By L. FEWTRELL. BSc. MSc. PhD*. D. KAY. BSc. PhD (Member)*. R. L. SALMON. MA. NIB. BS. MRCGP. MFPHM**.M.D. WYER BSc MSc. PhD (Member)***. G. NEWMAN. BSc. MSc (Member)****, and G. BOWERING, MIEH. MIH*****
IWEM. 1994, 8, February, p 97-101
Four studies were carried out at separate locations to investigate the relationship between health effects and low-contact water sports and intensive microbiological sampling was conducted in parallel to the health studies at each site. The two sports examined were marathon canoeing and rowing.
The extremes of water quality were at the estuarine sites on the River Torridge, where pollution levels varied from a geometric mean faecal coliform value of 62/100 ml at the App1edore/Instow site to 4613/l00 ml at the Bideford site.
A comparison of "exposed" and "unexposed" groups, 5-7 days after exposure, showed that the health effects of low-contact water sports are minimal, within the water quality ranges which were studied.
Key words : Epidemiology ; health ; recreation ; standards ; water quality.
The association between
water and disease has been recognized since ancient times and, until recently,
the majority of concern has (not surprisingly) been related to the quality of
drinking-water supplies. Recently, however, with the expansion in waterbased
recreational activity, concern has been shown over the microbiological quality
of recreational water : Some of the pathogens which have been isolated from
UK environmental water sources, along with the infections they can cause are
shown in Table I(2)
TABLE I : Pathogens
ISOLATED FROM UK ENVIRONMENTAL WATER
Pathogen........................ Infection
Aeromonas
spp................. Wound infection. gastro-enteritis
Campylobacter spp............. Enteritis
Cryptosporidium spp.......... Gastro-enteritis
Escherichia coli ............... Gastro-enteritis
Gardia ssp...................... Gastro-enteritis
Legionella ssp. .................Legionnaires disease, Pontiac fever
Leptospira spp. .................Leptospirosis
Pseudomonas spp. .............Otitis externa, follicular dermatitis
Salmonella spp................. Enteric fever, gastro-enteritis
ShigeIla opp.................... Bacillary dysentery
Trichobilharzia spp. .......... swimmers-itch
Vibrio spp...................... Cholera. gastro-enteritis. wound infection
Public concern
over marine bathing water has led to (a) improvements in marine water-quality
monitoring, (b) public information schemes through the provision of display boards
at beaches, and (c) the initiation of the national sea-bathing epidemiological
studies by the Department of the Environment (DoE). These studies, which took
place over the four-year period 1989-1992, will be reported in the near future.
The result of this recent research effort should be a credible dose-response relationship linking marine recreational water quality with any health effects identified in coastal swimmers. It cannot, however, provide information to define appropriate quality standards for other special-interest activities or types of water. Currently, marine water quality is assessed against standards set by the EC bathing water Directive (3), and these standards are shown in Table II.
TABLE II. EC BATHING WATER
DIRECTIVE MICROBIOLOGICAL STANDARDS
Parameter .........................Guide value ................Imperative value
Total
coliform/100 ml .............500 (80) ......................10 000 (95)
Faecal coliforms/100 ml.......... 100 (80).........................2 000 (95)
Salmonellae/l ..........................- ...................................0
faecal streptococci/ 100 .............(90)................................. -
Enteroviruses (pfu/10 l)............... - ..................................0
N B. Figures in brackets are the percentage of samples in which microbiological
counts must not be exceeded in order for the site to comply with the Directive
standards.
pfu = plaque-forming units
In 1991 the National
Rivers Authority (NRA) produced proposals for statutory water quality objectives(4).
The system is expected to consist of a set of "use"categories, each of which will
comprise a set of standards, and compliance will be assessed against these standards.
One of the proposed categories is "water contact activity". The DoE's response
to these proposals, in the form of a consultation paper, was made public in 1992
(5).
It has been recognized that the health risks associated with water quality may be different for different kinds of activity : therefore it has been suggested that a hierarchical classification may be appropriate for water-contact activities. Indeed, previous work has demonstrated that the risks of minor illness associated with a high-contact activity on polluted water can be considerable (6). Despite the imminent introduction of statutory water quality objectives and encouraging statementst (7), the competent authorities have yet to initiate research into the possible health risks to special-interest groups from the recreational use of controlled water, which could lead to scientifically defensible standards.
The present investigations were initiated to acquire data on the possible health effects caused by low water-contact activities.
Epidemiological cohort studies were performed at four sporting events over a period of 10 months : (i) the Banbury canoe marathon on the Oxford canal (10/11/91). (ii) the Galley canoe marathon on the Staffordshire and Worcester canal (l5/3/92). (iii) The App1edore/Instow rowing regatta on the River Torridge (26/7/92). and (iv) the Bideford rowing regatta on the River Torridge (5/9192).
A three-part questionnaire set was used to obtain data on health, recreational, social, dietary and demographic factors, and the 'unexposed' (spectator) and 'exposed" (recreator) groups were recruited from individuals present at each event. Individuals were approached by trained interviewers. and those who agreed to participate completed a one-to-one interview using the first questionnaire. The second structured interview was conducted, usually by te1ephone, 5-7 days after the exposure event. The final questionnaire was sent by post to all participants with a pre-paid return envelope, for completion between one week and four weeks after the initiaI exposure.
Intensive water-quality sampling took place at multiple locations throughout each sporting event, and samples were analysed by standard UK methods for a range of enteric indicator bacteria and enteroviruses.
Contingency tables (8) were used to examine differences in demographic and social factors, diet, general health and the symptoms reported by the 'exposed' and 'unexposed' groups, using the Yates-corrected chi-square test2. Where an expected cell count was less than five, Fisher's 2 x 1 tailed p3was used.
Results were deemed significant at alpha ² 0,05 4
For symptoms, relative risk (RR) values and 95% confidence intervals (CI) were calculated from 2 x contingency tables. The analysis for confounder was performed with the Mantel-Haenszel test. Differences in the microbiological quality between individual sites were compared using Student's t-test and comparisons between all four sites were made using Student-Newman-Keul's multiple-range test.
The results of the water-quality analyses are shown in Table III, and statistical differences are discussed below.
Fresh-water canal sites
Water quality was generally similar at the two canal sites (9) : however, faecal streptococci levels were significantly elevated (p<0.05) at Galley, while total staphylococci levels were higher at Banbury (p<0.01). Samples analysed for Sa/monella spp., Cryptosporidium spp. and enterovirus presented negative results.
Estuarine sites
Levels of the indicator bacteria faecal coliform and faecal streptococci were both significantly elevated (p<0.00l) at the Bideford site compared to the App1edore/Instow site. For levels of total staphylococci this pattern is reversed, with the App1edore/Instow site demonstrating significantly higher (p<0.001) levels than at Bideford.There was no significant difference between reported levels of Pseudomonas aeruginosa, Salmonellae and Cryptosporidium spp were not detected at either of the sites. Enterovirus was isolated from three samples taken from the Bideford site. the highest level reported being 8 plaque-forming units(pfu) /l0 l.
Bactérial
indicator---------- Geometric mean -------Log10 SD ------Min. -------
Max------------ N
Banbury
Faecal coliform ---------------------547 ---------------0.708----------- 33--------------
6000---------- 36
faecal streptocci ----------------------36--------------- 0.516 ------------1.67
------------320---------- 36
Total staphylococci -----------------30 ---------------0.326 -------------0
------------------95----------- 36
Ps. aeruginosa -----------------------2----------------0.379--------------
0 ------------------12----------- 36
Gailey
Faecal coliform --------------------675--------------- 0.643 ------------51------------
84000----------- 56
faecal streptocci--------------------- 61 ---------------0.418 -------------4
----------------374 -----------56
Total staphylococci ------------------7---------------- 0.610 -------------1------------------
97---------- 56
Ps. aeruginosa ---------------------- 2 ----------------0.129 --------------1------------------
12 ----------56
Appledore/Instow
Faecal coliform ---------------------62---------------- 0.458 ---------------7
-------------1090 ----------54
faecal streptocci --------------------13 ----------------0.172 --------------6
-----------------50 ----------54
Total staphylococci ----------------59---------------- 0.657 -------------10-------------
1230 ----------54
Ps. aeruginosa --------------------12 -----------------0.178-------------
10 ---------------160 ----------54
Bideford
Faecal coliform ------------------4613----------------- 0.355 -----------500
----------19212 ---------45
faecal streptocci -------------------528 ----------------0.254 -----------206------------
6279--------- 50
Total staphylococci----------------- 16 ----------------0.260 ------------10
--------------200--------- 50
Ps. aeruginosa ---------------------12 -----------------0.114 ------------10
---------------30 ----------50
SD= standard deviation
N= no.
of sample
Fresh-water canal sites
-----------------------------Freshwater
sites ---------------Estuarine sites
Category ------------Banbury --------Galey ---------Instow ----------Bideford
'Exposed'--------------137 -------------122 -----------129 ----------------131
'Unexposed'---------- 178------------- 154 -----------123---------------- 112
Total --------------------315 --------------276 -----------252---------------
243
% one week return ---94,3 -----------94,6 ----------86,9 ---------------95,1
At the Oxford canal
site, the 'exposed' group was more likely to be male (p<0.0001) and younger
than the 'unexposed' group (26 v. 37, p<0.001). There were no significant differences
in the reporting of individual symptoms in the three weeks prior to the study,
although the 'exposed' group was more likely to report 'any'symptom (p<0.05).
In the week following exposure, no significant differences in symptom reporting
were detected.
At Gailey, on the Staffordhire and Worcester canaI, the 'exposed' group was. again more likely to be male (p<0.001) and younger than the 'unexposed' group (27 v. 337, p<0,01). In both the three weeks prior to the study and the week following exposure, no significant differences were detected in the rates of reporting of either individual or grouped symptoms.
Statistical comparisons of the 'exposed' group at Banbury with that at Gailey showed no significant differences in the gender composition or the mean age of the groups, and there was no significant difference in symptom reporting between the two groups one week after exposure.
Both estuarine studies
took place on the River Torridge in North Devon, the first at Appledore/Instow
towards the mouth of the estuary (at the confluence of the River Taw) and the
second at Bideford. A total of 495 subjects was recruited from
these two sites (Table IV), of which 450 recruits (90.9%) completed the one-week
questionnaire.
At the Appledore/Instow site, the 'exposed' group was more likely to be male
(p<0.0001), younger (25 v. 34, p<0.001) and a day-tripper
or local (p<0.00l) than the 'unexposed' group. The 'unexposed' group had
a greater number of smokers (29/123 v. 12/129, p<0.001), and was more likely
to report the symptoms 'headache' (p<0.05) and 'sore throat' (p<0.001)
in the three weeks prior to the study. There were a number of differences in
dietary habits and alcohol consumption between the two groups. In the week following
exposure there was no significant difference in symptom reporting.
As with the previous site, the 'exposed' group at the Bideford regatta was significantly male biased (p<0.01) and younger (28 v. 32. p<0.05) than the 'unexposed' group, and there were also differences between dietary habits and alcohol consumption. There were no significant differences in symptom reporting between the two groups, either in the three weeks prior to exposure or in the week following exposure.
Recreators at Appledore/lnstow and Bideford were very similar in both mean age, gender composition ad pre-exposure symptom reporting. Bideford recreators were more likely to be hoIiday-makers than those from Appledore/Instow. There were also differences between the groups in dietary habits and alcohol intake. At one week, postexposure recreators at Bideford were more Iikely to report the symptom 'dry cough' (p<0.05, RR 3.17Cl : 1.07 9.34). FoIlowing stratified analysis controlling for hoIiday-maker status, the symptom 'dry cough' was no longer significant using the conservative chi square statistic (p=0.06). although the relative risk value was not much changed (RR 3.17, Cl : 1.01 9.95). Differences in dietary habits and alcohol consumption did not affect the results.
Although rowing and marathon
canoeing are essentially low water-contact activities, there is stilI some risk
of water ingestion through capsizing and splashing. At the freshwater sites
about 8 % of the canoeists reported capsizing, while approximately 16
% of rowers reported ingesting some water . At the estuarine sites 6
% of rowers reported ingestinig water.
Analysis comparing individuals from the 'exposed' groups (who reported ingesting
water with the combined non-ingesters) and the 'unexposed' groups from all four
sites, demonstrated that ingesters are significantly more likely to report a
number of grouped symptoms including 'gastro-intestinal' (p<0.01, RR 2.20,
CI : 1.35 3.58) and 'any' symptom (p<0,001, RR 1.75, CI : 1.32 2.32). The
results for the grouped symptoms are shown in Fig : 1 in the form of a relative
risk diagram. Statistical significance is ascertained by comparing the lower
95 % confidence interval to a value of l. If this value is greater
than unity, the relative risk value is significant at alpha = 0.05, and the
relative risk of an ingester reporting a symptom is significantly greater than
a non-ingester reporting a symptom. The results were unaffected by stratification
controlling for water type (i.e. fresh or estuarine water).
Fig. 1 Post-exposure relative
risk and 95% confidence intervals for water ingesters v. non-ingesters
Only one site achieved the imperative standards defined in the EC bathing water Directive (3). This site at Appledore/Instow, the more saline of the two estuarine sites, also achieved the guide level for fæcal streptococci. Interestingly, however, it showed the highest levels of total staphylococci and Ps. aeruginosa.
An examination of the symptom reporting between Banbury and Gailey, and also between the App1edore/Instow and Bideford sites, shows that there are very few differences, despite significant differences in water quality : therefore it is possible that the health effects from low-contact water activities are minimal, even in significantly polluted water.
Although both rowing and marathon canoeing are essentially low water-contact activities, a number of individuals report water ingestion.Within this subgroup of high water-contact, individuals show signilicantly elevated symptom reporting one week after exposure. This is in agreement with the results obtained from a similar study which examined white-water canoeing (6). In this instance significantly elevated levels of influenza, respiratory and gastro-intestinal symptoms were reported by the recreators one week after exposure, despite faecal coliform levels being significantly lower than those reported at Banbury, Gailey or Bideford.
l. The results of the present study, in conjunction with previous work, lend support for the hierarchical designation of water-use classes as suggested by the NRA.
2. The apparent lack of identifiable health effects in these studies suggests that it may be appropriate to use a relatively polluted water for low-contact recreational activities. However, it is imperative that the appropriate UK agencies further extend these studies to provide clear dose-response relationships linking fresh recreational water quality and health outcomes.
3. Such relationships now exist for marine waters, and they are urgently required to underpin the wholly appropriate policy of statutory water quality objectives which has been outlined by the NRA.
The authors are grateful to Bill Dodwell, Carol Parkes, Peter Reveley. BARC and BAAC, the organizers of the canoeing and rowing events. British Waterways provided funding to complete the freshwater canal events and the National Rivers Authority, South West Water and Torridge District Council co-funded the estuarine events. The questionnaire forms were adapted from a set originally designed by Dr R. Stanwell-Smith of the Communicable Disease Surveillance Centre, Colindale. Microbiological analyses were conducted by AItwell Ltd and supervised by Alan Godfree (Bacteriology) and Helen Merrett (Virology). Paula Hopkins and Jerome Whittingham were responsible for computer input and database management. Many interviewers have assisted, and the authors gratefully acknowledge their enthusiasm and dedication. Finally, the authors wish to thank the participants in all four studies, without whom none of this would have been possible.
(1) FEWTRELL
L. Freshwater recreation : a cause for concern ? Appl. Geog. , 1991,
11, 215-226.
(2) FEWTRELL L., GODFREE A., JONES F., KAY D., and MERRIT H.
Pathogenic Micro-organisms in temperature Environmental Water. Samara Press,
Dyfeld, 1993 (In press).
(3) COUNCIL OF EUROPEAN COMMUNITIES, Directive concerning the
quality of bathing water (76/160/EEC). Official Journal L31/1, 5 February 1976.
(4) NATIONAL RIVERS AUTHORITY, Recommandations for a scheme
of Water Quality Classification for Setting Water Quality Objectives, Water
Quality Report N° 5, National Rivers Authority, Bristol, UK, 1991.
(5) DEPARTMENT OF THE ENVIRONMENT, River Quality, The Government's
Proposals : A Consultation Paper. Department of the Environment, London, 1992.
(6) FEWTRELL L., GODFREE A., JONES F., KAY D., SALMON R. AND
WYER M. D. Health effects of white-water canoeing, Lancet, 1992, 339, 1587-1589.
(7) VAUGHAN J. Recreational Water quality standards. In
: Recreational Water Quality Management. Volume II : Fresh Waters, (D. Kay
and R. Hanbury Eds) Ellis Horwood, Chichester, 1993, Chapter 4.
(8) DEAN A.G., DEAN J.A., BURTON J.H. AND DICKER R. C. Epi
Info version 5 manual, US Department of Health and Human Services Public, Health
Service, Centers of Disease Control, Atlanta, Georgia, USA, 1990.
(9) FEWTRELL L, KAY D., NEWMAN G., SALMON R. AND WYER M. Results
of epidemiological pilot investigations. In : Recreational Water Quality
Management. Volume II : Fresh Waters (D. Kay and R. Hanbury Eds) Ellis Horwood,
Chichester, 1993, Chapter 6.
1*
Centre for Research into Environment and Health, Leeds Environment Centre, University
of Leeds.
** PHLS Communicable Disease Surveillance Centre, Welsh Unit, Cardiff.
*** Centre for Research into Environment and Health, Lampeter, Wales.
**** British Waterways, Gloucester.
***** Torridge District Council, Bideford.
2
A non-parametric measure of categorical association
3 The p level is the exact probability of a
given test statistic
4 alpha is the significance level at wi the
null hypothesis is rejected. Alpha=0,05 indicates that there is a 95% chance
that any interred relationships were not produced by chance.
