Fat Emulsion Tolerance in Very Low Birth Weight Neonates: Effect on Diffusion of Oxygen in the Lungs and on Blood pH

(1)

Fat Emulsion

Tolerance

in Very

Low Birth

Weight

Neonates:

Effect

on Diffusion

of

Oxygen

in the Lungs

and on Blood

pH

Yves W. Brans, MD, Elizabeth B. Dutton, BSN, Donna S. Andrew, MS,

Elizabeth M. Menchaca, BA, and Donna L. West, BSN

From the Perinatal Research Laboratory, Departments of Pediatrics and of Obstetrics and Gynecology, The University of Texas Health Science Center, San Antonio

ABSTRACT. Forty-one very low birth weight neonates

(820 to 1,510 g and 27 to 34 weeks of gestation) requiring

total parenteral nutrition were randomly assigned to one

of three regimens of administration of fat emulsion for a

period of eight days. Groups I and II received the

emul-sion at a constant rate for, respectively, 24 and 16 hours,

beginning with a daily dosage of 1 g/kg and increasing

daily by 1 g/kg to a maximum of4 g/kg. Group III received

the emulsion at a constant rate or 4 g/kg for 24 hours.

Blood pH and alveolar-arteriolar gradient of oxygen

dif-fusion in the lungs were measured at regular intervals.

The various regimens and rates of fat infusion appeared

to have no deleterious effect on blood pH and

alveolar-arteriolar oxygen diffusion gradient. Infusion rates as

used in the study for appropriate for gestational age very

low birth weight neonates appear to be safe, although

caution is always warranted when dealing with tiny

neo-nates whose pulmonary reserve is minimal. In view of

data from other studies, it is suggested to infuse fat at a

constant rate for 24 hours to avoid overloading the

clear-ance mechanisms of fat particles from plasma. Pediatrics

1986;78:79-84, fat emuLsion, prematurity, oxygen diffusion gradient, blood pH.

substrate, infusion of fat emulsions may give rise

to considerable increases in plasma concentrations

of the various lipid fractions.4#{176} Coupled with the fact that those tiny neonates who most need fat

emulsions to ensure adequate caloric intakes

sel-dom have normal pulmonary function, serious

com-plications could potentially occur. Indeed, a syn-drome of pulmonary fat overload has been de-scribed.7 The frequency and severity of this

syn-drome

have

not yet been

properly

defined

and

nei-ther have the clinical situations in which it occurs, its pathophysiology, and whether in fact it is related to excessive intake of fat emulsion or to the under-lying disease. This study was designed to compare

the effects of three commonly used regimens for

infusion of fat emulsions. We report the effect of

fat emulsions on oxygen diffusion in the lungs and on blood pH homeostasis. Changes in plasma lipid patterns will be the object of a separate communi-cation.

Healthy adults who receive a fat emulsion par-enterally respond with a slight decrease in Pao2. This decrease in oxygen-diffusing capacity through the alveolar-arteriolar membrane in the lung par-allels the increase in plasma concentrations of

tn-glycenides.13 In healthy individuals, this side effect

is probably of little clinical importance. In view of

the limited metabolic tolerance of very low birth

weight neonates to any type of exogenous metabolic

Received for publication Sept 3, 1985; accepted Nov 11, 1985.

Reprint requests to (Y.W.B.) Department of Pediatrics, The

University of Texas Health Science Center, 7703 Floyd Curl Dr,

San Antonio, TX 78284.

American Academy of Pediatrics.

METHODS

Tolerance to parenterally administered fat

emul-sions was studied in neonates weighing 1,500 g or

less at birth. Informed parental consent was

ob-tamed

in all cases.

Birth

weights

were

recorded

to

the nearest

10 g. Gestational

ages were

determined

from the mother’s menstrual history, checked in

most cases by sonographic determination of the

biparietal diameter, and confirmed by physical ex-amination of the neonate.12 All neonates were

nor-mally grown, ie, their birth weights were between

the tenth and the 90th percentiles for gestational maturity, sex, and race. Separate birth weight-ges-tational age curves were used for white Latin-Amen-ican (C. E. Gibbs, unpublished data), white

(2)

less

than

750 g at birth or with estimated gesta-tional ages of less than 27 weeks were excluded

from the study.

The

neonates were managed in accordance with

the usual practices for our nursery. Total parenteral nutrition was started on the third postnatal day

unless otherwise indicated. The solution provided 2.5 g/kg of crystalline amino acids, a maximum of 15 g/kg of dextrose (adjusted in cases of dextrose

intolerance), and 100 to 130 mL/kg of water each

day. Fat emulsion (Intralipid 10% or 20%, Kabi-Vitrum, Berkeley, CA) was infused according to one of three randomly allocated regimens. Neonates in

group

I received

the

fat emulsion

at constant

rate

for 24 hours, beginning with a daily dosage of 1 g/

kg and increasing by 1 g/kg on each successive day

up to a daily

maximum

of

4 g/kg. Neonates in group

II received

the

fat emulsion

at a constant

rate

for

16

hours followed by eight hours without infusion

of fats; daily dosage was the same as in group I.

Neonates in group

III received

the

fat emulsion

at

a constant

rate

for 24 hours

with

daily

dosage

of

4

g/kg from the beginning of the infusion. All fluids

were

administered

through

a catheter

placed

in the

umbilical artery or through a peripheral vein. The study was discontinued if a neonate’s plasma

ap-peared

frankly

creamy.

The study

period

lasted

eight

days. Arterial

blood

samples were obtained immediately before

begin-fling the lipid infusion or when changing the dosage

and

every

12 hours thereafter. Blood Pao2, Paco2,

and

pH

were

determined

with

an AVL

blood

gas

analyzer. The alveolar-arteriolar gradient of oxygen

diffusion (A -

aDo2)

in the lungs was calculated

from

the formula:

A-aDo2

(PB -

PH2O)

Fi02

-

PAco2

Fi02 +

1 -

Fi02\

R

)PaO2

where PB = atmospheric pressure; PH2O = partial

pressure of water in inspired gas; Fi02 = proportion

of oxygen in inspired gas; PAco2 = partial pressure

of carbon dioxide in alveoli, assumed to equal the

partial pressure of carbon dioxide in arterial blood

(Paco2);

R

= respiratory quotient = 0.8;

and

Pa02

= partial pressure of oxygen in arterial blood.

In groups

I and

III,

because

two

values

were

obtained each day while the infants received the same rate of lipid infusion, the average of the two

values was used. Each variable was analyzed by

means of one-way analysis of variance to detect

differences between the three groups at a given time

during the study. When main effect differences

were

detected

(F test, P < .05), Duncan’s multiple

range test was applied to locate differences between

means for various groups. A significance level of P

< .05 was chosen to define statistical significance.

Within each group, the mean value for each day of

study was compared with the preinfusion value by means of Student’s paired t test. Because eight

comparisons within a set of data were made, P <

.05/8

or

P < .006 was needed to indicate statistically

significant differences. In addition, for group II, the

values obtained on each day of study while fat

emulsion was infused were compared with the

preinfusion values for that day by means of

Stu-dent’s paired t test; a level of P < .006 was needed to indicate statistical significance.

RESULTS

Forty-one neonates were studied and are de-scribed in the Table. Birth weights ranged from 820

to

1,510 g, gestational ages ranged from 27 to 34

weeks, and postnatal ages ranged from one to nine

days. There were no statistically significant

differ-ences between the three groups in mean birth

weights, mean gestational ages, and mean postnatal

ages. Ten neonates

did

not complete the eight-day

study. In group I, one baby died of hyaline

mem-brane disease. In group II, one baby died of hyaline membrane disease, in two necrotizing enterocolitis developed and surgical treatment was required, one

became

severely

hyperlipemic,

one

was

removed

from the study at the request of the parents, and in

one

baby

the

lipid

infusion

was

interrupted

by

mistake.

In group

III, two babies

became

severely

hyperlipemic and one baby was removed from the

study when enteral feedings were started by

mis-take.

Data

from

these

neonates

until

the

time

of

discontinuation of the study were included in the

statistical analysis. Toward the end of the study

period,

data

were missed

when

blood

samples

could

not be obtained

in a timely

manner

or when

there

was no clinical justification to obtain samples of arterial blood.

Mean

A-aDo2

values

are

depicted

in Fig

1 for

each of the study groups. Preinfusion A-aDo2

val-ues ranged from 8 to 504 mm Hg and their distni-bution was similar in the three groups, with 60% of

neonates in group I, 57% of neonates in group II, and 58% of neonates in group III having preinfusion A-aDo2 values of 100 mm Hg or less. In groups I

and II, all but two neonates

had A-aDo2

values

that

decreased or stabilized during the period of the

study as would be expected from their improving or stable respiratory status. Two neonates in each

group

had

increasing

A-aDo2

values:

one

due

to

worsening of hyaline membrane disease and one to

(3)

t

TABLE. Characteristics of the 41 Neonates Studied

Characteristic Group I Group II Group III

No.ofneonates 15 14 12

Sex (M/F) 11/4 8/6 5/7

Race (Latin-American/Anglo- 13/1/1 9/3/2 9/3/0

American/Black)

Birth wt (g)* 1,190 ± 190

(820-1,480)

1,160 ± 217

(820-1,500)

1,190 ± 214

(840-1,510)

Gestational age (wk)* 29 ± 1.7

(27-34)

29 ± 1.4

(27-31)

29 ± 1.4

(28-31)

Postnatal age (d)* 4 ± 1.8

(1-9)

3 ± 1.4 (2-7)

3 ± 0.9 (2-5)

* Mean ± SD (range).

4

0

400-rJrJ

I

rJrJll1

I

1111

300-200

l00

fl= 15151412 II 7 5 5 4 1313139 7 7 3 4 3 12121010107653

Fig 1. Mean (±2 SE) alveolar-arteriolar gradient (mm Hg) of oxygen diffusion in lungs

before infusion of lipids and during infusion of lipids on each of eight days of study.

Corresponding rate of lipid infusion (mg/kg/24 h) is shown at top. Symbols denote

statistically significant differences at a given time between groups I and II (t)-and groups

II and III (1:) (analysis of variance and Duncan’s multiple range test, P < .05).

the data presented in Fig 1, mean A-aDo2 would

progressively decrease with duration of lipid infu-sion in all three groups. In group III, A-aDo2 values

decreased or stabilized in all neonates. Within each

group, A-aDo2 values during infusion of lipids were

not statistically different from preinfusion values,

ie, the mean differences between gradients on a given day or on a given rate of lipid infusion and

preinfusion gradients were not statistically

differ-ent from zero. On a given day during the study,

mean

A-aDo2

values

were

similar

in

all

study

groups, except on the very last day when the

A-aDo2 result

was significantly

higher

in group

II (168

± 74.4 mm

Hg) than

in either

group

I

(37 ± 7.4 mm

Hg)

or group

III

(41 ± 8.2 mm

Hg)

(P < .05).

In

group II, the mean differences between infusion

and

preinfusion

A-aDo2

values

during

each

day of

the study were not statistically different from zero

(Fig 2). Of the two neonates who died during the study, only one had an autopsy performed, and there was no sign of fat accumulation in the lungs.

Mean blood pH values before infusion of lipids

and

during

each

day of lipid

infusion

are depicted

in Fig 3. The values ranged from 7.08 to 7.55 during

the

study.

Although

blood

pH

remained

within

normal limits during most of the study, values less

than 7.20 occurred transiently in six of 15 group I neonates, seven of 14 group II neonates, and five of 12 group III neonates. The occurrence of such low

pH

values

appeared

to be related

to a neonate’s

respiratory and perfusion status and not to changes

(4)

300

-200

10

0-

7.50-7.40’

7.30-7.20

7.50-pH was affected

neither

by duration

of lipid

infu-sion nor by the regimen of lipid infusion. In group

II,

the

mean

differences

between

infusion

and

preinfusion pH values during each day of the study

were

not statistically

significant

from

zero

(Fig

4).

DISCUSSION

Since Sundstrom et a!’ reported a small and not

statistically significant decrease in

A-aDo2

values

in healthy adults who received 0.15 g/min of fat emulsion during a 20-minute period, there have been other reports. Investigations in sheep and rabbits have suggested that accumulation of tn-glycenides in the lungs was not the primary

of-]

,nHHHHH

400-fender, but that fat emulsions (or one or several of

their components) may trigger

prostaglandin-me-diated alterations of vascular tone leading to a state

of pulmonary hypertension and decreased oxygen-ation.14’6 This prostaglandin response may be

blocked by indomethacin15”7 and may be

poten-tiated by preexisting lung damage.’6 This would, of course, be of key importance in the case of very low birth weight neonates who seldom escape some

degree of pulmonary injury. Alternatively, it has

been

proposed

that

particles

of fat emulsions

may

saturate the reticuloendothelial system, by which

they are removed from circulation, may compro-mise further function of macrophages, and may thereby produce a barrier to oxygen diffusion in the lungs.16 Although the effect of higher rates of lipid

infusions are not known, infusions of as much as

4

0

r-iflflflflflfl

Fig 2. Group II only. Mean (±2 SE) alveolar-arteriolar

gradient (mm Hg) of oxygen diffusion in lungs before and

during infusion of lipids on each of eight days of study.

Corresponding rate of lipid infusion (g/kg/24 h) is shown

at top.

]

rJrJI]

Fig 4. Group II only. Mean (±2 SE) blood pH before

and during infusion of lipids on each of eight days of

study. Corresponding rate of lipid infusion (mg/kg/24 h)

is shown at top.

1

II’

7.40 -7.30

-7.20

- = IS IS 4 2 II 7 5 6 5 13 13 13 9 7 7 3 4 3 12 12 $0 10 10 7 6 5 5

Fig 3. Mean (±2 SE) blood pH before infusion on lipids and during infusion of lipids on

each of eight days of study. Corresponding rate of lipid infusion (mg/kg/24 h) is shown at

(5)

0.08 g/kg/h appear to

be compatible

with

normal

macrophage function.18

Several neonates and infants were described at

autopsy as having fat embolisms in the capillaries

of the lungs and lipid-laden macrophages.8 Dahms

and Ha1pin described three infants who had ar-tenial lipid lesions characterized by a wide foamy layer of intima partially occluding the limina of small muscular arteries in the lungs as well as by infiltration of the intima, the media, and the adven-titia by lipids. All three of these infants were very ill for weeks prior to their deaths: one had repeated episodes of septicemia, and the other two suffered from severe congenital heart disease. Other inves-tigators have reported that, under certain

circum-stances, infusion of fat emulsions resulted in

con-current decreases in Pao2 values and increases in plasma concentrations oftniglycerides.7”#{176} It is

stnik-ing that most cases of pulmonary side effects were

associated with high rates of lipid infusion, all in excess of 0.14 g/kg/h and as high as 0.70 g/kg/h. With the exception of the three infants described

by Dahms and Halpin, whose degree of illness has already been noted, only one baby of 38 received fat emulsions at a rate less than 0.17 g/kg/h and only three were exposed to a rate less than 0.25 g/ kg/h before hypoxia was noted or death occurred.

In fact,

it has been stated that fat emulsion admin-istered at rates of 0.20 g/kg/h had no effect on oxygenation of neonates weighing less than 1,500 g.1#{176}The data from our study support this conten-tion. Lipids were infused at rates as great as 0.17 g/kg/h in groups I and III and 0.25 g/kg/h in group

II without

demonstrating

any deleterious

effects

on

A-aDo2

values.

Although the rates of lipid infusion used in this

study appeared to be safe, some warnings are in

order. Very immature neonates (mostly less than

27 weeks of gestation but sometimes more mature by dates but not metabolically) may have difficul-ties clearing lipid particles from plasma and storing the particles in the reticuloendothelial system; im-paired diffusion of oxygen could conceivably occur. Small-for-dates neonates clear lipid particles less effectively and, because this group of neonates was not addressed in our study, no recommendation as to “safe” infusion rates may be made. No data on lipid tolerance of neonates with chronic lung

dis-eases, such as bronchopulmonary dysplasia, are

available. Clinically, they appear to tolerate

infu-sion rates of 4 g/kg/d without obvious problems, but caution should be observed in view of the ani-mal experience mentioned earlier in this discussion. Certainly, on the basis of our knowledge at this time, no neonate should be deprived of the caloric benefit of fat emulsions administered at a

reason-able and cautious rate.

Fat emulsions have anecdotally been blamed for producing metabolic acidosis, although supporting evidence has been conspicuously lacking. Our data suggest that pH was affected neither by the dura-tion oflipid infusion nor by the regimen of infusion.

Low

pH

values

that

were

occasionally

observed

were readily explained on the basis of the neonate’s respiratory and perfusion status. Infusion of lipids neither worsened the pH status nor delayed

recov-ery.

SPECULATION AND RELEVANCE

These data suggest that infusion of fat emulsions in appropriate for gestational age, very low birth weight neonates, at a maximal hourly rate of 0.17 or 0.25 g/kg, appears to have no deleterious effect on either diffusion of oxygen in the lungs or blood

pH.

In view

of the data

in the literature,

use of an

infusion rate of 0.17 g/kg/h (therefore infusing fat emulsion at a constant rate for the full 24-hour period) may decrease the risk of untoward side effects.

SUMMARY

Very low birth weight neonates were randomly allocated to one of three regimens of administration of fat emulsions to compare continuous v intermit-tent administration and stepwise increase of dosage V constant dosage. The effect of these various reg-imens on acid-base status and on oxygen-diffusing capacity in the lungs was particularly scrutinized. The data suggest that continuous administration of fat emulsions at a constant rate for 24 hours may be safer than intermittent infusions. Beginning at

a daily

dosage

of 4 g/kg

rather

than

increasing

the

dosage stepwise does not appear to have adverse effects on pulmonary function in appropriate for gestational age neonates. Data on other aspects of tolerance of fat emulsion must be examined before recommending this method of infusion as a safe procedure.

ACKNOWLEDGMENT

This work was supported, in part, by grant HD 15967

from the National Institute of Child Health and Human

Development, Bethesda, MD, and by KabiVitrum,

Berke-ley, CA.

REFERENCES

1. Sundstrom G, Zauner CW, Arborelius M: Decrease in

pul-monary diffusing capacity during lipid infusion in healthy

men. J Appi Physiol 1973;34:816-820

2. Greene HL: Effects of Intralipid on the lung, in Winters

(6)

High Risk Infant. New York, John Wiley & Sons, 1975, pp

369-380

3. Greene HL, Hazlett D, Demaree R: Relationship between

Intralipid-induced hyperlipemia and pulmonary function.

Am J Clin Nutr 1976;29:127-135

4. Brans YW: Parenteral nutrition of the very low birth weight

neonate: A critical view. Clin Perinatol 1977;4:367-376 5. Hilliard JL, Shannon DL, Hunter MA, et al: Plasma lipid

levels in preterm neonates receiving parenteral fat emul-sions. Arch Dis Child 1983;58:29-33

6. Brans YW: Erratum: Values for plasma glycerol, FFA, and

triglycerides. J Pedizztr 1984;105:855

7. Sun SC, Ventura C, Verasestakul 5: Effect of

Intralipid-induced lipaemia on the arterial oxygen tension in preterm infants. Resuscitation 1978;6:265-270

8. Barson AJ, Chiswick ML, Doig CM: Fat embolism in infancy

after intravenous fat infusions. Arch Dis Child 1978;53:218-223

9. Levene MI, Wigglesworth JS, Desai R: Pulmonary fat

ac-cumulation after Intralipid infusion in the preterm infant.

Lancet 1980;2:815-818

10. Pereira GR, Fox WW, Stanley CA, et al: Decreased

oxygen-ation and hyperlipemia during intravenous fat infusions in

premature infants. Pediatrics 1980;66:26-30

11. Dahms BB, Halpin TC Jr: Pulmonary arterial lipid deposit

in newborn infants receiving intravenous lipid infusion. J

Pediatr 1980;97:800-805

12. Dubowitz LMS, Dubowitz V, Goldberg C: Clinical

assess-ment of gestational age in the newborn infant. J Pedwtr

1970;77:1-10

13. Freeman MG, Graves WL, Thompson RL: Indigent Negro

and Caucasion birthweight-gestational age tables. Pediatrics

1970;46:9-15

14. Inwood RI, Gora P, Hunt CE: Indomethacin inhibition of Intralipid-induced lung dysfunction. Prostaglandin Med

1981;6:503-514

15. Hageman JR, McCulloch K, Gora P, et al: Intralipid

alter-ations in pulmonary prostaglandin metabolism and gas

ex-change. Grit Care Med 1983;11:794-798

16. Friedman Z, Marks KH, Maisels MJ, et al: Effect of

paren-teral fat emulsion on the pulmonary and reticuloendothelial

systems in the newborn infant. Pediatrics 1978;61:694-698

17. McKeen CR, Brigham KL, Bowers RE, et al: Pulmonary vascular effects of fat emulsion infusion in unanesthetized

sheep: Prevention by indomethacin. J Gun Invest 1978;

61:1291-1297

18. Strunk RC, Murrow BW, Thilo E, et al: Normal macrophage

function in infants receiving Intralipid by low-dose inter-mittent administration. J Pediatr 1985;106:640-645

19. Cashore WJ: Growth and transcutaneous oxygen transport

in very low birthweight infants receiving intravenous fat

emulsion: Clinical conference on pediatric nutrition: The

role of Neopham and Intralipid in TPN. Acta Paediatr

Scand 1982;517(suppl):123-134

VIEWS THAT DESERVE TO BE HEARD

I have

recently

had

occasion

to interview

15 mothers of 17 severely mentally

handicapped

young

adults

aged

between

19 and 25 years. . . .

Three

mothers

thought

all medical

means

should

be used

to keep

such

infants

alive.

. . . 12

mothers

took

the

opposite

view.

. . .

It is clear

that

most

of these

mothers

do

not look upon

a lifetime

spent

caring

for the severely

mentally

handicapped

as

time

well

spent,

even

though

they

love

their

children,

have

compassion

for

them,

and

want

to do the

best

for them

that

they

can.

Those

who

have

had

twenty

years’

experience

caring

for young

people

classified

as severely

mentally

handicapped

have

views

that

perhaps

deserve

to be heard

more

than

most.

Submitted by Student

(7)

1986;78;79

Pediatrics

L. West

Yves W. Brans, Elizabeth B. Dutton, Donna S. Andrew, Elizabeth M. Menchaca and Donna

Oxygen in the Lungs and on Blood pH