Dose-response effects in healthy volunteers Marion C Blonk, Henk JG Bib, Cees Mulder, andAb JMDonker

Jos JP Nauta,

We performed on the dose-response effects

1 .5, 3, and

6 g of the

marine

Corrie

Popp-Snzjders,

a randomized, controlled ofdaily supplementation of

ABSTRACT

study

of fish-oil supplementation

fatty

acids

eicosapentaenoic

acid

(EPA, 20:5n-3) and docosahexaenoic acid (DHA, 22:6n-3)as their ethyl esters for 12 wk in 45 healthy normotriglyceridemic male volunteers. Significant dose-related increases of the n-3 fatty acids 20:5, 22:5, and 22:6 in plasma phospholipids (p < 0.0001) were found, corresponding roughly to decreases of the n-6 fatty acids 18:2 and 20:4 (p < 0.001). Serum triglycerides

and

HDL3-cholesterol

pendent reduction (p < 0.05). Results

(p < 0.05) for 3 and

No dose-dependent

and total ing

effects

that

to

observed

increased

acids

in the

blood

were

similar.

VLDL-,

LDL-,

pressure;

bleed-

or capacity

kill

3 g n-3

a dose-dc-

cholesterol

fatty

deformability;

leukocytes

indicates

HDL2

subfractions;

erythrocyte

phonuclear

and

showed

6 g n-3

were

HDL-cholesterol

time;

study

concentrations

of polymor-

Staphylococcus

ethyl

ester

aureus.

fatty

acids

This

appears

to

be the appropriate

supplementation dose in humans, at least regarding lipid-profile changes and the ability to incorporate such fatty acids in the plasma phospholipids. Am J Clin Nutr 1 990;52: 120-7.

literature on n-3 PUFAs. This can be explained partly by the lack of proper control groups in many studies, which makes correction for time trends, conditions of measurements, and laboratory drift impossible. The majority ofthe studies showed a reduction in serum triglyceride concentrations, with the most striking reductions in hypertriglyceridemic patients(4, 6, 7, 10, 15, 19-21).

Reported tein

cosity

Fish oil, blood pressure,

oferythrocytc

eicosapentaenoic

lipid

suspensions,

profile,

leukocyte

acid,

bleeding

docosa-

time,

vis-

vascular-risk

docosahexaenoic

acid

(DHA,

22:6n-3).

investiga-

tions.

attributed

Beneficial

effects

ofn-3

fatty

acids

have

been

to an improved lipid profile (1, 2, 4-7), a reduced platelet aggregability (2, 4, 5, 8, 9), a prolonged bleeding time ( 10, 1 1 ), a reduced blood viscosity (12-14), and a fall in blood pressure 15) as well ( 16,

inconsistent, recently

120

very-low-

profile

remains

open

for discussion.

data arc not available regarding the dose of n-3 fatty acids necessary for producing most of the described beneficial effects. The purpose ofthc present study was to examine doseresponse relationships with blood pressure, lipid and lipopropatterns,

bleeding

time,

and

function in normotriglyceridemic the appropriate dose. and

as to anti-inflammatory

1 7). However, especially

confirmed

results regarding

by Leaf

and

and

from

the

lipoprotcin Weber

immunological

various

studies

patterns,

J C/in

were as was

(1 8) in a review Am

Forty-five range

crythrocyte

healthy

and volunteers

leukocyte

to find

methods

Nutr

of the

No

healthy

male

y) were

investigated.

volunteers

drugs

were

taken

from

aged 33.7 ± 6.2 y (1± Body

weight

was

77.7

SD, ± 9.3

wt/ht2) was 23.2 ± 2.0 kg/rn2. consent. The study was apand Ethical Committee of the 4 wk before

the

start

until

the

I From the Departments oflnternal Medicine; Theory of Medicine, Epidemiology and Biostatistics; Endocrinology; and Ginical Chemistry; Free University Hospital, Amsterdam. 2 Supported by the Dutch Kidney Foundation (Nier Stichting Nederland) grant C86.60 1. 3 Address reprint requests to MC Blonk, Department of Internal Medicine, Free University Hospital, Dc Boelelaan I 1 17, 1081 HV Amsterdam, The Netherlands. Received March 2, 1989. Accepted for publication August 9, 1989.

1990;52:120-7.

Downloaded from https://academic.oup.com/ajcn/article-abstract/52/1/120/4695368 by University of Glasgow user on 03 April 2018

22-48

kg and a body mass index (BMI; All participants gave informed proved by the Human Research Free University.

Presum-

ably Eskimos ingest -5-10 g n-3 fatty acids/d (3). Growing interest in the subject has led to extensive

effects

and

Subjects

The observations in the mid-l970s by Bang ct al (1) and Dyerberg et al(2) ofthe low incidence ofcardiovascular disease among Greenland Eskimos led to the idea ofa possible protective role oftheir diet, which consists predominantly of marine fish and seal. Marine fatty fish is rich in n-3 polyunsaturated fatty acids(PUFAs), especially eicosapentacnoic acid (EPA, 20:

(10,

(LDL),

Good

Subjects

and

of high-density-lipopro-

function

Introduction

5n-3)

in concentrations

low-density-lipoprotein

density-lipoprotein (VLDL) cholesterol arc not consistent. Cholesterol-lowering effects were reported primarily in studies with extreme doses of n-3 PUFAs of 30 g/d (4, 6, 21-23). However, the Zutphcn study suggests that the consumption of small amounts of fish (a low n-3 PUFA intake) over a long period may reduce cardiovascular risk (24). Whether a high daily intake ofn-3 PUFAs is necessary to influence the cardio-

tein

KEY WORDS hcxaenoic acid,

changes

(HDL),

Printed

in USA.

C 1990 American

Society

for Clinical

Nutrition

DOSE-RESPONSE

STUDY

end ofthe study. Average alcohol intake was 1.4 units/d 0-3 units/d; 1 unit is equivalent to 15 mL pure alcohol). two

participants

were

nonsmokers.

The

other

(range Thirty-

13 smoked

an

WITH

FISH

presence

100 mL

of

viable

bacteria

The killing

121

OIL

were

capacity

pooled

human

counted

after

(KC)

was defined

average of 7 cigarettes/d (range 1-20 cigarettes/d). The fish consumption before the study never exceeded one fish meal per week. Subjects were excluded from the study when fasting serum triglycerides were > 2.0 mmol/L at baseline. Subjects kept their normal diets during the experiment and a 72-h dietary recall was completed in the first month ofthe study. The

Control tubes without PMNLS ment. Killing always was found

subjects

killing

were

randomly

assigned

to one

offour

groups

and

took

0, 3, 6, or 12 capsules/d, respectively, ofa marine-lipid concentrate (Super EPA, Pharmacaps Inc. Marlow, Buckinghamshire, UK) for 12 wk. Each capsule provided 300 mg EPA and 200 mg DHA as their ethyl esters and I mg vitamin E as described in a previous

study

(25).

This

implied

that

daily

doses

ofO,

1.5,

3, and 6 g n-3 ethyl ester fatty acids (EEFAs) were administered daily. Measurements were performed before, at 12 wk of, and at 12 wk after supplementation. Subjects fasted overnight from 2100 and venous blood samples were drawn with minimal venous occlusion between 0800 and 1000 the next morning. Before samples were drawn, blood pressure was measured in duplicate with a London School of Hygiene mercury sphygmomanometcr by the same observer after the subjects had rested 15 mm in supine position and immediately after the subjects stood up. Blood samples were used for measurement of hematologic and biochemical variables, lipid profile, plasma phospholipids, and leukocyte

killing

capacity.

Erythrocyte

and bleeding-time measurements and at the end ofthc 12-wk period

deformability

were performed oflipid-concentrate

studies

only

before admin-

istration.

Methods The

fatty

acid composition

ofplasma

sessed by capillary gas chromatography (26). Plasma triglyceride concentrations

phospholipids

was as-

as described previously were measured enzy-

matically (GPO-PAP method) after removal of chylomicrons as described by Terpstra (27) as was the total amount of free fatty acids (FFAs; Nefa-C test, Wako Ltd. Tokyo). Lipoprotein fractionation was carried out by density-gradient ultracentrifugation

(28).

Total

of the fractions (Boehringer

plasma

were

cholesterol

analyzed

Mannheim

GmbH,

and

with

cholesterol

content

the CHOD-PAP

method

Mannhcim,

FRG).

Bleeding-

time measurements were performed as described by Mielkc et al (29) with the Simplate II device (General Diagnostics, Turnhout, Belgium). Erythrocyte filterability was studied with the recently

introduced

St George’s

Filtrometer

(Cassi-Med

Ltd,

Dorking, UK), which can discriminate between cell deformability and filter occlusion (30-32). An crythrocytc concentration of 10% was used for filtration. The white cell contamination in the eiythrocyte suspension was < 0. 1 X 109/L and platelet contamination was not detectable. The erythrocyte suspensions were filtered through a vertical filter (batch 5404 C34, Nucleopore, Pleasanton, CA; nominal pore diameter 5 aim, filter diameter The filter-to-filter sured in duplicate.

13 mm, effective filtration area 0.78 cm2). variation was < 5%. All samples were mcaErythrocyte dcformability was expressed as

erythrocyte

time.

transit

The killing capacity of polymorphonuclear leukocytes (PMNLs) was measured as described in detail elsewhere (33). Staphylococcus aureus was incubated with PMNLs at 2: 1 in the

Downloaded from https://academic.oup.com/ajcn/article-abstract/52/1/120/4695368 by University of Glasgow user on 03 April 2018

bacteria

The

60 mm

remaining incubation.

as

at 0 mm

=

KC

serum/L. 30 and

bacteria

at 30 mm (or 60 mm)

>< l00

bacteriaat0min

was observed

without

were included in the presence

in each experiof PMNLs; no

PMNLS.

Statistics In this paper hypotheses ofthe form H1 : t6 P12 or H1 : Pi P3 P6 P12 (with at least one strict inequality) were investigated, with pn denoting the theoretical mean of a population on a diet regimen ofn (n = 0, 3, 6, or 12) capsules of lipid

concentrate

daily.

These

hypotheses

imply

relations of the treatment on the variables null hypotheses of the form H: p p alternative

hypotheses

response

relations were considered p value was < 0.05. For the

Jonckheere-Terpstra triglyceride

test

was

concentrations

Observations baseline

at weeks

as absolute

the

order

changes

when

24 were compared with the EEFAs can be expressed changes. Absolute changes were of n-3

were

independent

of baseline

-

with

changing

changes

were

studied

baseline

values.

applied

to intraindividual

doses.

when

In those

the

cases

By

ranges

to allow

values.

test was (eg, week 12 increased or

contrast,

changes

were

percentage dependent

the Jonckheere-Terpstra

quotients

[eg,

(week

week 0J. Although nonparametric tests were used, data were summarized by means and SEMs rather and

that

EEFAs.

Under such circumstances, the Jonckhcerc-Terpstra applied to the absolute intraindividual changes week 0) to investigate whether sample means decreased

the twoone-tailed

it is known

n-3

test

Dose-

12 and

Effects

or as percentage when

test. the

because with

against the

(34). This

significant

applied

To test

P12

is specified,

triglycerides

decrease

observations.

studied

in which

P6

test was applied to the data test related to the Kruskal-Wallis

Jonckheere-Tcrpstra is a nonparametric tailed

(H1)

dose-response

of interest.

comparisons

with

other

12

-

on

test was week 0)/

for each group than medians studies

of n-3

EEFAs.

Results The four groups of subjects were comparable for age, BMI, smoking habits, and alcohol consumption. The amounts of fat, carbohydrate, and protein in the diets ofthe groups were cornparable as shown by the 72-h dietary recalls. The lipid-concentrate capsules were well-tolerated although some subjects described a fishy aftertaste. Body weight did not change during the study period (0.76 ± 0.49 kg, i ± SD). All scheduled visits were cessfully

completed. with

the 0-(control),

Supplementation marked changes

The

45 volunteers

10, 1 1, 10, and 3-, 6-, and

completed

14 participants,

12-capsule

the

study

respectively,

suc-

in

groups.

with n-3 fatty acids in fatty acid composition

for 12 wk resulted in ofthc plasma phospholipids in all groups except for the control group (Table 1). There was a significant dose-related increase in the proportion of n-3 EEFAs 20:5n-3, 22:5n-3, and 22:6n-3 with the use

122

BLONK

ET

AL

TABLE

1 acid composition

Fatty

ofplasmaphospholipidsafter

various

0 capsules

doses

offish

oil5 12 capsules

t

2 1.49 ± 0.58 17.78 ± 0.76 [-3.71] 2 1.98 ± 0.60 [0.18]