ARTICLES
16 Pear R. President to order drug makers to conduct pediatric studies.
New York Times, August 13, 1997: section A, 17.
17 Fontaine O. Antibiotics in the management of shigellosis in children:
what role for the quinolones? Rev Infect Dis 1989; 11 (suppl 5):
S1145–49.
18 Bennish ML, Wojtyniak BJ. Mortality due to shigellosis: community
and hospital data. Rev Infect Dis 1991; 13 (suppl 4): S245–51.
19 Bennish ML, Levy SB. Antimicrobial resistance of enteric pathogens.
In: Blaser JM, Smith PD, Ravdin JI, Greenberg HB, Guerrant RL,
eds. Infections of the gastrointestinal tract. New York: Raven Press,
1995: 1499–523.
20 Bennish ML, Salam MA, Khan WA, Khan AM. Therapy of shigellosis
III: comparison of one- or two-dose ciprofloxacin with standard 5-day
therapy: a randomized, blinded trial. Ann Intern Med 1992; 117:
727–34.
21 Salam MA, Seas C, Khan WA, Bennish ML. Treatment of shigellosis
IV: cefixime ineffective in shigellosis in adults. Ann Intern Med 1995;
123: 505–08.
22 Khan WA, Seas C, Dhar U, Salam MA, Bennish ML. Treatment of
shigellosis V: comparison of azithromycin and ciprofloxacin: a double-
blind, randomized, controlled trial. Ann Intern Med 1997; 126:
697–703.
23 Hewitt W, Vincent S. Theory and application of microbiologic assay.
San Diego: Academic Press, 1989.
Atopic eczema and domestic water hardness
N J McNally, H C Williams, D R Phillips, M Smallman-Raynor, S Lewis, A Venn, J Britton
Summary
Background The environment plays an important part in
the aetiology of atopic eczema, but specific causes are
unknown. Exposure to hard water is thought to be a risk
factor for eczema. We undertook an ecological study of
the relation between domestic water hardness and the
prevalence of eczema among Nottinghamshire
schoolchildren.
Methods Questionnaire details of 1-year period and lifetime
prevalence of eczema were obtained from parents of 4141
randomly selected primary-school children and 3499
secondary-school children in southern Nottinghamshire.
Geographical information systems (GIS) were used to link
the geographical distribution of eczema prevalence with
domestic water-hardness data (four categories).
Adjustment was made for potential confounding by sex ,
age, socioeconomic status, and access to health care.
Findings Among the primary-school children there was a
significant direct relation between both 1-year period and
lifetime prevalence of eczema and water hardness, both
before and after adjustment for confounders. The 1-year
period prevalence was 17·3% (261/1509) in the highest
water-hardness category and 12·0% (94/786) in the
lowest (adjusted odds ratio 1·54 [95% CI 1·19–1·99] p for
trend <0·001). The corresponding values for lifetime
Health Research Group, Department of Geography, University of
Nottingham, University Park, Nottingham NG7 2RD, UK
(N J McNally BA , Prof D R Phillips PhD, M Small-man-Raynor PhD);
Department of Dermatology, University Hospital, Queen’s Medical
Centre, Nottingham (H C Williams PhD); and Division of Respiratory
Medicine, City Hospital, Nottingham (S Lewis PhD, A Venn MSc ,
J Britton MD)
Correspondence to: Mr N J McNally
THE LANCET • Vol 352 • August 15, 1998
24 Krol GJ, Beck GW, Benham T. HPLC analysis of ciprofloxacin and
ciprofloxacin metabolites in body fluids. J Pharm Biomed Anal 1995;
14: 181–90.
25 Blackwelder WC. Proving the null hypothesis in clinical trials.
Control Clin Trials 1982; 3: 345–53.
26 Peltola H, Väärälä M, Renkonen OV, Neuvonen PJ. Pharmacokinetics
of single-dose oral ciprofloxacin in infants and small children.
Antimicrob Agents Chemother 1992; 36: 1086–90.
27 Rubio TT, Miles MV, Lettieri JT, Kuhn RJ, Echols RM, Church DA.
Pharmacokinetic disposition of sequential intravenous/oral
ciprofloxacin in pediatric cystic fibrosis patients with acute pulmonary
exacerbation. Pediatr Infect Dis J 1997; 16: 112–17.
28 Richard DA, Nousia-Arvanitakis S, Sollich V, Hampel BJ,
Sommerauer B, Schaad UB. Oral ciprofloxacin vs intravenous
ceftazidime plus tobramycin in pediatric cystic fibrosis patients:
comparison of anti-pseudomonas efficacy and assessment of safety
with ultrasonography and magnetic resonance imaging: Cystic Fibrosis
Study Group. Pediatr Infect Dis J 1997; 16: 572–78.
29 Ronsmans C, Bennish ML. Current practices for treatment of
dysentery in rural Bangladesh. Rev Infect Dis 1991; 13 (suppl 4):
S351–56.
30 Caudill TS, Johnson MS, Rich EC, McKinney WP. Physicians,
pharmaceutical sales representatives, and the cost of prescribing.
Arch Fam Med 1996; 5: 201–06.
prevalence were 25·4% (384/1509) and 21·2% (167/786;
adjusted odds ratio 1·28 [1·04–1·58], p for trend=0·02).
Eczema prevalence trends in the secondary-school
population were not significant (adjusted odds ratio for
highest compared with lowest hardness category for 1-year
prevalence 1·03 [0·79–1·33], p for trend=0·46; for lifetime
prevalence 0·99 [0·83–1·23], p for trend=0·93). Eczema
prevalence in primary-school children increased in relation
to chlorine content of water, but the trend across four
chlorine-content categories was not independently
significant after adjustment for confounders.
Interpretation Ex posure to hard water in the home may
increase the risk of eczema in children of primary-school
age.
Lancet 1998; 352: 527–31
Introduction
There is evidence to suggest that environmental and
lifestyle factors play an important part in the aetiology of
atopic eczema, in common with other atopic diseases.1–3
One factor that may be particularly relevant, and which
has not been investigated in epidemiological studies, is
the domestic water supply. Water hardness is believed to
be important in the development of eczema, and water-
softening units are recommended by manufacturers and
by some physicians4 for the management of eczema. The
effect of exposure to hard water and the independent
effects of the main causes of water hardness on the
occurrence of eczema have not, however, been
established. Chlorine in domestic water may be a trigger
factor for established atopic eczema,5 but this effect has
not been studied in the general population. We tested the
hypothesis that domestic hard water is associated with an
increased risk of atopic eczema in children, and we
527
ARTICLES
investigated the role of high chlorine content in the local
water supply as a potential additional risk factor.
Methods
Participants
Data on eczema occurrence were taken from surveys of
respiratory and other atopic disorders done between 1994 and
1996, in children aged 4–16 years attending primary schools
(ages 4–11) and secondary schools (ages 11–16) in a defined
postcode area of Nottinghamshire, which included the city of
Nottingham. A short questionnaire was distributed for
completion by parents of all primary-school children, and for
personal completion by all secondary-school children.6,7 This was
followed by a second more detailed questionnaire sent to the
parents of a one in four random sample of all children for whom
the initial questionnaires were returned. The second
questionnaire included questions on the lifetime occurrence and
Primary-school sample
N
Secondary-school sample
N water content, and adjustment for potential confounders age, sex,
Figure 1: Districts of southern Nottinghamshire and
geographical distribution of primary-school and secondary-
school study populations
Each point represents one child or more.
528
1-year-period prevalence of atopic eczema, defined as an itchy,
flexural skin rash. (“Has your child ever had an itchy skin rash
which has affected the skin creases, for example, in the folds of
the elbows or behind the knees, at some stage?”; “If yes, has this
rash been present in the past year?”). Similar questions were used
in the International Study of Asthma and Allergies in Childhood
(ISAAC) study, and have been validated previously.8–10 The
Nottingham City Hospital Ethics Committee approved the
surveys.
Design
We used the postcode of residence given on the second
questionnaire to allocate each child to an Ordnance Survey
National Grid reference, by use of the Central Postcode
Directory (CPD).11 Grid references were translated by a factor of
50 m for accurate allocation to water-supply zones.12,13 Water-
supply information for the study area for 1995 was obtained from
Trent Water Limited. The company gave mean values of four
water-hardness readings and between 12 and 120 chlorine-
content readings for each of 33 geographically distinct water-
quality zones. A boundary map of these zones was digitised
manually by use of Laser-Scan software and the Lites 2 vector-
mapping system. The data were imported into a geographical
information systems (GIS) package (ARC/INFO version 7.0.3).
Values for total water hardness, expressed as total concentration
(mg/L) of calcium and magnesium salts (the main causes of water
hardness), and chlorine content were linked to the digitised
boundaries of water-quality zones. We then grouped water-
quality zones into four categories, taking advantage of natural
breaks in the range of water-hardness values encountered in the
Nottingham area.14,15 Similar categories were constructed for
calcium, magnesium, and chlorine separately.
Data for enumeration districts from the 1991 census16 were
used to calculate values for the Townsend Index of
socioeconomic deprivation.17 Access to health care was also
estimated for each child, in terms of distance from the nearest
health centre, by use of the unit postcodes for each of the health
centres in the study area (obtained from the Nottingham Health
Authority and the North Nottinghamshire Health Authority). An
ARC/INFO map was created in the same way as for the postcode
of residence for the survey population, and used to estimate the
distance of each child to the nearest health centre in the study
area. Using point-in-polygon procedures within ARC/INFO, we
then linked the geographical distribution of the survey
populations to the water-supply information. Each water-quality
zone (a polygon) was superimposed on the complete set of grid-
referenced point data (points). The survey population within
each particular water-quality zone was then given the water-
supply characteristics of that zone. The same procedure gave
socioeconomic status values for the survey population.
Analysis
Data from the primary and secondary schools were collected in
different ways, and there was a much higher response rate to the
questionnaire given to the subsample of parents of the primary-
school children than to the other questionnaire. We therefore
analysed the primary-school and secondary-school datasets
separately. Estimates of the lifetime prevalence and 1-year period
prevalence of eczema were calculated for each of the water
hardness and chlorine categories, by use of the spatially matched
survey data. Comparisons of eczema prevalence in relation to
Townsend Index, and distance from the nearest health centre,
were done by multiple logistic-regression with the EGRET
statistical package (version 1.00). Significance in all multiple
logistic-regression analyses was defined as probability given for
the likelihood ratio statistic in EGRET, with one degree of
freedom for tests for trend, and with three degrees of freedom for
categorical comparisons between water-hardness categories. The
main hypothesis tested was the relation between total water
hardness and prevalence of eczema in primary-school and
secondary-school children in Nottinghamshire.
THE LANCET • Vol 352 • August 15, 1998
 ARTICLES

Soft Categor y 1 (118-135 mg/L)
Categor y 2 (151-157 mg/L)
Categor y 3 (172-214 mg/L)
Hard Categor y 4 (231-314 mg/L)
N highly significant both before and after adjustment for
Figure 2: Water-hardness categories in districts of southern
Nottinghamshire
Results
The first questionnaire was returned by 22 968 primary-
school parents and 27 826 secondary-school children
(response rates 83% and 86%, respectively). Second
questionnaires from parents in the subsample were
returned for 80% of primary-school children and for 58%
of secondary-school children. No data on eczema were
collected in the initial questionnaire, but there was no
evidence of response bias between water-hardness
categories in relation to questions about asthma. Of the
4643 responses to the second primary-school
questionnaire, 1079 (23·2%) reported eczema symptoms
at some stage in the child’s lifetime, and 692 (14·9%)
reported symptoms in the previous year. Of 3894 parents
of secondary-school respondents, 874 (22·4%) reported
eczema symptoms at some point during the child’s
lifetime, and 525 (13·5%) reported symptoms in the
previous year. At the georeferencing stage, roughly
10% of each dataset was excluded from further analysis
because the data fell outside the study area, because of
incomplete responses, or because of postcode
inaccuracies on the questionnaire or in the CPD. Thus,
4141 of the primary-school pupils surveyed and 3499 of
the secondary-school pupils surveyed were georeferenced
within the study area (figure 1). The categories of water
hardness were defined as: category 1 (118–135 mg/L
salts); category 2 (151–157 mg/L salts); category 3
(172–214 mg/L salts); and category 4 (231–314 mg/L
salts). The hard-water areas are in southern and western
parts of the study area, with soft water in the north and
east, but all four categories of water hardness are found in
Nottingham city, where most of our study population
lived (figure 2).
Among the primary-school population, there was an
increase in both lifetime and 1-year period prevalence of
eczema with increasing water hardness. This relation was
confounders. Unadjusted lifetime prevalence of eczema
was 4·2% higher for areas with the hardest water than for
areas with the softest water (crude odds ratio 1·27
[95% CI 1·02–1·56], p for trend=0·03), and the
unadjusted 1-year-period prevalence was 5·3% higher for
areas with the hardest water than for areas with the
softest water (1·54 [1·19–2·00], p for trend <0·001). The
absolute differences in unadjusted prevalence of eczema
between the category 1 and category 4 water-hardness
areas decreased with children’s age, from 6·9% among
5-year-olds to 2·2% at age 11. Adjustment for age, sex,
socioeconomic status, and distance from nearest health
centre had little effect on the size or significance of the
odds ratios for eczema in the four categories of water
hardness (adjusted odds ratio for lifetime prevalence of
eczema in category 4=1·28 [1·04–1·58], 1-year period
prevalence of eczema 1·54 [1·19–1·99]; table 1) or on
the significance of the linear trend through the four
categories of water hardness (p for trend=0·02, <0·001,
respectively). There was no correlation between water
hardness and Townsend Index. There was a significant
relation between 1-year period prevalence of eczema
and calcium content of water (adjusted odds ratio for
highest vs lowest calcium category 1·29 [1·00–1·66],
p for trend=0·01), but there was no significant trend in
the relation between calcium content and lifetime
prevalence of eczema (adjusted odds ratio for highest vs
lowest category 1·13 [0·91–1·39], p for trend=0·09).

Water- Number of Lifetime reported eczema 1-year reported eczema

hardness children
Number reporting Prevalence (%) Adjusted odds Number reporting Prevalence (%) Adjusted odds
category
eczema ratio* (95% CI) eczema ratio* (95% CI)
1 786 167 21·2 1·00 94 12·0 1·00
2 731 171 23·4 1·13 (0·88–1·45) 103 14·1 1·19 (0·88–1·62)
3 1115 253 22·7 1·16 (0·92–1·46) 163 14·6 1·32 (1·00–1·76)
4 1509 384 25·4 1·28 (1·04–1·58) 261 17·3 1·54 (1·19–1·99)
*Adjusted for age, sex, socioeconomic status, and distance from nearest health centre. Significance of trends in odds ratios: lifetime prevalence p=0·02; 1-year prevalence p<0·001.
Table 1: Water hardness and prevalence of eczema among primary-school children

Chlorine Number of Lifetime reported eczema 1-year reported eczema

content children
Number reporting Prevalence (%) Adjusted odds Number reporting Prevalence (%) Adjusted odds
category
eczema ratio* (95% CI) eczema ratio* (95% CI)
1 1296 288 22·2 1·00 181 14·0 1·00
2 1228 287 23·3 1·10 (0·91–1·32) 183 14·9 1·10 (0·88–1·38)
3 815 188 23·1 1·04 (0·84–1·28) 111 13·6 0·96 (0·74–1·25)
4 802 212 26·4 1·23 (1·00–1·52) 146 18·2 1·33 (1·04–1·70)
*Adjusted for age, sex, socioeconomic status, and distance from nearest health centre. Significance of trends in odds ratios: lifetime prevalence p=0·10; 1-year prevalence p=0·08.
Table 2: Chlorine content of water supply and prevalence of eczema among primary-school children

THE LANCET • Vol 352 • August 15, 1998 529

ARTICLES
Water- Number of Lifetime reported eczema
hardness children
Number reporting Prevalence (%)
category
eczema
1 807 184 22·8
2 670 147 21·9
3 795 177 22·2
4 1227 281 22·9
*Adjusted for age, sex, socioeconomic status, and distance from nearest health centre. Significance of trends in odds ratios: lifetime prevalence p=0·93; 1-year prevalence p=0·46.
Table 3: Water hardness and prevalence of eczema among secondary-school children
There was also no relation between either measure of
eczema prevalence and magnesium content (lifetime
prevalence, adjusted odds ratio 0·92 [0·73–1·15],
p for trend=0·79; 1-year period prevalence 1·03
[0·80–1·34], p for trend=0·99). An increase in the
prevalence of eczema with increasing chlorine content in
the domestic water supply was also shown for the
primary-school population, but these trends were not
significant after adjustment for potential confounders
(table 2).
Among secondary-school children the differences in
eczema prevalence between water-hardness categories
were much less apparent (table 3), and there was no
significant relation between any measure of eczema
prevalence and any water-content variables. Inclusion of
estimates of water hardness for each school had no
appreciable effect on these results. However, if data from
the primary-school and secondary-school populations
were combined, the relation between 1-year period
prevalence and water hardness remained significant
(adjusted odds ratio for highest vs lowest hardness
category 1·26 [1·05–1·51], p=0·02).
Discussion
Our ecological study suggests an association between
total water-hardness and the prevalence of atopic eczema
in primary-school children, which appeared attributable
to calcium rather than magnesium content of the water.
No significant association was shown between these
variables for the secondary-school population. A possible
association between atopic eczema in primary-school
children and the chlorine content of the water supply was
suggested by our unadjusted findings, but this association
was not independently significant when adjusted for
potential confounders. To our knowledge, this is the first
study to investigate the potential effects of the domestic
water-supply on eczema, although there has been
research into regional variation in the prevalence of
eczema in the UK, with the lowest prevalence in soft-
water areas such as Wales and Scotland.18 We studied this
relation in the Nottingham area because we had eczema-
prevalence data from a large, representative sample of
children from an area with a wide range of water hardness
typical of that of the UK as a whole,19,20 but with little
local change in water quality over time. Our findings are
therefore likely to be representative of the UK population
as a whole.
The study used data on water-supply characteristics
and socioeconomic data aggregated over quite large
geographical areas. Selection and information bias may
have affected the estimates of eczema prevalence, but are
unlikely to have affected the aetiological analysis. The
strongest associations between eczema and water
hardness were shown in the primary-school survey
population, in which the response rates for both stages of
the survey were high. Since we did not collect
530
1-year reported eczema
Adjusted odds Number reporting Prevalence (%) Adjusted odds
ratio* (95% CI) eczema ratio* (95% CI)
1·00 112 13·9 1·00
0·93 (0·72–1·20) 76 11·3 0·78 (0·57–1·07)
0·97 (0·76–1·24) 111 13·9 1·00 (0·75–1·33)
0·99 (0·80–1·23) 175 14·3 1·03 (0·79–1·33)
information on eczema in the first questionnaire we do
not know whether there was any response bias in the
second questionnaire in relation to eczema, although
there was no bias in relation to questions about asthma.
Misclassification of water supply and enumeration-
district areas at the georeferencing stage was minimised
by appropriate translation of grid references.12,13
The use of an area-based measure of socioeconomic
status may have limited our ability to remove
socioeconomic confounders, although the fact that there
was no evidence of correlation between water hardness
and Townsend Index score makes residual confounding
unlikely. Birth order and maternal atopy also affect the
prevalence of eczema, but these are unlikely to be
important confounders since they are not independently
related to water hardness. A further possibility is that the
observed effect of water hardness on eczema prevalence is
explained by another environmental factor with a spatial
distribution similar to that of water hardness in this study
area. Mean population density in our study area was
higher in the hard-water areas, and traffic density may
well also be greatest in these areas. However, most of the
participants live in Nottingham City, and although
exposure to most environmental factors is likely to be
homogeneous, their water supply is not homogeneous.
Moreover, the effect of water hardness on eczema
prevalence is stronger than the reported effects of traffic
pollution or any other spatially disparate suggested risk
factors for eczema. Nonetheless, we are currently
investigating the independent effects of road-vehicle
traffic flow on the prevalence of wheeze reported in the
first questionnaire in this study.
If the association between water hardness and eczema
is real, how can it be explained? Water hardness may act
more on existing eczema—exacerbating the disorder or
prolonging its duration—than as a cause of new cases.
The stronger association with recent than with lifetime
atopic eczema symptoms supports this argument.
Calcium and magnesium may act as direct chemical
irritants, or may modify the effects of other chemicals.
Skin irritation and dryness caused by exposure to such
chemicals may lead to mucocutaneous-barrier defects,
which in turn allow invasion of antigens.21 Alternatively,
the association may arise indirectly from the need for
more soap and shampoo to obtain a lather when washing
and bathing in hard water. Increased exposure to soaps or
their additives may have an irritant effect on the skin,
which could cause eczema in predisposed children; we
believe that a topical effect of hard water on the skin is a
more likely explanation than ingestion of increased
amounts of calcium or magnesium, since drinking water
contributes only a small part of the total dietary intake of
these cations. Other constitutents of tap water linked to
water hardness could also contribute to the occurrence of
eczema, through either a direct effect on the skin or an
interaction with other dietary components, but we have
THE LANCET • Vol 352 • August 15, 1998

not been able to study this possibility from the data
available to us. A generic effect of water hardness on the
prevalence of atopy seems unlikely, since there was no
evidence of a relation between the prevalence of self-
reported asthma or hayfever.
The lack of a relation between eczema prevalence and
water hardness among secondary-school children suggests
that the effect may be age-related, and the fact that
differences in eczema prevalence with water hardness were
greater among the younger children than older children
within the primary-school sample supports this conclusion.
The lack of association between water hardness and eczema
among secondary-school children may be explained by
exposure to a broad mix of water hardness through washing
and showering after physical recreation at school, or
because older children spend more time away from home
and are therefore exposed to a variety of hard water types.
However, inclusion of secondary-school location in the
analysis had no appreciable effect on the results, which
suggests that exposure to water at school is not important.
Greater total exposure to water in primary-school children,
through supervised washing by parents in early childhood,
might explain the stronger trends in prevalence among
primary-school children than in secondary-school children.
Adolescents’ skin may be less susceptible to the drying
effects of hard water. If chlorine in the local water supply
does affect eczema, it may be an irritant.22 The impaired
barrier function of the skin in atopic eczema is likely to
increase the effects of irritants such as chlorine.21 The role
of chlorine as an irritant has been discussed elsewhere.5
If our results are replicated in other localities, further
investigation of the effects of water hardness on eczema in
individual patients is also required. This ecological study
is an important starting point for further investigation of
the risk of atopic eczema from exposure to hard or soft
water supplies.
Contributors
Nick McNally did the GIS analyis and wrote the first draft of the paper.
Hywel Williams, David Phillips, and Matthew Smallman-Raynor
developed the original hypothesis, guided the analysis, and commented on
drafts of the paper. Sarah Lewis, Andrea Venn, and John Britton collected
the eczema data, checked the statistical analyses, and edited the paper.
Acknowledgments
We thank Trent Water Limited for the water-supply information;
Peter Housden (Director of Education), Julia Swan (Assistant Director)
and Tony Dessent (Senior Assistant Director) of Nottinghamshire County
Council for their support of the survey; the teaching and secretarial staff at
participating schools; and Marie Cooper, Jennifer Hill, Marilyn Antoniak,
Andrea Goldsmith, Chris Smith, and Nicola Williamson for help with
THE LANCET • Vol 352 • August 15, 1998
ARTICLES
data collection. The study was funded by the British Skin Foundation, the
National Asthma Campaign, and the Department of Health.
References
1 Williams HC. Atopic eczema: we should look to the environment.
BMJ 1995; 311: 1241–42.
2 Strachan DP. Time trends in asthma and allergy: ten questions, fewer
answers. Clin Exp Allergy 1995; 25: 791–794.
3 Taylor B, Wadsworth J, Wadsworth M, Peckham C. Changes in the
reported prevalence of childhood eczema since the 1939–45 war.
Lancet 1984; ii: 1255–57.
4 Atherton DJ. Eczema in childhood: the facts. Oxford: Oxford
University Press, 1994.
5 Morren M-A, Przybilla B, Bamelis M, Heykants B, Reynaers A,
Degreef H. Atopic dermatitis: triggering factors. J Am Acad Dermatol
1994; 31: 467–73.
6 Venn A, Lewis S, Cooper M, Hill J, Britton J. Questionnaire survey of
the effect of sex and age on the prevalence of wheeze and asthma in
adolescence. BMJ 1998; 316: 1945–46.
7 Venn A, Lewis S, Cooper M, Hill J, Britton J. Increasing prevalence of
wheeze and asthma in Nottingham primary schoolchildren:
1988–1995. Eur Respir J 1998; 11: 1324–28.
8 Williams HC, Burney PGJ, Pembroke AC, Hay RJ. Validation of the
UK diagnostic criteria for atopic dermatitis in a population setting.
Br J Dermatol 1996; 135: 12–17.
9 Asher MI, Keil U, Anderson HR, et al. International study of asthma
and allergies in childhood (ISAAC): rationale and methods. Eur Respir
J 1995; 8: 483–91.
10 Popescu CM, Popescu R, Williams HC, Forsea D. Community
validation of the UK diagnostic criteria for atopic dermatitis in
Romanian schoolchildren. Br J Dermatol 1998; 138: 436–42.
11 Office of Population Censuses and Surveys. OPCS Central Postcode
Directory, user guide. Fareham: Titchfield, 1985.
12 Gatrell AC. On the spatial representation and accuracy of address-
based data in the United Kingdom. Int J Geographical Information
Systems 1991; 3: 395–48.
13 Gatrell AC, Dunn CE, Boyle PJ. The relative utility of the Central
Postcode Directory and Pinpoint Address Code in applications of
geographical information systems. Environment and Planning A 1991;
23: 1447–58.
14 Gray NF. Drinking water quality: problems and solutions. Chichester:
John Wiley and Sons, 1994.
15 WHO guidelines for drinking water quality vol 1: recommendations,
2nd edn. Geneva: WHO, 1993.
16 OPCS. 1991 Census of population. London: HMSO, 1991.
17 Townsend P, Phillimore P, Beattie A. Health and deprivation:
inequality and the north. London: Croom Helm, 1988.
18 Golding J, Peters TJ. Eczema and hay fever. In: Butler NR, Golding J,
eds. From birth to five: a study of the health and behaviour of Britain’s
5-year-olds. Oxford: Pergamon Press, 1986: 171–86.
19 Giggs JA, Bourke JB, Katschinski B. The epidemiology of primary
acute pancreatitis in Greater Nottingham: 1969–1983. Soc Sci Med
1988; 26: 79–89.
20 Giggs JA, Ebdon DS, Bourke JB. The epidemiology of primary acute
pancreatitis in the Nottingham defined population area. Trans Inst Br
Geographers 1980; NS5: 229–42.
21 Ogawa H, Yoshiike T. A speculative view of atopic dermatitis: barrier
dysfunction in pathogenesis. J Dermatol Sci 1993; 5: 197–204.
22 Wahlgren CF. Pathophysiology of itching in urticaria and atopic
dermatitis. Allergy 1992; 47: 65–75.
531