January 2015 E-Journal Club


We do not host a traineeship program in January in order to avoid any weather-related travel misadventures, but now that the holidays are over, we are back on track with our journal club.  Our article was a recent one about the effects of Omega-3 fatty acids in parenteral lipids.  Although we may not have much of a choice with lipid emulsions in the United States (yet), I keep hearing rumors that FDA consideration of new lipid products are in the works.

January Citation:

Grau-Carmona T, Bonet-Saris A, García-de-Lorenzo A, et al.  Influence of n-3 Polyunsaturated Fatty Acids Enriched Lipid Emulsions on Nosocomial Infections and Clinical Outcomes in Critically Ill Patients: ICU Lipids Study.  Crit Care Med. 2015 Jan;43(1):31-39.


This was a multicenter, double-blind, randomized controlled study of the effect of replacing 10% of the long-chain triglycerides (LCT) in a MCT/LCT lipid emulsion with Omega-3 fish oils in a mixed medical/surgical adult ICU population.  Patients who were expected to receive parenteral nutrition (PN) for at least 5 days were randomized to receive 40% of their calorie requirements as either a 50% MCT/50% LCT lipid emulsion or 50% MCT/40% LCT/ 10% fish oil lipid emulsion.  Total calories were estimated at 25 kcals/kg for nonseptic, 30 kcals/kg for septic patients, with 1.5 gm protein/kg and a lipids up to 1.5 g/kg. Enteral nutrition with standard formulas, up to 50% of daily calorie expenditure was allowed.

The primary outcome of the study was the prevalence of nosocomial infections during 28 days of ICU stay (from the initiation of TPN, finalized before day 28 in the event of death or ICU discharge). Time free of infection (TFI) was calculated as timeframe between the first treatment day and the onset of the first episode of nosocomial infection, death, or ICU discharge (maximal TFI, 28 d).  Antibiotic-free time was also recorded.

Secondary endpoints were ICU mortality, length of ICU stay, days of MV, nutritional efficacy, and liver function.  Patients were followed 6 months after ICU discharge for length of hospital stay, hospital mortality, and 6-month mortality.

Inclusion and Exclusion Criteria were:

Inclusion criteria:

Adults admitted to the ICU with an APACHE II score ≥ 13, who were expected to require PN for at least 5 days according to the ASPEN guidelines.

Exclusion criteria:

Malignant disease with metastasis and life expectancy < 6 months, Morbid obesity (BMI ≥ 39), Liver disease/failure, Chronic renal insufficiency, Severe hyperlipidemia, Recent TPN administration, Propofol (or other lipid-based medication) infusion > 24 hours, Chemo or radiation therapy within the previous month, Chronic corticosteroids in the past month (prior to admission only), Severe chronic neurological disease, Infectious diseases transmitted by blood products, or urine (hepatitis B, C or HIV), Pregnancy, Enrolled in another clinical trial, Refusal to participate.

Major Results:

A total of 3,610 ICU patients who required initiation of PN were evaluated, with 175 patients randomized. Seven patients withdrew informed consent and 9 were withdrawn by the investigators after randomization. There were 159 patients included in what was called the “intent-to-treat” analysis: 81 fish oil, 78 control (4.4% of all patients that received PN).  Further, there were 23 patients that received PN for less than 5 days, with ultimately only 59 control and 58 fish oil patients that completed the full protocol.  The demographic and illness characteristics between the fish oil and control groups were similar in most factors, except that the fish oil group had nearly 3X as many patients (n=14) with pancreatitis, compared to the control group (n=5).

EN was started at some point in 53% of the fish oil and 55% of the control patients, but PN was administered for 78% of the days on nutrition support in both groups.  Average parenteral lipids actually received was just over a gram/kg in both groups. PN provided an average of 23.4 kcals/kg in the control and 22.7 kcals/kg in the fish oil group.  The amount of lipid and calorie provision actually received from EN were not reported.

There were significantly fewer patients with a nosocomial infection in the fish oil group (17 of 81 patients, 21%) compared to the control group (29 of 78 patients, 37.2%) p= 0.04.  The nosocomial infection risk of patients receiving the MCT/LCT/FO lipid emulsion was reduced by more than 40% compared with the risk of patients receiving the standard MCT/LCT lipid emulsion (risk ratio [RR] = 1.8; 95% CI, 1.06–2.96).

In the subgroup of patients that were without signs of infection on the first and second day, 13/71 (18.3%) in the fish oil and 25/68 (36.8%) in the control group experienced episodes of nosocomial infections (p = 0.022). Time free of infection (TFI) in the subgroup without signs of infection on the first and second day was significantly longer: 21 ± 2 days in the fish oil group compared to 16 ± 2 days in the control group (p = 0.03).

Length of mechanical ventilation and ICU and hospital stay were numerically less, but not statistically different between the groups. There were no statistically significant differences for ICU, hospital, and 6-month mortality. Patients with pancreatitis receiving fish oil had a nonsignificant, but clinically relevant mortality (20% in the MCT/LCT group vs 50% in the FO group). Mortality from pancreatitis represented the 27% of the overall mortality in the FO group.

Author’s Conclusions:

Administration of ~0.1 g fish oil/kg BW per day in combination with MCT and LCT in a lipid emulsion reduces the risk of nosocomial infections and increases the predicted time to first infection in critically ill medical and surgical ICU patients.


The strengths of this study include the multi-center and double-blind methodology.  One of the weaknesses of this study is the relatively small number of subjects enrolled, because the study was terminated early due to slow recruitment and less than anticipated incidence of overall nosocomial infections.

The relatively small number of subjects may have led to an imbalance between the fish oil and control group right from the start, because there were significantly more patients with pancreatitis in the fish oil group.  Another possible contributor to an imbalance in patients was the use of a randomization process that stratified patients by APACHE score and presence of sepsis.  Forcing the allocation of the randomization process for these 2 known factors may have contributed to an imbalance in the groups with some other unknown, or less appreciated factor(s).

Although the total number of patients with pancreatitis is small (5 control, 14 fish oil), the fact that there was 50% mortality of patients that had pancreatitis who received fish oil is concerning.  It would take a much larger study of parenteral fish oil in pancreatitis to know if this morality was just a random event, or truly an adverse effect of parenteral fish oil during pancreatitis.

We discussed how the macronutrient breakdown in this study differed from our usual practice, and thus the results may not apply to our patients.  The investigators gave 40% of calories as fat, while we generally provide < 25% of parenteral calories from lipid (even when not in the midst of a shortage).  The average amount of parenteral lipid actually received was in the range of 1 gm/kg, but many patients received additional lipids from enteral nutrition (amounts not disclosed).

The fact that more than ½ of the patients in this study received EN also suggests that these patients are unlike our population, because it would be rare for us to begin PN in patients likely to tolerate EN so quickly, and we would have stopped PN in patients able to receive EN.  The amount of enteral calories and lipid actually received should have been disclosed in this study.

Overall, due to exclusion criteria and inability to obtain timely consent, only 4.4% of patients who received PN were enrolled.  It is unclear if these results can be extrapolated to the average ICU patient that receives PN.

Although the authors reported that intention to treat analysis was used, ultimately, not everyone who was randomized was analyzed.  The investigators did not analyze data from subjects that were withdrawn by the investigators, so this was not true intention to treat analysis.

Our Take Home Message (s)

1.   Parenteral fish oil as a component of parenteral lipid administration during lipid-rich PN may reduce the incidence of nosocomial infections compared to MCT/LCT lipid emulsion.

2.   Larger studies with true randomization and intention to treat analysis will be required before we can know the full potential and risks of parenteral fish oil as a component of PN.

3.   When enteral feedings are used in a nutrition study, the details of the actual amount of nutrition provided should be reported.

Other News on the UVAHS GI Nutrition Website: (www.ginutrition.virginia.edu):

–Weekend Warrior Mini-Nutrition Support Traineeship—March 21-22, 2015

–Openings available in the April 13-17, 2015 week-long Nutrition Support traineeship

–2015 webinars to be announced shortly

Check out What’s New:

–“Nutrition Support Blog”

–“ Resources for the Nutrition Support Clinician

Latest Practical Gastroenterology article:

Had a break in January 🙂

Joe Krenitsky MS, RD

Carol Rees Parrish MS, RD

PS – Please feel free to forward on to friends and colleague