Hypertriglyceridaemia in extremely preterm infants receiving parenteral lipid emulsions

Lipid emulsions (LE) are a vital component of parenteral nutrition (PN) in preterm infants providing a low-volume source of essential fatty acids and energy [1, 2]. The timing and rate of commencement LE has long been debated. The European Society for Paediatric Gastroenterology Hepatology and Nutrition (ESPGHAN) 2005 guideline recommended that LE could be commenced on day 1, but no later than day 3 in order to avoid essential fatty acid deficiency [1]. A number of systematic reviews have shown no increase in side effects with early commencement of LE [2‐4]. Potential benefits include improved nitrogen balance, anabolism and growth [4], better control of hyperglycaemia and reduced rates of retinopathy of prematurity (ROP) and necrotizing enterocolitis (NEC) [5].

Plasma triglyceride (TG) concentrations were performed as a safety variable in a number of clinical trials evaluating LE in neonates [6‐11]. TG thresholds varied in these trials and adverse clinical outcomes were not reported in cases of hypertriglyceridemia (HT). Lack of reporting and the variation in TG thresholds for LE titration make it difficult to guide clinical practice in neonatal intensive care units (NICU). Adamkin et al. [12] suggested that preterm infants can tolerate serum triglyceride levels of up to 2.8 mmol/L (250 mg/dl) without any undesirable consequences. This was based on a study of only 10 infants with birth weights 1424 g ± 177 (SEM) started on 0.4 g/kg/day of LE increased up to 1.6 g/kg/day in the first week of life. ESPGHAN 2005 Guidelines recommend monitoring of triglycerides in preterm and term infants and suggest a triglyceride level of 2.8 mmol/L as the upper limit [1].

In Australasia, a consensus group of tertiary NICUs in the region has been standardizing newborn PN formulations culminating in an evidence based consensus statement in 2012 [13]. One of the recommendations was the timing of commencement of parenteral LE would be determined by individual NICUs with day 1 administration as an option (LOE 1, GOR C). No consensus could be reached on the need and/or frequency of plasma triglyceride (TG) measurements. In February 2015, as part of the ANZNN Parenteral Nutrition Network Meeting, a survey of participating NICUs in the group found that 98% commenced LE within 2 h of birth in extremely preterm infants but only 8% routinely monitored plasma TG concentrations (unpublished).

Our NICU introduced the consensus group guidelines and PN formulations in July 2012. As part of the new guidelines, we introduced routine titration of LE infusion according to plasma TG concentrations. The objectives of this quality improvement study were to measure compliance to the unit policy, report the incidence of HT and evaluate any association between HT and neonatal mortality and morbidity.

This was a retrospective case note review of all infants less than 29 weeks gestation admitted to a tertiary neonatal unit between 1st July 2012 and 30th June 2016 (Infants were divided into two groups based on gestation [23–25 + 6 weeks GA; 26–28 + 6 weeks GA] and birth weight [BW < 1000 g; BW ≥1000 g] for the purpose of multivariate analysis). Total intravenous fluids were commenced at 60 ml/kg/day at birth comprising AA/Dextrose solution aiming to deliver 2 g/kg/day of protein. AA/Dextrose solution contained 33 g/L of Primene 10% (Baxter Pharmaceuticals Pty Ltd) as the amino acid source and 100 g/L dextrose. LE was commenced at birth. The starting dose was 1 g/kg/day and every 24 h the dose was increased by 1 g/kg/day until a maximum of 3 g/kg/day was reached [14]. Plasma triglycerides were measured 24 h after commencement of lipids and then 24 hourly until 3 g/kg/day was reached. If plasma triglycerides were > 2.8 mmol/L, the dose was reduced by 1 g/kg/day. There was a general policy of running a minimum of 0.5 g/kg/day of lipids irrespective of plasma triglycerides to prevent essential fatty acid deficiency, but clinical teams had the choice of ceasing lipids altogether if HT was considered severe. Once 3 g/kg/day of lipid intake was reached triglycerides were measured after 24–48 h and then weekly if they remained on LE infusion. LE infusions were ceased when infants’ enteral feed volume had reached 100–120 ml/kg/day and PN infusions were ceased when enteral feeds had reached 120–140 ml/kg/day.

During the study period, 248 infants born < 29 weeks gestation at birth were admitted. Fifty two were excluded: 12 due to death occurring prior to commencing parenteral nutrition or prior to any levels being taken. Forty infants were admitted after the first day of life. The remaining 196 infants were included in the study (Fig. 1).

Over the past decade numerous studies have shown that commencing lipids shortly after birth are safe with numerous potential benefits [5, 16, 17]. These studies regularly monitored TG concentrations and LE intakes were adjusted accordingly. The majority of units in Australia and New Zealand are now commencing early lipids but regular monitoring is only being practised by a small number of units.

The incidence of HT varies among studies largely due to different TG thresholds, variable patient demographics and different infusion rates of LE. Holtrop et al. [16] commenced infants < 1000 g on 0.5 g/kg/day of Intralipid on day 1 of life and increased by 0.5 g/kg/day measuring lipids when they were on 1 g/kg/day, 2 g/kg/day and then weekly thereafter. They defined HT as TG levels ≥2.3 mmol/L (200 mg/dl) and noted an incidence of 26.7%. Vlaardingerbroek et al. [17], in a randomized controlled trial in infants < 1500 g compared two different AA and lipid doses from birth. Both intervention arms received lipids at a starting dose of 2 g/kg/day on day 1 and then 3 g/kg/day on day 2, whereas the control arm received lipids at a starting dose of 1.4 g/kg/day on day 2 and then 2.8 g/kg/day on day 3. HT (defined as > 3 mmol/L) occurred frequently in all 3 arms with no significant difference (44% in controls vs 27 and 45% in the intervention arms). Drenckpohl et al. [5] commenced the control group on 0.5 g/kg/day and the intervention group on 2 g/kg/day on day one of life and increased by 0.5 g/kg/day until all infants reached 3 g/kg/day. Infants between 750 g and 1500 g were included. They reported HT (TG ≥ 2.27 mmol/L) incidence of 4 and 15% in the control and treatment groups respectively. Neither Vlaardingerbroek [17] nor Drenckpohl [5] published the characteristics of the infants that developed HT or if the infants with HT had increased mortality or morbidity.

Similar to previous studies, our study found that gestation 23–25^+ 6 weeks GA and lower birth weight were significant risk factors for the development of HT. HT was significantly higher at 26.9% in the < 1000 g group in comparison to only 3 infants (3.9%) in the ≥1000 g group. The high incidence of HT among infants < 1000 g in our study is similar to the findings of Holtrop et al [16]. They found that with each 100 g decrease in BW there was almost twice the odds of an elevated TG. Brans et al. [18], also demonstrated infants < 1000 g are especially at risk of HT. We also found that SGA was a significant risk factor. Despite small numbers, there was also a significantly higher incidence of severe HT in growth restricted infants. Based on these findings, we recommend regular monitoring for HT among infants 23–25^+ 6 weeks GA (or < 1000 g).

Unlike Holtrop our study found that HT was associated with higher mortality and severe ROP in univariate analysis. This could be due to a lower TG threshold and a smaller sample size used in their study. On multivariate logistic regression analysis our study found that HT was not significant for increased risk of mortality or severe ROP with only lower gestation being an independent risk factor. It is likely that our practice of regular monitoring with prompt titration of lipid intake as per TG thresholds reduced the chance of plasma TG reaching and/or staying at harmfully high level. In cases of severe HT specifically, TG levels returned to normal 24 h after ceasing LE. This may have mitigated any true adverse effect of sustained HT on clinical outcomes in our study. A well-designed large clinical trial is required to evaluate the effect of regular TG monitoring and titration of LE intake on clinical outcomes.

Compliance rate over the study period was 89%. Since this policy has been introduced TG monitoring and titration of LE has become accepted practice among nursing and medical staff particularly in infants < 1000 g. Although compliance was high, we initiated further education sessions to the staff to reach the goal of 100% compliance. LE appears to be well tolerated in infants ≥1000 g. Given the low occurrence in this group and that brief periods of HT appears safe, it seems reasonable to measure TG levels when the infant reaches 3 g/kg/day and then weekly thereafter in infants ≥1000 g. Following review of our guideline, a further audit to check compliance would be completed.

HT is common in extreme preterm infants, particularly among infants 23–25^+ 6 weeks GA. Regular monitoring and prompt titration of LE resulted in only short periods of HT potentially avoiding any harmful outcomes. With a lack of evidence that prolonged or severe HT is safe, regular monitoring is recommended especially among extreme preterm infants 23–25^+ 6 weeks GA (or < 1000 g). Larger studies are needed to clarify the impact of HT on morbidity and mortality.