European Journal of Clinical Nutrition (2000) 54, 632±642
ß 2000 Macmillan Publishers Ltd All rights reserved 0954±3007/00 $15.00
www.nature.com/ejcn
De®ciencies of essential fatty acids, vitamin A and E and changes
in plasma lipoproteins in patients with reduced fat absorption
or intestinal failure
PB Jeppesen1*, C-E Hùy2 and PB Mortensen1
1
2
Department of Medicine CA, Section of Gastroenterology, Rigshospitalet, University of Copenhagen, Copenhagen, Denmark; and
Department of Biochemistry and Nutrition, The Technical University of Denmark, Lyngby, Denmark
Objective: De®ciencies of essential fatty acids (EFA), vitamin A (retinol) and vitamin E (a- and g-tocopherol)
were studied in non-HPN patients with different degrees of fat malabsorption (managing without home
parenteral nutrition (HPN)), and in HPN-patients receiving HPN with and without parenteral lipids.
Design and subjects: Phospholipid fatty acids (including EFA), cholesterol, retinol, a- and g-tocopherol in
plasma and the intestinal absorption of fatty acids and energy (balance-studies) were measured in 40 non-HPN
patients, 44 HPN patients and 35 controls. Subgroups were non-HPN patients with fat:total energy absorption
ratios >25% (A), between 15 and 25% (B), and <15% (C), and HPN patients receiving (D) and not receiving
parenteral lipids (E).
Results: Intestinal absorption of the EFA linoleic acid was 8.2, 4.4, 3.8, 0.5 and 0.5 g=day and corresponding
plasma concentrations were 17.3%, 15.5%, 13.1%, 12.1% and 8.9% in groups A ± E, respectively (P < 0.001).
De®ciencies in EFA, de®ned as a Holman index >0.2 (20:3n-9=20:4n-6 ratio), were con®ned to 42% of the
patients in group E. Plasma cholesterol was decreased in groups B ± E. Plasma retinol was reduced (below the
lower 2.5% con®dence interval of controls) in 7% of non-HPN patients and in 20% of HPN patients. Plasma
a-tocopherol was reduced in 64% of patients from groups B ± E. Plasma g-tocopherol was decreased in 33% of
the patients, except in HPN-patients receiving parenteral lipids.
Conclusions: Plasma linoleic acid may decrease considerably (from 26% to 8 ± 10%) as fat absorption decreases
before secondary signs of essential fatty acid de®ciencies occur (an increase in 20:3n-9 and the Holman index).
In this study this was con®ned to patients on lipid-free HPN. Vitamin A de®ciencies were mainly seen in HPN
patients. Vitamin E de®ciencies were common in both HPN and non-HPN patients, but administration of
parenteral lipids normalized plasma g-tocopherol.
Sponsorship: The study was sponsored by Rigshospitalet.
Descriptors: fat malabsorption; de®ciencies; essential fatty acids; cholesterol; retinol; tocopherol
European Journal of Clinical Nutrition (2000) 54, 632±642
Introduction
Suf®cient intake and absorption of a diet is required to meet
or exceed the nutritional needs of the individual, thereby
maintaining the nutritional equilibrium, body composition,
function and health (Jeejeebhoy et al, 1990a). This equilibrium is challenged in patients with diseases that impair the
integrity of the intestinal epithelium, either through in¯ammation, in®ltration, surgical resection or ischemia, causing
malabsorption. Short bowel patients frequently compensate
for the malabsorption by hyperphagia (Cosnes et al, 1985,
1990; DiCecco et al, 1987), but patients who do not meet
their nutritional needs suffer from intestinal failure, impos-
*Correspondence: PB Jeppesen, Department of Medicine, Section of
Gastroenterology CA 2121, Rigshospitalet, Blegdamsvej 9, 2100
Copenhagen, Denmark.
E-mail [email protected]
Guarantor: PB Jeppesen.
Contributors: All the investigators contributed to the study design. PBJ
initiated and supervised the clinical part of the study, collected the data and
drafted the manuscript. All authors were involved in the revision of the
manuscript.
Received 15 October 1999; revised 27 March 2000; accepted
12 April 2000
ing the need for parenteral support (Fleming & Remington,
1981).
De®ciencies may evolve in the intermediate phase from
moderate malabsorption to the condition of intestinal failure according to the balance between the dietary intake and
the pre-existent body stores on one side, and the metabolic
needs and losses in stools and urine the on the other. Some
de®ciencies may have long latency periods before becoming clinically manifest, whereas others such as electrolyte
depletion and dehydration may evolve rapidly and necessitate parenteral support.
In short bowel patients the malabsorption of fat is more
predominant than the malabsorption of carbohydrate and
protein (Messing et al, 1991). This is most pronounced in
short bowel patients with a preserved colon, where carbohydrate and protein malabsorbed in the small bowel are
fermented in the colon, thus salvaging energy through the
absorption of short chain fatty acids (Nordgaard et al, 1994,
1996). Therefore the energy contribution from fat absorption in relation to the total energy absorption may decrease
with an increase in the overall energy malabsorption.
Accordingly, the patients who manage in a situation with
severe malabsorption without parenteral support by a
compensatory hyperphagia may cover a low percentage
De®ciencies related to fat malabsorption
PB Jeppesen et al
of their total energy requirements with energy supplied
from fat absorption. The reduced absorption of fat, and the
concurrent reduced absorption of essential fatty acids
(EFAs) and fat-soluble vitamins, may lead to depleted
body stores or localized tissue de®ciencies and ultimately
clinical signs of de®ciency of EFAs and fat-soluble vitamins A, D, E and K. In these patients the occurrence
of subclinical or clinical de®ciencies may evolve more
insidiously and receive less attention than in the patients
who receive parenteral support, in whom attention is
addressed to the maintenance of the nutritional equilibrium
through the parenteral supplementation of fat and fatsoluble vitamins.
In this study the ratio between the energy absorption
from fat and the total energy absorption was measured in
balance studies in patients with severe malabsorption and
related secondary biochemical signs of a reduced fat
absorption by measurement of fatty acid composition of
plasma phospholipids, plasma cholesterol (total cholesterol
high-density lipoprotein (HDL) cholesterol and low-density
lipoprotein (LDL) cholesterol), plasma vitamin A (retinol)
and plasma vitamin E (a- and g-tocopherols). One group
of patients with functional or anatomical short bowel
syndrome managed without home parenteral nutrition
(non-HPN patients), whereas the other group of patients
had intestinal failure and received HPN (HPN patients).
Methods
Patients and control subjects
The two groups of patients were recruited by mail for the
study. Both groups had continuity of care in the program
of intestinal insuf®ciency and failure at Rigshospitalet,
Copenhagen. One group consisted of 76 non-HPN patients
with intestinal insuf®ciency, de®ned as either a faecal
energy excretion of more than 2.0 MJ=day (measured at a
previous admission), or a remnant small intestine measured
intraoperatively from the ligament of Treitz of 200 cm or
less, or consecutive small intestinal resections exceeding
150 cm. The other group consisted of the 58 patients with
intestinal failure, corresponding to approximately 75% of
patients receiving HPN in Denmark (Jeppesen et al,
1998a), who received or initiated HPN in the period from
the beginning of the study in November 1995 until the end
in March 1997. Forty-®ve of the 76 non-HPN patients with
intestinal insuf®ciency and 46 of the 58 HPN patients were
willing to participate in the study. These patients were
representative of the total regarding employment status.
One of the non-HPN patients was excluded from the study,
as she experienced a gastroenteritis and vomiting during
admission. One HPN patient with a protein-losing enteropathy was excluded from the study, as she only used her
catheter for intravenous calcium supplements. Five HPN
patients had to be excluded from the study because the
appropriate blood tests for vitamin and fatty acid analysis
were not taken. Forty-four non-HPN and 40 HPN patients
remained for investigation.
Intestinal transit was measured as the time from oral
administration of a brilliant blue marker to its appearance
in the faeces in non-HPN patients and HPN patients with a
preserved colon prior to inclusion in the study. Only three
of these patients had a transit time of more than 10 h. In our
laboratory the reference interval of the intestinal transit
time measured by the brilliant blue test in normal subjects
has been established to be between 10 and 36 h.
Parenteral nutrition and ¯uids
All HPN patients had parenteral supplements of nutrients or
saline at least three times a week. The composition and
energy content in parenteral supplements was calculated
from information given by the manufacturers. Energy and
electrolyte supplements were adjusted to maintain a normal
body weight, hydration, urine output and normal levels of
plasma albumin and plasma electrolytes. Glucose was
supplied as glucose potassium phosphate 27% (The Counties Medicine Registrations Of®ce in Denmark, Copenhagen) or glucose 5, 10, 20 or 50%. Patients were supplied
with synthetic amino acids (Vamin 14 or Vamin glucose,
Pharmacia, Copenhagen). Electrolytes were given as a
hypertonic electrolyte solution containing calcium, potassium, sodium, magnesium, zinc, copper, acetate, chloride,
iodide and glucose. Attempts were made to use `standard'
solutions if at all possible (1 l of glucose 27% potassium
phosphate, 1 l of Vamin 14, 1 l of electrolytes daily), but the
composition of the HPN solution was often adjusted to
meet the patient's individual requirement. The rhythm of
administration was generally cyclic nocturnal, but four
patients with large stoma volumes had supplements of
saline during daytime. The recommended infusion time of
standard 3 l HPN bags was 10 h. Lipids (10 or 20%
Intralipid, Pharmacia, Copenhagen) were supplied in separate 500 ml bottles once or twice a week, mainly to patients
with large energy demands or poor nutritional status in
whom energy requirements were not met by the energy
content of the standard 3 l bags (6.8 MJ). Sixteen of the
40 HPN patients included in this study received parenteral
lipids. Fat-soluble vitamins (Vitalipid Adult, Pharmacia,
Copenhagen) were added to the parenteral lipid emulsions
in 12 of the 16 HPN patients or given once a week as a
10 ml bolus dose in 16 of the 24 HPN patients not receiving
parenteral lipids. Thus, lipid infusion and supplements of
fat-soluble vitamins is not obligatory in the Danish home
parenteral nutrition regime, but is based on a clinical
evaluation of the need. Therefore, the exclusion of lipids
and vitamins in some patients was not due to intolerance or
adverse effects. None of the patients in this study suffered
from clinical signs of EFAD, vitamin A or E de®ciency.
633
Study protocol
Prior to admission the patients were contacted by phone
and informed about the exact study-procedures. Patients
were instructed to bring special food items in duplicate
(snacks, candy, soft drinks etc;), if they had dietary requests
that could not be provided for by the hospital. The patients
were admitted for 212 days. At arrival the ®rst afternoon the
patients were given three containers and an electronic
precision balance. Two of the containers were for collection of faeces and urine, respectively. In the third container
the patients were told to collect a duplicate of their oral
intake Ð beverages and diet. Using the precision balance,
which had a scale in grams, the patients were instructed to
weigh out individual food items and beverages separately; a
portion for themselves and an equal portion for the container. The patients were instructed that they could eat just
as they pleased from a continental style breakfast, lunch
and supper buffet, each containing a wide range of food
items. Beverages included water, tea, coffee, milkproducts, soft-drinks, juice, glucose-saline solutions etc.
Sandwiches, biscuits and beverages were available in the
kitchen between meals, and patients were allowed to use
European Journal of Clinical Nutrition
De®ciencies related to fat malabsorption
PB Jeppesen et al
634
the hospital cafeteria, as long as double portions of the diet
were collected.
The patients were told to abstain from food intake from
10 pm to 8 am during the day of admission. This was done
in order to avoid some of the faecal excretion during the
®rst 24 h of study being a re¯ection of the intake prior to
the balance period. The study- and collection-period began
at 8 a.m. on the second day of admission, where fasting
blood test were taken and patients were requested to
urinate, defecate or empty their stoma-bags. During the
next 48 h the patients collected the faeces, urine and the
duplicate diet in the three containers. The restriction on the
diet intake from 10 pm the last night of admission ensured a
complete collection of the faeces derived from diet intake
during the previous 48 h balance period. The faeces were
collected on ice and immediately frozen at 720 C until the
analyses.
The patients were interviewed regarding composition
and volume of their parenteral support. The patients had
their usual parenteral supplements and medication during
admission. The height and the fasting body weight of the
patients were measured during admission. Fasting whole
blood samples (10 ml) for fatty acid, lipoprotein and vitamin analyses were collected in EDTA. Patients receiving
parenteral lipids and vitamins were instructed not to take
these 48 h before the blood test.
The protocol was approved by the Ethics Committee for
Medical Research in Copenhagen, Denmark. Procedures
followed were in accordance with the ethical standards of
the Helsinki Declaration of 1975, as revised in 1983. All
participants signed informed consent before beginning the
study.
Dietary and faecal analyses
Analyses of the diet and faeces were done on homogenized
and freeze-dried samples. The dietary and faecal energy
content was determined by bomb calorimetry with approximately 1 g of freeze-dried samples ignited in an IKA
adiabatic calorimeter, model C 4000 A (IKA-Analysentechnik, Heitersheim, Germany). The dietary and faecal
fatty acids were determined by combined gas ± liquid
chromatography and mass spectrometry as previously
described (Jeppesen et al, 1998b). In general the peaks
were identi®ed from their retention time. However, in some
instances, especially in the faeces of patients with a preserved colon, positive identi®cation of the fatty acid peaks
was based on mass spectrometry. The energy conversion
factor for fatty acids was 39.1 kJ=g. The intestinal energy
and fat absorptions were calculated as the difference
between the ingested and excreted energy and fat.
Basis of the calculation of parenteral supplements
In general the content of the parenteral nutrition was
calculated from information given by the manufacturers
of the products. Intralipid 10% and 20% contained 52%
linoleic acid and 8% linolenic acid. The triglyceride content
was 115 and 230 mmol=l, respectively. Trilineoleat content
was 52.6 and 105.2 g=l, whereas trilinolenat content was
8.0 and 16.1 g=l. Energy contents in the two emulsions
were 4600 and 8400 kJ=l, respectively. Information regarding the content of fat-soluble vitamins in the Intralipid
solutions could not be provided by the manufacturer.
However, studies by Kelly et al (1989) determined the a,
b ‡ g, and d-tocopherol isomer concentrations in the Intralipid emulsion to 12.0, 91.0 and 43.2 mg=l, respectively,
European Journal of Clinical Nutrition
and these values were used for the estimate of parenteral
vitamin E supply in this study. Thus, the majority of
vitamin E present in Intralipid was as the b, g and dtocopherol isomers with a-tocopherol representing only
about 8% of the total vitamin E concentration. The relative
potencies of the tocopherols have been estimated to be 1.0,
0.2 and 0.01 for a, b ‡ g and d-tocopherol, respectively.
Therefore the plasma total a-tocopherol equivalent may be
calculated as a-tocopherol ‡ 0.2 g-toco-pherol ‡ 0.01dtocopherol (Century & Horwitt, 1965). d-Tocopherol was
not measured in this study and the plasma total a-tocopherol equivalent was calculated as a-tocopherol ‡ 0.2 gtocopherol. The retinol and a-tocopherol content of the
10 ml vitamin supply was 990 mg and 9.1 mg, respectively,
based on information from the manufacturer.
Plasma essential fatty acid analysis
Fatty acid analysis was performed as described earlier
(Jeppesen et al, 1997). The total lipid fraction from
plasma samples was extracted according to the method of
Folch et al (1956). The phospholipid fraction was isolated
by thin layer chromatography and saponi®ed and methylated using BF3 (Morrison & Smith, 1964). The fatty acid
methyl esters were analysed by gas ± liquid chromatography using a Hewlett-Packard 5890, series II, chromatograph equipped with a fused silica column (SP2380,
60 m60.25 mm i.d., Supelco Inc., Bellefonte, PA).
Cholesterol in plasma lipoproteins
Cholesterol, HDL cholesterol and triglycerides were measured by enzymatic colorimetric tests (cholesterol CHODPAP-method 1 491 458, HDL cholesterol precipitation
reagent 543 004, and triglycerides GPO-PAP 1 730 711,
respectively, Boehringer Mannheim Systems, Ingelheim
am Rhein, Germany; Burstein et al, 1970). LDL cholesterol
was calculated from the equation, LDL cholesterol ˆ total
cholesterol 7 (triglycerides=2.2) 7 HDL cholesterol.
Analysis of fat-soluble vitamins A and E
Vitamin A and E were analysed as described by Kaplan
et al (1990). All operations were performed in darkness.
Plasma samples were added a-tocopherol acetate as internal
standard prior to extraction twice with hexane. Following
evaporation of solvent the residue was redissolved in
acetonitril=chloroform, 80=20, ®ltered, and transferred to
coloured vials. HPLC was performed on a Supelcosil LC18 column (25 cm64.6 mm, 5 mm, Supelco, Bellefonte,
PA, USA) using a ¯ow-rate of 2 ml=min, detection at
292 nm and a mobile phase of acetonitril= chloroform=isopropanol=water, 780=160=35=25.
Statistics
Non-parametric testing was performed as observations
were not sampled from a population with a normal distribution, and the assumptions of equal variances for a
parametric test between groups were not met. Differences
between groups were assessed by non-parametric Kruskal ±
Wallis one-way ANOVA on ranks. A Tukey's test adjusting for multiple comparisons was used as the post hoc test
for multiple pairwise comparisons, and the subscripts a, b,
c and d given in tables denote signi®cant differences (Pvalues < 0.05) among groups. Calculations were performed
using the Sigmastat statistical program package (Jandel
Corp, Erkrath, Germany).
De®ciencies related to fat malabsorption
PB Jeppesen et al
Results
Patient characteristics
The non-HPN patients were divided into three groups
according to their energy absorption from fat absorption in relation to their total energy absorption: group A,
>25%; group B, 15 ± 25%; and group C <15%, respectively. The HPN patients were divided into those who
received parenteral lipids and those who did not: group
D, ‡ i.v. lipids; and group E, 7 i.v. lipids, respectively
(Table 1). The blood tests of the patients were compared
with those of 35 control subjects (laboratory and staff
members). Characteristics of the patients and the controls
are shown in Table 1. The median age of the control
subjects was lower than in patients in groups B, C and E,
respectively. The majority of patients in all groups had
Crohn's disease, but none of the patients with in¯ammatory
bowel disease had evidence of active disease at the time of
the study.
Diet and intestinal absorption
The dietary intake and intestinal absorption of energy, fat
and essential fatty acids in the ®ve groups of patients is
shown in Table 2. The non-HPN patients in group C had
the largest compensatory hyperphagia compared with the
patients in the rest of the groups, and the HPN patients in
group D consumed less than the non-HPN patients. The
energy absorption was higher in the non-HPN patients
compared to the HPN patients, and the energy absorption
as percentage of intake was higher in the non-HPN patients
in group A compared with HPN patients.
The diet fat intake was higher in the non-HPN patients
compared with the HPN patients in group D. The fat
absorption was higher in the non-HPN patients in groups
A and B compared with the HPN patients and the non-HPN
patients in group C. The fat absorption in the non-HPN
patients in group A was higher compared with non-HPN
patients in group B. The fat absorption as a percentage of
intake was higher in the non-HPN patients in group A
635
Table 1 Patient characteristics
Non-HPN
HPN
Group 0 (n ˆ 35); Group A (n ˆ 14); Group B (n ˆ 15); Group C (n ˆ 15); Group D (n ˆ 16); Group E (n ˆ 24);
controls
FA=TEA > 25% FA=TEA > 15 ± 25% FA=TEA < 15%
‡ iv. lipids
7 iv. lipids
P-Value*
Gender (women=men)
27=8ab
9=5ab
Age (y)
34 (31 ± 45)a
48 (44 ± 52)abc
Height (cm)
171 (165 ± 177) 168 (160 ± 177)
Weight (kg)
68 (61 ± 79)a
65 (55 ± 73)a
22.0 (20.3 ± 25.5) 22.7 (18.8 ± 25.6)
Body mass index
(kg=m2)
Diagnosis (CD=MT=
Ð
13=0=1=0=0
Surg=RE=ID)
Remnant small
Ð
195 (165 ± 255)a
bowel (cm)
Remnant colon (%)
Ð
29 (0 ± 57)
Number of patients with
35
8
remnant part of colon
45
165
61
21.9
10=5ab
(41 ± 55)bc
(162 ± 176)
(54 ± 67)a
(19.4 ± 23.7)
50
172
64
22.2
12=2=1=0=0
5=10b
(44 ± 61)bc
43
(164 ± 178)
169
(57 ± 70)a
56
(19.6 ± 23.9) 19.3
12=1=2=0=0
a
250 (223 ± 276)
8=1=2=1=4
ab
160 (126 ± 230)
0 (0 ± 79)
6
57 (0 ± 79)
9
9=7ab
(31 ± 54)ab
53
(163 ± 175)
165
(49 ± 65)a
59
(17.8 ± 22.2) 21.7
20=4a
< 0.05
(45 ± 63)c
< 0.001
(161 ± 167)
0.11
(52 ± 65)a
0.04
(19.3 ± 24.5)
0.14
14=1=5=2=2
ab
Ð
b
140 (100 ± 235)
0 (0 ± 28)
8
75 (40 ± 145)
0.005
28 (0 ± 93)
10
0.21
Ð
Results are expressed as median with 25% and 75% percentiles given in parentheses. FA=TEA ˆ fat absorption=total Energy absorption (%). Total energy
supply corresponds to the sum of the intestinal energy absorption and the parenteral energy supply. CD ˆ Crohn's disease; MT ˆ mesenteric thrombosis;
Surg ˆ complications to surgery; RA ˆ radiation enteritis; ID ˆ intestinal dysmotility. *Kruskal ± Wallis one-way Anova on ranks. The superscripts a ± d
denote statistical difference, P < 0.05, by all pairwise multiple comparison procedures (Tukey's test) for multiple comparisons.
Table 2 Diet intake and intestinal absorption of energy, fat and essential fatty acids
Non-HPN
Group A (n ˆ 14); Group B (n ˆ 15);
FA=TEA > 25% FA=TEA > 15 ± 25%
Energy
Diet (MJ=day)
Absorption (MJ=day)
Absorption (percentage
Fat
Diet (g=day)
Absorption (g=day)
Absorption (percentage
18:2n-6
Diet (g=day)
Absorption (g=day)
Absorption (percentage
18:3n-3
Diet (g=day)
Absorption (g=day)
Absorption (percentage
HPN
Group C (n ˆ 15);
FA=TEA < 15%
Group D (n ˆ 16);
‡ iv. lipids
11.01 (9.55 ± 12.25)b 9.68 (8.46 ± 12.94)b 13.52 (11.26 ± 17.23)c 6.46 (4.27 ± 8.35)a
7.70 (7.48 ± 9.09)a 7.76 (5.74 ± 8.92)a
8.39 (6.20 ± 9.99)a 3.00 (2.00 ± 4.76)b
70 (63 ± 75)ab
57 (54 ± 71)ab
51 (44 ± 71)b
of intake)
78 (74 ± 81)a
of intake)
87 (66 ± 99)bc
57 (54 ± 68)a
74 (68 ± 83)a
10.4 (8.3 ± 11.4)c
8.2 (6.6 ± 9.4)a
of intake)
80 (73 ± 85)a
of intake)
1.3 (1.2 ± 1.5)b
1.2 (1.0 ± 1.4)a
91 (77 ± 100)a
Group E (n ˆ 24);
7 iv. lipids
P-Value*
8.34 (6.93 ± 10.36)ab < 0.001
4.31 (3.25 ± 5.55)b
< 0.001
45 (30 ± 80)b
0.005
67 (61 ± 86)bc
37 (30 ± 44)b
54 (38 ± 62)ab
91 (61 ± 128)c
20 (8 ± 29)c
25 (16 ± 33)b
35 (24 ± 51)a
7 (73 ± 19)c
24 (20 ± 46)b
59 (47 ± 75)ab
17 (11 ± 24)c
28 (3 ± 73)b
< 0.001
< 0.001
< 0.001
7.2 (5.3 ± 9.7)bc
4.4 (3.6 ± 5.9)b
64 (52 ± 83)b
7.7 (4.6 ± 13.5)bc
3.8 (1.6 ± 4.4)bc
32 (21 ± 58)bc
2.8 (1.7 ± 5.0)a
0.5 (70.5 ± 1.7)cd
16 (711 ± 47)c
3.8 (2.2 ± 6.1)ab
0.5 (70.3 ± 1.7)d
40 (717 ± 74)bc
< 0.001
< 0.001
< 0.001
1.0 (0.7 ± 1.3)b
0.7 (0.5 ± 1.2)ab
79 (58 ± 100)a
1.2 (0.5 ± 2.0)b
0.4 (0.3 ± 0.5)bc
36 (13 ± 95)ab
0.2 (0.0 ± 0.6)a
0.0 (0.0 ± 0.6)a
0.0 (70.2 ± 0.1)c 7 0.1 (70.3 ± 0.1)d
0 (0 ± 14)b
0 (77 ± 24)b
< 0.001
< 0.001
< 0.001
Results are expressed as median with 25% and 75% percentiles given in parentheses. FA=TEA ˆ fat absorption=total energy absorption (%). Total energy
supply corresponds to the sum of the intestinal energy absorption and the parenteral energy supply. *Kruskal ± Wallis one-way Anova on ranks. The
superscripts a ± d denote statistical difference, P < 0.05, by all pairwise multiple comparison procedures (Tukey's test) for multiple comparisons.
European Journal of Clinical Nutrition
De®ciencies related to fat malabsorption
PB Jeppesen et al
636
compared with the HPN patients and the non-HPN patients
in group C.
The linoleic acid (18:2n-6) intake was higher in nonHPN patients in group A compared with the HPN patients.
The absorption of linoleic acid was higher in the non-HPN
patients in groups A and B compared with the HPN
patients. The absorption of linoleic acid in the non-HPN
patients in group A was higher compared with non-HPN
patients in group B.
When calculating the energy absorption from linoleic
acid in relation to the total energy supply the median (25 ±
75%) values in groups A ± E were 3.7% (2.9 ± 4.3%), 2.1%
(1.5 ± 4.1%), 1.6% (0.8 ± 2.1%), 0.2% (70.2 ± 0.7%), and
0.3% (70.2 ± 0.8%). Thus, the energy absorption from
linoleic acid in relation to the total energy supply was
higher in the non-HPN patients in group A compared with
the HPN patients. The absorption of linoleic acid as a
percentage of intake was higher in the non-HPN patients in
groups A and B compared with the HPN patients in group
D. The absorption of linoleic acid as a percentage of intake
was higher in the non-HPN patients in group A compared
with non-HPN patients in group B.
Diet linolenic acid (18:3n-3) intake was higher in the
non-HPN patients compared with the HPN patients. The
linolenic acid absorption was higher in the non-HPN
patients in groups A and B compared with the HPN
patients. The linolenic acid absorption was higher in the
non-HPN patients in group A compared with the patients in
group C.
When calculating the energy absorption from linolenic
acid in relation to the total energy supply the median (25 ±
75%) values in groups A ± E were 0.6% (0.4 ± 0.7%), 0.5%
(0.2 ± 0.7%), 0.2% (0.1 ± 0.3%), 0.0% ( 7 0.1 ± 0.0%), and
0.0% ( 7 0.2 ± 0.1%). Thus, the energy absorption from
linolenic acid in relation to the total energy supply was
higher in the non-HPN patients in groups A and B compared with the HPN patients and the non-HPN patients in
group C. The absorption of linolenic acid in percentage of
intake was higher in the non-HPN patients in groups A and
B compared with the HPN patients.
Parenteral support of energy, fat, essential fatty acids,
retinol and tocopherol
The parenteral support to the two groups of patients
receiving HPN is given in Table 3. The need for parenteral
energy was signi®cantly higher in patients in group D who
received parenteral lipids. These patients received a median
of 7.5 and 1.15 g=day of linoleic and linolenic acids,
respectively, as a part of the parenteral lipid support.
Thus, the parenteral lipids given to HPN patients in
group D provided these patients with the same amount of
the essential fatty acids, linoleic and linolenic acids, as the
non-HPN patients in group A, absorbed by the enteral
route. The parenteral supply from lipids and fat-soluble
vitamins led to a signi®cantly larger supply of retinol, and
a- and g-tocopherol to HPN patients in group D compared
with the HPN patients in group E.
Fatty acid composition of plasma phospholipids
The fatty acid composition of plasma phospholipids is
given in Table 4. A progressive decline in the level of
linoleic acid (18:2n-6) over the groups 0 and A ± E could be
demonstrated. The level of linoleic acid was 21.8% in the
control subjects, 17.3%, 15.5%, 13.1%, 12.1%, and 8.9% in
patients in groups A ± E, respectively. The level of linolenic
acid (18:3n-3) was signi®cantly lower in the HPN patients
compared with the non-HPN patients and the controls. The
level of linoleic acid (18:2n-6) in the plasma phospholipids
correlated positively with the intestinal absorption of linoleic acid in the non-HPN patients and in the HPN not
receiving parenteral lipids, as evidenced in Figure 1. Thus,
a low absorption of linoleic acid was re¯ected in a low
level of linoleic acid in the plasma phospholipids.
The relation between the level of selected fatty acids and
the concentration of linoleic acid in plasma phospholipids
is given in Figure 2. A characteristic sign of essential fatty
acid de®ciency (EFAD) is the increase in the level of
20:3n-9 (eicosatrienoic acid), the numerator of the
Holman (20:3n-9=20:4n-6) index (Holman et al, 1979).
The level of this fatty acid was 16-fold higher in the
HPN patients not receiving parenteral lipids (group E)
compared with the level in the control subjects (1.6%
compared with 0.1%, P < 0.001, Table 4). Ten of the 24
HPN patients in group E had a Holman index above 0.2, the
classically used biochemical evidence of EFAD, whereas
none of the patients in the other groups had EFAD based on
this de®nition. Figure 2A illustrates the relation between
the level of 20:3n-9 and 18:2n-6 in the plasma phospholipids in the non-HPN and HPN patients, respectively. In the
non-HPN patients the maximum elevation of 20:3n-9 was
to a concentration of 0.3% in spite of levels of 18:2n-6
ranging from 9.4% to 26.0%. In the HPN patients all
patients with an 18:2n-6 level above 10% had a 20:3n-9
level below 1.1%. All eight HPN patients with a level of
18:2n-6 below 8% had elevated 20:3n-9 above 1.8% (range
1.8 ± 6.3%). Seven of the eight patients did not receive
parenteral lipids. In the 10 HPN patients with 18:2n-6
between 8 and 10% the 20:3n-9 level ranged from 0.6 to
3.6%. Eight of the 10 patients did not receive parenteral
lipids. Figure 2B illustrates that no clear relation existed
between 18:2n-6 and the denominator of the Holman index,
20:4n-6 (arachidonic acid). Similar ®ndings were evident
for the other members of the n-6 family (ie 18:3n-6, 20:3n6, 22:4n-6 and 22:6n-6, not shown) and the members of the
n-3 family given in Table 4 (ie 18:3n-3, 20:5n-3, 22:5n-3
and 22:6n-3, not shown). Figure 2C illustrates the compen-
Table 3 Parenteral support of energy, fat and essential fatty acids
Parenteral
Energy (MJ=day)
18:2n-6 (g=day)
18:3n-3 (g=day)
Retinol
a-tocopherol (mg=day)
g-tocopherol (mg=day)
Group D (n ˆ 16); ‡ iv. lipids
Group E (n ˆ 24); 7 iv. lipids
P-Value*
5.74 (4.48 ± 7.58)
7.5 (7.5 ± 15.0)
1.15 (1.15 ± 2.30)
141.4 (70.7 ± 141.4)
2.6 (2.2 ± 6.0)
45.5 (45.5 ± 89.2)
2.71 (1.43 ± 3.84)
0.0 (0.0 ± 0.0)
0.0 (0.0 ± 0.0)
141.4 (0 ± 141.4)
1.3 (1.3 ± 1.3)
0.0 (0.0 ± 0.0)
< 0.001
Ð
Ð
0.002
< 0.001
< 0.001
Results are expressed as median with 25% and 75% percentiles given in parentheses. *Mann ± Whitney rank sum test.
European Journal of Clinical Nutrition
De®ciencies related to fat malabsorption
PB Jeppesen et al
637
Table 4 Fatty acid composition of plasma phospholipids (percentage by weight of total fatty acids)
Non-HPN
HPN
Fatty acids
Group 0 (n ˆ 35); Group A (n ˆ 14); Group B (n ˆ 15); Group C (n ˆ 15); Group D (n ˆ 16); Group E (n ˆ 24);
controls
FA=TEA > 25% FA=TEA > 15 ± 25% FA=TEA < 15%
‡ iv. lipids
7 iv. lipids
P-Value*
14:0
16:0
18:0
0.4 (0.3 ± 0.5)a
0.5 (0.4 ± 0.5)a
0.5 (0.4 ± 0.5)a
0.5 (0.4 ± 0.6)a
0.4 (0.4 ± 0.5)a
0.5 (0.4 ± 0.5)a
0.007
30.1 (29.1 ± 30.8)a 31.8 (30.0 ± 32.7)ab 31.6 (30.0 ± 33.3)ab 32.3 (31.1 ± 33.6)b 30.7 (29.5 ± 32.5)ab 31.5 (30.0 ± 32.6)ab < 0.001
b
a
ab
a
c
ab
15.5 (14.0 ± 15.9) 13.3 (12.6 ± 14.7) 13.3 (12.6 ± 14.7)
12.9 (11.8 ± 13.2) 15.8 (14.5 ± 16.9) 13.4 (12.3 ± 15.9)
< 0.001
Total saturated
FA
46.2 (45.6 ± 46.9)a 46.2 (45.8 ± 46.6)a 47.1 (45.8 ± 47.5)a
16:1 n-7
18:1 n-7
Total n-7
16:1 n-9
18:1 n-9
20:1 n-9
20:3 n-9
Total n-9
n-7 ‡ n-9
18:2 n-6
18:3 n-6
20:3 n-6
20:4 n-6
22:4 n-6
22:5 n-6
Total n-6
18:3 n-3
20:5 n-3
22:5 n-3
22:6 n-3
Total n-3
n-6 ‡ n-3
0.5
1.6
2.9
0.4
12.1
0.4
0.1
12.9
15.8
21.8
0.0
2.5
7.8
0.3
0.0
32.9
0.3
0.9
0.7
3.0
4.6
37.8
FA
FA
FA
FA
FA
FA
Total unsaturated
FA
(0.4 ± 0.6)a
(1.6 ± 1.8)ab
(2.7 ± 3.1)a
(0.4 ± 0.5)a
(11.2 ± 12.6)a
(0 ± 4 ± 0.4)a
(0.0 ± 0.2)a
(12.1 ± 13.5)a
(15.0 ± 16.3)a
(20.3 ± 23.0)a
(0.0 ± 0.0)a
(2.3 ± 2.9)a
(7.3 ± 8.7)a
(0.2 ± 0.3)a
(0.0 ± 0.0)a
(32.0 ± 34.0)a
(0.2 ± 0.3)a
(0.6 ± 1.2)c
(0.6 ± 0.9)a
(2.5 ± 3.5)
(4.2 ± 5.6)ab
(37.1 ± 38.5)a
0.9
1.6
2.9
0.4
11.8
0.3
0.2
12.8
15.6
17.3
0.2
3.3
9.8
0.0
0.2
31.6
0.2
1.5
1.1
3.1
5.7
37.2
(0.7 ± 1.2)ab
(1.5 ± 2.1)ab
(2.6 ± 3.5)a
(0.3 ± 0.4)ab
(11.0 ± 12.4)ab
(0.3 ± 0.4)a
(0.2 ± 0.3)a
(12.0 ± 13.3)a
(14.5 ± 17.2)ab
(15.9 ± 19.1)b
(0.1 ± 0.2)b
(2.9 ± 3.8)b
(8.0 ± 10.6)ab
(0.0 ± 0.0)c
(0.1 ± 0.2)b
(30.2 ± 32.7)a
(0.2 ± 0.3)a
(1.1 ± 1.6)a
(0.8 ± 1.2)b
(2.7 ± 3.8)
(5.4 ± 7.6)bc
(36.4 ± 38.4)a
1.1
1.9
3.5
0.4
12.8
0.4
0.2
13.8
17.1
15.5
0.2
3.6
8.7
0.0
0.2
30.1
0.3
1.3
1.0
2.9
5.4
35.7
(0.9 ± 1.5)b
(1.6 ± 2.2)bc
(3.0 ± 3.6)ab
(0.3 ± 0.5)ab
(12.0 ± 13.7)ab
(0.3 ± 0.5)a
(0.2 ± 0.2)a
(12.9 ± 15.0)a
(16.4 ± 19.6)abc
(13.9 ± 18.2)bc
(0.1 ± 0.2)b
(2.8 ± 3.9)b
(8.0 ± 9.5)ab
(0.0 ± 0.0)c
(0.0 ± 0.3)b
(27.0 ± 31.2)ab
(0.2 ± 0.3)a
(0.9 ± 1.7)bc
(0.9 ± 1.3)b
(2.6 ± 3.7)
(4.8 ± 6.3)cd
(33.3 ± 36.8)ab
46.3 (45.7 ± 47.1)a 47.8 (46.1 ± 48.8)a 46.3 (45.4 ± 47.6)a
1.5
2.5
4.2
0.4
13.9
0.4
0.2
15.0
18.8
13.1
0.2
3.6
8.5
0.0
0.2
27.8
0.2
1.3
1.1
2.6
5.3
33.6
(1.3 ± 1.9)bc
(1.9 ± 2.8)cd
(3.6 ± 5.1)bc
(0.3 ± 0.4)ab
(11.9 ± 15.6)bc
(0.3 ± 0.5)a
(0.2 ± 0.3)a
(12.8 ± 16.2)a
(17.2 ± 22.5)bc
(12.4 ± 14.6)cd
(0.1 ± 0.2)b
(3.2 ± 4.0)b
(7.5 ± 10.2)a
(0.0 ± 0.1)c
(0.1 ± 0.4)bc
(25.4 ± 29.6)ab
(0.1 ± 0.3)a
(1.1 ± 1.5)b
(0.8 ± 1.4)ab
(2.1 ± 3.3)
(4.8 ± 6.4)bc
(30.2 ± 35.0)b
1.7
2.8
5.1
0.3
12.7
0.2
0.5
14.5
19.6
12.1
0.2
4.0
11.2
0.0
0.3
28.1
0.1
0.8
0.9
2.6
4.5
32.7
(1.3 ± 1.9)bc
(2.5 ± 3.3)cd
(4.4 ± 6.0)c
(0.3 ± 0.4)b
(11.8 ± 14.2)ab
(0.2 ± 0.3)a
(0.3 ± 0.6)a
(12.5 ± 16.6)a
(17.1 ± 22.2)c
(10.3 ± 14.0)de
(0.1 ± 0.2)b
(3.0 ± 4.4)b
(9.2 ± 12.5)b
(0.0 ± 0.1)b
(0.0 ± 0.4)b
(25.6 ± 30.3)bc
(0.1 ± 0.2)b
(0.6 ± 1.3)c
(0.8 ± 1.0)a
(2.0 ± 2.9)
(3.8 ± 5.3)ab
(29.5 ± 34.7)bc
2.0
3.1
5.3
0.4
15.1
0.3
1.6
17.3
23.9
8.9
0.2
3.2
8.9
0.0
0.3
22.9
0.1
1.0
1.0
2.6
5.0
28.5
(1.4 ± 2.8)c
(2.6 ± 3.7)d
(4.4 ± 6.6)c
(0.4 ± 0.5)a
(13.6 ± 19.5)c
(0.2 ± 0.4)a
(0.5 ± 3.5)b
(15.3 ± 23.8)b
(20.4 ± 29.0)d
(7.5 ± 12.5)e
(0.1 ± 0.2)b
(2.7 ± 4.0)b
(8.0 ± 10.3)ab
(0.0 ± 0.0)c
(0.3 ± 0.4)c
(20.6 ± 27.5)d
(0.1 ± 0.2)b
(0.6 ± 1.3)c
(0.7 ± 1.3)a
(2.1 ± 3.1)
(3.8 ± 5.7)ab
(24.4 ± 32.4)d
53.7 (53.0 ± 54.4)a 53.2 (52.7 ± 53.5)ab 52.0 (51.8 ± 53.5)ab 52.9 (51.9 ± 53.4)ab 50.3 (50.3 ± 53.0)b 52.9 (51.6 ± 53.8)ab
n-7 ‡ n-9=n-6 ‡ n-3 0.42 (0.39 ± 0.44)a 0.43 (0.37 ± 0.48)ab 0.48 (0.43 ± 0.59)ab 0.56 (0.49 ± 0.76)ab 0.61 (0.49 ± 0.76)b 0.85 (0.66 ± 1.28)c
20:3n-9=18:2n-6
0.00 (0.00 ± 0.01)a 0.01 (0.01 ± 0.02)a 0.01 (0.01 ± 0.02)a 0.02 (0.02 ± 0.02)a 0.04 (0.02 ± 0.07)a 0.19 (0.05 ± 0.47)b
Holman index
0.01 (0.00 ± 0.02)a 0.03 (0.02 ± 0.03)a 0.02 (0.02 ± 0.02)a 0.03 (0.02 ± 0.03)a 0.05 (0.02 ± 0.08)a 0.17 (0.05 ± 0.37)b
No. with Holman
0
0
0
0
0
10
index > 0.2
0.03
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
0.001
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
0.06
< 0.001
< 0.001
0.001
< 0.001
< 0.001
< 0.001
Results are expressed as median with 25% and 75% percentiles given in parentheses. FA=TEA ˆ fat absorption=total energy absorption (%). Total energy
supply corresponds to the sum of the intestinal energy absorption and the parenteral energy supply. *Kruskal ± Wallis one-way Anova on ranks. The
subscripts a ± d denote statistical difference, P < 0.05, by all pairwise multiple comparison procedures (Tukey's test) for multiple comparisons.
satory increase in 16:1n-7 with a decrease in the level of
18:2n-6. A similar relation existed for 18:1n-7 and 18:1n-9,
respectively (not shown), and these compensatory changes
were demonstrated in both the HPN and the non-HPN
patients. A similar relation between 18:2n-6 and 20:1n-9
(Figure 2D) or 18:2n-6 and 16:1n-9 (not shown) could not
be demonstrated. Comparing the responsiveness or sensitivity of various indicators of EFAD, Figure 3 illustrates the
relation between the Holman index (20:3n-9=20:4n-6
ratio), 20:3n-9, 20:3n-9=18:2n-6, and n-9 ‡ n-7=n-6 ‡ n-3
plotted according to an arbitrary range of 0 ± 1.00, where 0
represents the patient with the lowest value, 1.00 the patient
with the highest value, and the patients in between according to their proportional value.
Figure 1 The relationship between the intestinal absorption of linoleic
acid (18:2n-6) and the plasma phospholipid (PL) level of 18:2n-6.(e)
Patients in group A; (D) patients in group B; (m) patients in group C; (u)
patients in group E.
Cholesterol in plasma lipoproteins
The cholesterol content of the plasma lipoproteins is given
in Table 5. Total cholesterol was lower in groups B ± E
compared with the control subjects. A trend of a parallel
reduction in both HDL and LDL cholesterol in the HPN
patients made the HDL=total cholesterol ratio unchanged
compared the control subjects, whereas an increase in the
European Journal of Clinical Nutrition
De®ciencies related to fat malabsorption
PB Jeppesen et al
638
Figure 2 The relationship between the level of selected fatty acids and 18:2n-6 in plasma phospholipids in non-home parenteral nutrition (HPN) and HPN
patients. (e) Patients in group A; (D) patients in group B; (m) patients in group C; (u) patients in group E.
HDL=total cholesterol ratio was observed in the non-HPN
patients.
Plasma fat-soluble vitamins A and E
The concentration of the fat-soluble vitamins A and E is
given in Table 6. No differences in the plasma retinol
concentration could be demonstrated among the groups.
The lower limit of the 2.5 ± 97.5% con®dence interval of
the retinol concentration in the control subjects was
0.52 mg=ml (range, 0.50 ± 2.29 mg=ml). Eight of 11 patients
(73%) with a retinol concentration below 0.52 mg=ml were
among those who received HPN. Only two of these patients
did not receive parenteral vitamin A supplements.
Plasma a-tocopherol concentration was signi®cantly
lower in the non-HPN patients in groups B and C and in
the HPN patients group D and E compared with the control
European Journal of Clinical Nutrition
subjects (Table 6). The lower limit of the 2.5 ± 97.5%
con®dence interval of the plasma a-tocopherol concentration in the control subjects was 4.25 mg=ml (range, 4.18 ±
9.94 mg=ml). As evidenced in Table 6 the majority of
the non-HPN patients in groups B and C had low plasma
a-tocopherol concentrations. Low plasma a-tocopherol
concentrations were also demonstrated in the HPN patients
in spite of the parenteral supplements of fat soluble vitamins or lipids given to all except eight HPN patients. Three
of these eight patients had low plasma a-tocopherol concentrations. The g-tocopherol concentration was signi®cantly higher in the HPN patients receiving parenteral
lipids compared with the controls, the non-HPN patients
in groups B and C and the HPN patients in group E (Table
6). Thus, none of the HPN patients receiving parenteral
lipid had a plasma g-tocopherol concentration below
De®ciencies related to fat malabsorption
PB Jeppesen et al
0.28 mg=ml, which was the lower limit of the 2.5 ± 97.5%
con®dence interval in the control subjects. Figure 4 demonstrates a close and signi®cant correlation between plasma
total a-tocopherol equivalents and the total cholesterol in
plasma.
639
Discussion
Figure 3 A comparison of different indices of essential fatty acid
de®ciency. The range of the indices are plotted according to a arbitrary
range of 0 ± 1, where 0 represents the patient with the lowest value, 1 the
patient with the highest value and the patients in between according to
their proportional value. Bold lines represent patients with a Holman index
above 0.2.
We previously demonstrated that a proportion of patients
with severe fat malabsorption have low concentrations of
essential fatty acids in plasma (Jeppesen et al, 1997), which
also is encountered in patients receiving HPN (Jeppesen
et al, 1998c). In the present study, which included both
non-HPN and HPN patients, identi®cation of the individual
fatty acids in the diet and faeces by combined gas ± liquid
chromatography and mass spectrometry rendered it possible to measure the absorption of the essential fatty acids,
linoleic and linolenic acids, and a correlation between the
intestinal absorption of linoleic acid and the concentrations
in the plasma phospholipids was demonstrated (Figure 1).
Essential fatty acid de®ciency (EFAD) is traditionally
de®ned as a 20:3n-9=20:4n-6 ratio (Holman index) above
0.2. However, the lack of coherence between clinical
Table 5 Cholesterol in plasma lipoproteins
Non-HPN
HPN
Group 0 (n ˆ 35); Group A (n ˆ 14); Group B (n ˆ 15); Group C (n ˆ 15); Group D (n ˆ 16); Group E (n ˆ 24);
controls
FA=TEA > 25% FA=TEA > 15 ± 25% FA=TEA < 15%
‡ iv. lipids
7 iv. lipids
P-Value*
Total cholesterol
4.9
(mmol=l)
HDL cholesterol
1.1
(mmol=l)
LDL cholesterol
3.6
(mmol=l)
HDL=total cholesterol 0.23
ratio
(4.4 ± 5.5)a
4.7 (3.6 ± 5.2)ab
3.6 (3.0 ± 4.4)b
3.4 (2.7 ± 3.9)b
2.6 (2.4 ± 3.7)b
3.5 (2.9 ± 4.4)b
< 0.001
(0.8 ± 1.2)b
1.5 (1.0 ± 1.8)a
1.0 (0.9 ± 1.3)ab
1.1 (0.9 ± 1.3)ab
0.5 (0.5 ± 0.7)c
0.8 (0.7 ± 1.1)bc
< 0.001
(3.2 ± 4.2)a
2.3 (1.7 ± 2.6)b
1.7 (1.6 ± 2.2)b
1.8 (1.3 ± 2.4)b
1.8 (1.6 ± 2.5)b
2.0 (1.6 ± 2.8)b
< 0.001
(0.16 ± 0.27)b 0.31 (0.24 ± 0.40)a 0.36 (0.25 ± 0.39)a 0.30 (0.26 ± 0.35)a 0.22 (0.16 ± 0.31)ab 0.20 (0.14 ± 0.22)b < 0.001
Results are expressed as median with 25% and 75% percentiles given in parentheses. FA=TEA ˆ fat absorption=total energy absorption (%). Total energy
supply corresponds to the sum of the intestinal energy absorption and the parenteral energy supply. *Kruskal ± Wallis one-way Anova on ranks. The
subscripts a ± d denote statistical difference, P < 0.05, by all pairwise multiple comparison procedures (Tukey's test) for multiple comparisons.
Table 6 Vitamins A and E in plasma
Non-HPN
HPN
Group 0 (n ˆ 35); Group A (n ˆ 14); Group B (n ˆ 15); Group C (n ˆ 15); Group D (n ˆ 16); Group E (n ˆ 24);
controls
FA=TEA > 25% FA=TEA > 15 ± 25% FA=TEA < 15%
‡ iv. lipids
7 iv. lipids
P-Value*
Retinol (mg=ml)
Retinol < 0.52 mg=ml,
n ˆ x (%)
Alfa-tocopherol (mg=ml)
Alfa-tocopherol < 4.25
mg=ml, a n ˆ x (%)
Gamma-tocopherol
(mg=ml)
Gamma-tocopherol < 0.28
mg=ml, n ˆ x (%)
Total alfa-tocopherol
equivalents (mg=ml)
Total alfa-tocopherol
equivalents < 4.34 mg=
ml, n ˆ x (%)
1.0 (0.9 ± 1.2)
0 (0%)
1.1 (0.8 ± 1.4)
2 (14%)
1.1 (0.7 ± 1.2)
1 (7%)
0.8 (0.6 ± 1.1)
0 (0%)
0.6 (0.4 ± 1.0)
5 (31%)
1.0 (0.7 ± 1.2)
3 (13%)
6.1 (5.1 ± 7.7)a
1 (3%)
7.2 (3.7 ± 8.8)ab
4 (29%)
3.7 (3.0 ± 4.5)c
11 (73%)
2.0 (1.4 ± 2.8)c
12 (80%)
3.1 (2.5 ± 4.2)c
10 (81%)
4.1 (3.4 ± 6.4)bc
12 (50%)
0.6 (0.4 ± 0.7)a
0.8 (0.3 ± 1.2)ab
0.7 (0.2 ± 0.9)a
0.4 (0.2 ± 0.6)a
0.9 (0.5 ± 1.2)b
0.4 (0.2 ± 0.7)a
1 (3%)
6.2 (5.3 ± 7.8)c
1 (3%)
3 (21%)
7.5 (3.9 ± 9.0)bc
4 (29%)
6 (40%)
5 (33%)
0 (0%)
7 (29%)
3.8 (3.1 ± 4.6)a
2.3 (1.5 ± 2.9)a
4.2 (3.5 ± 6.4)b
3.5 (2.7 ± 4.5)a
10 (67%)
11 (73%)
10 (81%)
12 (50%)
0.11
Ð
< 0.001
Ð
0.006
Ð
< 0.001
Ð
Results are expressed as median with 25% and 75% percentiles given in parentheses. FA=TEA ˆ fat absorption=total energy absorption (%). Total energy
supply corresponds to the sum of the intestinal energy absorption and the parenteral energy supply. *Kruskal ± Wallis one-way Anova on ranks. The
subscripts a ± d denote statistical difference, P < 0.05, by all pairwise multiple comparison procedures (Tukey's test) for multiple comparisons. Total alfatocopherol equivalents is calculated as alfa-tocoherol ‡ 0.2 gamma-tocopherol. # The value corresponds to the lower limit of the 2.5% ± 97.5% con®dence
interval in control subjects.
European Journal of Clinical Nutrition
De®ciencies related to fat malabsorption
PB Jeppesen et al
640
Figure 4 The relationship between total a-tocopherol equivalents (calculated as a-tocopherol ‡ 0.2 g-tocopherol) and total cholesterol. (s)
Control subjects; (e) patients in group A; (D) patients in group B; (m)
patients in group C; (u) patients in group E.
manifestations and this biochemical marker makes a clear
demarcation between the de®ciency and suf®ciency state
dif®cult, and other fatty acid concentrations and ratios have
been proposed as alternative indicators of EFAD (Siguel
et al, 1987). EFAD causes a reduction in some of the
essential n-6 and n-3 fatty acids and is often accompanied
by an elevation in some of the non-essential n-7 and n-9
fatty acids (Jeppesen et al, 1997), which was mainly
encountered in the numerator of the Holman index,
20:3n-9, and in 16:1n-7, 18:1n-7 and 18:1n-9 with decreasing concentrations of linoleic acid (Figure 2 and Table 4).
An omission of the denominator, 20:4n-6, in the Holman
index (20:3n-9=20:4n-6) or its replacement by linoleic acid,
18:2n-6, selected the same 10 patients with an index above
0.2, demarcated with bold lines in Figure 3, whereas the
ratio between the total amounts of essential n-6 and n-3
fatty acids and nonessential n-7 and n-9 fatty acids selected
differently for some of the patients. The Holman index
nominator, 20:3n-9 was only elevated in patients with
concentrations of linoleic acid, 18:2n-6, below 10% and
20:3n-9 was not increased in a single non-HPN patient
(Figure 2A). In accordance with the ®ndings of Mascioli
et al (1996), a 20:3n-9 >2% was only encountered in HPNpatients not receiving parenteral lipids (Figure 2A), who
also were the only patients with biochemical signs of
EFAD de®ned as a Holman index >0.2, which nevertheless
was a very common ®nding in nearly half of the patients
from this group (Table 4). Therefore, patients with very low
concentrations of linoleic acid, <10%, are probably
candidates to receive treatment with parenteral essential
fatty acids (contained in the ordinary products used for
parenteral lipid supplementation), especially in the presence of a coexisting secondary rise in 20:3n-9 >2% and
the Holman index. This limited strategy would virtually
exclude non-HPN patients as candidates (only two had
concentrations of linoleic acid <10%), but include the
majority of HPN-patients (Figure 2). In our view, the
clinical signi®cance of an intermediately decreased
plasma linoleic acid is uncertain and it is usually not
associated with secondary changes in fatty acid composition and an increase in the Holman index. One may argue
that half (20 out of 44) of the non-HPN patients in this
study should receive parenteral lipids as well, if treatment
is guided by a desire to increase plasma linoleic acid above
European Journal of Clinical Nutrition
15%, which still is well below concentrations in healthy
controls. However, in non-HPN patient recommended a
low-fat diet, as it is conventionally done in patients who
malabsorb large quantities of fat, a low plasma linoleic acid
<15% may bring speculation about whether such a diet is
still optimal and whether dietary fat could include more
essential fatty acids.
Jeejeebhoy et al advocated daily lipid infusions of
approximately 500 ml of 10% Intralipid in order to prevent
a linoleic acid de®ciency of the phospholipids of membranes (Jeejeebhoy et al, 1990b). The practice in our unit
has been to limit parenteral lipids to patients in need of a
signi®cant parenteral supply of energy. Patients primarily
in need of parenteral saline who only require low energy
parenteral nutrition were less frequently supplemented with
lipids. Accordingly, the HPN-patients receiving lipids
absorbed less energy (3.00 vs 4.31 MJ=day, Table 2) by
the enteral route, and received more energy by the parenteral route (5.74 vs 2.71 MJ=day, Table 3) compared with
patients not supplemented with lipids. Our impression over
30 y of experience has been that this regimen has not
caused clinical problems which could be attributed to
EFAD. Nevertheless, the results obtained in this study
have changed our policy and we now supply lipids to the
majority of our HPN population except those only treated
with parenteral saline. We still use moderate amounts of
lipids, eg 20% intralipids, 500 ml once or twice a week,
which seems to be adequate to avoid the increase in the
20:3n-9 fatty acids and a rise in the Holman index. As
mentioned, we do not ®nd it suf®ciently justi®ed to aim at a
normalization of plasma linoleic acid, which would require
a considerable increase in the administered amounts of
parenteral lipids compared to our current practice. The fact
that linoleic acid in plasma was signi®cantly lower in the
HPN patients treated with parenteral lipids in group D in
comparison with the non-HPN patients in group A (12.1%
vs 17.3%, Table 4) despite the amount of linoleic acid
received by the organism was identical (7.5 and 0.5 g=day
by the parenteral and enteral route, respectively, in group D
and 8.2 g=day by the enteral route in group A, Tables 2 and
3), may also lend support to the view that the dosage of
parenteral linoleic acid has to include other considerations
than a moderate reduction in the plasma concentration.
Cholesterol was lowered in both the non-HPN and the
HPN patients (Table 5), probably re¯ecting the reduced
intestinal absorption of fat (and cholesterol), the primary
determinants of plasma cholesterol and LDL cholesterol
levels (Dietschy, 1998). The resection of the terminal ileum
and a corresponding loss of bile acids caused by the
interruption the enterohepatic bile acid circulation may
also have contributed. LDL cholesterol was low and HDL
cholesterol was normal in the non-HPN patients Ð changes
also described after intestinal bypass operations for hyperlipidemia (Goldberg et al, 1983; Sorensen et al, 1982;
Moore et al, 1980). LDL and HDL cholesterol were
reduced in HPN patients receiving parenteral lipids, as
earlier reported by Levy et al (1987) with HDL=total
cholesterol ratios comparable to values in healthy controls
(Table 5).
Vitamin A de®ciency has been described in short bowel
patients whether or not parenteral nutrition was required
(Gans & Taylor, 1990; Howard et al, 1980), although Smith
& Lindenbaum (1974) found normal mean serum concentrations of vitamin A and retinol-binding-protein (RBP) in
22 adults with a moderate fat malabsorption (22 g=day). It
De®ciencies related to fat malabsorption
PB Jeppesen et al
is suggested that hepatic stores of vitamin A are mobilised
when supplies are reduced, and Underwood et al calculated
that a drop in the plasma retinol concentration would not be
anticipated for up to 220 days in a complete absence of
dietary vitamin A (Underwood et al, 1970). We found that
three out of 44 non-HPN-patients (7%) and eight out of
40 HPN-patients (20%) had retinol concentrations below
the lower 2.5% con®dence limit found in healthy controls
(Table 6), suggesting that sporadic cases of vitamin A
de®ciencies may occur especially in patients on HPN,
although the median plasma retinol in the HPN- and nonHPN-patients did not differ signi®cantly from concentrations in healthy controls (Table 6). Shils et al (1985) found
that the AMA-FDA daily dose of 3300 mg maintained most
blood vitamin levels within a normal range in HPNpatients, which was con®rmed by Davis et al (1987). In
this study HPN-patients with normal plasma retinol
received a median parenteral dose of 141 mg=day (range
0 ± 141 mg=day), which was equivalent to the amount given
to the eight HPN-patients with low plasma retinol (median
141 mg=day, range 0 ± 141 mg=day). Thus, the few HPNpatients with low plasma levels of retinol (vitamin A) may
be advised to receive increased supplementation according
to repeated measurements.
Vitamin E de®ciency has also been reported in intestinal
disorders with malabsorption (Goransson et al, 1973;
Ghalaut et al, 1995). Vitamin E is mainly located in the
adipose tissue in the bulk lipid stores but is not readily
mobilized (Traber & Kayden, 1987; Schaefer et al, 1983).
In this study vitamin E was measured as plasma concentrations of a- and g-tochopherol, which may be inappropriate
because plasma vitamin E concentrations depend on circulating lipid concentrations (Horwitt et al, 1972) in accordance with the correlation found between a-tocopherol
equivalents and cholesterol (Figure 4). Vitamin E primarily
functions as a membrane-bound, chain-breaking antioxidant, and it has been suggested that the tocopherol concentration in membranes of red blood cells may be a better
index of vitamin E status (Kelly & Sutton, 1989). The
majority of the HPN-patients and the non-HPN patients
with the more severe fat malabsorption had plasma concentrations of a-tocopherol below the lower 2.5% con®dence limit in healthy controls in spite of parenteral
supplementations with 2.6 and 1.3 mg=day in groups D
and E, respectively. This is less than the recommended dose
of 10 mg=day (Americal Medical Association Department
of Food and Nutrition, 1979), and our ®ndings may support
this recommendation and, if so, one may similarly argue
that treatment with vitamin E also should be extended to
non-HPN patients with severe fat malabsorption. Likewise,
a considerable proportion of the patients had g-tochopherol
concentrations below the lower limits found in controls
(Table 6), except patients treated with Intralipid, known to
contain 6 ± 10 times more g-tochopherol than a-tocopherol
due to its content of soybean oil (Gutcher et al, 1984;
Vandewoude et al, 1986).
In conclusion, abnormalities in the fatty acid composition of plasma phospholipids, lipoproteins and of plasma
concentrations of vitamins A and E were demonstrated,
both in short bowel patients with intestinal failure necessitating parenteral support and in patients with severe fat
malabsorption managed without parenteral support. Subclinical damage and unexplained symptoms cannot be
excluded as being associated with the various de®ciencies
found in these patients, but the clinical signi®cance is not
quite clear to us, because cases where we have been able to
identify symptoms associated with these biochemical indicators have been extremely rare. Signs of essential fatty
acid de®ciencies resulting in an increase in the Holman
index were only encountered in HPN-patients not treated
with parenteral lipids, which we now recommend is
avoided, although we do not aim at a normalization of
the plasma linoleic acid concentration. The few, mainly
HPN, patients with low plasma levels of retinol (vitamin A)
may also be advised to receive increased supplementation,
whereas it is unclear if it is necessary to normalize the
plasma levels of g- and a-tocopherol (vitamin E), which
was low in the majority of the patients.
641
Acknowledgements ÐThe technical assistance of Grete Peitersen, Dorte
Christensen, Anne Birgitte Larsen, Bodil Petersen and Jette Christiansen is
greatly appreciated.
References
American Medical Association Department of Foods and Nutrition (1979):
Multivitamin preparations for parenteral use. A statement by the
Nutrition Advisory Group. J. Parenter. Enteral Nutr. 3, 258 ± 262.
Burstein M, Scholnick HR & Mor®n R (1970): Rapid method for the
isolation of lipoproteins from human serum by precipitation with
polyanions. J. Lipid. Res. 11, 583 ± 595.
Century B & Horwitt MK (1965): Biological availability of various forms
of vitamin E with respect to different indices of de®ciency. Fed. Proc.
24, 906 ± 911.
Cosnes J, Gendre JP, Evard D & Le Quintrec Y (1985): Compensatory
enteral hyperalimentation for management of patients with severe short
bowel syndrome. Am. J. Clin. Nutr. 41, 1002 ± 1009.
Cosnes J, Lamy P, Beaugerie L, Le Quintrec M, Gendre JP & Le Quintrec
Y (1990): Adaptive hyperphagia in patients with postsurgical malabsorption. Gastroenterology 99, 1814 ± 1819.
Davis AT, Franz FP, Courtnay DA, Ullrey DE, Scholten DJ & Dean RE
(1987): Plasma vitamin and mineral status in home parenteral nutrition
patients. J. Parenter. Enteral Nutr. 11, 480 ± 485.
DiCecco S, Nelson J, Burnes J & Fleming CR (1987): Nutritional intake of
gut failure patients on home parenteral nutrition. J. Parenter. Enteral
Nutr. 11, 529 ± 532.
Dietschy JM (1998): Dietary fatty acids and the regulation of plasma
low density lipoprotein cholesterol concentrations. J. Nutr. 128,
444S ± 448S.
Fleming CR & Remington M (1981): Intestinal failure. In: Nutrition and
the Surgical Patient, ed. GL Hill, pp 219 ± 235. New York: Churchill
Livingstone.
Folch J, Lees M & Sloane Stanley GH (1956): A simple method for
isolation and puri®cation of total lipids from animal tissues. J. Biol.
Chem. 226, 497 ± 509.
Gans M & Taylor C (1990): Reversal of progressive nyctalopia in a patient
with Crohn's disease. Can. J. Ophthal. 25, 156 ± 158.
Ghalaut VS, Ghalaut PS, Kharb S & Singh GP (1995): Vitamin E in
intestinal fat malabsorption. Ann. Nutr. Metab. 39, 296 ± 301.
Goldberg RB, Weinberg RB & Landau RL (1983): Changes in plasma
apolipoproteins A-I, A-II, and B, and lipoprotein cholesterol after
jejunoileal bypass. Gastroenterology 84, 732 ± 736.
Goransson G, Norden A & Akesson B (1973): Low plasma tocopherol
levels in patients with gastrointestinal disorders. Scand. J. Gastroenterol. 8, 21 ± 25.
Gutcher GR, Lax AA & Farrell PM (1984): Tocopherol isomers in
intravenous lipid emulsions and resultant plasma concentrations.
J. Parenter. Enteral. Nutr. 8, 269 ± 273.
Holman, RT, Smythe, L & Johnson, S (1979): Effect of sex and age on
fatty acid composition of human serum lipids. Am. J. Clin. Nutr. 32,
2390 ± 2399.
Horwitt MK, Harvey CC, Dahm CH Jr & Searcy MT (1972): Relationship
between tocopherol and serum lipid levels for determination of nutritional adequacy. Ann. NY Acad. Sci. 203, 223 ± 236.
Howard L, Chu R, Feman S, Mintz H, Ovesen L & Wolf B (1980):
Vitamin A de®ciency from long- term parenteral nutrition. Ann. Intern.
Med. 93, 576 ± 577.
Jeejeebhoy KN, Detsky AS & Baker JP (1990a): Assessment of nutritional
status. J. Parenter. Enteral Nutr. 14, 193S ± 196S.
European Journal of Clinical Nutrition
De®ciencies related to fat malabsorption
PB Jeppesen et al
642
Jeejeebhoy KN (1990b): Does twice a week lipid infusion prevent essential
fatty acid (EFA) de®ciency during total parenteral nutrition (TPN).
[Editorial.] Nutrition 6, 193.
Jeppesen PB, Christensen MS, Hùy C-E & Mortensen PB (1997): Essential
fatty acid de®ciency in patients with severe fat malabsorption. Am. J.
Clin. Nutr. 65, 837 ± 843.
Jeppesen PB, Staun M & Mortensen PB (1998a): Adult patients receiving
home parenteral nutrition in Denmark from 1991 to 1996: who will
bene®t from intestinal transplantation? Scand. J. Gastroenterol. 338,
839 ± 846.
Jeppesen PB, Staun M, Tjellesen L & Mortensen PB (1998b): Effect of
intravenous ranitidine and omeprazole on intestinal absorption of water,
sodium, and macronutrients in patients with intestinal resection. Gut 43,
763 ± 769.
Jeppesen PB, Hùy C-E & Mortensen PB (1998c): Essential fatty acid
de®ciency in patients receiving home parenteral nutrition. Am. J. Clin.
Nutr. 68, 126 ± 133.
Kaplan LA, Miller JA, Stein EA & Stampfer MJ (1990): Simultaneous,
high-performance liquid chromatographic analysis of retinol, tocopherols, lycopene, and alpha- and beta-carotene in serum and plasma.
Meth. Enzymol. 189, 155 ± 167.
Kelly FJ & Sutton GL (1989): Plasma and red blood cell vitamin E status
of patients on total parenteral nutrition. J. Parenter. Enteral Nutr. 13,
510 ± 515.
Levy Y, Shils ME, McNamara DJ & Shike M (1987): Serum lipoproteins
in home total parenteral nutrition patients. J. Parenter. Enteral Nutr. 11,
471 ± 474.
Mascioli EA, Lopes SM, Champagne C & Driscoll DF (1996): Essential
fatty acid de®ciency and home total parenteral nutrition patients.
Nutrition 12, 245 ± 249.
Messing B, Pigot F, Rongier M, Morin MC, Ndeindoum U & Rambaud JC
(1991): Intestinal absorption of free oral hyperalimentation in the very
short bowel syndrome. Gastroenterology 100, 1502 ± 1508.
Moore RB, Buchwald H & Varco RL (1980): The effect of partial ileal
bypass on plasma lipoproteins. Circulation 62, 469 ± 476.
European Journal of Clinical Nutrition
Morrison WR & Smith LM (1964): Preparation of fatty acid methyl esters
and dimethylacetals from lipids with boron ¯uoride-methanol. J. Lipid.
Res. 5, 600 ± 608.
Nordgaard I, Hansen BS & Mortensen PB (1994): Colon as a digestive
organ in patients with short bowel. [See comments.] Lancet 343,
373 ± 376.
Nordgaard I, Hansen BS & Mortensen PB (1996): Importance of colonic
support for energy absorption as small-bowel failure proceeds. Am. J.
Clin. Nutr. 64, 222 ± 231.
Schaefer EJ, Woo R, Kibata M, Bjornsen L & Schreibman PH (1983):
Mobilization of triglyceride but not cholesterol or tocopherol
from human adipocytes during weight reduction. Am. J. Clin. Nutr.
37, 749 ± 754.
Shils ME, Baker H & Frank O (1985): Blood vitamin levels of long-term
adult home total parenteral nutrition patients: the ef®cacy of the AMAFDA parenteral multivitamin formulation. J. Parenter. Enteral. Nutr. 9,
179 ± 188.
Siguel EN, Chee KM, Gong JX & Schaefer EJ (1987): Criteria for
essential fatty acid de®ciency in plasma as assessed by capillary
column gas ± liquid chromatography. Clin. Chem. 33, 1869 ± 1873.
Smith FR & Lindenbaum J (1974): Human serum retinol transport in
malabsorption. Am. J. Clin. Nutr. 27, 700 ± 705.
Sorensen TI, Andersen B & Damgaard Pedersen F (1982): Plasma
cholesterol fractions after jejunoileal bypass with 3:1 or 1:3 jejunoileal
ratio. Scand. J. Gastroenterol. 17, 199 ± 203.
Traber MG & Kayden HJ (1987): Tocopherol distribution and intracellular
localization in human adipose tissue. Am. J. Clin. Nutr. 46, 488 ± 495.
Underwood BA, Siegel H, Weisell RC & Dolinski M (1970): Liver stores
of vitamin A in a normal population dying suddenly or rapidly from
unnatural causes in New York City. Am. J. Clin. Nutr. 23, 1037 ± 1042.
Vandewoude MG, Vandewoude MF & De Leeuw IH (1986): Vitamin E
status in patients on parenteral nutrition receiving Intralipid. J. Parenter. Enteral Nutr. 10, 303 ± 305.