Omega-6 sparing effects of parenteral lipid emulsions—an updated systematic review and meta-analysis on clinical outcomes in critically ill patients

Critical illness is often characterized by an imbalanced immune response, which can lead to excessive cytokine release and accumulation of reactive oxygen species. Systemic inflammation and oxidative stress ultimately result in tissue damage, multi organ failure and high mortality rates [1‐4]. Parenteral lipid emulsions provide fatty acids (FAs) as a source of calories and cellular building blocks [5] and have been under investigation because of their immunomodulating features [6]. Traditionally, parenteral lipid emulsions are derived from plant—and especially soybean oil (SO) to provide the patient with essential long-chain triglycerides (LCTs). However, SO’s ratio of omega-6 polyunsaturated FAs to omega-3 polyunsaturated FAs (7:1) is being regarded critically [6]. In vivo, linoleic acid (omega-6) is converted to arachidonic acid and pro-inflammatory eicosanoids, such as prostaglandins, thromboxanes and leukotrienes. On the other hand, α-linolenic acid (omega-3) is the metabolic precursor of docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA), which both have anti-inflammatory and anti-oxidative properties like inhibitory effects on various innate and adaptive immune cells and the transcription of inflammatory cytokines [7‐9]. Unfortunately, the pathway of α-linolenic acid conversion to DHA and EPA appears to be highly ineffective in most humans [10] and an abundance of omega-6 FAs will furthermore suppress the balancing effects of omega-3 FA conversion, as both FAs (omega-3 and omega-6) compete for the same enzymes [11]. Therefore, omega-6 FA reducing strategies have been introduced to clinical nutrition. A reduction in omega-6 FAs can be achieved with the addition of medium-chain triglycerides (MCT), olive oil (OO) or fish oil (FO) to SO-based lipid emulsions. FO supplements, as a major source of DHA and EPA, can simultaneously provide adequate levels of omega-3 FAs, which is regarded as an attractive immunomodulatory treatment option with potential clinical benefits [12]. Hitherto, meta-analyses solely focused on the FO aspect, resulting in limited evidence regarding the overall picture of omega-6 FA reduction. After aggregating 10 randomized controlled trials (RCTs) in 2015, Manzanares et al. found that lipid emulsions with a FO component may reduce infectious complications, duration of mechanical ventilation (MV) and hospital length of stay (LOS) in critically ill patients [13]. In the most recent meta-analysis by Pradelli et al., FO containing parenteral nutrition (PN) again reduced the rate of infections and hospital as well as ICU LOS [14]. Current nutrition guidelines state that DHA and EPA (FO dose of 0.1–0.2 g/kg/d) can be provided in patients receiving PN. Ambiguous results regarding other omega-6 sparing strategies are delineated, however without a clear recommendation [15, 16]. In recent years, several new RCTs on omega-6 sparing effects in general and on FO-containing lipid emulsions in particular have been published. This systematic review and meta-analysis aims to give a broad and comprehensive update on the emerging topic.

Overall, our systematic review and meta-analysis covered 26 trials in critically ill patients, of which 16 studies with 1046 patients were newly included in comparison with the previous meta-analysis by Manzanares et al. [13]. We hereby provide an update and additional insights not only on FO but also on general omega-6 FA reducing strategies in intensive care. Omega-6 sparing effects included a significant decrease in hospital LOS and trends towards reduction in 28-day mortality, ICU LOS and mechanical ventilation. Among different omega-6 FA reducing PN regimens, FO containing lipid emulsions reduced the length of intensive care and rate of nosocomial infections. Potential signals on mortality rates were observed with stand-alone use of FO, which encourages further research in this area.

In the past decades, clinical and technological advances promoting physiologic and metabolic resuscitation as well as early diagnosis and optimal treatment have significantly decreased hospital mortality in critical illness. In this context, nutritionally derived compounds exhibiting pharmacological effects contribute to an improved clinical outcome [46]. However, due to low quality of evidence, the potential benefits of anti-inflammatory and immunomodulating effects in critically ill patients remain unproven [47]. While a balance between pro- and anti-inflammatory response mechanisms is crucial for an adequate immune response, critically ill patients often exhibit immune dysregulation [48‐50]. Imbalance arises from either excessive production of pro-inflammatory cytokines and reactive oxygen species or a lack of their physiological anti-inflammatory and anti-oxidative counterparts. A similar concept applies for omega-6 and omega-3 FAs or rather their metabolites. By reducing prostaglandins, leukotrienes and thromboxanes via MCT, OO or FO on the one hand and increasing DHA and EPA levels by FO on the other, pharmaconutrition might play a role in regaining inflammatory balance in critically ill patients. This basic concept is tempting, even though clinical evidence is scarce and ambiguous. There are—if at all—only minor effects of sole omega-6 FA reduction on inflammatory parameters and markers of oxidative stress [42, 51, 52], which is in line with our results, where the impact of SO/MCT and SO/OO on the clinical course was rather small. However, an RCT, in which OO was significantly associated with reduced ICU LOS and MV, had to be excluded, due to an unusual lipid- versus glucose-based treatment regimen [53]. In the case of FO, omega-6 sparing effects are complemented by a simultaneous increase in omega-3 FAs, which may be even more beneficial for inflammatory control.

In fact, FO has been studied for decades, after realizing that Inuits in Greenland with a fish-based diet had low rates of cardiovascular events [54]. This led to general recommendations for FO supplements in patients with coronary heart disease [55] and multiple promising investigations in autoimmune and inflammatory disorders [56‐58]. Intravenous FO administration immediately affects physiological parameters, which was demonstrated by a significant increase in DHA and EPA concentrations in blood cell phospholipids within 60 min after infusion [59]. The rapid incorporation into various cellular membranes is crucial for an acute modulation of the immune response and suppression of pro-inflammatory cytokine release [60‐62]. In addition, it has been recently discovered that DHA- and EPA-derived metabolites like resolvins, protectins, and maresins are required for the resolution of inflammation in the post-acute phase [7], as they regulate neutrophil recruitment and initiate macrophages to clear apoptotic cells and microorganisms [63].

Despite significant bias risk and rather small patient populations in the underlying studies [64‐67], ESPEN guidelines state that parenteral FO can be provided for critically ill patients, as it likely decreases LOS and infections [16]. These conclusions align nicely with our results, including potential benefits on 28-day mortality and duration of MV. Hence, omega-6 FA reduction seems to hasten the recovery of patients requiring PN, which might justify the higher cost of these lipid emulsions in an economic evaluation. Pradelli et al. have recently conducted a systematic review and cost-effectiveness analysis on the use of omega-3 containing parenteral nutrition in ICU patients, which clearly demonstrated considerable cost savings due to a significant reduction in the risk of infections, and length of hospital or ICU stay. One main difference to our meta-analysis is that the authors included surgical patients without critical illness, for example postoperative surveillance after elective and cancer surgery [14]. Patterns of inflammation and the extent of organ injury might be very different in surgical patients with a defined tissue trauma, compared to systemic inflammation due to infection and sepsis. Our approach potentially provides a more homogenous collective of critically ill patients with a strong focus on sepsis and acute respiratory distress syndrome as well as additional insights into further omega-6 FA reducing strategies beyond FO. In comparison with Manzanares et al., we also excluded enteral nutrition protocols to further reduce heterogeneity [13]. Despite the distinctive differences between the meta-analyses, it is reassuring, that the results compare well.

We observe a trend towards improved 28-day survival, both for overall omega-6 FA reducing strategies as well as for FO containing PN alone. However, this potential signal could not be confirmed in terms of overall mortality. It has to be said, that multiple other factors and treatments beyond PN influence survival in critically ill patients, so that it is often difficult to interpret observed mortality benefits of a single intervention in modern ICU settings [68, 69]. In this light, the significant mortality benefits suggested for Omegaven and FO as a stand-alone medication have to be considered carefully.

We acknowledge several limitations of our meta-analysis. First, only five trials fulfilled the criteria of a “level I” study, which are concealed randomization, blinded outcome adjudication and intention-to-treat analysis. The median methodological score of all included studies was 9 out of 14 points, highlighting a considerable bias risk and mirroring the mediocre quality of most trials in the field of immunonutrition. This is a well-known problem in ICU research, as authors are faced with challenges of resources, recruitment and early randomization [70]. Second, the number of included studies and overall sample sizes for certain endpoints are too small to draw strong conclusions. Patient populations are still rather heterogeneous, despite our efforts to reduce variety. Third, overall mortality is an unsharp endpoint. Aggregating as many trials as possible comes at the price of including the whole variety of reported mortalities. Four studies for example only reported “other mortality” with a range from 15 to 109 days. Last, all included studies predominantly focused on “hard outcomes” and neglected the patient centered perspective. This requires more attention in future research.

On the other hand, this systematic review also has distinctive strengths. In contrast to previous meta-analyses, which merely focused on FO, we integrated additional omega-6 FA reducing PN strategies like SO/MCT and SO/OO [13, 14]. We only included RCTs with a relative specific patient population of critically ill ICU patients to reduce heterogeneity. We were not limited to any language barriers so that all internationally available evidence could be considered. Published abstracts, which did not undergo peer-review, were considered for inclusion after careful assessment. This approach might potentially reduce publication bias, we do, however, acknowledge the ongoing controversy on this topic [71‐73]. We report the largest and the most current aggregation of PN lipid trials conducted in the critical care setting. To guarantee the reproducibility of this systematic review and meta-analysis, study eligibility decisions are provided in detail and data were abstracted using two independent experienced reviewers and referees.

Taken together, the present meta-analysis in critically ill patients suggests omega-6 sparing effects attributed to parenteral immunonutrition. Among different lipid emulsions, clinical benefits were most pronounced for FO, either in combination with PN or as a stand-alone treatment.