Parenteral lipid emulsions in the preterm infant: current issues and controversies

Abstract

Parenteral lipid emulsions are a necessary component of nutrition for extremely low gestational age newborns until adequate levels of enteral intake are established. Historically, Intralipid, a 100% soybean oil emulsion, has filled this role. Newer multicomponent lipid emulsions containing a mixture of other oils, including olive oil and fish oil, are now available as options, although the regulatory approval for use in neonates varies worldwide. When dosed at currently published recommendations, each of these lipid emulsions meets total fat and energy requirements without a risk of essential fatty acid deficiency. Thus, when choosing which lipid emulsion to provide, the answer must be based on the metabolic differences induced as a result of these fatty acid-rich emulsions and whether the emulsions provide a health advantage or pose a health risk. The questions of induced fatty acid profiles, health benefit and health risk are discussed sequentially for multicomponent lipid emulsions. Despite the growing acceptance of multicomponent lipid emulsions, there is concern regarding changes in blood fatty acid levels and potential health risk without strong evidence of benefit. There remains no ideal parenteral lipid emulsion option for the preterm infant. Standardising future animal and human studies in lipid delivery with the inclusion of lipid metabolism data will iteratively provide answers to inform the optimal lipid emulsion for the preterm infant.

INTRODUCTION

The intake and balance of lipids and fatty acids are central to fetal and neonatal development. The reduced adiposity of preterm infants and resultant low fatty acid reserves make them highly vulnerable to inadequate nutritional delivery and rapid development of postnatal deficits of long-chain polyunsaturated fatty acids (LCPUFAs). Deficits in these fatty acids are linked to poor short-term and long-term health outcomes.^1–3 Timely and adequate replenishment of lipids and fatty acids require both parenteral and enteral strategies and likely vary based on gestational age and clinical status. This review focuses on the choice of parenteral lipid emulsions after birth as part of maintenance nutritional delivery and does not discuss the parenteral lipid choices after established parenteral nutrition-associated cholestasis (PNAC) or intestinal failure-associated liver disease.

Parenteral lipid emulsions were first developed in the 1960s and were designed for critically ill adults, not preterm infants. In adults, parenterally administered lipids are converted into distal LCPUFAs and their terminal metabolites. In contrast, the immature metabolic system of the preterm infant is unable to adequately process the upstream lipids provided by these emulsions. This metabolic inadequacy, coupled with baseline requirements and ongoing consumption, leads to fatty acid deficits and imbalances that impact health and risk of disease. Prompt and safe delivery of a parenteral lipid emulsion soon after birth that is aligned with the unique requirements of the preterm infant is essential to meet their nutritional requirements and optimise health.

Multicomponent lipid emulsions contain a mixture of oils, including olive oil and fish oil. When dosed at published recommendations, each of these lipid emulsions meets total fat and energy delivery requirements without risking essential fatty acid deficiency (EFAD). When choosing which lipid emulsion to provide, the answer should be based on the metabolic differences induced by administration of these fatty acid-rich emulsions and whether these emulsions provide a health advantage or pose a health risk. The questions of induced fatty acid profiles, health benefit and health risk are discussed sequentially for multicomponent lipid emulsions.

CURRENTLY AVAILABLE PARENTERAL LIPID EMULSIONS OFFER A DIVERSE SPECTRUM OF FATTY ACID CONTENT

Intralipid (Fresenius Kabi, Uppsala, Sweden), a 100% soybean oil based lipid emulsion, is routinely administered to neonates worldwide and was approved for neonates by the US Food and Drug Administration (FDA) in 1981.^4,5 Soybean oil has a high content of the essential fatty acids α-linolenic acid (ALA, 18:3 n-3) and linoleic acid (LA, 18:2 n-6), which are necessary to prevent EFAD (table 1).^5,6 ClinOleic (outside the USA)/Clinolipid (Baxter SA, Lessines, Belgium), a soybean oil-containing emulsion, consists of 20% soybean oil and 80% olive oil with a reduced concentration of LA and ALA relative to Intralipid due to the higher proportion of olive oil (table 1).

Table 1.

Lipid emulsion composition

	Intralipid 20%5	ClinOleic/Clinolipid 20%5,6		Lipidem/Lipoplus 20%7	SMOFlipid 20%8	Omegaven 10%10
Oil source (%)
Soybean	100	20		40	30	0
Fish	0	0		10	15	100
Coconut	0	0		50	30	0
Olive	0	80		0	25	0
Fatty acid content (% of fatty acids)
n-6
Linoleic (18:2 n-6)	44–62	14–22	24.4	14–25
Arachidonic 20:4 n-6)	0	0	NR	1.0
n-3
α-Linolenic (18:3 n-3)	4–11	0.5–4.2		3.3	1.5–3.5	1.1
Eicosapentaenoic (20:5 n-3)	0	0		3.1	1.0–3.5	13–26
Docosahexaenoic (22:6 n-3)	0	0		2.1	1.0–3.5	14–27
ARA:DHA ratio*	--		--	--	1:3.5

^*Ratio calculated form the upper limits reported for both ARA and DHA. Data from references.^5,7,8,10

ARA, arachidonic acid; DHA, docosahexaenoic acid; NR, not reported.

Lipidem/Lipoplus (B. Braun, Melsungen, Germany) contains 40% soybean oil, 50% coconut oil and 10% fish oil (table 1).⁷ SMOFlipid (Fresenius Kabi, Bad Homburg, Germany) was FDA approved for use in adults in 2016 although off-label use does occur in neonatal intensive care units (NICUs) across the USA.⁸ It is also used in NICUs across Canada, Europe and Asia.⁹ SMOFlipid contains 30% soybean oil, 30% medium-chain triglycerides (as coconut oil), 25% olive oil and 15% fish oil (table 1).⁸ Relative to Intralipid, SMOFlipid provides higher concentrations of arachidonic acid (ARA, 20:4 n-6) and docosahexaenoic acid (DHA, 22:6 n-3), as well as the n-3 fatty acid eicosapentaenoic acid (EPA, C20:5 n-3) with an ARA:DHA ratio of 1.0:3.5 (table 1).

Omegaven (Fresenius Kabi) is a 100% fish oil emulsion approved by the US FDA in 2018 for paediatric patients with PNAC.¹⁰ Omegaven has the highest levels of DHA, ARA and EPA (table 1). The ARA:DHA ratio of 1.0:13.5 is highly skewed towards DHA. For the remainder of the article, the term ‘fish oil-containing’ refers to lipid emulsions that have a non-dominant component of fish oil, which includes SMOFlipid and Lipidem/Lipoplus.

The increasing array of parenteral lipid emulsions and the growing appreciation of the role of fatty acids in neonatal health prompted a rapid increase in clinical studies evaluating the impact of lipid emulsions on neonatal outcomes. Many of these clinical studies vary in study design and in outcome metrics, which leads to difficulty comparing studies and interpreting results. In our analysis of primarily human studies, we will pose a series of questions followed by a review of the literature and then a conclusion to help guide both practice and design of future studies.

Question 1: do currently available parenteral lipid emulsions prevent acquired deficits in DHA and ARA in preterm infants?

Preterm infants rapidly develop circulating deficits in critical LCPUFAs such as DHA and ARA.^1,11,12 These postnatally acquired deficits occur regardless of the type of parenteral lipid emulsion used. One of the original studies to make this observation was a small study of moderate preterm infants almost 30 years ago.¹¹ In this cohort, in infants receiving advancing breast milk or formula feedings and Intralipid, the percentage of DHA and ARA in plasma choline phosphoglycerides fell by 40%–50% by the second postnatal week. Almost 20 years later, an analysis of whole blood fatty acid levels within the first postnatal week for infants receiving early enteral nutrition and Intralipid demonstrated that preterm infants, by the first postnatal week, experience a 30% decline in ARA and a 40% decline in DHA accompanied by a 2.7-fold increase in LA and a 19.2-fold increase in ALA.¹ The accumulation of these essential fatty acids, in conjunction with the decrease in the downstream fatty acids DHA and ARA, is multifactorial and possibly due to a combination of inadequate stores,¹³ increased demand and immature enzymatic pathways of fatty acid metabolism (figure 1) combined with insufficient fatty acid content in current parenteral and enteral diets.¹⁴ In select studies, it has been shown that preterm infants can convert LA and ALA into plasma phospholipid ARA and DHA, respectively^15,16; however, the adequacy of these enzymatic pathways in meeting the LCPUFA requirements of preterm infants has not been proven and is complicated by the fact that these requirements, including usage rates, remain undefined in critically ill infants.

Figure 1.

Enzymatic pathways in n-6 and n-3 fatty acid metabolism.

Recognising the importance of LCPUFAs in fetal and neonatal development, combined with the link between postnatal deficits and an increased risk of neonatal morbidities,^1,2,17 as well as the lack of DHA and ARA in Intralipid (table 1), made the neonatal community ripe for accepting alternatives in parenteral lipid emulsions. Unfortunately, the newer generation of parenteral emulsions containing fish oil (DHA and EPA) and ARA fails to make significant improvements in postnatal fatty acid levels in preterm infants.

Studies quantifying whole blood,¹⁸ plasma¹⁹ and plasma phospholipids,²⁰ fatty acid levels in neonates receiving 100% soybean oil, soybean plus olive oil or fish oil-containing lipid emulsions have identified consistent changes relative to birth levels (table 2). All emulsions lead to an increase in LA; however, fish oil-containing lipid emulsions increase levels of LA less dramatically than soybean oil emulsions, consistent with a dose–response effect in relation to the proportion of soybean oil present. None prevent the postnatal ARA deficit with fish oil-containing lipid emulsions increasing this deficit compared with soybean oil emulsions. Although fish oil-containing emulsions increase the postnatal DHA levels above what is observed with soybean oil lipid emulsions, none prevent the postnatal deficits. EPA levels from birth remain relatively constant with soybean oil emulsions, while these levels substantially increase by an average of 3.5-fold with the use of fish oil-containing lipid emulsions.^18–21 As a result of these changes, fish oil-containing lipid emulsions decrease the n-6:n-3 ratios, including the ARA:DHA ratio in preterm infants when compared with birth levels, leading to a divergence from the ratios found in utero and at birth for infants born to mothers consuming Western diets.¹³

Table 2.

Qualitative summary of changes in blood fatty acid levels early postnatal levels in neonates receiving soybean oil and fish oil-containing emulsions

Fatty acids	Intralipid	ClinOleic Clinolipid	SMOFlipid
n-6
Linoleic	⇈	↑	↑
ARA	↓	↓	⇊
n-3
DHA	⇊	⇊	↓
EPA	--/↓	--/↓	⇈
ARA:DHA ratio	--/↑	--/↑	⇊

Number of arrows indicates the qualitative degree of increase (↑) or decrease (↓) in fatty acid from baseline from values reported in the literature.^18–24 -- indicates no change from baseline. Baseline indicates birth or very early postnatal levels.

ARA, arachidonic acid; DHA, docosahexaenoic acid; EPA, eicosapentaenoic acid.

Studies measuring fatty acid levels in the serum,²¹ plasma^22,23 and red blood cell membranes²⁴ of infants receiving soybean oil/olive oil emulsions (ie, ClinOleic/Clinolipid) demonstrate similar changes in fatty acid levels compared with infants receiving soybean oil-based lipid emulsions (ie, Intralipid) (table 2). There is a general pattern of increased LA levels from early postnatal levels with both emulsions, although less of an increase with the soybean oil/olive oil mixtures.^21–24 There is also a decline in ARA and DHA levels as well as either a decrease or no change in EPA in infants receiving either emulsion.^21–24 There are no studies examining fatty acid levels in preterm infants receiving Lipidem/Lipoplus.

Changes in blood fatty acid levels can largely be predicted by the content of the lipid emulsions. Less LA leads to less dramatic increases in LA levels; more EPA content increases EPA more significantly. However, there are two interesting but paramount changes that are not as simply explained: (1) the worsened ARA deficit in infants receiving fish oil-containing lipid emulsions despite increased ARA provision and (2) the inadequate repletion of DHA with fish oil-containing lipid emulsions despite a 1.0%–3.5% content relative to none in 100% soybean oil. It is known that both the n-3 and n-6 pathways use the same enzymes to desaturate and elongate the essential fatty acids to the distal LCPUFAs (figure 1), and there is competitive inhibition for these metabolic pathways and in the incorporation of these fatty acids into cell membranes.²⁵ This may explain the reduction in ARA with increasing levels of EPA when exposed to fish oil-containing lipid emulsions.²¹ These complex dynamics must be considered in design of future lipid emulsions for preterm infants.

• Conclusion 1: None of the currently available parenteral lipid emulsions prevent the postnatal deficits in DHA and ARA in preterm infants. Fish oil-containing lipid emulsions increase the ARA deficit relative to 100% soybean oil, increase EPA levels, and have a reverse ARA:DHA ratio compared with the ratio observed at birth, and thus in utero.

Question 2: does the use of fish oil-containing lipid emulsions reduce comorbidities in preterm infants?

With increased ARA and DHA content in fish oil-containing lipid emulsions, numerous studies have examined whether these emulsions decrease neonatal morbidities. Small, retrospective studies have reported potential benefit; however, they are limited by their study design, sample size and use of historical controls, with the latter often lacking nutritional detail when comparing cohorts. A recently published meta-analysis found no benefit of fish oil-containing lipid emulsions for prevention of bronchopulmonary dysplasia (BPD).²⁶ A 2019 Cochrane review of 29 studies with 2037 neonates born at <37 weeks of gestation found no difference between lipid emulsions with or without fish oil in a variety of outcomes including growth, mortality, BPD, sepsis, necrotising enterocolitis or severe retinopathy of prematurity (ROP); however, overall quality of evidence was low, given the absence of large randomised controlled studies.²⁷

Analysis of blood fatty acid levels induced by these fish oil-containing lipid emulsions (described earlier and summarised in table 2) would predict that these emulsions are unlikely to improve outcomes as deficits of these fatty acids are not prevented and may be exacerbated, as seen in ARA levels and in the ARA:DHA ratio. Thus, it is critical that prior studies be critically analysed to optimise future study design.

Controversy remains around the use of fish oil-containing lipid emulsions for the prevention of PNAC, which is defined as a conjugated bilirubin level of ≥2 mg/dL (34.2 μmol/L) in infants receiving parenteral nutrition for 14 or more days without an alternative aetiology for the hyperbilirubinaemia. Small studies suggest a benefit in the reduction of cholestasis,^28,29 but these findings are not supported by meta-analyses or larger studies.^27,30,31 Thus, to date, there are no data to support the use of fish oil-containing lipid emulsions, SMOFlipid or Omegaven for the prevention of PNAC. A randomised clinical trial comparing SMOFlipid to Intralipid in the prevention of PNAC in high-risk neonates has recently finished enrolment, although only 204 of the originally planned 400 infants were enrolled (Clinicaltrials.gov: NCT02579265).

• Conclusion 2: Studies on the benefit of fish oil-containing lipid emulsions in reducing neonatal morbidities are mixed, and meta-analyses of these studies do not support their routine use for the reduction of preterm morbidities.

Question 3: is use of fish oil-containing lipid emulsions without risk in preterm infants?

There is a frequent misperception that nutrients or nutrition-related supplements carry only potential benefit without a risk of harm, which is dangerous when considering the developing preterm infant. A randomised trial of enteral DHA supplementation to preterm infants to reduce the risk of BPD found that the infants who were supplemented, compared with the placebo group, had a greater risk of BPD and an increased risk of the composite outcome of BPD or death by 36 weeks.³² The DHA supplementation group had significantly higher blood DHA and EPA levels and significantly lower levels of LA and ARA. The amount of DHA supplementation was designed to match third trimester fetal accretion rates and was equivalent to what was delivered with fish oil-containing lipid emulsions.

Using a similar approach of augmenting DHA delivery to infants born before 29 weeks of gestation but through maternal supplementation of DHA, we found that a multicentre, randomised control trial was stopped early due to safety concerns.³³ A planned interim analysis revealed an increase in severe BPD with maternal DHA supplementation. Infant blood fatty acids levels were not reported, so it cannot be determined whether this effect was due to a reduction in ARA, which had been implicated with other fatty acid replacement studies. However, it does support the findings observed in other studies of the potential harm in providing a single fatty acid or n-3 dominant lipid supplement to preterm infants.

In a trial comparing the effect of ClinOleic/Clinolipid with SMOFlipid on growth and the incidence of morbidities including ROP, no significant differences were seen.³¹ A secondary analysis irrespective of study assignment found that a reduction in serum ARA levels was linked to an increased risk of ROP.³² Although this risk was not based on lipid group, the likelihood of reducing ARA levels is greater with fish oil-containing lipid emulsions as described previously.

In two retrospective studies, fish oil-containing lipid emulsions were associated with an increase in late-onset sepsis³⁴ and lower oxygen saturation to inspired oxygen ratio, implying worse lung function and injury, at 36 weeks of postmenstrual age.³⁵ The latter finding is consistent with the role of ARA in alveologenesis³⁶ and the potential to reduce alveologenesis with fish oil-containing lipid emulsions in animal studies.³⁷ The importance of ARA in numerous physiological processes cannot be underemphasised (box 1) and the use of fish oil-containing lipid emulsions compromises ARA status.

Box 1. Biological functions of arachidonic acid (ARA).

• ARA is a predominant fatty acid in the developing brain.

• Higher levels are associated with improved growth and neurodevelopment.

• Lower levels are associated with increased risk of nosocomial sepsis.

• Lower levels are associated with increased risk of retinopathy of prematurity.

• Administration of a distal metabolite of ARA, lipoxin A₄, improves alveologenesis in a murine hyperoxia model of lung injury.

The higher levels of EPA induced by fish oil-containing emulsions are of unclear clinical significance, although increased EPA levels are linked to reduced ARA levels,^21,38 which can be harmful to preterm infants (box 1), and thus further studies are needed.

• Conclusion 3: Fish oil-containing lipid emulsions may not be without risks in the preterm population.

PRACTICAL CLINICAL APPROACH IN THE USE OF PARENTERAL LIPID EMULSIONS

The provision of adequate total energy is critical during the initial days of life and has long-term implications for the neonate.³⁹ Lipids are essential in this process, with a higher energy density than proteins and glucose. Lipid infusions decrease energy use,⁴⁰ promote gluconeogenesis,⁴¹ improve nitrogen balance⁴² and increase albumin synthesis in neonates.⁴³

The European Society for Paediatric Gastroenterology Hepatology and Nutrition (ESPGHAN) recommends initiation of lipid infusion between birth and day 2, with a maximum infusion rate of 4 g/kg/day for infants.⁴⁴ There is debate as to whether a graduated approach to lipid advancement is necessary. At a minimum, the dose of lipid emulsion should support recommended macronutrient balance and total energy goals for the specific infant.

Current product availability presents a choice in which parenteral lipid emulsion to provide routinely, after birth. Unfortunately, currently available parenteral lipid emulsions do not meet the unique fatty acid requirements of the preterm infant. Both soybean oil and fish oil-containing lipid emulsions carry potential risks in their use. Omegaven is highly skewed to the n-3 pathway and should never be used as a maintenance lipid emulsion but does have a role after infants have developed PNAC. Although Intralipid does not have readily available ARA and DHA, it, in fact, preserves ARA levels and thus ARA:DHA ratios to a greater extent compared with fish oil-containing lipid emulsions. The substantial rise in EPA seen in fish oil-containing lipid emulsions introduces an unknown added risk to their use. As a result, the authors prefer to use Intralipid as the first line option for parenteral lipid nutrition until further data are accrued, demonstrating superior nutritional and health outcomes with fish oil-containing lipid emulsions.

Our recommendation to use Intralipid does contrast with the current recommendation by ESPGHAN to switch to a composite lipid emulsion (with or without fish oil) if parenteral nutrition is needed for longer than several days. The time-oriented approach is presumably to reduce the incidence of PNAC. The authors of the ESPGHAN guidelines highlight that this recommendation is not supported by randomised clinical trials and that alterations in the fatty acid balance induced by fish oil-containing emulsions need to be addressed.⁴⁵ We anticipate that as enteral nutritional strategies advance reducing the total duration of parenteral nutrition exposure, the dilemma will be less about minimising the occurrence of PNAC in otherwise healthy preterm infants but rather more about the immunonutrient programming and subsequent host responses that are induced by the various options available.

PHYSIOLOGICAL APPROACH IN CLINICAL TRIALS OF LIPID EMULSIONS AND FATTY ACIDS

Although none of the currently available emulsions adequately meet the fatty acid requirements in preterm infants or optimise health benefits, general physiological principles have been identified through basic and clinical studies that can inform future clinical studies:

1. Adult principles of fatty acid metabolism should not be applied to the preterm infant. Responses to fatty acid administration and fatty acid requirements for preterm infants are uniquely distinct from other populations and require separate investigations. The metabolic impact of dietary lipids is evident in preterm infants within days versus weeks in adults.

2. The postnatal deficits in fatty acids occur early in the postnatal period and any replacement strategies to prevent deficits must be delivered as early as possible after birth.^45,46

3. Fatty acids in both the n-3 pathway and the n-6 pathway are critical, and single DHA replacement strategies in the early neonatal period should be avoided as they compromise ARA content, which comes at a health risk (box 1).

4. The ratio of ARA:DHA should approach 2:1.^46,47

KNOWLEDGE GAPS AND RESEARCH PRIORITIES

The effort to restore lipid and fatty acid delivery to the preterm infant is an important one, and despite over 30 years of research, the translation to the bedside with demonstrated benefit remains elusive. A strategy that aims to increase one fatty acid or one pathway, often DHA and the n-3 pathway, is overly simplistic and ignores the complex interplay of LCPUFA metabolism and the need for both n-6 and n-3 fatty acids for optimal outcomes. Given the bioactivity of these nutrients in swiftly changing the infant’s metabolism and extensive animal and human data demonstrating potential risks when levels and/or balance of these nutrients are not suitable for the preterm infant, future studies should aim to apply the principles of rigour and reproducibility, conduct adequate and detailed measures and read-outs of the metabolic response, and adhere to definitions of nutritional and health outcomes. Applying these principles will enable collective and comparative data and make forward progress towards the goal of optimising health through lipid and fatty acid delivery. Research gaps that have not been resolved in the field of lipids and fatty acids in the preterm infant are elaborated further in table 3.

Table 3.

Research gaps in the field of lipids and fatty acids in the preterm infant

Optimal targets	What are the postnatal target levels, circulating and within membranes/tissues, for specific long-chain polyunsaturated fatty acids and the balance among these fatty acids to support optimal health?
Context specificity	What are the competing risks across dosing strategies, timing of administration and end-organ responses? As potent regulators of inflammation, the ideal n-6:n-3 balance may depend on the physiological goals, such as supporting development and immune reactivity during periods of vulnerability to acute events, such as sepsis versus assisting in the termination of the inflammatory response during periods of chronic, unmitigated inflammation.
Formulation	What are the oil sources and fatty acid composition of a parenteral lipid emulsion that achieve the desirable effect in total lipid and fatty acid delivery? What is the impact of the other factors in lipid emulsions that have their own distinct metabolic responses such as phytosterol content, vitamin E and phospholipid species and concentration?
Systems biology integration	What is the impact of lipid and fatty acid delivery beyond changes in fatty acid levels? How does it impact lipogenesis in other lipid classes, protein use, glucose production and cellular energy? How does it regulate the production of proinflammatory and anti-inflammatory mediators, host immune cell responsiveness, adipokines and other endocrine axes?

Continued efforts in applying basic and translational methods in the conduct of prospective, randomised clinical trials are needed. With the emergence of advanced technological and data analytical tools, these vital questions are surmountable. We urge investigators of future fatty acid studies to follow published guidelines for best practices as recommended by Brenna et al.⁴⁸ Of particular importance, fatty acid analysis of the dietary intervention and of blood levels in the infant should be quantified. A convergence in standards for conducting and reporting results will enable data synthesis, appropriate comparisons and potentially shorten the timeline for bedside translation.