Clinical Science (1992) 82, 7 1-76 (Printed in Great Britain)
71
Increased energy expenditure in cystic fibrosis is
associated with specific mutations
A. O’RAWE’, I. MclNTOSH2a3,J. A. DODGE’, D. J. H. BROCK2,A. 0. B. REDMOND4,R. WARDS
and A. J. S. MACPHERSONS
I Nuffield
Department of Child Health, Queen’s University of Belfast, Institute of Clinical Science,
Belfast, U.K., 2Human Genetics Unit, University of Edinburgh, Western General Hospital,
Edinburgh, U.K., 4Royal Belfast Hospital for Sick Children, Belfast, U.K., and SDepartment of
Clinical Biochemistry, King’s College School of Medicine and Dentistry, London, U.K.
(Received 29 April129 July 199 I ; accepted 8 August 199 I )
1. Measurements of resting metabolic rate were made by
open-circuit indirect calorimetry in 78 unrelated cystic
fibrosis patients and 30 healthy control subjects. The
aims of this study were: (i) to determine the range of
variability in resting metabolic rate in cystic fibrosis, (ii)
to relate this to pulmonary function and body size, and
(iii) to investigate the hypothesis that, in cystic fibrosis,
genotype exerts a significant influence on energy requirements.
2. There was no significant difference in age or body
weight between patients with cystic fibrosis and control
subjects. Resting metabolic rates for control subjects fell
within f 10%of predicted values. Fifty-nine per cent of
patients with cystic fibrosis had elevated resting metabolic rates (i.e. > l l l%
of predicted). Genotype analysis
divided the patients with cystic fibrosis into three groups:
AF508 homozygotes, AF508 heterozygotes and others.
Patients homozygous for the AF508 allele had a significantly higher resting metabolic rate (121%of predicted,
95% confidence interval 116-126%), compared with
other genotypes ( P < 0.005).
3. There were significant differences in pulmonary
function between the groups ( P < 0.005). However, after
adjustment of individual resting metabolic rates for differences in pulmonary function by using analysis of covariance, resting metabolic rates remained significantly
higher for AF508 homozygotes than for other genotypes
(P< 0.05).
4. We conclude that there is a significant contribution to
resting metabolic rate in cystic fibrosis associated with
specific mutations that is not explained by declining
pulmonary function. The increase in resting metabolic
rate in patients homozygous for mutations involving a
nucleotide-binding fold, which may result from a disruptive effect on ATP binding, indicates a practical implication of genotype identification with the need for effective
nutritional intervention and support in this patient subgroup.
INTRODUCTION
Although an elevation in resting metabolic rate (Rh4R)
is documented in cystic fibrosis (CF)[l-31, this is subject to
wide variability in onset and in magnitude, and it is
becoming clear that this is a multi-factorial trait determined by genetic and non-genetic factors and their interaction. The CF gene is large, extending over 250 kb of
genomic DNA, with over 70 mutations having been identified in the gene product, cystic fibrosis transmembrane
conductance regulator (CFTR). In CF, homozygotes or
compound heterozygotes for two abnormal alleles are
clinically affected. The common CF mutation, which
involves a phenylalanine deletion at position 508 of the
gene product (AF508), is identified in 71% of CF
chromosomes in the U.K. [4], and is located in a proposed
nucleotide-binding fold (NBF) of CFTR [5]. Recent
evidence from Thomas et al. [6] demonstrates that this
region binds adenosine nucleotides. The protein contains
two proposed nucleotide-binding domains, and both sites
may be required for ATP catalysis. The structural alteration induced by mutation in a NBF is therefore of potential importance for the functional regulation of cellular
energy metabolism.
Heritability is believed to account for between 11 and
40% of the total variance in RMR between healthy subjects [7, 81. We have suggested a genetic influence on
energy expenditure in CF [9], recognizing two clinical
subgroups of CF patients, those with elevated Rh4R and
those whose RMR is normal. Although a weak correlation
has been shown between Rh4R and pulmonary function
[l. 21, the recognition of increased RMR in patients with
Key words: cystic fibrosis genotype, resting metabolic rate, variability.
Abbreviations: BMP. body mass percentile; CF, cystic fibrosis; CFTR cystic fibrosis transmembrane conductance regulator; CN, Chrispin-Norman score; FEV,, forced
expiratory volume in I s; FVC. forced vital capacity; MAMC, mid-arm muscle circumference; NBF, nucleotide-bindingfold; RMR, resting metabolic rate; RQ, respiratory
quotient.
’Present address: Center for Medical Genetics, Johns Hopkins University Medical School, Baltimore MD 21 205, U.S.A.
Correspondence: D r A. O’Rawe, Nuffield Department of Child Health, Queen’s University of Belfast, Institute of Clinical Science, Grosvenor Road, Belfast
BT 12 681, U.K.
A. ORawe et al.
72
normal lung function and growth parameters suggests
additional factors, including a response to the basic cellular defect. We hypothesized that mutations in the NBFs of
the gene product exert an influence on R M R in CF.
METHODS
Subjects
Seventy-eight unrelated CF patients (52 males, 26
females) with a mean age of 11.2 years (range 1.5-28
years) were studied. We used a group of 30 healthy control subjects with a mean age of 14.9 years (range 8-24
years) to standardize our calorimetric methods. All subjects were of British Caucasian origin. Body weight in
light indoor clothing, without shoes, was measured to the
nearest 0.1 kg (model 707;Secca Delta, Hamburg, Germany) and height was recorded to the nearest 0.1 cm by
using a wall standiometer. To assess nutritional status,
these parameters were plotted on the growth and
development charts of Tanner & Whitehouse [lo],and
body mass percentile (BMP) was calculated by dividing
the body mass index (weight/height2)for each subject by
the index calculated for a subject of the same age, but at
the 50th centile for both height and weight, and multiplying by 100 to give the BMP [ll]. Th'IS measure was
chosen to allow comparison of stature at different ages
allowing for expected weight and height changes with
growth. Triceps, biceps subscapular and supra-iliac skinfold thicknesses were measured by a single observer, with
a Harpenden skinfold caliper. Mid-arm muscle circumference (MAMC) was calculated as an indirect indicator of
protein reserve and was expressed as a percentage of the
expected standard [12].
The proportions of body fat were
calculated [ 131 from the sum of skinfold measurements at
four sites.
Tests of ventilatory capacity [forced expiratory volume
in 1 s (FEV,) and forced vital capacity (FVC), expressed
as percentage of predicted for age, height and sex] were
made in patients old enough to permit pulmonary function testing, and the scoring system of Chrispin &
Norman (CN score) [14]was used to assess the chest
radiographs. This score measures changes in the lung
fields with points awarded, to a maximum of 38, according to the severity of radiographic features.
Genomic DNA analysis
DNA was extracted from the leucocytes of CF patients
RMR
R M R was measured by continuous computerized
open-circuit indirect calorimetry, using a ventilated hood
(Datex Metabolic Monitor; Datex Instrumentation Corporation, Helsinki, Finland). Subjects aged 5 years and
over had measurements made whilst they were supine in a
warm room, 12 h post-prandially between 08.00 and
10.00 hours. There was no history of acute infection for at
least 4 weeks (although a number of those studied were
colonized by Pseudomonas).No medication was given for
12 h before the study. An adaptation period of 15-30
min was allowed before measurement was started. R M R
was calculated from 0, consumption and COz production
by using the equation of Weir [17],
with measurements at
1 min intervals for a minimum of 30 min. Tests were
carried out under close monitoring by a single observer
and where extra movements were recorded the subjects
were excluded from analysis. In children less than 5 years
old ( I I = 18), measurements were made at home at least
4 h post-prandially, within 2 h of the onset of sleep. These
patients were excluded from subsequent analysis of
respiratory quotient (RQ).To determine intra-individual
variability in R M R and RQ, replicate studies were made
on 30 subjects (including 15 control subjects) within 3
months. The measured R M R is expressed as a percentage
of predicted, as derived from the Harris-Benedict equations [ 17a],and the RQ was determined from gas fractions
(CO,/O,). The individual variability in R M R measurements in the individual subject on repeat assessment is
given by a coefficient of variation [ 181 of 4% for R M R and
3% for RQ.
Statistical analysis
Values in the text and Tables are expressed as means
with 95% confidence limits in parentheses. As there was
evidence of heterogeneity of variance in R M R values
between patients and control subjects, the non-parametric
Kruskal-Wallis test and an associated multiple-range
procedure [ 191 were used for this comparison. All other
group comparisons (including comparisons of RMR
between the patient groups) were performed by using
analysis of variance with Newman-Keuls multiple-range
procedure [20].
Analysis of co-variance was also used to
adjust R M R comparisons between patient groups for the
possible confounding effects of BMP, age and pulmonary
function. The square of the multiple-correlation coefficient ( r 2 )was used to assess the contribution of genotype
to RMR.
[15].Samples were randomly coded and sent to the
Human Genetics Unit, Western General Hospital, Edinburgh, for analysis. Genomic DNA was amplified by the
polymerase chain reaction and the AF508 mutation was
identified by hybridization of the amplified sequence with
allele-specific oligonucleotides or by polyacrylamide-gel
electrophoresis as previously described [4,161.In patients
without this mutation on both chromosomes, subsequent
analysis was undertaken for the presence of five other
mutations (AI507,G551D, R553X,G542X,R117H)[4].
RESULTS
We studied 108 subjects (including 30 control subjects).
Of the 78 C F patients, 31 (40%)were homozygous for the
AF508 mutation (AF508/AF508), 29 (37%) were
heterozygous for the AF508 mutation (AF508/N) and 18
(23%) carried other mutations (N/N genotype). Eight of
these latter patients had mutations characterized subsequently. An abnormal R M R has been defined as a
Energy requirements in cystic fibrosis
73
Table I. Anthropornetric data and pulmonary function in 108 subjects. Values are means with 95% confidence intervals
in parentheses. N denotes an uncharacterized mutation. Statistical significance: *P (0.005 by analysis of variance on four group
means; **significance difference from values for other genotype groups (P (0.01) by the Newman-Keuls multiple-range
procedure.
Group (n)
Age (years)
BMP (%)
FEV, (%)*
FVC (%)*
AF5081AF508 (31)
AF5081N (29)
NIN ( I 8)
Control subjects (30)
I3.0( 10.6-15.4)
9.7(7.1-12.3)
10.9(8.3-13.5)
l4.9( 13.6-16.1)
95(91-99)
98(94-101)
97(92- 102)
103( 100-105)
56(48-64)
63(52-74)
97(93-102)**
I03(99- 107)
67(59-74)
79(69-88)
104(98-109)**
I 10(104-115)
deviation of greater than 11% from the predicted RMR
[7]. Fifty-nine per cent of patients had an elevated RMR.
The range of RMR values for the control subjects fell
within f 10% of predicted values.
There were no significant differences in age or body
size (as indicated by BMP) between genotype groups and
control subjects (Table 1). Differences in ventilatory
capacity (FEV,, FVC) were significant between the four
groups ( P < 0.005); however, there was no significant difference in pulmonary function between the AF508/
AF508 genotype and the AF508/N genotype using
multiple-range procedure [ 191. The CN score (0-38
according to severity) and ventilatory capacity were significantly different between the three patient groups
( P < 0.01). The CN score for both the AF508/AF508 and
the AF508/N genotypes was 9 (95% confidence interval
7-11) and that for the N/N genotype was 5 (4-6). Our
data suggest that AF508 heterozygotes had slightly less
percentage body fat and less lean body mass (as expressed
by MAMC) than other genotype groups (Table 2). Of the
78 CF patients studied, 29 were chronically infected with
Pseudomonas aeruginosa and two with P. cepacia. There
was a significant difference in the rate of colonization
between genotype groups, AF508 homozygotes having
the highest frequency. Bacteriological data did not, however, appear to correlate with RMR of 31 patients culturing Pseudomonas in sputum, 13 had normal RMRs. In the
analysis of RMR (Table 3), results were significantly
higher for the AF508/AF508 genotype compared with
those for other patient groups and control subjects
(P<0.005). The differences in RMR between patient
groups alone were significant ( P < 0 . 0 0 5 ) , with a mean
RMR for AF508 homozygotes of 121% of predicted, for
AF508 heterozygotes of 109% of predicted and for those
without the AF508 mutation on either chromosome (N/N
group) of 104% of predicted. The Newman-Keuls multiple-range procedure confirmed that the significant difference lay in the AF508/AF508 group comparison(s).
There was no significant correlation shown between the
decline in pulmonary function (FEV,) and RMR ( r = 0.31)
in the 48 CF patients old enough to permit testing. In view
of differences in pulmonary function (including that indicated by the CN score), and the differences in body size
(BMP), percentage body fat, MAMC and age between
groups which might influence mean RMR values for
genotype groups, we used analysis of co-variance to
adjust RMR for differences in these co-variates
( 3= 0.14). RMR values remained significantly higher for
AF508 homozygotes than for the other patient groups
Table 2. Clinical data for patients with CF. Values are means+sD.
Statistical significance: *P (0.05.
Body composition
Fat (% of body weight)
Male
Female
MAMC (percentile)
Pseudomonas
in sputum
AF5081AF 508
AF 508/N
NIN
13f4
20f5
20 26
+
15f2
21 f5
37+28*
16+4
17+4
25+22
I713 I
12/29
2118
Table 3. Influence of the AF508 mutation on RMR and RQ.
Values are means with 95% confidence intervals in parentheses. N
denotes an uncharacterized mutation. Statistical significance: *P <0.005
by the Kruskal-Wallis test on patient and control groups, and by analysis
of variance on three patient groups only; **P<O.OOl by analysis of
variance on four means; $significantly different from patient and control
groups (P (0.01) by the Kruskal-Wallis multiple-range procedure;
+significantly different from other patient groups (P (0.01) by the
Newman-Keuls procedure.
Group
RMR* (% of predicted)
RQ**
AF5081AF508
AF5081N
NIN
Control subjects
I21 (I 16-1 26)t:
109(103-116)
104(97-1 I I )
10 I(98-104)
0.88(0.87-0.89)
0.86(0.85-0.87)
0.86(0.85-0.87)
0.80(0.79-0.82)t
after adjustment ( P < 0.05). When genotype was added to
the regression model that predicted RMR, and additional
20% of the variation in RMR was explained (3= 0.34).
Significantly higher RQ values (Table 3) were measured
for all CF patient groups compared with control subjects
( P < O.Ooi), with mean values of 0.87 and 0.80, respectively. These were not accounted for by differences in
FEV, and FVC between the four groups; however, differences in RQ between the three patient groups were
removed when adjustment was made for pulmonary function by analysis of co-variance.
Genomic DNA analysis detected mutations other than
AF508. We compared the results of extended genotype
analysis with clinical and metabolic manifestations of the
disease (Table 4). Each combination of compound heterozygote demonstrates a wide variability in the individual
patient. The functional consequences of AF508 heterozygosity, where the other mutation is not known, vary
widely. The second most common mutation amongst
Caucasian CF chromosomes is G551D, which is also
A. O'Rawe et al.
74
Table 4. Influence of CF genotype on BMP, CN score and RMR.
Values are means with 95% confidence intervals, except where n < 2 .
when individual values are given. N denotes an uncharacterized
mutation.
Genotype (n)
BMP
AF5081AF508 (31)
AF5081G55 ID (I)
AF5081RI 17H (3)
AF508IR553X (2)
AF508IG542X (2)
G551DIR117H (I)
AF508IN (20)
G55 I DIN (5)
A15071N (I)
G542XIN (2)
NIN (I0)
95 (9 1-99)
93
98(97-101)
95,89
79,97
I03
100 (97-104)
100(91-110)
79
92.89
96(88- 105)
CN score*
9(8-I I)
10
5(3-6)
12, I I
13, 3
I
8(6-1 I)
7(4-1 I)
14
9, 3
4(2-6)
RMR (%)
I21 (I16-1 26)
121
IIl(100-122)
91, 106
91,94
86
I10(101-118)
I21 (I12-1 34)
I33
96,87
102(91-112)
*Severity o f radiographic features according t o the system o f Chrispin & Norman
[ 141, with points awarded t o a maximum of 38 (best score=O).
situated in a proposed nucleotide-binding domain [23],
and was identified on six chromosomes in six patients.
This mutation had a similar effect on RMR to the AF508
mutation; however, patients had slightly better CN scores
and pulmonary function. Patients ( w = 8) who were compound heterozygotes for a mutation within a nucleotidebinding domain with a second mutation either outsidc this
functional region (R117H) [21] or a nonsense mutation
(R553X [22] or G542X [23]) demonstrated a normal
RMR of 100% of predicted (91-109). Four of the five
lowest BMPs were recorded in patients carrying the
G542X allele, either in combination with AF508 or
another undetermined allele. RMR was not increased in
this sub-group. The basis for their energy imbalance is
unclear, but high faecal energy losses may be one explanation.
DISCUSSION
We have identified a correlation between genotype and
phenotype in CF. Homozygosity for mutation(s) in a proposed nucleotide-binding domain of the gene product
CFTR has a significant influence on RMR in CF. Moreover, although our data confirm that the lowest recorded
values for FEV, and FVC, and the highest chronic infection rate by Pseudomonas, were found in the AF508/
AF508 genotype group, there was no direct correlation
between RMR and these measures of pulmonary function, and no direct association between colonization of
the respiratory tract with Pseudomorias and an increased
RMR. In attempting to relate metabolic requirements to
lean body mass, we were constrained by the availability of,
methods with which to assess body composition. Assessment of body composition in CF has demonstrated a
decrease in both body fat and LBM [24], and inadequate
accretion of total body potassium has been shown in both
presymptomatic infants detected by screening and in
patients growing within normal centiles [25]. As prediction equations to calculate the body-fat content of
children with C F appear to be invalid [26], we did not use
them to derive values for lean body mass, but used
MAMC as an index of lean body mass. Including MAMC
and percentage body fat in analysis of co-variance, differences in body composition between the CF groups
derived from anthropometric data did not explain the variation in RMR. Differences between genetic group means
were significant after adjustment for differences in pulmonary function and anthropometric data, with significantly elevated values for RMR in AF508 homozygotes
irrespective of FEV,, FVC, percentage body fat and
h4AMC. Other studies, while recognizing a contribution
from the decline in pulmonary function and differences in
body composition to measured increases in RMR, have
suggested the influence of a basic defect in energy metabolism at the cellular level [l, 21. The variable clinical
course in CF patients has been attributed, at least in part,
to specific genotypes at the C F locus [27]. Using multipleregression analysis we have shown that 20% of the
variance in RMR is explained by the presence of the
AF508 mutation. This value is similar to the percentage
variation in RMR attributed to heritability in family
studies of healthy subjects [7, 81. Although exerting a significant influence on RMR independent of pulmonary
function, genotype is not, however, the sole determinant
of energy requirement in CF and other factors (genetic
and/or environmental) are recognized to play a role in the
course of the disease.
The AF508 mutation has been localized to a putative
NBF in the CFTR molecule [5]. Recent studies with synthetic oligopeptides have shown that this region of CFTR
does indeed possess adenine-nucleotide-binding activity.
Circular dichroism spectroscopy predicts that the deletion of the phenylalanine residue at position 508 would
disrupt a B-sheet structure in this region [6]. It has been
assumed, but not proven, that bound ATP is hydrolysed
and the energy released is used to permit a conformational change in the molecule. Other mutations in the
NBF (e.g. G551D) may also have a disruptive effect on
ATP binding. It has recently been suggested that CFTR is
a tightly regulated chloride channel and that ATP binding
and hydrolysis is necessary for conformational change
before chloride transport [28].Our evidence from indirect
calorimetry, that specific mutations in the nucleotidebinding domain alter cellular energy requirement, supports this hypothesis.
The recessive nature of C F means that both alleles play
a role in determining phenotype and both characters in a
compound heterozygote are expected to exert specific
effects on the CFTR protein. The functional consequences of rare allelic variants vary widely. That values
for some of the AF508/N group fell within the range of
the AF508/AF508 group suggests that the other mutation may exert a similar influence to AF508 (i.e. that it is
situated within the NBF). In particular, the G551D mutation has a similar effect on RMR to the AF508 mutation,
and both mutations lie within exons 10-11 of the gene
and relate to the same functional domain [5. 231. Differences in pulmonary and anthropometric assessment
support the suggestion of independent genetic determinants for various clinical components of the phenotype
[29]. Certain known mutations do not confer an increase
Energy requirements in cystic fibrosis
in RMR. For example, the R117H mutation is near a
transmembrane region of the protein and is probably not
involved in nucleotide binding [21]. If the observed
phenotype is the result of aberrant ATP binding o r hydrolysis, null mutations which prevent synthesis of the CFTR
protein (e.g. the stop mutations G542X and R553X)
would not result in this phenotype.
Our results have shown a significant influence on R Q in
all C F patient groups compared with control subjects,
indicating a shift towards an increase in carbohydrate
metabolism. This was most marked in patients homozygous for AF508, although differences in pulmonary
function explained the difference in RQ between genotype groups. A n increase in R Q in C F patients has previously been reported [30]. The possibility that this might
be a diet-induced response [31] was explored by analysing
7-day weighed intake data on 37 of the patients studied.
The average food quotient was 0.87 for patient and control groups and no correlation was shown between RQ
and food quotient in regression analysis. Two possible
explanations for this increase in R Q are: (i) the glycogen
reserve in C F is larger (although there is no evidence of
this), or (ii)glycogen reserves are depleted more slowly by
earlier mobilization of fat in CF. Evidence for an
increased turnover of fatty acids in C F [32, 331 would be
consistent with this hypothesis.
In summary, the variable nutritional requirements in
C F patients can be partly attributed to specific genotypes
at the locus of the C F gene. An increase in R Q indicating
altered substrate utilization, however, seems to be a metabolic consequence of CF, irrespective of genotype. At a
clinical level these results indicate that specific genotypes
are at a greater risk of nutritional compromise. A high
RMR combined with elevated fasting RQs resulting in
RQ/food quotient imbalance has implications not only for
intervention, but in planning current and future strategies
for nutritional supplementation in patients with CF.
ACKNOWLEDGMENTS
We thank Dr C. Patterson, Department of Medical
Statistics, Queen’s University of Belfast, and Dr F.
Stanford, Mrs B. Martin and Dr K. McCracken for help
and advice. A.OR. was supported by the British
Paediatric Association (Cow and Gate Research Fellowship). I.M. was supported by the Cystic Fibrosis Research
Trust.
REFERENCES
I. Buchdahl, R.M., Cox, M., Fulleylove, C. et al. Increased resting energy
expenditure in cystic fibrosis. J. Appl. Physiol. 1988; 64, 1801-6.
2. Shepherd, R.W., Holt, T.L., Vasques-Velasquez, L., Coward, W.A., Prentice,
A. & Lucas, A. Increased energy expenditure in children with cystic fibrosis.
Lancet 1988; i, 1300-3.
3. Vaisman, N., Pencharz, P.B., Corey, M., Canny, G.J. & Hahn, E. Energy
expenditure of patients with cystic fibrosis. J. Pediatr. 1987; I I I, 496-500.
4. Shrimpton, A.E., Mclntosh, I. & Brock, D.J.H. The incidence of different
cystic fibrosis mutations in the Scottish population: effects on prenatal
75
diafnosis and fenetic counsellinn. I. Med. Genet. 1991; 28, 317-21.
5. RiGdan, J.R., iommens, J.M., K&m, B.4. et al. Identification of the cystic
fibrosis gene: cloning and characterisation of complementary DNA. Science
(Washington, D.C.) 1989; 245, 1066-73.
6. Thomas, P.J., Shenbagamurthi, P., Ysern, X. & Pedersen, P.L. Cystic fibrosis
transmembrane conductance regulator: nucleotide binding to a synthetic
peptide. Science (Washington, D.C.) 1991; 251,555-7.
7. Bogardus. C.. Lillioja, S., Ravussin, E. et al. Familial dependence of the resting
metabolic rate. N. Engl. J. Med. 1986; 315,96-100.
8. Bouchard, C.. Tremblay, A,, Nadeau, A. et al. Genetic effect in resting and
exercise metabolic rates. Metab. Clin. Exp. 1989; 38, 364-70.
9. ORawe. A., Dodge, J.A., Redmond. A.O.B., Mclntosh, I. & Brock. D.J.H.
Genelenergy interaction in cystic fibrosis. Lancet 1990; 335, 552-3.
10. Tanner, J.M. & Whitehouse, R.H. Clinical longitudinal standards for height,
weight, height velocity and weight velocity and stages of puberty. Arch. Dis.
Child. 1976; 51, 170.
I I. Coates, A.L., Boyce, P., Muller, D., Mearns, M. & Godfrey, S. The role of
nutritional status, airway obstruction, hypoxia, and abnormalities in serum
lipid composition in limiting exercise tolerance in children with cystic
fibrosis. Acta. Paediatr. Scand. 1980; 69, 353-8.
12. Frisancho, A.R. New norms of upper limb fat and muscle areas for
assessment of nutritional status. Am. J. Clin. Nutr. 1981; 34,2540-5.
13. Durnin, J. & Rahaman, M.M. The assessment of the amount of fat in the
human body from measurements of skinfold thickness. Br. J. Nutr. 1967; 21,
681-9.
14. Chrispin, A.R. & Norman, A.P. The systematic evaluation of the chest radiograph in cystic fibrosis. Pediatr. Radiol. 1974; 2, 101-6.
15. Bell, G.I., Karam, J.H. & Rutter. W.J. Polymorphic DNA region adjacent to
the 5’ end of the human insulin gene. Proc. Natl. Acad. Sci. U.S.A. 1981; 78,
5759-63.
16. Mclntosh. I., Raeburn, J.A.. Curtis, A. & Brock, D.J.H. First trimester prenatal
diagnosis of cystic fibrosis by direct gene probing. Lancet 1989; ii, 972-3.
17. Weir, J.B., de V. New methods for calculating metabolic rate with special
reference to Drotein metabolism. ,I. Phvsiol.
(London) 1949: 109. 1-9.
,
17a. Roza, A.M. & Shizgal. H.M. The Harris Benedict equation reevaluated
resting energy requirements and the body cell mass. Am. J. Clin. Nutr. 1984;
40, 168-82.
18. Bland, M. An introduction to medical statistics. Oxford: Oxford Medical
Publications, 1989.276-7.
19. Conover, W.J. Practical nonparametric statistics. 2nd ed. Dallas: Texas
Technical University, 1980.
20. Snedecor, G.W. & Cochran, W.G. Statistical methods. Ames, I A Iowa State
University Press, 1967.
21. Dean, M., White, M.B., Amos, J. et al. Multiple mutations in highly conserved
residues are found in mildly affected cystic fibrosis patients. Cell 1990; 61,
863-70.
22. Cutting, G.R., Kasch, L.M., Rosenstein, B.]. et al. A cluster of cystic fibrosis
mutations in the first nucleotide-binding fold of the cystic fibrosis conductance regulator protein. Nature (London) 1990; 346,366-9.
23. Kerem, B.-S.. Zielenski, J., Markiewicz, D. e t al. Identification of mutations in
regions corresponding t o 2 putative nucleotide (ATP)-binding folds in the
cystic fibrosis gene. Proc. Natl. Acad. Sci. U.S.A. 1990; 87,8447-5 I.
24. Soutter, V.L., Kristidis, P., Gruca, M.A. & Gaskin, K.J. Chronic undernutritionlgrowth retardation in cystic fibrosis. Clin. Gastroenterol. 1986; 15,
137-46.
25. Shepherd, R.W., Holt, T.L., Greer, R.. Cleghorn, G.J. & Thomas, B.J. Total
body potassium in cystic fibrosis. J. Paediatr. Gastroenterol. Nutr. 1989; 9,
200-5.
26. Johnston,J.L., Leong, M.S., Checkland, E.G., Zuberbuhler, P.C., Conger, P.R.
& Quinney, H.A. Body fat assessed from body density and estimated from
skinfold thickness in normal children and children with cystic fibrosis. Am. J.
Clin. Nutr. 1988; 48, 1362-6.
27. Kerem, E., Corey, M., Kerem, B. et al. The relation between genotype and
phenotype in cystic fibrosis - analysis of the most common mutation
(AF508). N. Engl. J. Med. 1990; 323, 1517-22.
28. Kartner, N., Hanrahan, J.W.. Jensen, T.J. et al. Expression of the cystic
fibrosis gene in non-epithelial invertebrate cells produces a regulated anion
conductance. Cell I99 I ; 64,68 1-9 I.
29. Santis, G., Osborne, L., Knight, R.A. & Hodson, M.E. Independent genetic
determinants of pancreatic and pulmonary status in cystic fibrosis. Lancet
1990; 336, 1081-4.
30. Hirsch. J.A., Zhang, S.-P., Rudnick, M.P., Cerny, F.J. & Cropp, G.J. Resting
oxygen consumption and ventilation in cystic fibrosis. Pediatr. Pulmonol.
I
76
A. O’Rawe et al.
1989; 6, 19-26.
31. Goldstein, S.. Askanazi, J., Weissman, C., Thomashow, B. & Kinney. J.M.
Energy expenditure in patients with chronic obstructive pulmonaty disease.
Chest 1987; 91,222-4.
32. Dodge, J.A., Custance, J.M., Goodchild, M.C., Laing, S.C. & Vaughan, M.
Paradoxical effects of essential fatty acid supplementationon lipid profiles
and sweat electrolytes in cystic fibrosis. Br. 1. Nutr. 1990; 63,259-71.
33. Rogiers, V., Dab, I., Michotte, Y., Vercruysse, A,, Crockart, R. & Vis, H.L.
Abnormal fatty acid turnover in the phospholipids of the red blood cell
membranes in cystic fibrosis patients. Pediatr. Res. 1984; 18, 704-9.