Papers: Breast Feeding and Obesity: Cross Sectional Study

Download as pdf or txt
Download as pdf or txt
You are on page 1of 4

Papers

Breast feeding and obesity: cross sectional study


Rdiger von Kries, Berthold Koletzko, Thorsten Sauerwald, Erika von Mutius, Dietmar Barnert,
Veit Grunert, Hubertus von Voss

Abstract
Objective To assess the impact of breast feeding on
the risk of obesity and risk of being overweight in
children at the time of entry to school.
Design Cross sectional survey
Setting Bavaria, southern Germany.
Methods Routine data were collected on the height
and weight of 134 577 children participating in the
obligatory health examination at the time of school
entry in Bavaria. In a subsample of 13 345 children,
early feeding, diet, and lifestyle factors were assessed
using responses to a questionnaire completed by
parents.
Subjects 9357 children aged 5 and 6 who had
German nationality.
Main outcome measures Being overweight was
defined as having a body mass index above the 90th
centile and obesity was defined as body mass index
above the 97th centile of all enrolled German
children. Exclusive breast feeding was defined as the
child being fed no food other than breast milk.
Results The prevalence of obesity in children who
had never been breast fed was 4.5% as compared with
2.8% in breastfed children. A clear dose-response
effect was identified for the duration of breast feeding
on the prevalence of obesity: the prevalence was 3.8%
for 2 months of exclusive breast feeding, 2.3% for 3-5
months, 1.7% for 6-12 months, and 0.8% for more
than 12 months. Similar relations were found with the
prevalence of being overweight. The protective effect
of breast feeding was not attributable to differences in
social class or lifestyle. After adjusting for potential
confounding factors, breast feeding remained a
significant protective factor against the development
of obesity (odds ratio 0.75, 95% CI 0.57 to 0.98) and
being overweight (0.79, 0.68 to 0.93).
Conclusions In industrialised countries promoting
prolonged breast feeding may help decrease the
prevalence of obesity in childhood. Since obese
children have a high risk of becoming obese adults,
such preventive measures may eventually result in a
reduction in the prevalence of cardiovascular diseases
and other diseases related to obesity.

Introduction
In industrialised countries obesity and being overweight are the most frequent nutritional disorders in
BMJ VOLUME 319

17 JULY 1999

www.bmj.com

children and adolescents, and there is a continuing


increase in their prevalence.1 Overweight children have
a high risk of being overweight in adulthood 24 and
therefore are also at risk from the associated health
complications such as hypertension and coronary
heart disease.5 Since therapeutic interventions aimed
at encouraging weight loss in obese children are costly
and have long term success rates that are less than satisfactory,6 the identification of strategies for the
effective prevention of obesity is particularly attractive.
Simple strategies without potential side effects are
the most appealing. Breast feeding would fulfil these
criteria. However, the impact of breast feeding on
obesity has only been studied in comparatively small
cohorts.79 These small studies failed to find a
protective effect possibly due to a lack of statistical
power, whereas a protective effect has been reported in
a Canadian cross sectional study of 1320 adolescents
born in the late 1960s.10 The rates of breast feeding
were low in Canada at that time, and lifestyles in industrialised countries have changed considerably over the
past three decades. We have therefore studied the
impact of breast feeding on the prevalence of being
overweight or obese in children born in the early
1990s.

Subjects and methods


Study population and data sources
The 1997 obligatory health examination before school
entry evaluated 134 577 children in Bavaria, southern
Germany. At the examination, the parents of 13 345
children seen in two rural regions were asked to complete a questionnaire about risk factors for atopic
diseases.11 Data collected by this questionnaire were
linked with the data from the school health
examination. Our analysis was confined to children
aged 5 and 6 who had German nationality.
The childrens height and weight were measured as
part of the routine examination. Body mass index was
calculated as weight (kg)/(height (m)2). The age specific
and sex specific distribution of the body mass index
among all children with German nationality in Bavaria,
which had been investigated during the 1997 school
health examination, was used as the reference for
being overweight (defined as body mass index above
the 90th centile) or obese (defined as body mass index
above the 97th centile) because these centiles were
higher than other European reference values.12

Institute for Social


Paediatrics and
Adolescent
Medicine, Ludwig
Maximilians
University,
Heiglhofstr 63,
D-81377 Munich,
Germany
Rdiger von Kries,
professor of
paediatrics
Dietmar Barnert,
statistician
Veit Grunert,
statistician
Hubertus von Voss,
professor of
paediatrics
Dr von
Haunersches
Kinderspital,
Ludwig
Maximilians
University,
Lindwurmstr 4,
D-80337 Munich
Berthold Koletzko,
professor of
paediatrics
Thorsten
Sauerwald,
senior house officer
Erika von Mutius,
reader in paediatrics
Correspondence to:
R von Kries
ag.epi@lrz.
uni-muenchen.de
BMJ 1999;319:14750

147

Papers

Table 1 Duration of breast feeding and prevalence (95%


confidence interval) of being overweight (body mass index above
the 90th centile) or obese (body mass index above the 97th
centile) among 5 and 6 year olds living in rural Bavaria
Prevalence (%) of:
Duration of breast feeding

Being overweight

Being obese

Never breast fed (n=4022)

12.6 (12.4 to 12.9)

4.5 (4.4 to 4.6)

Ever breast fed (n=5184*)

9.2 (9.0 to 9.3)

2.8 (2.7 to 2.8)

Exclusively breast fed for:


<2 months (n=2084)

11.1 (10.6 to 11.6)

3.8 (3.6 to 4.0)

3-5 months (n=2052)

8.4 (8.1 to 8.8)

2.3 (2.2 to 2.4)

6-12 months (n=863)

6.8 (6.1 to 7.6)

1.7 (1.6 to 1.9)

>12 months (n=121)

5.0 (1.1 to 8.8)

0.8 (0.2 to 1.5)

*Duration of breast feeding was not reported for 64 children who had ever
been breast fed.

Questionnaire
Parents were asked: Was your child breast fed? If they
answered yes they were then asked: For how long was
your child exclusively breast fed: (a) 2 months or less,
(b) 3 to 5 months, (c) 6 to 12 months or (d) more than
a year? Exclusive breast feeding was defined as the
child being fed no food other than breast milk.
To assess potential confounders additional questions
were asked about the number of older siblings the child
had and the parents ages, the childs health (for
example, was the child born prematurely or at low birth
weight?), early feeding (for example, when solid foods
were introduced), and the actual frequency of eating
selected foods. The highest level of education attained by
either parent was used as a marker for social class.
Statistical analyses
The prevalences of overweight and obese children
were calculated according to the duration of breast
feeding. The appropriate 2 tests were used to compare
several items in breastfed and non-breastfed children
and their association with the child being overweight
or obese. Logistic regression models were used to
assess the impact of variables that were significantly
associated (P < 0.05) with both breast feeding and
being overweight or obese. Confounding was assumed
to have occurred if the odds ratio changed by >10%.
Confounders and independent risk factors were
included in the final logistic regression model. All

calculations were carried out with the SAS software


package, version 6.12.

Results
The overall response rate to the questionnaire was
76.7% (10 240/13 345) Parents whose children had
been examined before 1 February did not receive questionnaires and were classed as non-responders. We
could not differentiate between true non-responders
and children whose parents had not been given
questionnaires so both were included as nonresponders for calculations of well baby visits and
immunisations because similar results had been
obtained when these calculations were confined to those
offices where all parents had received questionnaires.
A total of 9357 questionnaires were completed for 5
and 6 year old German children. Information on breast
feeding and its duration was available for 9206 children.
A total of 4022 children had never been breast fed and
5184 had ever been breast fed. The duration of breast
feeding was not reported for 64 children.
Responders were more likely than non-responders
to have attended all well baby visits (70.6% (6524/
9238) v 64% (4511/7044)) and to have had their children vaccinated against measles and Haemophilus
influenzae type b (72.3% (6764/9357) v 64%
(4596/7178)), but mean body mass index and the 90th
and 97th centiles for body mass index were similar
between the groups (mean 15.36 v 15.34; 90th centile
17.70 v 17.75; and 97th centile 20.12 v 20.07).
There was a clear dose dependent effect of the
duration of breast feeding on the prevalence of being
overweight or obese in children at the time of entry to
school (table 1). Similar effects of the duration of breast
feeding on the prevalence of being overweight or
obese were observed when different definitions of
being overweight or obese were used (that is, above the
90th centile or above the 97th centile for weight; above
110% or above 120% of the median weight for height
categories in the total population (data not shown)).
Several indicators of the familys lifestyle and make
up (for example, whether the child had his or her own
bedroom, the amount of time spent playing outside in
winter and summer, whether the mother smoked during

Table 2 Prevalence of independent risk factors associated with breast feeding and being overweight or obese in 5 and 6 year old
children in rural Bavaria
Prevalence (%)
Non-breast fed
children (n=4022)
High level of parental education (>10 years)*

41.4

Odds ratio (90% CI) for:

Breast fed children


(n=5184)

Being overweight

Being obese

66.7

0.77 (0.67 to 0.89)

0.62 (0.49 to 0.79)

Maternal smoking during pregnancy

12.8

4.2

1.51 (1.20 to 1.89)

1.82 (1.28 to 2.58)

Prematurity

13.8

9.0

0.78 (0.62 to 0.98)

0.69 (0.46 to 1.03)

Birth weight <2500 g

10.4

6.6

0.69 (0.48 to 0.84)

0.78 (0.54 to 1.10)

Own bedroom

45.6

54.4

1.19 (1.03 to 1.37)

1.22 (0.96 to 1.56)

Consumes margarine >3 times/week

35.3

32.4

1.22 (1.05 to 1.41)

1.21 (0.94 to 1.56)

Consumes butter >3 times/week

60.5

69.2

0.73 (0.63 to 0.83)

0.70 (0.56 to 0.88)

Consumes full fat milk >3 times/week

50.8

59.6

0.69 (0.60 to 0.80)

0.54 (0.42 to 0.68)

Consumes low fat milk >3 times/week

31.9

28.8

1.72 (1.49 to 1.99)

1.77 (1.38 to 2.25)

Consumes full fat quark or yogurt >3 times/week

28.8

36.1

0.66 (0.56 to 0.78)

0.52 (0.38 to 0.70)

Consumes low fat quark or yogurt >3 times/week

25.9

23.8

1.42 (1.22 to 1.66)

1.32 (1.02 to 1.71)

Consumes whipped cream >1 time/week

18.6

24.7

0.65 (0.54 to 0.79)

0.58 (0.41 to 0.81)

Consumes breakfast cereals >3 times/week

25.6%

35.3

0.80 (0.68 to 0.93)

0.76 (0.58 to 0.99)

Consumes sweet desserts >3 times/week

54.4%

57.8

0.84 (0.74 to 0.97)

0.82 (0.66 to 1.03)

*This variable changed the odds ratio for breast feeding and being overweight or obese by at least 10%, so confounding was assumed to have occurred.

148

BMJ VOLUME 319

17 JULY 1999

www.bmj.com

Papers

Table 3 Odds ratios for independent risk factors associated with


being overweight or obese in the final logistic regression model
for 9206 5 and 6 year old children in rural Bavaria
Odds ratio (90% CI) for:
Being overweight

Being obese

Breast feeding

0.79 (0.68 to 0.93)

0.75 (0.57 to 0.98)

High level of parental education


(>10 years)*

0.81 (0.70 to 0.95)

0.75 (0.58 to 0.98)

Maternal smoking during


pregnancy

1.52 (1.81 to 1.95)

1.85 (1.26 to 2.71)

Birth weight <2500 g

0.68 (0.53 to 0.88)

0.81 (0.54 to 1.20)

Own bedroom

1.24 (1.05 to 1.46)

1.20 (0.91 to 1.59)

Consumes butter
>3 times/week

0.72 (0.62 to 0.84)

0.69 (0.54 to 0.90)

*This variable changed the odds ratio for breast feeding and being overweight
or obese by at least 10%, so confounding was assumed to have occurred.

pregnancy, the number of older siblings, and the age of


the parents), early feeding habits, and the frequency of
the consumption of several products in the childs diet
differed significantly between children who had been
breast fed and those who had not. The prevalence of
variables which were also significantly associated with a
child being overweight or obese and which were
independent risk factors in a logistic regression model
with breast feeding is shown in table 2.
Higher levels of parental education (>10 years),
premature birth, and low birth weight were inversely
associated with being overweight or obese, whereas
maternal smoking during pregnancy and the child
having his or her own bedroom were positively correlated. Full fat milk products and sweet desserts were
less frequently consumed by overweight children, and
the consumption of low fat milk products was higher.
Overweight children also ate less butter and breakfast
cereals than children who were not overweight.
The level of parental education was the only factor
that accounted for a shift of the odds ratio towards
unity by at least 10% which related breast feeding to
being overweight or obese. In the final logistic
regression, which estimated the influence of breast
feeding on the dependent variable of being overweight
or being obese, the independent risk factors for maternal smoking, low birth weight, own bedroom, and
frequent consumption of butter were included in addition to the confounding variable of higher level of
parental education (table 3).
Crude and adjusted odds ratios for the dose
dependent impact of breast feeding on being
overweight or obese are shown in table 4. In children
who had been breast fed for at least 6 months or more
the risks of being overweight or obese were reduced by
> 30 % and > 40%, respectively.

Discussion
Epidemiological evidence for reduced risk
To our knowledge this is the largest epidemiological
study on the impact of breast feeding on the risk of
school age children being overweight or obese. The
most remarkable finding is a consistent, protective, and
dose dependent effect of breast feeding on different
definitions of obesity or being overweight. Some relevant
questions, however, were not asked in our questionnaire
because it had been designed originally to identify atopy.
The impact of breast feeding on body mass increase
(catch up growth) in low birthweight infants, which is
associated with an additional risk of coronary heart disease,13 could therefore not be analysed. A positive family
history of being overweight is an important indicator of
the genetic risk for obesity and being overweight,9 10
although it is not a confounder of the association
between breast feeding and obesity or being overweight
as found in a previous study.10
Breast feeding was associated with family make up
and lifestyle, premature birth, low birth weight, and
early and current diet. There was an inverse relation
between the consumption of butter and margarine and
consumption of low fat and full fat dairy products. The
inverse relation between the actual consumption of full
fat dairy products and obesity or being overweight
probably reflects avoidance of these products by
children who are overweight. Because of these strong
relations and the apparent reverse causation with
regard to full fat dairy products, only low birth weight
and the consumption of butter were added to the final
logistic regression model.
The protective effects of a higher level of parental
education and low birth weight accord with the results of
other studies.9 10 13 Family income or social class might be
better indicators of socioeconomic status. Unfortunately,
in Germany there is no accepted equivalent to the British categories of social class, and respondents to written
questionnaires are reluctant to report income.
A similar dose dependent reduction in the risk of
being overweight or obese as associated with breast
feeding was observed in Canadian adolescents born in
the 1960s.10 Only 18.5% of these children had been
breast fed exclusively as compared with 56% of those
born in Bavaria in the 1990s; this suggests that mothers
with different sociodemographic characteristics have
chosen to breast feed their children in Bavaria now. If
this dose dependent protective effect had been caused
by lifestyle factors associated with breast feeding, similar
confounding factors should have been operative during
different times in different societies. These factors would
also have to be closely related to the duration of breast

Table 4 Crude and adjusted odds ratios (95% confidence intervals) of the dose dependent impact of breast feeding on being
overweight or obese in children aged 5 or 6 in rural Bavaria
Being overweight

Being obese

Adjusted odds ratio*

Crude odds ratio

Adjusted odds ratio*

Crude odds ratio

<2 months (n=2084)

0.89 (0.73 to 1.07)

0.87 (0.74 to 1.02)

0.90 (0.65 to 1.24)

0.84 (0.64 to 1.10)

3-5 months (n=2052)

0.87 (0.72 to 1.05)

0.64 (0.53 to 0.76)

0.65 (0.44 to 0.95)

0.50 (0.36 to 0.69)

6-12 months (n=863)

0.67 (0.49 to 0.91)

0.51 (0.38 to 0.67)

0.57 (0.33 to 0.99)

0.38 (0.22 to 0.64)

>12 months (n=121)

0.43 (0.17 to 1.07)

0.36 (0.16 to 0.82)

0.28 (0.04 to 2.04)

0.18 (0.03 to 1.28)

Ever breast fed (n=5184)

0.79 (0.68 to 0.93)

0.70 (0.61 to 0.80)

0.75 (0.57 to 0.98)

0.61 (0.50 to 0.76)

Exclusively breast fed for:

*Odds ratios adjusted for level of parental education, maternal smoking during pregnancy, low birth weight, own bedroom, and frequent consumption of butter.

BMJ VOLUME 319

17 JULY 1999

www.bmj.com

149

Papers
Key messages
+ Obesity is the most frequent nutritional disorder in children, and is
an important risk factor for cardiovascular disease in adulthood
+ Preventing obesity in children should be a useful strategy in
preventing later heart disease because weight loss interventions in
obese children are costly and rarely successful
+ Data from a cross sectional study in Bavaria suggest that the risk of
obesity in children at the time of school entry can be reduced by
breast feeding: a 35% reduction occurs if children are breastfed for
3 to 5 months
+ Preventing obesity and its consequences may be an important
argument in the drive to encourage breast feeding in industrialised
countries

Contributors: RvK coordinated and designed the study, analysed the data, wrote the paper and is guarantor for the paper.
BK had the original idea for the study and wrote the nutritional
aspects of the discussion. TS contributed to the discussion of the
results. EvM designed the questionnaire for the study of atopy
and made important suggestions about the epidemiological and
statistical analyses and the writing of the paper. DB managed the
dataset on all children enrolled in the 1997 school entry health
examination. VG checked all statistical procedures and calculations in SAS, and is also a guarantor for the study. HvV initiated
the research project in collaboration with the public health
offices in Bavaria.
Funding: Bayrisches Staatsministrium fr Arbeit und Sozialordnung, Familie, Frauen und Gesundheit and Stiftung
Kindergesundheit.
Competing interests: None declared.
1

feeding for each child in both populations to explain the


dose effect related to the duration of breast feeding
which was observed both in the Canadian study and our
study. The protective effect of breast feeding therefore is
more likely to be related to the properties of breast milk
or the breastfeeding process than to lifestyle factors
associated with breast feeding.
Biological plausibility
In addition, it is plausible that breast feeding might
indeed have a programming effect in preventing obesity
or becoming overweight in later life. Lucas and
colleagues found significantly higher plasma concentrations of insulin in infants who had been bottle fed than
in infants who had been breast fed; these higher
concentrations would be expected to stimulate fat deposition and the early development of adipocytes.14 15
Breast milk also contains bioactive factors which may
modulate epidermal growth factor and tumour necrosis
factor , both of which are known to inhibit adipocyte
differentiation in vitro.16 17 The amount of energy
metabolised and the protein intake of breastfed children
is considerably lower than previously assumed and
significantly below the intake of infants who are fed formulas.18 19 In longitudinal studies a significant relation
was found between dietary protein intake at the age of
10 months and later body mass index and the
distribution of body fat2022; this suggests that a high
intake of protein in early childhood might increase the
risk of obesity later. Indeed, in animal studies the
availability of protein during fetal and postnatal
development was found to have long term effects on the
metabolic programming of glucose metabolism and
body composition in adult life.2325
Conclusion
Prolonged exclusive breast feeding reduced the risk of
being obese or overweight among school age children
in Bavaria who were born in the early 1990s. This effect
is more likely to be related to the composition of breast
milk than to lifestyle factors associated with breast
feeding. Preventing childhood obesity and its consequences may be an important argument in the drive to
encourage breast feeding in industrialised countries.
We thank all parents for answering the questionnaires, the doctors and their assistants in the public health offices for distributing and collecting the questionnaires, for encouraging the
parents to answer the questionnaires and for measuring and
recording the weight and height of all children attending the
school entrance health examination

150

3
4

5
6

7
8

9
10
11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

Freedman DS, Srinivasan SR, Valdez RA, Williamson DF, Berenson GS.
Secular increases in relative weight and adiposity among children over
two decades: the Bogalusa Heart Study. Pediatrics 1997;99:420-6.
Charney E, Goodman HC, McBride M, Lyon B, Pratt R. Childhood antecedents of adult obesity: do chubby infants become obese adults? N Engl
J Med 1976;295:6-9.
Stark O, Atkins E, Wolff OH, Douglas JW. Longitudinal study of obesity in
the National Survey of Health and Development. BMJ 1981;283:13-7.
Abraham S, Collins G, Nordsieck M. Relationship of childhood weight
status to morbidity in adults. Health Services and Mental Health Administration Health Report 1971;86:273-84.
Power C, Lake JK, Cole TJ. Measurement and long-term health risks of
child and adolescent fatness. Int J Obes Relat Metab Disord 1997;21:507-26.
Canadian Task Force on the Periodic Health Examination. Periodic
health examination, 1994 update. I. Obesity in childhood. Can Med Assoc
J 1994;150:871-9.
Wilkinson PW, Parkin JM, Pearlson J, Philips PR, Sykes P. Obesity in childhood: a community study in Newcastle upon Tyne. Lancet 1977;i:350-2.
Baranowski T, Bryan GT, Rassin DK, Harrison JA, Henske JC. Ethnicity,
infant-feeding practices, and childhood adiposity. J Dev Behav Pediatr
1990;11:234-9.
Poskitt EM, Cole TJ. Nature, nurture, and childhood overweight. BMJ
1978;i:603-5.
Kramer MS. Do breast-feeding and delayed introduction of solid foods
protect against subsequent obesity? J Pediatr 1981;98:883-7.
Von Ehrenstein O, von Mutius E, Illi S, Bhm O, Hachmeister A, von
Kries R. Reduced risk for atopic diseases in farmers children. Clin Exp
Allergy 1999 (in press.)
Rolland-Cachera MF, Cole TJ, Sempe M, Tichet J, Rossignol C, Charraud
A. Body mass index variations: centiles from birth to 87 years. Eur J Clin
Nutr 1991;45:13-21.
Eriksson JG, Forsen T, Tuomilehto J, Winter PD, Osmond C, Barker DJP.
Catch-up growth in childhood and death from coronary heart disease:
longitudinal study. BMJ 1999;318:427-31.
Lucas A, Sarson DL, Blackburn AM, Adrian TE, Aynsley-Green A, Bloom
SR. Breast vs bottle: endocrine responses are different with formula feeding. Lancet 1980;1:1267-9.
Lucas A, Boyes S, Bloom SR, Aynsley-Green A. Metabolic and endocrine
responses to a milk feed in six-day-old term infants: differences between
breast and cows milk formula feeding. Acta Paediatr Scand 1981;70:195200.
Hauner H, Rohrig K, Petruschke T. Effects of epidermal growth factor
(EGF), platelet-derived growth factor (PDGF) and fibroblast growth
factor (FGF) on human adipocyte development and function. Eur J Clin
Invest 1995;25:90-6.
Petruschke T, Rohrig K, Hauner H. Transforming growth factor beta
(TGF-beta) inhibits the differentiation of human adipocyte precursor
cells in primary culture. Int J Obes Relat Metab Disord 1994;18:532-6.
Whitehead RG. For how long is exclusive breast-feeding adequate to satisfy the dietary energy needs of the average young baby? Pediatr Res
1995;37:239-43.
Heinig MJ, Nommsen LA, Peerson JM, Lonnerdal B, Dewey KG. Energy
and protein intakes of breast-fed and formula-fed infants during the first
year of life and their association with growth velocity: the DARLING
Study. Am J Clin Nutr 1993;58:152-61.
Rolland-Cachera MF, Deheeger M, Akrout M, Bellisle F. Influence of
macronutrients on adiposity development: a follow up study of nutrition
and growth from 10 months to 8 years of age. Int J Obes Relat Metab Disord 1995;19:573-8.
Rolland-Cachera MF, Deheeger M, Bellisle F. Nutrient balance and
android body fat distribution: why not a role for protein? [letter;
comment]. Am J Clin Nutr 1996;64:663-4.
Deheeger M, Akrout M, Bellisle F, Rossignol C, Rolland-Cachera MF.
Individual patterns of food intake development in children: a 10 months
to 8 years of age follow-up study of nutrition and growth. Physiol Behav
1996;59:403-7.
Burns SP, Desai M, Cohen RD, Hales CN, Iles RA, Germain JP, et al. Gluconeogenesis, glucose handling, and structural changes in livers of the
adult offspring of rats partially deprived of protein during pregnancy and
lactation. J Clin Invest 1997;100:1768-74.
Desai M, Byrne CD, Zhang J, Petry CJ, Lucas A, Hales CN. Programming
of hepatic insulin-sensitive enzymes in offspring of rat dams fed a
protein-restricted diet. Am J Physiol 1997;272:G1083-90.
Desai M, Hales CN. Role of fetal and infant growth in programming
metabolism in later life. Biol Rev Camb Philos Soc 1997;72:329-48.

(Accepted 4 May 1999)

BMJ VOLUME 319

17 JULY 1999

www.bmj.com

You might also like