Omega-3 fatty acid-containing parenteral nutrition in ICU patients: systematic review with meta-analysis and cost-effectiveness analysis

Abstract

Background: Omega-3 (ω-3) fatty acid (FA)-containing parenteral nutrition (PN) is associated with significant improvements in patient outcomes compared with standard PN regimens without ω-3 FA lipid emulsions. Here, we evaluate the impact of ω-3 FA-containing PN versus standard PN on clinical outcomes and costs in adult intensive care unit (ICU) patients using a meta-analysis and subsequent cost-effectiveness analysis from the perspective of a hospital operating in five European countries (France, Germany, Italy, Spain, UK) and the US.

Methods: We present a pharmacoeconomic simulation based on a systematic literature review with meta-analysis. Clinical outcomes and costs comparing ω-3 FA-containing PN with standard PN were evaluated in adult ICU patients eligible to receive PN covering at least 70% of their total energy requirements and in the subgroup of critically ill ICU patients (mean ICU stay > 48 h). The meta-analysis with the co-primary outcomes of infection rate and mortality rate was based on randomized controlled trial data retrieved via a systematic literature review; resulting efficacy data were subsequently employed in country-specific cost-effectiveness analyses.

Results: In adult ICU patients, ω-3 FA-containing PN versus standard PN was associated with significant reductions in the relative risk (RR) of infection (RR 0.62; 95% CI 0.45, 0.86; p = 0.004), hospital length of stay (HLOS) (- 3.05 days; 95% CI - 5.03, - 1.07; p = 0.003) and ICU length of stay (LOS) (- 1.89 days; 95% CI - 3.33, - 0.45; p = 0.01). In critically ill ICU patients, ω-3 FA-containing PN was associated with similar reductions in infection rates (RR 0.65; 95% CI 0.46, 0.94; p = 0.02), HLOS (- 3.98 days; 95% CI - 6.90, - 1.06; p = 0.008) and ICU LOS (- 2.14 days; 95% CI - 3.89, - 0.40; p = 0.02). Overall hospital episode costs were reduced in all six countries using ω-3 FA-containing PN compared to standard PN, ranging from €-3156 ± 1404 in Spain to €-9586 ± 4157 in the US.

Conclusion: These analyses demonstrate that ω-3 FA-containing PN is associated with statistically and clinically significant improvement in patient outcomes. Its use is also predicted to yield cost savings compared to standard PN, rendering ω-3 FA-containing PN an attractive cost-saving alternative across different health care systems.

Study registration: PROSPERO CRD42019129311.

Keywords: Cost-effectiveness; Critically ill; Intensive care; Meta-analysis; Omega-3 fatty acid; Parenteral nutrition.

Fig. 1

Infection rates in a all ICU patients and b critically ill ICU patients. Risk of bias legend: +: high risk; −; low risk; blank: risk unclear. A: random sequence generation (selection bias); B: allocation concealment (selection bias); C: blinding of participants and personnel (performance bias); D: blinding outcome data (attrition bias); E: incomplete outcome data (attrition bias); F: selective reporting (reporting bias); G: other bias. CI, confidence interval; FA, fatty acid; ICU, intensive care unit; ω-3, omega-3; M–H, Mantel–Haenszel study weighting; PN, parenteral nutrition

Fig. 2

Sepsis in ICU patients. Risk of bias legend: +: high risk; −; low risk; blank: risk unclear. A: random sequence generation (selection bias); B: allocation concealment (selection bias); C: blinding of participants and personnel (performance bias); D: blinding outcome data (attrition bias); E: incomplete outcome data (attrition bias); F: selective reporting (reporting bias); G: other bias. CI, confidence interval; FA, fatty acid; ICU, intensive care unit; ω-3, omega-3; M–H, Mantel–Haenszel study weighting; PN, parenteral nutrition

Fig. 3

Thirty-day mortality in a all ICU patients and b critically ill ICU patients. Risk of bias legend: +: high risk; −; low risk; blank: risk unclear. A: random sequence generation (selection bias); B: allocation concealment (selection bias); C: blinding of participants and personnel (performance bias); D: blinding outcome data (attrition bias); E: incomplete outcome data (attrition bias); F: selective reporting (reporting bias); G: other bias. CI, confidence interval; FA, fatty acid; ICU, intensive care unit; ω-3, omega-3; M–H, Mantel–Haenszel study weighting; PN, parenteral nutrition

Fig. 4

HLOS in a all ICU patients and b critically ill ICU patients. Risk of bias legend: +: high risk; −; low risk; blank: risk unclear. A: random sequence generation (selection bias); B: allocation concealment (selection bias); C: blinding of participants and personnel (performance bias); D: blinding outcome data (attrition bias); E: incomplete outcome data (attrition bias); F: selective reporting (reporting bias); G: other bias. CI, confidence interval; FA, fatty acid; HLOS, hospital length of stay; ICU, intensive care unit; ω-3, omega-3; M–H, Mantel–Haenszel study weighting; PN, parenteral nutrition

Fig. 5

ICU LOS in a all ICU patients and b critically ill ICU patients. Risk of bias legend: +: high risk; −; low risk; blank: risk unclear. A: random sequence generation (selection bias); B: allocation concealment (selection bias); C: blinding of participants and personnel (performance bias); D: blinding outcome data (attrition bias); E: incomplete outcome data (attrition bias); F: selective reporting (reporting bias); G: other bias. CI, confidence interval; FA, fatty acid; ICU, intensive care unit; ω-3, omega-3; M–H, Mantel–Haenszel study weighting; PN, parenteral nutrition

Fig. 6

Length of mechanical ventilation in ICU patients. Risk of bias legend: +: high risk; −; low risk; blank: risk unclear. A: random sequence generation (selection bias); B: allocation concealment (selection bias); C: blinding of participants and personnel (performance bias); D: blinding outcome data (attrition bias); E: incomplete outcome data (attrition bias); F: selective reporting (reporting bias); G: other bias. CI, confidence interval; FA, fatty acid; ICU, intensive care unit; ω-3, omega-3; M–H, Mantel–Haenszel study weighting; PN, parenteral nutrition

Fig. 7

Scatterplots of 1000 ICER estimates in country-specific probabilistic sensitivity analyses. CI, confidence interval; ICER, incremental cost-effectiveness ratio; PSA, probabilistic sensitivity analysis; pt, patient