Liposyn infusion increases plasma prostaglandin concentrations

Liposyn Infusion Increases Plasma Prostaglandin Concentrations Carl E. Hunt, MD, Lauren M. Pachman, MD, Joseph R. Hageman, MD, Michael A. Cobb, MA, and Linda Klemka, BA

Summary. To determine whether Liposyn infusion results in increased plasma prostaglandin (PG) concentrations, the following study was performed in 33 adult rabbits with chronically implanted arterial and venous catheters. Plasma PG concentrations were determined by radioimmunoassay for two vasodilators, PGE, and PGI, (as measured by its metabolite 6-keto-PGF1a),and t w o vasoconstrictors, thromboxane (TX) A, and PGFla, as measured by their metabolites TXB, and PGFza-M, respectively. A I-hour infusion of Liposyn at 4 ml per kg resulted in statistically significant increases in arterial and venous concentrations of PGE, and 6-keto-PGF1a(p < 0.001) and of TXB, (p < 0.04). There were no significant changes in PGFza-M plasma concentrations. Liposyn infustion also resulted in a small but statistically significant increase in Pao, of 4.7 f 1.5 torr (p < 0.01). It is concluded that Liposyn infusion results in statistically significant increases in plasma concentrations of PGE,, 6-ket0-PGFla, and TXB,. (Key words: Increased plasma concentrations; PGE,; 6-ket0-PGFla; TXB,; Liposyn infusion; 33 adult rabbits.) Pediatr Pulmonol 1986; 2:754-158.

Intralipid is a 10%fat emulsion containing 56 mg per ml of linoleic acid, 7.7 mg per ml of linolenic acid, and 24 pg per ml of arachidonic acid. We have previously demonstrated that Intralipid infusion in rabbits with oleic acid-induced lung injury results in increased plasma concentrations of PGE, and 6-keto-PGFIa,associated with a mean decrease in Pao, of 12 torr.1.2In animals with normal baseline lung function, however, Intralipid infusion was not associated with any increases in plasma prostaglandin (PG)concentrations. Liposyn also is a 10% fat emulsion, but it differs from Intralipid in that its linoleic acid content is greater (77 mg per ml) and linolenic acid is essentially absent (< 0.5%).The following study was designed to determine whether Liposyn infusion in rabbits with normal baseline lung function results in increased plasma PG concentrations. We hypothesized that, with its increased linoleic and decreased linolenic acid concentrations, Liposyn would result in significantly increased PGE, and 6-keto-PGF1aplasma concentrations. Because Intralipid has been ~

~

~~

From the Department of Pediatrics, Divisions of Neonatology and Immunology, Northwestern University Medical School, Children's Memorial Hospital, Chicago, Illinois Received July 24, 1985, revision accepted for publication Decem ber 31, 1985 Supported by grants from Abbott Laboratories (Hospital Prod ucts Division) and NHLBl grant HL-24607 Address correspondence and reprint requests to Dr Hunt Divi sion of Neonatology, Children s Memorial Hospital, 2300 Children's Plaza, Chicago, Illinois 60614

154

reported to result in increased thromboxane (TX)p r o d ~ c t i o nwe . ~also measured plasma concentrations of TXB,. the metabolite of TXA,.

Materials and Methods These experiments were conducted in adult pertussis-free New Zealand albino rabbits weighing 2.4 to 4.9 kg. Each experiment required 2 days to complete. On the first day, the animal was anesthetized by an intraperitoneal injection of 35 mg per kg ketamine and 5 mg per kg xylazine. A lateral neck incision was made and the jugular vein and carotid artery isolated. An 8Fr Argyle umbilical catheter was inserted 7.5 cm into the jugular vein so that the tip was in the right atrium. A 5-Fr Argyle umbilical catheter was inserted 4 cm into the carotid artery so that the tip was in the thoracic aorta. The catheters were subcutaneously tunneled posterolaterally behind the right ear to exit through the scalp at the top of the head. Postoperatively, 100 mg per kg of chloramphenicol were administered intravenously. Each rabbit was then placed in a small animal sling (AliceKing Chatham. Medical Arts. Los Angeles. CA) for the duration of the experiment. The catheters were connected to a dualchamber infusion pump (Harvard, Model 975). and each catheter was continuously infused overnight at a rate of 3 ml per hour with lactated Ringer's solution containing 0.5 units per ml of heparin. On the second day, all 33 rabbits received 4 ml per kg over 1 hour of 10%Liposyn (Abbott).

pediatric pulmonology

155

Vol. 2, No. 3-May-Jun. 1986

Intra-arterial blood pressure was monitored continuously beginning 1 hour prior to the Liposyn infusion, using a Gould-Statham P23ID transducer and a Hewlett-Packard 80300A monitor. Arterial pH. Pco,. and PO, were measured using 0.1 ml of heparinized whole blood and a Radiometer ABL2 blood gas analyzer. End-tidal or alveolar CO, tension (PAco,) was determined using a Cavitron CO, analyzer (Model PM-1OR). At the time of each arterial blood gas sample. the end-tidal gas sampling catheter was manually held in the nasal air stream and a representative continuous tracing of expired Pco, was obtained. Serum triglyceride measurements were obtained from 1.0ml of whole blood samples and analyzed as previously reported.' Four PGs were measured, two vasodilators and two vasoconstrictors. The vasodilators were PGE, and prostacyclin (PGI, as measured by its metabolite 6-ket0-PGF~a).The two vasoconstrictors were thromboxane (TX)A, and PGFza, as measured by their metabolites TXB, and 13.14dihydr0-15-keto-PGF~~ (PGFza-M). respectively. At each sampling interval, 10 ml of arterial and venous blood was drawn into syringes containing 0.4 M NaEDTA as an anticoagulant and 0.02 M aspirin as a PG synthetase inhibitor, at final concentrations of 4 mg per ml and 100 pg per ml of whole blood, respectively. Samples were spun at 3200 x g for 10 minutes in a refrigerated centrifuge. The plasma was separated except for the final 0.5ml and frozen in liquid nitrogen pending PG analysis. Organic extracts of plasma samples were subjected to silicic acid chromatography to separate PGs for subsequent radioimmunoassay (RIA).1.5 One ml of plasma was spiked with approximately 3500 cpm of 'H-6-keto-PGF1a(New England Nuclear, Boston, MA) to monitor extraction efficiencies (2 60%). Neutral lipids were removed by first extracting the plasma with five volumes of heptane:isopropanol: 1N H 2 S 0 4 (240:236:6).After centrifugation (1500 x g, 5 minutes). the aqueous layer was extracted with 7 ml of chloroform. The chloroform layer was removed after centrifugation (1500 x g, 5 minutes) and evaporated to dryness under nitrogen. The residue was dissolved in 0.5 ml of methano1:chloroform (2:98)and applied to mini-columns containing 1 gm of washed silicic acid (100 mesh, Mallinckrodt). Fatty acids were eluted first with 7 ml of methano1:chloroform (2:98). PGs were then eluted from the columns with 7 ml of methano1:chloroform (10:90). The PG fraction was evaporated to dryness under nitrogen and dissolved in 1 ml of ethyl acetate. A 0.2 ml aliquot

of the ethyl acetate was then removed to measure extraction efficiency. Duplicate 0. l-ml aliquots of the ethyl acetate sample were dispensed into polypropylene RIA tubes (Sarstedt, West Germany) and evaporated and the residue dissolved in 0.05-M phosphate buffer (pH 6.8).RIA of PGE,. 6-keto-PGF1a,PGFza-M, and TXB, was performed on each phosphate buffer-dissolved sample. The assays of PGE, and TXB, were carried out using specific antibody (Seragen, Boston, MA) and standard PGE, and TXB, (Upjohn. Kalamazoo, MI). For the PGE, assay, [5.6,8.11.12.14.15 (n)-'H] PGE, specific activity 120-170 Ci/mmol was used, and for the TXB, assay (Amersham. Arlington Heights, IL), [5.6.8.9,11.12.14,15 (n)'HI TXB, was used. An RIA kit (NewEngland Nuclear) was used to assay levels of 6-ket0-PGF~a. The RIA of PGF,a-M was performed using standard PGFza-M antibody raised in our laboratory rabbits using bovine serum albumin conjugated with PGFza-M (1:7 ratio). Specific PGFla-M was separated from serum components by purification of IgG utilizing diethylaminoethyl (DEAE) cellulose (Whatman,England) column chromatography. Cross reactivity of the PGFza-M antibody was 6-keto-F1a.< 0.1%; PGE,. < 0.1%; and PGF,,. < 0.1%. The following measurements were obtained immediately prior to and at the end of the 1-hour infusion: serum triglyceride; arterial pH. Pco,. and Po,; PACO,;arterial and venous PGE,, 6keto-PGFla, TXB,. and PGFla-M. The arterialalveolar gradient for Pco, [P(a-A)CO,]was calculated. For each rabbit, the baseline l-hour differences were analyzed by paired t-test. All intergroup differences were analyzed by unpaired t-test. All group results were expressed as the mean f S.E.M.

ResuIts Baseline

The baseline values for all parameters are summarized in table 1. Liposyn Infusion

The mean serum triglyceride level following infusion was 1050 f 80 mgM1. a statistically significant increase compared to baseline (p< 0.01). There were no significant blood pressure changes. There was a small but significant Pao, increase, 4.7 f 1.5 torr (p < 0.01).Mean Paco, decreased by 0.5 f 0.6 torr. and mean P(a-A)CO, decreased by 0.9 f 1.3torr; neither change from baseline was statistically significant.

Increased Prostaglandin Levels with Liposyn-Hunt et a/.

156 table 1-Baseline Blood gases

Values for Each Measurement (mean f S.E.M.) Result

Prostaglandins

91 f 2

PGE,

28

6-keto-PGF,,

f 1

3 f 1

PGF,-M

TXB,

The increases in plasma PG concentrations from baseline are summarized in figure 1. For each PG, the mean increase was greater in venous than in arterial plasma. Liposyn infusion resulted in significant increases in both arterial and venous plasma concentrations of PGE,, 6keto-PGFla,and TXB,. There were no significant changes in PGFza-M concentrations. The magnitude of the PG increases was comparable for TXB, and 6-keto-PCF1a.both being significantly less than the increase in PGE,.

Discussion The primary objectives of this Liposyn study were to determine whether infusion results in increased plasma PG concentrations and whether pulmonary gas exchange is thereby impaired. Liposyn infusion resulted in significant increases in both vasodilator PGs measured, the increases being significantly greater for PGE, than for 6-keto-PGFla. Of the two vasoconstrictor metabolites measured, TXB, concentrations increased significantly, but PGFza-M concentrations did not increase. Overall, there was a significantly greater increase in vasodilatory PGs (E, plus 6-keto-Fla)than in vasoconstrictive PC plasma concentrations (TXB, plus Flu-M). The net increase in vasodilator relative to vasoconstrictor PGs likely explains the associated Po, and Pco, changes. The Liposyn-related increase in vasodilator PG concentrations may have resulted in improved ventilation/perfusion (VA/Q)balance owing to improved perfuion.'.^ causing the small but statistically significant Pao, increase. The mean Paco, and P(aA)CO, changes, trends that were not statistically significant, are also consistent with vasodilator PG-mediated VA/Q changes resulting from enhanced perfusion. Although a consideration of PG-related alterations in VA/Q may be of physiologic importance, none of the Liposyn-related Po, or Pco, changes is of sufficient magnitude to have any clinical significance.

Result (pg/ml) arterial venous arterial venous arterial venous arterial venous

910 f 170 710 90 210 f 50 170 f 20 300 f 70 240 + 50 30 f 30 140 f 140

The Liposyn-related PG increases were significantly greater than in our previous study with Intralipid,1,2 even though the infused volume was identical. In contrast to Intralipid infusion, which resulted in Statistically significant PG increases only with oleic-acid-induced lung injury,2 Liposyn infusion resulted in statistically significant increases in PGE,. 6-keto-PGF1a,and TXB, (see figure 1)in rabbits with normal baseline lung function. Intralipid infusion in rabbits with baseline oleic acid lung injury resulted in proportionately greater increases in arterial than in venous PG concentrations, suggesting increased pulmonary production.2 Unrelated to baseline lung function, however, Liposyn infusion resulted in significant and comparable increases in both venous and arterial PG concentrations. Although Liposyn infusion clearly results in increased PG production, we cannot draw any conclusions regarding pulmonary PG production since the increased arterial levels may simply reflect overloading of the pulmonary clearance capability. There are two potential reasons for significant increases in PG production with Liposyn in rabbits with normal lung function that did not occur with Intralipid. Although Liposyn and Intralipid are both 10% emulsions, Liposyn contains significantly more linoleic acid, 77 % versus 54%. Because the linoleic acid can be rapidly converted in vivo to arachidonic acid,',6it is not unexpected that Liposyn infusion results in higher concentrations of arachidonic acid metabolites such as these PGs. A second factor potentially causing proportionately greater PG increases with Liposyn is related to linolenic acid. Intralipid contains 8% linolenic acid, whereas Liposyn contains less than 0.5%.Because linolenic acid had been shown to inhibit the conversion of linoleic to arachidonic acid,' the linoleic acid present in Liposyn may be more effectively converted to arachidonic acid. The early studies of intravenous fat emulsions attributed the physiologic consequences of infusion to the resulting hyperlipemia.8Our previ-

pediatric pulmonology

9

Vol. 2, No. 3-May-Jun. 1986

157

figure 1-Changes from baseline in arterial and venous plasma concentrations o f PGEZ, 6-keto-

PGFZa, TXB,, a n d PGFZa-M in response t o Liposyn in-

CHANGE IN PLASMA LEVEL (pglrnl)

fusion. The change from baseline is statistically significant for each PG exc e p t PGFZa-M.

ous studies, however, showed no relationship between serum triglyceride levels and the resulting blood gas changes.'.' This study further confirms the lack of relationship between hyperlipemia and impaired ventilation or oxygenation, since Liposyn infusion yielded the greatest increase in serum triglyceride concentrations but no decrease in Pao, and no significant decrease in Paco,. The effects of Liposyn infusion on increased PG production in humans should be qualitatively similar to our results in rabbits. Adults receiving 1.5gm per kg per 8 hours of fatty acids as Liposyn, for example, resulted in increased urinary concentrations of both PGE, and PGI,. the only difference from our rabbits being a proportionately greater increase in 6-keto-PGFI, compared to PGE.9The increased PG production occurring with Liposyn may thus have clinically significant physiologic consequences that have previously been unrecognized. Circulating PGE, concentrations appear to have a significant role in regulating ductal patency in premature lambs.'O Although clinical studies have not yet demonstrated an increased frequency and/or severity of symptomatic patent ductus arteriosus in preterm infants receiving parenteral fat emulsions, we have now observed two ventilatordependent preterm infants with respiratory distress syndrome whose patent ductus arteriosus remained closed for several days following indomethacin treatment but then reopened 1 to 2 days after Intralipid infusion was begun (unpublished obskrvation). A net increase in vasodilating PGs subsequent to infusion of a fat emulsion may also increase the potential risk for retrolental fibroplasia, as retinal vasoconstriction may

be an important physiologic mechanism for protecting the immature retina from the damaging effects of hyperoxia." Indomethacin has been shown to reduce significantly the incidence of intraventricular hemorrhage in beagle puppies, presumably by preventing PG-mediated increases in cerebral blood flow." Further studies will be necessary to assess the clinical implications of increases in circulating PG concentrations such as can occur with Liposyn infusion. To Stephen Lee for assisting in the data analysis and to Judy Carbone, Susan Seidler, and Judy Lach for assisting in preparation of this article

References 1 Hageman JR, McCulloch K, Gora F! et al lntralipid alterations in pulmonary prostaglandin metabolism and gas exchange Crit Care Med 1983, 11 794-798 2 Hageman JR, McCulloch K, Hunt CE, et al Oleic acid lung injury increases plasma PG levels Prostaglandins Leukotrienes Med, in press 3 Gurtner GH, Knoblauch A, Smith PL, et al Oxidant and lipid in duced pulmonary vasoconstriction mediated by arachidonic acid metabolites J Appl Physiol 1983, 55 949-954 4 lnwood RI, Gora p Hunt CE lndomethacin inhibition of IntralipidR-inducedlung dysfunction Prostaglandin Med 1981, 6 503-51 4 5 Hsueh W, Jordan RL, Harrison HH, et al Serum and plasma stimu late prostaglandin production by alveolar macrophages Prostaglandins 1983, 25 793-808 6 Huang YS, Mitchell J, Jenkins K, et at Effect of dietary depletion and repletion of linoleic acid on renal fatty acid composition and urinary prostaglandinexcretion Prostaglandins Leukotrienes Med 1984, 15 223-228 7 Hwang DH, Carroll AE Decreased formation of prostaglandins derived from arachidonic acid by dietary linolenate in rats Am J Clin Nutr 1980, 33 590-597 8 Greene HL, Hazlett D, Demaree R Relationship between intralipidinduced hyperlipemia and pulmonary function Am J Clin Nutr

Increased Prostaglandin Levels with Liposyn-Hunt

158 1976, 29 127-1 35

9 Epstein M, LifschitzM, Rappaport K Augmentation of prostaglandin production by linoleic acid in man Clin Sci 1982, 63 565-571 10 Clyman RI, Mauray F, Roman C, et al Effect of gestational age on ductus arteriosus response to circulating prostaglandin E J kdiatr 1983, 102 907-911

et a/.

11 Flower RW, Blake DA, Wajer SD, et al Retrolental fibroplasia Evi dence for a role of the prostaglandin cascade in the patho genesis of oxygen induced retinopathy in the newborn bed gle Pediatr Res 1981, 15 1293-1302 12 Ment LR, Stewart WB, Scott DT, et al Beagle puppy model of intra ventricular hemorrhage Randomized prevention trial Neurology 1983, 33 179-184