Edited by RAD and WFB Effects of Soybean Lipid Infusion on Unbound Free Fatty Acids and Unbound Bilirubin in Preterm Infants

Abstract

Objective

To assess the effects of a soybean lipid emulsion infusions on levels of unbound (free) bilirubin (Bf) and unbound free fatty acids (FFAu) as well as changes in Bf and total serum bilirubin (TSB) during phototherapy in preterm infants.

Study design

Ninety-seven preterm infants (birthweight: 500-2000g; gestational age: 23- 34 weeks) were enrolled to investigate the effect of 0, 1, 2 and 3 g/kg/day of Intralipid (IL) infusion on Bf and FFAu. Pre- and post-phototherapy total serum bilirubin (TSB), FFAu and Bf were also analyzed in 91 infants to assess the effects of phototherapy. FFAu levels were measured using the fluorescent probe ADIFAB2 and Bf by the fluorescent Bf sensor BL22P1B11-Rh during IL infusion and at start and end of phototherapy. TSB and plasma albumin were measured by the diazo and Bromcresol Green techniques, respectively. Bilirubin-albumin dissociation constants (Kd) were calculated based on Bf and plasma albumin.

Results

Bf and FFAu increased with increasing IL dosage across all gestational ages. TSB and Bf were significantly correlated when infants received 0 or 1 g/kg/day of IL, but not at higher IL doses (2 and 3 g/kg/day). Although phototherapy effectively reduced both TSB and Bf in the total phototherapy group (by 32% and 12%, respectively), it reduced TSB, but not Bf, in infants below 28 weeks of gestation.

Conclusions

Increasing IL doses result in increasing FFAu levels, which are associated with increased Bf independent of TSB. In extremely preterm infants (<28 weeks gestation), phototherapy effectively reduces TSB but not Bf.

More than 80% of infants develop some degree of hyperbilirubinemia during the first week of life (1). Although the effects of jaundice are typically benign in full term infants, preterm infants are at increased risk for bilirubin encephalopathy (2, 3). Currently total serum bilirubin (TSB) levels are used as indicators for treatment with phototherapy to prevent progression to toxic levels. However, levels of free, unbound bilirubin predict the risk of bilirubin neurotoxicity more accurately than TSB. (4-6) Levels of free unbound bilirubin can be elevated relative to TSB by displacement of bilirubin from albumin binding sites by a variety of compounds including unbound free fatty acids (FFAu). Free unbound bilirubin crosses the blood brain barrier and exerts its toxic effect. Direct determination of free bilirubin may be important in assessing bilirubin toxicity risk and targeting therapy. Unfortunately, it is not routinely measured in the USA because of lack of commercially available assay methods but is available in other countries, such as Japan. (7, 8)

Most very low birth weight infants are treated with parenteral nutrition, including lipid infusions. Intralipid (IL) is an emulsion of soybean oil that is a major source of calories and essential fatty acids. Similar to bilirubin, serum free fatty acids are primarily bound to albumin in the circulation, but are in equilibrium with a small unbound fraction that exerts biologic effects. A concern about the use of IL is that elevated levels of FFAu can displace bilirubin from albumin binding sites, resulting in increased free bilirubin (Bf) levels relative to TSB (12, 13). This effect may be more apparent in very preterm infants because of lower albumin binding capacity for bilirubin. (14)

We have previously shown that both FFAu and Bf were highly elevated in extremely low birthweight infants treated with IL at 3 g/kg/day (15). In the present study, we hypothesized that 1) increasing doses of IL (from 1 to 3 g/kg/day) would be associated with increasing levels of FFAu and Bf, independent of TSB and 2) that phototherapy would be effective in lowering TSB but not Bf, particularly in extremely preterm (<28 weeks GA) infants during IL infusion.

Methods

Infants with birthweights between 500 and 2000g and gestational ages of 23 to 34 weeks treated with IL were eligible for this observational study. Recruitment was carried out in the neonatal intensive care unit at Rutgers Robert Wood Johnson Medical School. The study was approved by the institutional review boards of Rutgers and The Torrey Pines Institute for Molecular Studies. Parental informed consent was obtained. Because the laboratory policy is to store all blood samples for seven days it was possible to obtain research samples earlier than the age of the infant at consent.

IL administration was determined by the clinical team, usually commencing at a dose of 1 g/kg/day on the second day of life and advancing it daily to a maximum of 3 g/kg/day. The decision to withhold, reduce or terminate the lipid infusion was determined by the clinical staff and resulted in some infants receiving fewer than the projected doses. The decision to initiate and terminate phototherapy was determined by the clinical staff and based on published recommendations. (16) Phototherapy devices utilized were neoBLUE LED (Natus Medical Inc., Pleasanton, CA) positioned above the infants for those in incubators and fluorescent bulbs housed in the phototherapy attachment for those infants in warmer beds. Irradiance levels ranged between 15 and 30 μw/cm²/nm.

Blood Sampling

Plasma samples were obtained before IL infusion (IL0), during infusion of 1, 2 and 3 g/kg/day (IL1, IL2, IL3) and before and after phototherapy. Samples were obtained from residual blood drawn for clinical indications during those IL dosing windows. These samples were collected immediately after completion of the requested clinical test, processed and frozen. The lack of ordered clinical samples or inadequate residuals resulted in an incomplete set of research samples for some infants. De-identified samples were stored at -70°C and shipped to Fluoresprobe Sciences for determination of Bf and FFAu.

Bf, FFAu and TSB measurements

FFAu and Bf were measured in the same plasma sample using fluorescently labeled mutants of fatty acid binding proteins (17-19). The Bf probe (BL22P1B11-Rh) binds unconjugated bilirubin with high affinity (Kd=16 nM) but is poorly sensitive to FFA (Kd>3000 nM), conjugated bilirubin (Kd>300 nM), bilirubin photoisomers, bilirubin oxidation products, ibuprofen and indomethacin (17). FFAu concentrations were determined using the ADIFAB2 probe (15, 18). Fluorescence was assessed at two emission wavelengths: 457 and 550 nm for the FFAu probe and 525 and 580 nm for the Bf probe (excitation = 375 nm for both). The ratio of fluorescence at the two wavelengths, together with probe characteristics, yielded the FFAu and Bf concentrations. Fluorescence was measured at 22°C using handheld ratio fluorometers in which plasma sample volumes were 4 μL for FFAu and 25 μL for Bf measurements. After dilution the total volume was 200 μL so that FFAu was measured at 50 fold and Bf at 8 fold dilution of the plasma sample (17, 18).

The accuracy of the Bf probe measurements was confirmed previously by comparison with the peroxidase assay in adult plasma supplemented with bilirubin as well as in bilirubin-albumin in vitro measurements (17). The Bf probe has an average coefficient of variation of 3% for repeated measurements of aqueous bilirubin, bilirubin-albumin complexes and bilirubin spiked adult plasma over a Bf range from 1 to > 350 nmol/L (17). Measurements with the Bf probe in contrast to the peroxidase method, determines the Bf concentration directly in a single measurement and is insensitive to substances that can interfere with the peroxidase measurement (17).

TSB was measured using the diazo method at the Robert Wood Johnson Medical School clinical laboratory. Plasma albumin was measured in infants prior to IL infusion using the Bromocresol Green (BCG) Albumin Assay (Sigma) method which was standardized with fatty acid free human serum albumin (HSA) from 72 to 722 μM (coefficient of variation< 4%). Infant plasma (5 μL) was diluted into 200 μL of the BCG reagent and absorbance was measured at 620 nm. We calculated the Bilirubin –albumin equilibrium dissociation constants (Kd), which is the inverse of the binding constant Ka using equation (1),

$${\mathrm{K}}_{\mathrm{d}}=\frac{{\mathrm{B}}_{\mathrm{f}}{\mathrm{A}}_{\mathrm{t}}}{(\text{TSB}-{\mathrm{B}}_{\mathrm{f}})}-{\mathrm{B}}_{\mathrm{f}}$$

where At is the measured albumin concentration and Kd is in moles/L or nM.

Statistical analyses

Changes in TSB, Bf and FFAu as a function of increasing dose of IL or of phototherapy were analyzed by repeated measures analysis and verified by the generalized estimating equation semiparametric regression method. Multiple regression analysis was performed to evaluate the correlation between FFAu and TSB with Bf with increasing dose of IL, taking into account confounding variables including birth weight, gestational age, sex, race and 5 minute Apgar. The analysis was verified by non-parametric Spearman correlation.

The analyses were performed with XLSTAT (AddinSoft) and verified with SAS (SAS Institutes); p-values <.05 were considered statistically significant.

Results

Effects of increasing dose of IL on TSB, Bf and FFAu

The study population included ninety-seven infants with a mean birth weight of 1315+435 g and gestational age of 29.2+3.1 weeks. Fifty-two percent of infants were male. Fifty two percent of the infants were white, 30% were black, 17% were Asian and 1% were Hispanic. For 48 infants, measurements of Bf, FFAu and TSB were obtained before the IL infusion (IL0) and at all three IL dosages, 1, 2 and 3 g/kg/day (IL1, IL2 and IL3). Of the remaining infants, 42 had measurements at 3 of the 4 IL steps, 12 infants had measurements at 2 steps and 5 had a measurement at a single step. Thus, the data are cross sectional with reference to increasing doses of IL infusion.

The mean age at the start and advancement of IL infusion was 28.6+ 11.5 (SD) (range: 10-55) hours for IL1, 47.9+13.7 (range: 13-79) hours for IL2 and 77.6+22.3 (range: 37-200) hours for IL3. The mean age at initiation of blood sampling was 13±7 (range: 0.5-41) hours for IL0, 33±12 (range: 8-74) hours for IL1, 61±17 (range: 22-104) hours for IL2 and 110±39 (range: 59-311) hours for IL3. As shown in Figure 1, TSB increased between IL0 and IL2, consistent with the age-dependent natural history of hyperbilirubinemia of prematurity. TSB decreased to 5.8±2.0 mg/dL at IL3, likely related to the use of phototherapy in most (n=91) of the infants. Bf did not decrease at IL3 despite the phototherapy related reduction of TSB. The quadratic trend for TSB was significant (p<0.001). FFAu and Bf increased significantly (p<0.001) with each incremental increase in IL dose. It is notable that larger increases in free fatty acids and free bilirubin occurred in infants of less than 28 weeks gestation (Figure 2), but overall gestational age had little if any impact on the results of the study. This discrepancy may be due to the high variance of findings in the heavier infants, who constituted more (67%) of the overall group.

Figure 1.

Effects of increasing IL dose from 0-3 g/kg/day (IL dose of 1,2,3 g/kg/day, with n=77, 83, 77 and 86, respectively) on FFAu, Bf and TSB. Data are shown as individual points and mean values ± standard deviations. A linear correlation of FFAu (A) and Bf (B) with IL dose is significant (p<0.05) with or without IL0. The quadratic increase and decrease in TSB (C) is significant (p<0.001).

Figure 2.

Bf and FFAu with increasing dose of IL (IL0 to IL3). Average values ± standard deviations are shown. Bf increased with IL dose for both GA groups: GA≤28 wks, p<0.05, R²=0.99 and GA>28 wks, p<0.05, R²=0.8. FFAu increased with IL: GA≤28 wks, p<0.05, R²=0.84 and GA>28 wks, p<0.05, R²=0.85.

Bf was tightly correlated with TSB at doses of 0-1 g/kg/day of IL, but not at higher IL doses of 2-3 g/kg/day (Figure 3; available at www.jpeds.com). Seven infants had final FFAu levels above 100 nM. Of these, three above and four below or equal to 28 weeks gestational age and all (except one with birthweight 1505g) weighed less than 1500 g at birth. Three infants had final Bf levels above 40 nM, and were all <28 weeks of gestation and <1250 g birth weight. One infant of this group died, with no autopsy. Detailed auditory evaluations were not available for the two survivors. Bilirubin-albumin binding was highly variable among infants before IL infusion, with Kd values in 57 infants ranging from 15-75 nM. FFAu, but not bilirubin-albumin Kd, correlated with Bf levels at all IL infusion ratess, suggesting that FFAu overpower any intrinsic differences in binding affinity.

Figure 3.

online only. Correlations between Bf and TSB with increasing dose of IL (IL0 to IL3). Bf was well correlated with TSB at IL0 n=74 p<0.0001, R²=0.42 (A) and with IL at 1g/kg/day, n=80, p<0.0001, R=0.26 (B). At IL2 and IL3, Bf was not correlated with TSB, n=76, p=0.2, R=0.008 (C) and n=83, p=0.22, R=0.003 (D).

Multiple linear regression was performed at each IL step on Bf with the following selected variables: TSB, FFAu, birthweight, gestational age, sex, race and 5-minute Apgar score. The latter three variables did not have significant effects on the changes in Bf as a function of increasing dose of IL. At IL0 and IL1 only TSB contributed significantly, p<.05. At IL2 and IL3 both TSB and FFAu were significant, p<.05. Gestational age did not contribute significantly except at IL1 (p<0.05). At IL2 and IL3, Bf increased similarly for all gestational ages between 23 and 34 weeks (Figure 2).

Effects of phototherapy on TSB, Bf and FFAu

Phototherapy was initiated in 91 infants at 41.5+15.9 hours of age at a TSB level of 7.4+2.2 mg/dl and discontinued at 80.3+24.3 hours of age at a TSB level of 5.0+1.9 mg/dl with an exposure duration of 39.0+28.2 hours. The demographic characteristics of infants receiving phototherapy were similar to those of the entire group. For the entire group, TSB significantly decreased during the period of phototherapy. There was a slight but statistically significant decrease in Bf levels during phototherapy, but FFAu levels increased (Table I). When the changes in these variables were assessed with reference to gestational age, more mature infants (>28 weeks) demonstrated decreased TSB and Bf with phototherapy similar to the total group, as well as higher FFAu associated with advances to higher doses of IL (2-3g/kg/day) during phototherapy (Table I). In the less mature infants (< 28 weeks), phototherapy significantly reduced the TSB but not Bf. FFAu also increased significantly because of higher IL intake. The effect of FFAu on Bf was more powerful than phototherapy, as Bf was reduced by phototherapy at low FFAu but increased at high FFAu levels (Table II). We compared the effects of FFAu levels above and below the median (25nM) on the degree of phototherapy-mediated reduction of TSB and Bf. For FFAu levels ≥ 25 nM, TSB is reduced by 39% (p<0.0001) but Bf increased by 15% (p=ns). At FFAu < 25 nM, TSB drops 31 % (p<0.0001) and Bf 24% (p<0.0001).

Table 1. Effects of Phototherapy on TSB, Bf and FFAu.

	23-34 weeks			≤ 28 weeks			>28 weeks
	FFAu (nM)	TSB (mg/dl)	Bf (nM)	FFAu (nM)	TSB, (mg/dl)	Bf (nM)	FFAu, (nM)	TSB, (mg/dl)	Bf (nM)
N	87	91	91	33	33	33	54	58	58
Before PT	18.2+13.7	7.4+2.2	16.1+9.5	18.9+15.5	5.4+1.4	13.1+8.4	17.8+12.6	8.6+1.6	17.8+9.7
After PT	57.1+128.9	5.0+1.9	14.2+9.9	86.5+191.4	3.1+1.0	13.5+7.5	40.1+68.3	6.1+1.4	14.5+11.1
P-value	0.006	<0.001	0.021	0.051	< 0.001	0.823	0.021	<0.0001	0.001

Data are presented as the mean ± 1 SD

Table II. Effects of Phototherapy in Reducing Bf Decreases with Increasing FFAu.

FFAu increase	≥25 nM (n=23)			< 25 nM (n=64)
	FFAu, nM	TSB, mg/dl	Bf, nM	FFAu, nM	TSB, mg/dl	Bf, nM
Before phototherapy	18.7+13.1	6.5+2.3	17.5+15.1	17.7+13.9	7.7+2.1	16.1+6.3
After phototherapy	150.1+233.8	3.9+1.6	20.4+17.3	24.3+8.1	5.3+1.9	12.2+3.8
P-value	0.013	<0.0001	0.236	<0.001	<0.0001	<0.0001

Data are presented as the mean ± 1 SD

Discussion

Increasing doses of IL result in a significant increase in FFAu and Bf levels in preterm infants. This is consistent with the idea that IL increases FFAu, to levels that can displace bilirubin from its binding sites on albumin or inhibit its binding. These findings highlight the role of bilirubin-albumin binding equilibrium in the determination of Bf. Before IL and at 1 g/kg/day dose of IL, bilirubin is tightly bound to albumin, so Bf correlates well with TSB (Figure 3, A and B). In contrast, at high doses of IL, Bf is not well-correlated with TSB because of competition for, or inhibition of, albumin binding by higher FFAu (Figure 3, C and D). The strength of this effect is determined, in part, by the bilirubin-albumin binding constant, which we found to be widely variable between infants, consistent with previous reports (20, 21). The high degree of variability in FFAu and bilirubin-albumin Kd likely reflects developmentally determined alterations in metabolism and accounts for large infant-to-infant differences in Bf at specific doses of IL, as well as differences in the Bf in response to phototherapy in lower and higher gestational age (Table I).

Our findings are consistent with Amin et al, who previously reported that IL infusion in extremely low birth weight infants (<28 weeks gestation) results in increased FFA and Bf (22),(13). In contrast to Amin et al, we also found IL-dependent increases in FFAu and Bf in infants with gestation age between 28 and 34 weeks. This may be related to our earlier enrollment of infants, starting at average age of 13 hours versus the enrollment age of 3 days in the Amin et al study (22). Our Kd values were obtained before IL infusion, when bilirubin-albumin binding was dependent only on TSB and Kd, not on displacement by FFAu.

A unique feature of this study is the sensitive measurement of FFAu, the portion of FFA not bound to albumin. For healthy adults mean concentrations of FFAu in serum are about 1.5 nM, and even with severe physiologic insults FFAu rarely rise above 10 nM (23). Moreover, FFAu levels are generally consistent in adults in the fasting or postprandial state because FFA produced by hydrolysis of serum triglycerides are rapidly taken up by adipose, muscle and liver tissue (24). This suggests that the high levels of FFAu in the IL-treated infants are due to robust lipase activity and/or severely reduced FFAu uptake and metabolism. Due to these developmental phenomena, infants may not receive the full nutritional benefit of IL and may in fact be at risk for FFAu-mediated toxicity. This observation is of clinical importance because weight gain in extremely low birth weight infants is directly related to improved neurodevelopmental outcome at two years. (25) Because of these findings, clinicians are rightly aggressive in providing high energy intake by the use of higher dose of IL, a high caloric density nutrient during parenteral nutrition. Our findings of high probability of generating a high FFAu and Bf require a reassessment of this nutritional strategy to balance the benefits of nutritional value and the potential risk of bilirubin neurotoxicity. Futhermore, at high levels, FFAu are physiologically active, potentially inhibiting immune activity and exerting cardiotoxicity (23, 26).

Our finding that Bf and TSB are uncoupled at standard doses of IL infusion (2-3 g/kg/d) highlights a potential concern for clinicians, especially during phototherapy. The effectiveness of phototherapy, which converts TSB into water-soluble isomers that bypass hepatic conjugation to be excreted without further metabolism, (27) is presently judged by its impact on TSB with the assumption that there is direct correlation between TSB and Bf. However our data shows that when infants are also receiving high doses of IL (2-3g/kg/day) there is no correlation between TSB and Bf, such that Bf is unchanged in infants of < 28 weeks gestation (Table I). The reason for the lack of reduction of Bf may be due to a combination of a higher level of FFAu and lower bilirubin-albumin binding capacity and/or affinity in this population. (14)

A recent study by the Neonatal Research Network highlighted the danger of using TSB to guide therapy, showing that aggressive phototherapy reduced neurologic deficits but increased the rate of death in extremely low birth weight infants. (28) Not considered in this investigation was the possibility that FFAu derived from IL, which almost all extremely low birth weight infants receive during the first week of life, displaced albumin-bound bilirubin and subjected infants to bilirubin toxicity. (29) Our own previous study suggested that increases in FFAu and Bf can reach extreme and toxic levels, and that under such conditions phototherapy cannot reduce Bf sufficiently.(15) This investigation was not aimed at documenting the short or long-term toxicity of elevated Bf levels, thus auditory brainstem responses in the infants were not collected.(6) The standard of care in our neonatal intensive care unit is a hearing test by an automated auditory brainstem response test that provides only pass/fail data and infants who fail this screen are referred to the audiology program where the work-up may include a diagnostic auditory brainstem evoked response test.

In summary, increasing dose of IL infusion results in increase in FFAu and Bf, the latter independent of TSB. In extremely preterm infants (gestation < 28 weeks) receiving high dose (2-3 g/kg/day) administration, phototherapy reduces TSB but not Bf.

Abbreviations

TSB

total serum bilirubin

Bf

free bilirubin

FFAu

unbound free fatty acid

GA

gestational age

Kd

bilirubin-albumin dissociation constant

IL

Intralipid