia: Jeff Breeding, Priya Nair, Claire Reynolds, Karen Storer. Sydney Adventist Hospital, New South Wales: Roger Harris, Linley Shields, Hui (Whay) Yang. The Prince of Wales Hospital, New South Wales: Frances Bass, Michelle Campbell, Pam Edhouse, Naomi Hammond, Maryam Sana, Yahya Shehabi, Victoria Stockdale, Barb Trytko. The Queen Elizabeth Hospital, South Australia : Catherine Kurenda, Sandra Peake, Patricia Williams. Wellington Hospital, New Zealand: Lynn Andrews, Dick Dinsdale, Peter Hicks, Diane Mackle. Wollongong Hospital, New South Wales: Michael Davis, Michelle Gales, Francisco Hill, Bronwyn Johnson, Adam Purdon, Martin Sterba, Renee Xu.