&EPA
United States
Environmental Protection
Agency
Health Effects Research
Laboratory
Cincinnati OH 45268
EPA 600 1 80-006
January 1980
Research and Development
Epidemiologic
Studies of Virus
Transmission in
Swimming Waters
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL HEALTH EFFECTS RE-
SEARCH series. This series describes projects and studies relating to the toler-
ances of man for unhealthful substances or conditions. This work is generally
assessed from a medical viewpoint, including physiological or psychological
studies. In addition to toxicology and other medical specialities, study areas in-
clude biomedical instrumentation and health research techniques utilizing ani-
mals but always with intended application to human health measures.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/1-80-006
January 1980
EPIDEMIOLOGIC STUDIES OF
VIRUS TRANSMISSION IN SWIMMING WATERS
by
Donn J. D'Alessio, Theodore E. Minor, Donald B. Nelson,
Catherine I. Allen and Anastasios A. Tsiatis
University of Wisconsin
Madison, Wisconsin 53706
Grant No. R-804161
Project Officers
Elmer W. Akin and Victor J. Cabelli
Field Studies Division
Health Effects Research Laboratory
Cincinnati, Ohio 45268
HEALTH EFFECTS RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Health Effects Research
Laboratory, U.S. Environmental Protection Agency, and approved for publica-
tion. Approval does not signify that the contents necessarily reflect the
views and policies of the U.S. Environmental Protection Agency, nor does
mention of trade names or commercial products constitute endorsement or
recommendation for use.
ii
-------
FOREWORD
The U. S. Environmental Protection Agency was created because of
increasing public and government concern about the dangers of pollution
to the health and welfare of the American people. Noxious air, foul
water, and spoiled land are tragic testimony to the deterioration of our
natural environment. The complexity of that environment and the inter-
play between its components require a. concentrated and integrated attack
on the problem.
Research and development is that necessary first step in problem
solution and it involves defining the problem, measuring its impact, and
searching for solutions. The primary mission of the Health Effects
Research Laboratory in Cincinnati (HERL) is to provide a sound health
effects data base in support of the regulatory activities of the EPA.
To this end, HERL conducts a research program to identify, characterize,
and quantitate harmful effects of pollutants that may result from ex-
posure to chemical, physical, or biological agents found in the environ-
ment. In addition to valuable health information generated by these
activities, new research techniques and methods are being developed that
contribute to a better understanding of human biochemical and physio-
logical functions, and how these functions are altered by low level
insults.
This report provides an evaluation of the transmission of viral
disease among swimmers at recreational areas. The results indicated
that children who swam at public freshwater beaches or swimming pools
were more likely to be shedders of virus in their stools and more likely
to experience gastroenteritis, presumably of viral etiology, than were
nonswimmers. Although not studied extensively, water quality data sug-
gested that the infectious organisms had not originated from sewage
contamination of the water but rather were transmitted from infected
swimmers to other swimmers by the water medium.
These findings tend to underscore the everpresent potential of
fecally contaminated water to be a transmitter of infectious disease and
indicate that control measures may be required to prevent waterbome
disease transmission in areas of heavy recreational use even when
measures to prevent point source sewage^contamination have been effec-
tive.
R. J. Garner
Director
Health Effects Research Laboratory
iii
-------
ABSTRACT
Retrospective and prospective epidemiologic studies were conducted
during the summers of 1976 and 1977 to determine if swimming activities
increase the risk of acquiring enteroviral infection in children. The
retrospective study consisted of a surveillance of recent swimming activ-
ities and clinical histories in 3,774 children who visited a pediatric
clinic. Viruses consisting mainly of non-polio enteroviruses were recovered
from 63 of 157 and 138 of 262 children with clinically apparent acute viral
infections who were sampled during 1976 and 1977, respectively. On the
other hand, non-polio enteroviruses were isolated from only 2 of 28 well
children sampled during the second year. The relationship of swimming
activities to enteroviral and other illnesses showed similar trends in both
years, but because of more extensive sampling and surveillance efforts, the
1977 data showed greater significant associations. Most notably, there was
a highly statistically significant Increased rate of swimming activity
among children who had enterovirus associated illnesses as compared to the
well controls.
Additionally, in 1977, age-specific swimming rates of all ill children
with enteroviral isolates and ill children who were not sampled were
markedly higher than those of well children. The most striking difference
was seen in the birth to 3 year old ill children with an enterovirus isolate
who had an age-specific swimming rate that was 2.6 times greater than that
of the well controls. No significant differences were seen in those who
visited exclusively swimming pools, but a highly significant difference was
seen in exclusive beach goers between those with enterovirus isolates
and well children. The frequency of visits to swimming sites appeared to
have no relationship to illnesses. Weekly age adjusted swimming rates
generally showed greater differences during the periods when enterovirus
activity was high.
The prospective study examined the relationship between swimming
activities and enteroviral infections in 296 elementary school children.
Sera were collected from the children at the beginning and end of the summer
for neutralizing antibody tests for selected enterovirus serotypes. -When
the data were analyzed by two-week periods, swimmers had a two to more than
three fold greater frequency of reported illnesses compared to nonswimmers.
The difference was particularly striking during periods when enterovirus
activity was high. These differences, however, were not significant as «
there was an insufficient number of volunteers. Swimming rates for the
entire season showed no relationships -to reported illnesses or rates of anti-
body conversions to selected Coxsackievirus 8 and echovirus serotypes. This
lack of a relationship appeared to be the result of a failure to find enough
iv
-------
children who were infrequent or nonswimmers. Nevertheless, the trend toward
a decreased illness rate in children who refrained from swimming for two
weeks is consistent with the retrospective study results.
To our knowledge, this is the first study that has found a statisti-
cally significant association between exposure to recreational waters and
an increased risk of enteroviral disease. Various internal consistencies
of the data discussed in this report support the validity of the association
and suggest that water served as the transmission medium.
This report was submitted in fulfillment of grant number R-804161 by
the Enterovirus Research Laboratory, Department of Preventive Medicine,
University of Wisconsin at Madison under the sponsorship of the U.S.
Environmental Protection Agency. This report covers a period from February
1, 1976 to January 31, 1979, and work was completed as of April 30, 1979.
-------
CONTENTS
Disclaimer ii
Abstract iv
Tables ' vii
Acknowledgment . x
1. Introduction 1
2. Conclusions and Recommendations 2
3. Materials and Methods 4
Clinic Study 4
School Study 6
Statistical Analyses 5
4. Experimental Procedures g
Clinic Study 8
School Study 9
5. Results and Discussion X2
Clinic Study 12
School Study 3^
References 55
Appendices
A. Clinic study questionnaire 57
B. School study postcard questionnaire .... 53
C. School study swimmer activity profile .... 59
vii
-------
TABLES
Number Page
1 Virus isolated from children in Clinics A and B
(1976) and Clinic B (1977) 13
2 Categories of the Clinic B patient populations,
1976 and 1977 15
3 Age composition of the various diagnostic categories,
Clinic B, 1976 and 1977 17
4 Isolation rates among age groups at Clinic B, 1976
and 1977 18
5 Age specific distribution of virus isolates among ill
patients at Clinic B, 1976 and 1977 18
6 Syndromes of Clinic B sampled patients, 1976 19
7 Syndromes of Clinic B sampled patients and nonsampled
viral-like illness patients, 1977 20
8 Viral isolations by week, Clinics A and B (1976) and
Clinic B (1977) 23
9 Number of patients with nonpolio enterovirus positive
specimens during the summers of 1976 and 1977, State
Laboratory of Hygiene Virus Diagnostic Section ...» 24
10 Weekly age adjusted rates of swimming in various
diagnostic categories, Clinic B 1976 26
viii
-------
Number Page
11 Weekly age adjusted rates of swimming in various
diagnostic categories, Clinic B 1977 27
12 Age specific swimming rates by diagnostic category,
Clinic B, 1976 and 1977 28
13 Results of statistical anaylsis of swimming in various
diagnostic categories, Clinic B 1976 30
14 Results of statistical analysis of swimming in various
diagnostic categories, Clinic B 1977 32
15 Distribution of school study population of 218 children
responding to the activity questionnaire by average
season swimming score and the corresponding activity
score 1976 to 1977 33
16 The nature of swimming activity by 218 school study
children responding to the activity profile questionnaire
1976 to 1977 34
17 Frequency of illnesses by two-week periods in swimmers
and nonswimmers 1976 to 1977 35
18 The relationship of swimming frequency and illness
frequency. School study 1976 to 1977 37
19 Average frequency of illness per two-week period
associated with different types of swimming. School
study 1976 to 1977 38
20 Average frequency of illness per two-week period in
different age groups and in males and females. School
study 1976 to 1977 39
ix
-------
Number ' Page
21 Season frequency of illness and serologic conversions
to group B Coxsackieviruses and echoviruses. School
study 1976 to 1977 40
22 Ranking of beaches by total season attendance 41
23 Fecal coliform densities at Madison beaches,
6/7/76 to 9/6/76 43
24 Enterococci densities at Madison beaches,
6/7/76 to 9/6/76 44
25 Fecal coliform densities at Madison beaches,
6/13/77 to 8/29/77 45
26 Enterococci densities at Madison beaches,
6/13/77 to 8/29/77 » 46
-------
ACKNOWLEDGEMENTS
The cooperation of the Pediatric Departments at Dean Clinic and
Quisling Clinic, Madison, WI, is gratefully appreciated. We are
particularly grateful to the pediatricians, nurses, technicians and
clerical staff for their active participation during the three years of
the project.
The cooperation of Madison Public Schools is also gratefully
acknowledged. The interest and assistance of the school district's
External Research Committee, principals and teachers provided for
efficient recruitment and sampling of volunteers.
xi
-------
SECTION 1
INTRODUCTION
Virtually any virus that infects humans could be present in recrea-
tional waters and be transmitted to swimmers, but relatively few virus
groups appear to be good candidates for this mode of transmission. The
enteroviruses, which are the predominant viruses encountered during the
summer in temperate climates, are prime candidates for swimmer-acquired
waterborne infections. Certain enterovirus serotypes have been detected in
swimming pools (1, 2, 3) and in swimming beach waters (4,5) and concurrently
isolated from children who swam in these waters. This does not, however
indicate that water transmission occurred; e.g. virus may have passed from
infected individuals to the water but may not have been transmitted from
water to the swimmers.
Several investigations of outbreaks of pharyngoconjunctival fever,
caused by adenoviruses, have revealed common exposure to a swimming pool or
small lake (6, 7, 8, 9). Although most attempts to isolate virus from the
water were made long after exposure and therefore were unsuccessful, adeno-
virus was isolated from a water sample collected near a sewage outlet in a
lake where persons who experienced, pharyngoconjunctival fever had been
swimming (7).
A series of three field studies explored the question of potential
influence of bacteriologic quality of different bathing waters on the
frequency of reported swimmer illnesses (10). An appreciably higher
incidence of overall illness was observed in swimmers compared to non-
swimmers regardless of the bacterial quality of the water. Those who swam
in a river that had high coliformdensities had a significantly greater
incidence of gastrointestinal illnesses. More recently, Cabelli et al. (11)
also observed a significantly higher incidence of gastrointestinal symptoms
in swimmers compared to nonswimmers at a polluted New York City beach.
We conducted epidemiologic studies during two consecutive summers to
determine if swimming activities increase the risk of acquiring enteroviral
infection in children. These studies included both a retrospective and
prospective approach to the question. Children who visited a pediatric
clinic were studied retrospectively to compare recent swimming activities of
ill children with the activities of those who were well and to seek a viral
etiology of a selected sample of illnesses. Prospectively, elementary
school children were surveyed longitudinally to compare the rates of
illnesses and serologic conversions to certain enterovirus serotypes in
swimmers and nonswimmers.
-------
SECTION 2
CONCLUSIONS AND RECOMMENDATIONS
This Investigation, using both cohort and case-control epidemiologic
study designs, was meant to test the hypothesis that recreational water that
is not subject to sewage plant effluent serves as a transmission route for
viruses, in particular, the enteroviruses.
On the basis of the cohort study data, we think the following
conclusions are valid:
1. Children between 6 and 10 years of age who visited swimming
sites consistently reported more illnesses of apparent viral
etiology, both respiratory and gastrointestinal, than those
who refrained from these activities. The peak illness
frequency in swimmers coincided temporally with peak entero-
virus activity.
2. No association was apparent between the frequency of visits
to swimming sites or location of lake, beach or pool, and the
frequency of acute infectious illness.
The case-control study results permit somewhat more particular
conclusions:
1. A statistically significant association between swimming
activities and enteroviral illness was demonstrated. This
association held for children from birth to 15 years of age.
2. Features of the data suggest that water served as the entero-
virus transmission medium, although this study does not
provide direct evidence on this point.
3. The results also suggest, although less strongly, that the risk
of acquiring enteroviral illnesses may be greater at beaches
than at swimming pools.
4. As in the cohort study-, there was no apparent association
between the frequency of visits to swimming sites and illness.
This study was initiated with the realization that there was no
convincing evidence that recreational water plays a role in enterovirus
transmission. Our investigation was,, in part, meant to examine the
-------
feasibility of approaching this question by epidemiologic methods. We feel
that our findings constitute the strongest evidence currently available for
water transmission of the enteroviruses. They are not conclusive or
detailed enough to serve as the basis for specific control measures. This
step will require further investigation which appears to us to be fully
justified on the strength of our findings to date.
To be profitable, further study of this question must involve a more
detailed-definition of the shedding of virus by infected individuals,
chemical, bacteriologic and virologic characterization of specific swimming
sites, and specific etiologic investigation of illnesses in swimmers using
these sites.
1. Quantitation of virus shedding from infected Individuals:
The data reported here suggest that swimmers may serve as
the source of water contamination. Enteroviruses proliferate
and are shed from both the nasopharynx and gastrointestinal
tract. Quantitation of the extent of virus shedding by
infected children would be necessary to assess the likelihood
that water users are the source of contamination.
2. Definition of the character of the recreational water:
Additional study of this question would seem to require that
the bodies of water under examination be characterized in terms
of particulate and chemical composition. In swimming pools,
free chlorine levels would have to be carefully defined.
This kind of characterization would supplement careful
bacteriologic monitoring. Most importantly, attempts should
be made to recover and quantitate viruses from the recrea-
tional water.
3. An experimental epidemiologic study design: Assuming that
there is an association between swimming and an increased
risk of acquiring an enteroviral illness, one of the most
difficult questions to decide is whether actual water trans-
mission occurs or transmission is really via person-to-person
contact at the swimming site. Our current study design did
not allow .direct examination of this question, although we
feel there is evidence to suggest actual water transmission.
Decisive study of this question is very difficult with any
observational design. The most discerning approach to this
question would be a study in which participants agreed to a
random assignment to swimming or nonswimming categories.
Achieving such agreement would be difficult but would make
possible a crossover study and would be a very effective way
to clarify the mode of transmission.
In summary, the data we will present suggest that swimming plays an
Important role in enterovirus disease. This is of sufficient public health
importance to justify further study. If.enterovirus water transmission is
established, effective control measures might well be feasible.
-------
SECTION 3
MATERIALS AND METHODS
CLINIC STUDY
Virology Samples
Dual specimens consisting of throat and rectal swabs were collected
for isolation of viruses. Immediately after collection, the cotton swabs
were immersed in 2.5 ml of Hanks' Balanced Salt Solution plus 0.25% gelatin,
penicillin G (200 Units/ml), and streptomycin (100 Ug/ml) and wrung out.
The specimens were then refrigerated until the following morning when cell
cultures were inoculated. A 10 cc blood sample was collected by veni-
puncture and allowed to clot and express serum.
Isolation of Viruses
Two tubes each of Wlsl (a human embryonic lung diploid strain developed
by the Wisconsin State Laboratory of Hygiene Virus Section) and HEp-2, and
three tubes of primary rhesus monkey kidney (RMK) cell cultures were
inoculated per specimen. One tube of each cell type received 0.2 ml of
specimen while the others received 0.1 ml. The medium was changed after
2 hr and again after one week. Tubes were placed in a stationary position,
incubated at 37 C, and examined microscopically for cytopathic effect at
several intervals during a two-week period. One tube of BMK containing a
throat culture was tested for hemadsorption with 0.25 ml of 0.25 percent
guinea pig erythrocytes after one and two weeks of incubation. Suspected
viral isolates were passed in appropriate cell cultures.
Each specimen that did not yield a viral isolate in cell culture was
tested for the presence of group A Coxsackieviruses by inoculating six
albino mice (HA/ICR) that were less than 24 hr old with 0.05 ml subcutan-
eously and 0.02 ml intracerebrally. Mice were observed daily for two weeks
for development of paralysis.
Identification of Viruses
Viruses were identified with commercially obtained animal hyperimmune
antisera (Microbiological Associates, Bethesda MD). Bnterovirus-like agents
were tested for neutralization (tube method) by type-specific Coxsackie-
virus B 1 to 5 and poliovirus 1 to 3 antisera and by intersecting pools
containing echovirus and Coxsackievirus A9 antisera. Nine pools, each
containing 6 to 8 monotypic antisera, were formulated in our laboratory in a
-------
manner similar to the Benyesh-Melnick scheme. The isolates were diluted to
contain 32 to 320 TCID50/0.1 ml and were tested with 20 units of antiserum.
Agents not identified with any of these antisera were tested for stability
to ether and acidity (pH 3) and declared to be "enterovirus-like agents"
if they passed both tests.
Isolates from mice that developed a flaccid paralysis were declared to
be group A Coxsackieviruses and were not identified further. Isolates
obtained from throat specimens that were ether stable but acid labile were
assumed to be rhinoviruses.
Adenoviruses were identified by neutralization tests using type-
specific antisera (32 to 320 TCID50 of virus and 20 units of antiserum).
Herpes simplex virus was identified by the Wisconsin State Laboratory of
Hygiene Virus Section using a direct fluorescent antibody test. Hemad-
sorbing agents were tested for hemadsorption-inhibition with parainfluenza
1 to 3 antisera but were not tested further. Agents not identified by the
above procedures were tested for the presence of mycoplasma by the Wiscon-
sin State Laboratory of Hygiene Immunology Section and, if negative, were
declared to be "unidentified virus-like agents."
Serology
i
Available paired acute and convalescent sera from patients with isolates
were tested for serum neutralizing antibody rises to infecting serotypes.
In addition, paired sera from individuals without isolates vere tested for
neutralizing antibody rises to Coxsackieviruses B 1 to 5. The tube method
(12) was employed for all virus groups except the Coxsackievirus B group
which was used in a microtiter test that will be described below.
Statistical Analysisof Swimming vs. Infection
A log-linear model for the analysis of categorical data was used to
analyze swimming versus infection data. The data were categorized by four
variables.
-well
(i) Illness'
-ill
^ swim in the previous two weeks
(li) Swimming/^
^did swim (including beach or pool swimming)
-------
(iii)
^ female
to 3 years old
(iv)
years old
than 15 years old
Since a retrospective design was used,' the natural index in the
analysis is the odds ratio of illness and swimming. That is
e-
where e denotes odds ratio
PI denotes the probability of swimming among ill children
pw denotes the probability of swimming among well children
e=l implies no relationship between swimming and illness
e>l implies a positive relationship between swimming and illness
e
-------
containing 32 to 320 TCIDso of virus was added to each well. A single row
of serum dilutions containing no virus was prepared for each serum. A
virus back titration (100-0.01 TCID^Q/SO yl) was done in quadruplicate.
Four wells containing only growth medium were included in each plate. The
plates were gently agitated, then incubated at 37 C for 1 hr in a 5 percent
C02 atmosphere. A suitable .suspension of LLC-MK2 cells (one flask of cells
per five microtiter plates) was prepared in growth medium and 0.15 ml added
to each well. The plates were gently agitated, re-incubated for six days,
and examined microscopically for cytopathic effect.
-------
SECTION 4
EXPERIMENTAL PROCEDURES
CLINIC STUDY
Overall Study Design
All children who visited two private pediatric clinics in Madison,
Wisconsin were surveyed for histories of recent swimming activities, reasons
for visiting the clinic, and current clinical syndromes. Children with
clinically apparent acute viral infections and some well children were
sampled for virus isolation and serology.
Study Area
Madison (population: 175,000) is the site of the state capital and the
major campus of the University of Wisconsin System. Since the city is
nearly devoid of heavy industry and is not located near other cities, a
substantial portion of the population is comprised of state and university
employees. Residents, therefore, enjoy a stable and above average standard
of living.
Five lakes are located in Madison and the immediate surrounding area
and three of these are partially or wholly within the city limits. The city
alone has 14 municipally supervised swimming beaches in addition to numerous
public and private swimming pools. Discharge of raw or treated human sewage
into area lakes is prohibited, but run-off waters from farm land and storm '
sewer discharge reach the lakes and constitute the principal threat to
water quality. The municipal drinking water supply is drawn exclusively
from deep wells.
Study Periods^
Children were sampled from June 7, 1976 to September 9, 1976 and from
June 13, 1977 to September 1, 1977. The 1976 study began at just one clinic
(Clinic A). Because sampling at this clinic was quite sporadic, a second
clinic (Clinic B) joined the study on August 2, 1976. During 1977,
sampling was conducted exclusively at Clinic B.
Clinic Questionnaires
All children visiting the two clinics in 1976 were given questionnaires
that they or their parents were to complete if they were willing to cooper-
ate. In the second year, questionnaires were distributed only to patients
-------
of the two pediatricians directly participating in the study. The
surveillance was conducted by clinic staff nurses in the first year of the
study, but during the second year this responsibility was assumed by a
senior nursing student employed for the project. This resulted in more
uniform and complete clinical and epidemiclogic data collection.
One purpose of the questionnaire (Appendix A) was to determine if
swimming sites were visited during the two weeks immediately preceding the
clinic visit, where the activity occurred, and its frequency. For our
purposes, swimming activity was defined as the act of going to a swimming
site. The definition did not distinguish between the types of activity
that the children participated in at the swimming area.
The other major purpose of the questionnaire was to identify those
children visiting the clinic for reasons not related to infectious disease
(e.g. physical examinations, injuries, immunizations, atopic allergies,
chronic health problems, etc.) and distinguish them from children who had
apparent acute infectious illnesses. The latter included illnesses with
possible viral, bacterial, mycoplasmal, and other infectious etiologies.
A total of 3,962 children received questionnaires in 1976, including
1,497 at Clinic A and 2,465 at Clinic B. Questionnaires were distributed
to 1,561 children in the following year. The smaller number in 1977
resulted because all patients at Clinic B which includes seven pediatricians
were asked -to complete questionnaires in 1976, whereas in 1977 question-
naires were distributed only to patients of the two cooperating physicians.
Sampling
Nearly all of the children who presented with an apparent viral illness
were interviewed by the nursing student to obtain further clinical data.
The clinic pediatricians were responsible for judging whether a child had
signs and symptoms that would justify viral isolation attempts. Children
with syndromes that were suggestive pf non-enteroviral exanthems or
bacterial/mycoplasmal infections were excluded from this group. Specimens
were collected only if informed consent was provided.
A total of 157 patients, including 80 at Clinic A and 77 at Clinic B,
were sampled in the first year. Well children were not sampled during
this season. In 1977, 262 ill and 28 well children were sampled.
Throat specimens for viral isolation were collected from all of the 447
sampled children. Stool specimens were collected from all but 14 of these
individuals. Faired acute and convalescent sera were obtained from 53 of
the 157 patients (34 percent) in 1976 and 46 of the 290 patients (16 percent)
in 1977.
SCHOOL STUDY
Experimental Design
Sera were collected from elementary school children in late May and
9
-------
early September of 1976 and 1977. During the intervening periods, the
children or their parents kept a record of swimming activities and ill-
nesses. The purpose of the study was to determine whether children who went
swimming more frequently during the summer had more illnesses or serological
conversions to selected enteroviruses than children who went swimming less
frequently or not all.
Recruitment of Volunteers
Children in grades 1 through 5 (ages 6 through 10) were recruited from
Madison Public Schools, Madison, Wisconsin. The characteristics of the
study area have been described above. Three schools, each in a different
geographical area of the city, were approached during the first year,
whereas nine schools located throughout the city were involved in the second
year. During the first and second summers, 149 and 147 children were
recruited, respectively. A number of children participated in the study
during both summers.
Surveillance
Surveillance of swinaning activities and illnesses began just after the
end of the school year (June 13) and concluded just before the beginning of
the next school year (August 21). The 10 weeks of surveillance were
divided into five equal periods.
Parents completed and returned post card questionnaires at the end of
each two-week period (Appendix B). Reminders were sent whenever question-
naires were not returned. The questionnaire supplied information about
whether the child went swimming, the location and frequency of swimming,
and whether he or she had been ill during the two-week period.
Swimming again was defined as the act of going to a swimming site.
Swimming frequency was listed by category: 0, 1 to 3, 4 to 6, 7 to 9, and
greater than 9 days. These categories were assigned scores of 0 to 4.
When an illness was reported, the project nurse or physician telephoned
the parents to obtain a clinical history. Illnesses were defined as
acute episodes of at least 24 hr duration, consisting of signs and symptoms
compatible with a viral etiology (excluding non-enteroviral exanthems).
Each episode of illness was recorded only once and was assigned to the
period corresponding with its onset.
After the conclusion of the study, questionnaires were distributed to
obtain an approximation of the summer activities engaged in by volunteers
(Appendix C). One purpose of this questionnaire was to determine if
children with low average swimming scores tended to avoid group activities,
had little sibling contact, or came from small families. These situations
would greatly reduce opportunities for person-to-person transmission of
viruses and might lead to a false conclusion that the lower infection rate
was related to a low swimming exposure. Responses to the questionnaire
were assigned scores that were weighted to reflect the relative Importance
10
-------
of each factor in promoting person-to-person transmission (Appendix C). A
second purpose of the questionnaire was to determine the types of activity
that occurred at the swimming site.
BACTERIOLOGIC MONITORING OF SWIMMING WATERS
The City of Madison Public Health Department provided weekly reports of
fecal coliform and enterococci counts at municipal beaches during the 1976
and 1977 study periods. Fecal coliform densities were determined by the
membrane filter technique.
11
-------
SECTION 5
RESULTS AND DISCUSSION
CLINIC STUDY
Viruses Isolated
The case control study of the relationship of swimming to enterovirus
illnesses was carried out in two clinics, A and B, in 1976 but in only
Clinic B In 1977. Viruses were recovered from 63 of 157 children (40
percent) sampled In 1976. Fifty of these isolates (32 percent) were non-
polio enteroviruses. In 1977, 262 ill children were sampled and 138
viral isolations were made (53 percent). Of these, 120 (46 percent) shed
non-polio enteroviruses. During 1977, 28 well children were sampled and
nine shed viruses. However, only two (6 percent) of these isolates were
non-polio enteroviruses while the other seven were polioviruses assumed to
be vaccine strains (TABLE 1).
Group A Coxsackieviruses predominated in both years, but the other
non-polio enteroviruses recovered differed markedly between the two
seasons (TABLE 1). In 1976, among the Group B Coxsackieviruses, only types
3 and 4 were active whereas all five serotypes were isolated in 1977. The
major difference between the two years was in the number of echoviruses
isolated. There were four echovirus isolates in 1976, whereas in 1977, 40
isolates of six different serotypes were found. Isolation of agents other
than enteroviruses totaled just 31 for both years and the mixture of
adenoviruses, rhinoviruses, paramyxoviruses, and herpes simplex viruses
was not surprising in the population sampled.
Serologic Confirmations
Paired acute and convalescent sera were available from 21 of the
children in whom viruses were isolated, including 18 with enteroviral
isolates. Significant titer rises (4-fold or greater) occurred in 13
(62 percent) of these 21 patients. In a number of children who did not have
titer rises but had enteroviral isolates, the initial serum showed a fairly
high titer which may account for our inability to demonstrate a greater
percentage of significant rises. Acute and convalescent serum specimens
from 25 children who had no isolates were available. These 25 serum pairs
(all from children sampled in 1977) yielded no significant titer rises to
any of the five Coxsackievirus B serotypes.
12
-------
TABLE 1. VIRUSES ISOLATED FROM CHILDREN IN CLINICS A AND B (1976)
AND CLINIC B (1977)
Number of children with isolates*
1976 1977
Viruses
Coxsackievirus A
Mouse isolates
Type 9
Total
Coxsackievirus B
Type 1
2
3
4
5
Total
Echovirus
Type 3
5
6
7
11
18
22
30
Total
Poliovirus
Type 1
3
Mixture
Total
Enterovirus-like Agents,
untypeable
Adenovirus
Type 2
3
5
Total
111 eroup
32
2
34
0
0
3
7
0
10
0
0
0
0
0
1
0
3
4
1
0
0
1
2
2
0
0
2
(continued)
42
3
45
10
2
10
1
6
29
12
1
6
2
16
0
3
0
40
0
1
1
2
6
2
6
2
10
Well eroup
0
0
0
0
0
0
0
1
1
0
0
0
0
1
0
0
0
1
6
0
1
7
0
0
0
0
0
13
-------
TABLE 1 (continued)
Viruses
Rhinoviruses, nontyped
Myxovirus
Parainfluenza Type 3
Unidentified
Total
Herpes simplex virus
Number of children with isolates*
1976 1977
111 group
1
1
2
0
0
0
0
Well group
0
0
0
Type 1
2
Total
Unidentified virus-like agents
Total non-polio enteroviruses
Total isolates
0
0
0
1
50
63
1
1
2
4
120
138
0
0
0
0
2
9
*Number of children sampled:
well group, 28.
1976, 157; 1977 ill group, 262; 1977
Characteristics of the Study Population
Although the Clinic Study was begun in June 1976 in Clinic A, the
sampling in that clinic was so variable and incomplete that the data
derived do not bear analysis. Secondly, we excluded from analysis all
children 16 years of age or older because they were few in number and,
with a spread between 16 and the early 20's, differed significantly from
the bulk of our sample. Furthermore, very few virus isolations were made
in this group. Our study population then consisted of all children
between birth and 15 years of age who were seen at Clinic B in 1976 and
1977 and agreed to participate in the study. The 1976 study population,
which completed questionnaires at Clinic B from the first day of August
through the subsequent six weeks, consisted of 2,402 children. The 1977
population consisted of 1,372 children who completed similar question-
naires at Clinic B (TABLE 2).
In both years, well children who appeared at the clinic for physical
examinations or reasons other than illness comprised just under' 50 percent
of the study population (TABLE 2). Similarly about 22 percent of children
in 1976 and 27 percent of children in 1977 had illnesses which appeared
14
-------
TABLE 2. CATEGORIES OF THE CLINIC B PATIENT POPULATIONS, 1976 AND 1977
1976 Patients1977 Patients
Patient category No. Percent No. Percent
Well children 1154 (48) 679 (50)
111 children
Viral-like illness
Sampled 82 (34) 241 (18)
Not sampled 431 (18) 147 (11)
Nonviral-like illness - 219 (16)
Other 621 (26) 46 ( 3)
Health status unknown 114 (5) 40 ( 3)
Total questionnaires
completed 2402 1372
to be viral in etiology on the basis of symptomatology, the 513 viral-like
illnesses seen in 1976 at Clinic B again represent all of the participating
children seen by the entire pediatric group. Of these illnesses, 82
children who were patients of the two cooperating pediatricians had
samples taken for viral isolation. Sampling was much more complete and
extensive in 1977 because of the daily presence of the senior nursing
student in the clinic. This resulted in about two-thirds of the children
with viral-like illnesses having specimens taken for viral isolation.
Also, the more complete information in 1977 allowed the separation of the
remaining illnesses into nonviral infections (e.g., otitis media,
streptococcal pharyngitis, otitis externa) and non-infectious illnesses
(asthma-, atopic dermatitis and other acute and chronic diseases) .
Consequently, the 1976 "other" category in TABLE 2 contains a large number
of both non-viral and non-infections illnesses which we were not able to
distinguish on the basis of the information available. In 1977 we were
able to reduce this "other" group significantly to include just those
illnesses which were non-Infectious in etiology. The "health status
unknown" category includes patients who did not complete the questionnaire
in sufficient detail to make a determination.
15
-------
The age composition of the various diagnostic categories differed
markedly (TABLE 3). Children in the 10-15 year group comprised from 30
percent in 1976 to 40 percent in 1977 of the well children seen. This age
group comprised similar percentages of those who had viral samples but
no virus isolated, viral-like illnesses who were not sampled, and those
with nonviral-like illnesses. In contrast this age group comprised only
13 percent and 12 percent, respectively, of those who had viruses isolated
in 1976 and 1977. In our virus isolate population in 1977, 47 percent of
the children were in the birth to 3 year old group.
These marked differences In age composition between the virus
isolation group and the other diagnostic categories resulted primarily
because enterovirus infections and illnesses are most frequent in younger
children. This is strikingly apparent in the age-specific isolation rates
for 1976 and 1977 (TABLE 4). Isolation rates in the birth to 3 year old
age group were 80 percent and 63 percent, respectively, for the two years
with similarly high isolation rates in the 4 to 9 year olds. In contrast,
only one-fourth of the 10 to 15 year olds who were sampled had virus
isolates. Analysis of the distribution of virus isolates by age groups
documents the aggregation of enterovirus isolations of all types in those
children 9 years or younger, whereas the non-enterovirus isolates
distributed themselves fairly uniformly over the three age categories
(TABLE 5).
In 1976, the presenting syndromes of the sampled patients could only
be distinguished on the basis of the symptomatology listed on the
questionnaire (TABLE 6). Comparing those patients with a virus isolated
versus those without, we found that there was very little difference in
the percentage of children manifesting various syndromes. There were
slightly fewer patients with viral isolates who had nonfebrile illnesses.
The predominant syndrome involved the respiratory tract, Few illnesses
were completely localized to the gastrointestinal tract.
Because of the interviews that were conducted by the nursing student
in 1977, a somewhat finer clinical categorization of the patients and
identification of children not sampled but who had viral-like illnesses
was possible (TABLE 7). Again, there was little difference in the
distribution of clinical syndromes between those sampled patients who did
or did not have a virus isolate made. Infection of the respiratory tract
again accounted for the predominant syndrome and there were slightly more
nonfebrile patients among those who did not have a virus isolated. Of
course, the nonsampled viral-like illness group represents an unknown
mixture of etiologies. This group was notable In that it had a somewhat
lower frequency of respiratory, respiratory-gastrointestinal, and
gastrointestinal tract involvement compared to the other two groups and a
greater concentration of patients with exanthems, conjunctivitis, and
otitis media with possible viral involvement.
Seasonal Enteroviral Activity
Enteroviral activity in northern temperate climates is highly seasonal,
16
-------
TABLE 3. AGE COMPOSITION OF THE VARIOUS DIAGNOSTIC CATEGORIES*, CLINIC B 1976 AND 1977
Age in years All
0 to 34 to 910 to 15Children
1976 1977 1976 1977 1976 1977 1976 1977
Patient No. % No. % No. % No. % No. % No, % No. No.
Well children 340 (29) 162 (24) 442 (38) 228 (34) 372 (32) 289 (42) 1154 679
Sampled with
virus isolated 12 (31) 63 (47) 22 (56) 55 (41) 5 (13) 16 (12) 39 134
Sampled with no
virus isolated 3(7) 36 (34) 25 (58) 33 (31) 15 (35) 38 (35) 43 107
Viral-like illness
not sampled 182 (42) 48 (33) 170 (39) 52 (35) 79 (18) 47 (32) 431 147
Nonviral-like
illness4" - 64 (29) - 78 (36) - 77 (35) - 219
All categories 537 373 659 446 471 467 1667 1286
*Includes only well children and those ill children who had a suspected infectious disease.
Because of limitations of data collected in 1976, nonviral-like infectious illnesses could
not be distinguished.
-------
TABLE 4. ISOLATION RATES AMONG AGE GROUPS AT CLINIC B, 1976 AND 1977
No. of virus isolate/No, tested
Age groups
Clears)
0 to 3
4 to 9
10 to 15
All Ages
1976
No.
12/15
22/47
5/20
39/82
%
(80)
(47)
(25)
(48)
1977
No.
63/99
55/88
16/54
134/241
%
(64)
(63)
(30)
(56)
TABLE 5. AGE SPECIFIC DISTRIBUTION OF VIRUS ISOLATES AMONG ILL PATIENTS
AT CLINIC B, 1976 AND 1977
Number of Children
Ages 0 to 3
Viruses
Coxsackievirus A
Coxsackievirus B
Echovirus
En terovirus-1 ike
1976
9
1
1
1
1977
21
9
20
6
Ages 4 to 9
1976
13
9
0
0
1977
17
17
17
0
Ages 10 to 15
1976
2
0
1
0
1977
7
3
2
0
All Children
1976
24
10
2
1
1977
45
29
39
6
agents, untypeable
Other
0
15
All Viruses
12
63
22
55
16
39
134
18
-------
TABLE 6. SYNDROMES OF CLINIC B SAMPLED PATIENTS, 1976
Patients with
virus isolated
Syndrome
Respiratory
Febrile
Nonfebrile
Respiratory-
Gastrointestinal
Febrile
Nonfebrile
Febrile
Gastrointestinal
Undifferentiated
Unknown
No.
23
1
5
0
1
6
3
%
(59)
( 3)
(13)
( 3)
(15)
( 7)
Patients without
virus isolated
No.
25
3
3
2
1
7
2
%
(58)
( 7)
( 7)
( 5)
( 2)
(16)
( 5)
Total
Number
of Patients
48
4
8
2
2
13
5
All syndromes
39
100
43
100
82
19
-------
TABLE 7. SYNDROMES OF CLINIC B SAMPLED PATIENTS AND NONSAMPLED VIRAL-LIKE ILLNESS PATIENTS, 1977
Patients with
virus isolated
Syndrome
Respiratory
Febrile
Nonfebrile
Respiratory-
Gastrointestinal
Febrile
Nonfebrile
Gastrointestinal
Febrile
Nonfebrile
Uhdifferentiated
Central Nervous
System
Exanthem
Conjunctivitis
No.
51
14
26
0
15
0
26
2
0
0
%
(38)
(11)
(19)
(ID
(19)
( 2)
Patients without
virus isolated
No.
41
14
18
6
6
1
18
3
0
0
%
(38)
(13)
(17)
( 5)
( 5)
( 1)
(17)
( 3)
Patients with
viral-like illness
not sampled
No.
7
16
12
5
3
4
8
0
24
9
%
( 5)
(11)
( 8)
( 3)
( 2)
( 3)
( 5)
(16)
( 6)
Total Number
of Patients
99
44
56
11
24
5
52
5
24
9
(continued)
-------
TABLE 7 (continued)
Patients with Patients without
virus isolated virus isolated
Syndrome NO. 5£ Na. %
Otitis with
possible viral 0 0
Other possible
viral 0 0
Unspecified viral-
like illness 0 0
All syndromes 134 100 107 99
Patients with
viral-like illness
not sampled Total Number
No. "L of Patients
17 (12) 17
21 (14) 21
21 (14) 21
147 99 388
-------
with peak acitivity in July and August. This was apparent in the analysis of
weekly isolation rates from the clinics (TABLE 8). In 1976, between the
weeks of June 7 through July 26 when sampling was performed only in Clinic
A, relatively few cultures were taken per week and no enterovirus isolations
were made until the week of July 19. During the week of July 26, we
received no specimens for viral isolation from Clinic A. Reportedly, there
had been a marked increase in the number of ill children which so over-
loaded the staff that they felt they had no time to take cultures.
Consequently, we requested that Clinic B join the study August 2 and the
remainder of the 1976 isolation results in Table 8 reflect cultures taken
at both Clinics A and B. From the sudden marked increase in ill patients
during the week of July 26 at Clinic A and the peak number of viruses
isolated during the weeks of August 2 and 9, we assume that the enterovirus
season in 1976 was the period from July 26 through August 9 with a rapid
decrease in activity after that time.
The 1977 enterovirus season was much more active, both in terms of
number of cultures taken and the number and variety of enteroviruses
isolated. Enterovirus isolations began during the first week of sampling,
June 13, and activity persisted at steady levels through June and the first
week of July. There was a sudden increase in the number of children
sampled and isolations beginning July 11. Isolations persisted at high
levels through the first week of August and then showed a sharp decline
similar to that seen in 1976. However, during this period a 50 percent
decrease in cultures and isolations occurred in the week of July 18. This
resulted from one of the two participating physicians taking vacation that
week. In contrast to the enteroviruses, isolation of other viruses
persisted at low levels throughout the entire summer with little weekly
variation in either 1976 or 1977.
To assess how accurately these weekly variations in virus isolations
might represent the virus activity in Madison as a whole, we compared
our clinic results to those of the Wisconsin State Laboratory of Hygiene
Virus Diagnostic Section, which receives specimens for viral isolation
from throughout the state (TABLE 9). Patients tested by the State
Laboratory differ in a number of ways from ours. Specimens are received
from adults as well as children and the patients generally have more severe
illnesses which often require hospitalization. Despite these differences,
the Virus Section results indicate that the enterovirus season in
Wisconsin began about July 19, 1976, with few isolations prior to that
time. This preceded by a week the beginning that we saw in the clinics.
The 1976 season ran through August 23, or for two weeks following the
decline in clinic isolations. The 1977 statewide enterovirus season was
somewhat closer to the pattern that we saw in Clinic B. Enterovirus
isolations began in early June, with the peak incidence of July 11 through
the first week of August coinciding precisely with the peak isolation
frequency seen in our clinic_study. This comparison leads us to believe
that the variation in numbers of our weekly clinic isolations accurately
reflects enterovirus activity in Madison during the summers of 1976 and
1977 and that these data can be applied to the school study which will
be reported below.
22
-------
TABLE 8. VIRAL ISOLATIONS BY WEEK, CLINICS A AND B (1976) AND CLINIC B (1977)
to
No. of ill
Year Week children sampled
1976 June 7
14
21
28
July 5
12
19
26
August 2
9
16
23
30
Sept 6
1977 June 13
20
27
July 4
11
18
25
August 1
8
15
22
29
4
7
2
3
2
4
7
0+
35
51
20 "
11
5
6
13
24
25
20
38,
17T
35
38
16
16
12
8
Cox A
0
0
0
0
0
0
3
«
14
12
1
1
2
1
1
5
3
1
12
3
5
4
4
3
3
1
No.
Cox B
0
0
0
0
0
0
0
_
2
2
2
0
2
2
2
2
2
3
0
0
6
9
0
2
2
1
of children with isolates"
Echo
0
0
0
0
0
1
0
0
2
1
0
0
0
4
4
5
3
3
4
3
5
2
4
1
2
All NP
entero
0
0
0
0
0
1
3
17
17
4
1
4
3
7
11
10
7
18
7
16
18
g
v)
9
7
4
Other
viruses
1
0
2
0
1
1
2
1
4
0
1
0
0
1
0
2
1
2
2
3
3
0
3
0
1
All viruses
1
0
2
0
1 >
2
5
18
21
4
2
4
3
8
11
12
8
20
9
19
21
6
12
7
5
Abbreviations: Cox(Coxsackievirus); Echo(echovirus); NP entero(non-polio enterovirus).
+Patient loads increased so markedly in this week that clinic personnel didn't have time to
.sample children.
TThis marked decrease in specimens occurred because one of the cooperating physicians' took a
one-week vacation.
-------
TABLE 9. NUMBER OF PATIENTS WITH NONPOLIO ENTEROVIRUS POSITIVE SPECIMENS
DURING THE SUMMERS OF 1976 AND 1977, STATE LABORATORY OF
HYGIENE VIRUS DIAGNOSTIC SECTION
Week of Onset
1976
June 7
14
21
28
July 5
12
19
26
Aug 2
16
23
30
Sept 6
Total
1977
June 13
20
27
July 4
11
18
25
Aug 1
8
15
22
29
Number of
Cox A
0
0
2
0
2
0
3
1
0
1
1
3
0
0
13
1
0
1
1
3
3
1
1
1
0
0
0
patients with
Cox B
0
0
0
1
0
0
5
2
1
2
0
4
1
5
21
1
1
1
3
2
0
1
5
3
1
0
0
*
isolates
Echo
0
0
0
0
0
0
0
0
2
2
7
3
2
0
16
0
1
1
1
6
6
11
4
3
2
3
Total
0
0
3
3
J
5
10
0
J
5
50
2
*
2
&,
3
11
0
J
13
10
7
2
£
3
Total
12
18
40
70
24
-------
Association Between Swimming and Enteroviral Illness
With the weekly variation in enterovirus isolations as a background,
weekly age-adjusted swimming rates in the various diagnostic categories of
clinic patients can be compared. Age-adjusted rates were computed by
taking the weekly age-specific swimming rates for each diagnostic category
and applying them to a standard population comprised of the total popula-
tion surveyed during each summer. The 1976 weekly age-adjusted rates for
Clinic B (TABLE 10) present some difficulty in that the sampled patients
who did and did not have enteroviruses isolated were small in number.
Nevertheless, some trends do appear. Considering that the peak enterovirus
activity was already in progress during the first two weeks of August when
Clinic B joined the study, the adjusted swimming rate in the enterovirus
isolation group was 11 percent greater than the rate in the well group
during the week of August 2. The rate for the isolate group dropped below
the well controls during the week of August 9, and then exceeded the well
group by 12 and 13 percent during the weeks of August 16 and 23.
Similarly, the group of patients who had viral-like illnesses but had no
samples taken showed age-adjusted weekly swimming rates considerably
higher than the well controls throughout the entire test period. In
contrast, those patients who had virus specimens taken but from whom no
isolations were made had rates that were lower than those seen in the well
controls for four of the six weeks of the study.
Since the age-adjusted data for 1977 (TABLE 11) are based on consid-
erably larger numbers of sampled patients than in 1976, more consistent
patterns are evident. The 1977 data also have the advantage of covering
the entire summer season. Adjusted swimming rates of well children and
those with enteroviruses isolated show almost no differences until the week
of July-.-ll, when there is a marked increase in the percent of swimming
among the enterovirus isolate patients, exceeding the well controls by
about 19 percent. This increased frequency of swimming persists for the
enterovirus isolate group to a greater or lesser extent through the first
week of August, which encompasses the peak enteroviral isolation activity.
The weekly age-adjusted swimming rates among those sampled with no virus
isolated, and the nonsampled possible viral illness group, show much more
variation in comparison to the well controls and no consistent excess
swimming during the period of peak enteroviral activity.
Finally, some interesting differences emerge when comparing age
specific swimming rates by the various diagnostic categories over the
entire summer (TABLE 12). In 1976, the age-specific rates for all those
who were ill, those who had enteroviruses isolated, and those with viral-
like illnesses who were not sampled, consistently exceeded those of the
well children by as much as 10 to 14 percent. The only exception to this
generalization occurred in 10-15 year olds with enterovirus isolates
where the small number of children in that group went swimming about the
same number of times as the well controls. Children sampled but with
no isolations differed from the other ill categories. This group showed
age-specific swimming rates which were indistinguishable or slightly lower
than those of the well controls.
25
-------
TABLE 10. WEEKLY AGE ADJUSTED RATES OF SWIMMING IN VARIOUS DIAGNOSTIC CATEGORIES, CLINIC B 1976
Well patients
Week
August 2
August 9
August 16
August 23
August 30
September
Total no.
of children
198
232
360 .. + .
211
104
6 47
%
swimmers
46
51
42
43
47
28
Patients with
enterovirus isolates
Total no.
of children
11
14
5
1
3
3
%
swimmers
57
43
65
59
41
59
Sampled patients
with no isolates
Total no.
of children
9
18
7
5
1
3
%
swimmers
100
32
39
56
0
0
Nonsampled patients with
viral-like illnesses
Total no.
of children
118
113
84
60
35
14
%
swimmers
60
60
52
61
51
40
-------
TABLE 11. WEEKLY AGE ADJUSTED RATES OF SWIMMING IN VARIOUS DIAGNOSTIC CATEGORIES, CLINIC B 1977
N>
-J
Well patients
Week
June
July
Aug.
Total no. %
Of children swimmers
13
20
27
4
11
18
25
1
8
15
22
29
42
57
87
55
67
31
43
63
34
55
91
49
44
51
50
68
61
76
61
58
55
44
24
18
Patients with
enterovirus isolates
Total no.
of children
7
11
9
7
19
6
16
18
6
9
7
4
«HL<*.a
45
41
41
55
80
77
86
62
53
37
35
83
Sampled patients Nonsampled patients with
with no isolates viral-like illnesses
Total no.
of children
3
10
12
9
15
7
14
15
9
4
5
4
swimmers
23
47
43
74
72
51
43
68
63
0
24
25
Total no.
of children
5
5
12
17
22
7
21
19
7
14
11
7
swimmers
27
47
67
77
74
44
60
66
71
53
33
15
-------
TABLE 12. AGE SPECIFIC SWIMMING RATES BY DIAGNOSTIC CATEGOBY, CLINIC B 1976 AND 1977
oo
Age groups (years)
0 to 3
Patient Category
1976
Well children
All ill children
Sampled with
enterovirus isolate
Sampled with ^j<
no isolate
Viral-like illness
nonsampled
Other
1977
Well children
All ill children
Swimmers/
Total
66/340
58/197
' WW
4/12
0/3
54/182
65/261
27/162
59/204
%
swimmers
19
29
33
0
30
25
17
29
4 to 9
Swimmers/
Total
244/442
138/217
15/22
12/25
111/170
147/214
142/228
151/214
%
swimmers
55
64
68
48
65
69
62
71
10 to
Swimmers/
Total
239/372
71/97
2/3
10/15
59/79
99/146
175/289
124/174
15
%
swimmers
64
73
67
67
75
68
61
71
Total
Swimmers/
TotaJL
549/1154
267/511
21/37
22/43
224/431
311/621
344/679
334/592
%
swimmers
48
52
57
51
52
50
51
56
Sampled with
enterovirus isolate 24/56
43
35/51
69
9/12
75
68/119
57
(continued)
-------
TABLE 12 (continued)
NJ
vo
Age groups (years)
0 to
Swiimers/
Patient Category Total
Sampled with
no isolate 8/36
Viral-like illness
nonsampled 13/48
Non-viral-like
illness 14/64
3 4 to 9 10 to 15 Total
% Swimmers/ % Swimmers/ % Swimmers/ %
swimmers Total swimmers Total swimmers Total swimmers
22 20/33 61 27/38 71 55/107 51
27 37/52 71 33/47 70 83/147 57
22 59/78 76 55/77 71 128/219 58
-------
The same patterns were evident in the more extensive and consistent 1977
data. The most striking differences seen were in the group with enteroviral
isolates in the birth to 3 year old age group. In these children the age-
specif c swimming rate was 2.6 times that of the well controls. Similarly,
those in other age groups with viral isolations consistently exceeded the
controls in swimming frequency. This same excess of swimming was evident in
all ill patients, except for those who had samples taken but no virus isolated.
Only the 10 to 15 year olds in this group had swimming rates higher than
the well controls. The highest rate of excess swimming was seen among the
4 to 15 year old children who had nonviral-like illnesses. This seems
clearly related to the frequency of otitis externa in this group; an
infectious illness well known to be related to frequent swimming.
Statistical Analysis of Swimming versus Infection
In 1976, well children were compared with four illness categories:
those with enteroviral isolates, those sampled with no isolate, those with
viral-like illness not sampled, and all ill children (TABLE 13).
TABLE 13. RESULTS OF STATISTICAL ANALYSIS OF SWIMMING IN VARIOUS
DIAGNOSTIC CATEGORIES, CLINIC B 1976
Patient Categories
Swimming vs.nonswimming
Well children vs.
ill with enterovirus isolates
Well children vs.
ill with no isolate
Well children vs.
viral-like illness not sampled
Well children vs.
all ill children
E = 1.76
X2= 2.39
P = not significant
E - .91
X2- .08
P = not significant
E = 1.47
X2=11.85
P - <.005
E = 1.47
X2«11.85
P - <.005
Synbols:
estimate of odds ratio
value of chi square test statistic
P - the probability value
E
X2-
The greatest odds ratio was seen in well children versus those with entero-
viral isolates. However, because of small numbers, the difference was not
significant. In contrast, those children who had no isolates had an odds
ratio of less than one. Highly significant differences in the odds ratio
were seen in well children versus those with viral-like Illnesses not sampled,
and well children versus all ill children groups. No significant effect
modification by age or sex was apparent in any of the groups.
30
-------
In 1977, the numbers were sufficient to analyze the effect of
exclusive pool or beach swimming as well as swimming versus non swimming
(TABLE 14). In the well children versus those with enterovirus isolations,
there was no significant difference seen in exclusive pool swimming. How-
ever, there was a highly significant difference seen in beach swimming
between these two groups as well as with all swimming. The exclusive beach
swimmer analysis was the only category which showed an effect modification
for age, with a very large odds ratio for children between the ages of
birth and 3 years old who had enteroviral isolates (TABLE 14). In contrast,
the comparison of swimming and illness in well children versus those
sampled with no isolate, showed no significant difference in any of the
categories of swimming. In nonsampled children with viral-like illnesses
versus well children, no significant difference in odds ratios was seen for
exclusive pool or beach swimmers, but the overall swimming variable showed
a significant difference. Both exclusive beach and pool swimming differ-
ences were significant when all well children were compared to all ill
children.
The data also were analyzed with regard to the location and frequency
of swimming in the two weeks prior to the clinic visit. No differences
in frequency of swimming were apparent between well and ill swimmers.
Analyses of specific swimming sites were not meaningful because the total
number of different beaches and pools was so high that there were
insufficient numbers of children at any one site. However, sufficiently
large numbers resulted when swimming was tabulated for the three major
lakes within the city limits of Madison. Among the well children who swam
at a beach, about 40 percent swam at one of the beaches on Lake Mendota
which is the largest of the three city lakes. Thirty-four percent swam at
beaches on Lake Monona and 19 percent swam at the only beach on Lake Wingra,
the smallest and shallowest of the three lakes. On the other hand, 20
percent of the enterovirus isolate group swam at beaches on Lake Mendota
and 40 percent swam at Lake Wingra.
SCHOOL STUDY
Characteristics of the Population
Over 90 percent of participants returned biweekly questionnaires in
both years of the study. Activity profile questionnaires, which were
completed at the end of the study, were returned by 80 percent of the
children.
Respondents to the activity profile questionnaire were given a swim-
ming score by their average swimming frequency per two-week period over
the summer. About 50 percent swam at least 4 to 9 times per two weeks
(i.e. swimming score of 2.0 or more) and only 17 percent averaged less than
one swimming session per period (TABLE 15). On the basis of the average
activity score computed as explained in Appendix C, the infrequent swimmers
had as much.contact with siblings and other children as did those who swam
frequently. Information from questions related to the nature of swimming
(TABLE 16) revealed that most swimming was done with friends and that most
of the children considered themselves good swimmers. The profiles also
31
-------
TABLE 14. RESULTS OF STATISTICAL ANALYSIS OF SWIMMING IN VARIOUS DIAGNOSTIC CATEGORIES, CLINIC B 1977
ro
Patient Categories
Well children vs.
ill with enterovirus isolated
Well children vs.
ill with no isolate
- "* x
Well children vs.
viral-like not sampled
Well children vs.
all ill children
Pool swimming
vs non swimming
E - 1.58
IT** 2.64
P « not significant
E = 1.25
X2= .53
P = not significant
E = 1.49
X2= 2.61
P = not significant
E - 1.60
X2= 8.16
P = <.005
Beach swimming
vs non swimming
E = 3.4lt
X2= 8.07
P - <.005
E = 1.53
X2= .10
P = not significant
E = 1.53
X2* 1.91
P = not significant
E = 1.66
X2= 6.81
P - <.01
Swimming vs
non swimming
E = 2.17
X2=ll.ll
P - <.005
E = 1.28
X2= 1.13
P = not significant
E = 1.57
X2= 4.81
P = <.05
Symbols: E = estimate of odds ratio
X2= value of chi square test statistic
P = the probability value
tThis analysis was the only one showing a significant effect modification with age.
In particular the estimates of the different odds ratios are:
E for ages 0 to 3 = 10.63
E for ages 4 to 9 = 1.05
E for ages 10 to 15 = 3.93
-------
TABLE 15. DISTRIBUTION OF SCHOOL STUDY POPULATION OF 218 CHILDREN
RESPONDING TO THE ACTIVITY QUESTIONNAIRE BY AVERAGE SEASON
SWIMMING SCORE AND THE CORRESPONDING ACTIVITY SCORE*, 1976
AND 1977
Average
swimming
score
0 to 0.4
0.5 to 0.9
1.0 to 1.9
2.0 to 2.9
3 . 0 to 4.0
No. of
children
18
15
58
57
70
%ff
of
children
8
7
27
26
32
Average
activity
score
59
72
64
63
60
See Appendix C. The score is based on answers to questions 1-7 and
12, 13, 16.
33
-------
TABLE 16. THE NATURE OF SWIMMING ACTIVITY BY 218 SCHOOL STUDY CHILDREN
RESPONDING TO THE ACTIVITY PROFILE QUESTIONNAIRE 1976 AND 1977*
Question
8
9
10
11
No. of children
Answer choosing answer
a-group swimming
b-occasional group swimming
c-swims with friend
d-swims alone
a-can ' t swim
b-doesn't swim well
c good swimmer
a-seldom puts head under water
b -occasionally puts head under water
c-frequently puts head under water
a-spends most time out of water
b-spends most time in water
c-time spent half and half
49
22
128
8
9
41
167
5
22
190
6
177
35
See Appendix C.
showed that the large majority of children frequently put their heads under
water and that an equally large number spent most of the time during the
swimming activity in the water.
Frequency of Illness Per Two-Week Period
The school study population was comprised of children living throughout
Madison who used a large variety of different beaches and pools. The
following findings were correlated with enterovirus activity as defined by
the clinic study isolations which reflected the city-wide virus activity
for reasons described in the previous section (TABLE 8). For the entire
34
-------
summer, the frequency of reported illnesses in swimmers was two fold higher
in 1976 and more than three fold higher in 1977 (TABLE 17).
TABLE 17. FREQUENCY OF ILLNESSES BY TWO-WEEK PERIODS IN SWIMMERS AND
NONSWIMMERS 1976 AND 1977
Nonswimmers
Season
1976
1977
Period
June 13 to 26
June 27 to July 10
July 11 to 24
July 25 to Aug. 7
Aug. 8 to 21
Average
June 13 to 26
June 27 to July 10
July 11 to 24
July 25 to Aug. 7
Aug. 8 to 21
Average
Number
30
19
24
30
29
26
19
9
12
20
47
21
Percent
illness
Total GI*
-_t ._
5+ -
8
3* -
7 3
5 1
11* -
8* -
2* 2*
3 1
Swimmers
Percent
illness
Number
117
127
119
115
115
119
125
135
130
122
97
122
Total
9
6
6
13
13
10
9
8
13
14
7
10
GI*
6
2
3
4
4
4
2
2
5
1*
2
Gastrointestinal (either with or without accompanying respiratory symptoms),
Jno illnesses reported.
Represents only one illness.
The biweekly illness frequency data between the two groups revealed some
recurring patterns. During the period starting June 13, 1976, there were no
reported illnesses among nonswimmers but 6 percent of the swimming group
reported a gastrointestinal illness accounting for the majority of swimmer
illnesses during that period of time. At this time in the clinic study, we
were isolating few viruses of any kind. No clue as to the possible etiology
of this gastrointestinal illness could be determined from either our
isolation data or that of the State Laboratory. From June 27 through July 24
of the same year, illness frequency was identical in the swimmer and
nonswimmer groups. However, beginning July 25 through August 21, the illness
rate In swimmers was 2 to 3 times greater than that of nonswimmers. July 25
35
-------
also marked the increase of enteroviral activity noted in the clinic study.
Therefore, the peak illness period in the school study coincided closely with
the peak enteroviral activity in Madison in 1976. A similar pattern
occurred in 1977, with peaks of swimmer illness frequency coinciding with
clinic enterovirus isolations.
None of the"differences in illness frequency between swimmers and non-
swimmers either by two-week period or for the entire summer were statis-
tically significant. The major difficulty was with the small number of
nonswimmers, which averaged 26 per two-week period in 1976 and 21 per two-
week period in 1977. Many of the biweekly illness rates in nonswimmers
during both years were based on one ill person as noted in TABLE 17. A
finding which should be noted is that gastrointestinal illnesses were
reported by swimmers throughout virtually the entire summer season both
years. However, this syndrome was not reported by nonswimmers except for
the last two weeks of the study each year. Finally, swimmers reported a
greater variety of illnesses than did nonswimmers. In addition to in-
creased gastrointestinal and respiratory illnesses, there were also
instances of otitis media and conjunctivitis. Nonswimmer illnesses were
mostly confined to the respiratory tract.
Relationship of Swimming Frequency and Site to Illness Frequency
The frequency of swimming had no apparent effect on the rate of the
reported illnesses (TABLE 18). Dividing swimmers for each period through-
out both summers into those who swam 1 to 6 times per two weeks and those
who swam 7 or more times, showed that the percent reporting illnesses in
any period often varied widely but that the increased illnesses were just as
likely to occur in the less frequent swimmers as in the more frequent
swimmers.
Similarly, there was no consistent relationship between frequency of
reported illness and the kind of recreational water where swimming occurred
(TABLE 19). In 1976, the illness rate reported by children who swam exclus-
ively at beaches was similar to that of nonswimmers. The highest rate of
illness that year occurred in those who swam in both beaches and pools.
However, in 1977, the location of swimming had very little effect on the
frequency of reported illness. Differences fluctuated markedly by two-week
periods and showed no consistent trends in either summer.
Relationship of Age and Sex to Illness Frequency
Increased frequency of illnesses in swimmers compared with nonswimmers
held for every age category (TABLE 20). In 1976, the only illnesses report-
ed in nonswimmers were in the 6 to 8 year olds and the illness frequency in
swimmers was greatest in the 6 to 7 year olds, with a progressive decrease
in percent ill through the 10 year olds. This trend was not apparent in
1977. Therefore, age did not appear to consistently affect illness differ-
ences between swimmers and nonswimmers.
Some consistent differences did occur when the population was analyzed
by sex. Similar average illness rates were seen in male and female swimmers
36
-------
TABLE 18. THE RELATIONSHIP OF SWIMMING FREQUENCY AND ILLNESS FREQUENCY,
SCHOOL STUDY 1976 AND 1977
Swimming score
Season
1976
1977
Period
June 13 to 26
June 27 to July 10
July 11 to 24
July 25 to Aug 7
Aug 8 to 21
Average
June 13 to 26
June 27 to July 10
July 11 to 24
July 25 to Aug 7
Aug 8 to 21
Average
1 to
No. with
score
63
58
48
51
66
62
49
38
44
59 ).:
71
57
2
illnesses
10
7
8
12
17
8
12
5
11
12
9
11
3 to
No. with
score
54
69
71
64
49
70
76
97
86
63
26
61
4
illnesses
9
4
10
14
8
11
7
9
14
16
4
9
37
-------
TABLE 19. AVERAGE FREQUENCY OF ILLNESS PER TWO-WEEK PERIOD ASSOCIATED WITH
DIFFERENT TYPES OF SWIMMING, SCHOOL STUDY 1976 AND 1977
Average number
Type of of children per
Season swimming period
1976 None
Pool
Beach
Pool and Beach
1977 None
Pool
Beach
Pool and Beach
26
44
17
58
21
57
25
40
Percent
illnesses
5
8
5
13
3
10
11
11
38
-------
TABLE 20. AVERAGE FREQUENCY OF ILLNESS PER TWO-WEEK PERIOD IN DIFFERENT
AGE GROUPS AND IN MALES AND FEMALES, SCHOOL STUDY 1976-1977
Nonswimmers
Age (years)
Season or sex
1976 6
7
8
9
10
Males
Females
1977 6
7
8
9
10
Males
Females
Average
number
per period
5
6
6
4
6
13
13
<1
6
7
7
2
12
10
Percent
with
illnesses
8
11
3
*
6
3
3
3
11
5
Swimmers
Average
number
per period
16
18
25
30
29
57
62
3
31
37
33
18
63
59
Percent
with
illnesses
18
14
8
9
6
10
10
14
9
7
14
11
10
No illnesses reported.
39
-------
in each year of the study (TABLE 20). However, male nonswimmers had higher
illness frequencies than did female nonswimmers. This was especially
apparent in 1977 when there were no illnesses reported in female nonswimmers.
Consequently, the differences between swimmer and nonswimmer illness rates
were much greater in the females in both years of the study.
Relationship of Seasonal Swimming Frequency to Reported Illness and Serologic
Conversions
Very few children in the study refrained from swimming during the entire
10 weeks of either summer. Children who were nonswimmers during one two-
week period were likely to become frequent swimmers in the next period.
Furthermore, children who averaged less than one episode of swimming per
two-week period over the season represented less than 20 percent of the total
population. Because there were too few children who could be considered
nonswimmers for the season, children were divided into those with average
season swimming scores above and below the median in each summer (TABLE 21).
TABLE 21. SEASON FREQUENCY OF ILLNESS AND SEROLOGIC CONVERSIONS TO GROUP B
COXSACKIEVIRUSES (CB) AND ECHOVIRUSES (ECHO), SCHOOL STUDY
1976 AND 1977
Percent of children with one or more
Season
1976
1977
Average
swimming
score
median
median^
Number
of
Children,
70
69
69
69
illness
31
28
36
41
Serologic conversion to
CB 1-5 Echo 3,6rll CB or echo
6 6
10 10
26 25 39
41 26 39
"Median =
^Median =
1.8.
2.4.
Those children with below median swimming scores reported illnesses about as
frequently as those who were above the median. The frequency of serologic
conversions during 1976 in those abdve the median was almost double that of
the group below the median. This difference was not statistically
significant because of the very low number of total conversions. A much
higher rate of serologic conversions was seen in the 1977 study, but the
frequency of conversions was identical for those above and below the median
swimming frequency.
40
-------
BEACH ATTENDANCE AND BACTERIOLOGIC QUALITY
The 14 city beaches located on the three lakes either wholly or
partially within the city limits of Madison have varying usage (TABLE 22).
TABLE 22. RANKING OF BEACHES BY TOTAL SEASON ATTENDANCE*
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Beach
Vilas
Olbrich
Warner
Tenney
B.B. Clarke
Spring Harbor
Marshall
Willows
South Shore
Olin
James Madison
Brittingham
Esther
i
Lake Front
Lake
Wingra
Monona
Mendota
Mendota
Monona
Mendota
Mendota
Mendota
Monona
Monona
Mendota
Monona
Monona
Monona
Percent
total attendance
32
16
8
8
7
5
4
3
3
3
3
3
2
1
Based on total attendance at all beaches from the first Saturday in June
through Labor day, 1969 to 1973 (data obtained from the City of Madison
Parks Department).
As determined by a weekly survey between 1969 and 1973, Vilas Beach is by far
the most heavily used beach in the city. This is the sole swimming site on
Lake Wingra which is the smallest and shallowest of the city lakes. Olbrich
Beach, located on Lake Monona, is the second most frequently used with half
the attendance of Vilas. Despite the wide variation in usage of the other
beaches, when viewed collectively, each lake services almost exactly one-
third of Madison beach swimmers, albeit with widely varying population
densities per beach.
41
-------
The City of Madison Health Department monitors each beach at least
weekly for fecal coliforms and enterococci and provided us with these data
for the summers of 1976 and 1977 (TABLES 23, 24, 25, 26). In 1976 (TABLES
23, 24) the counts of these organisms varied widely by week at the various
beaches, with the highest geometric mean titers occurring in the first
several weeks of June and generally low counts occurring during late July
and the month of August. In 1977, the mean counts were noticeably higher
than in 1976 and some individual beaches had periods of very high bacterial
counts. In this year, the highest mean counts were seen during the weeks
of July 18 and August 8. In 1977, Lake Wingra had persistently high counts
which increased to very large numbers during the week of July 18 and the
last two weeks of August. For the entire summer, Vilas Beach had the high-
est mean count of any of the beaches.
None of these lakes receive sewage effluent but are subject to the
products of agricultural runoff and storm sewer drainage. The origin of the
fecal coliforms and enterococci which repeatedly lead to the closing of
individual beaches for short periods is not clear despite considerable
investigation.
Discussion
There is continuing concern over the potential role of both drinking
and recreational water in the transmission of viral infections, despite the
paucity of evidence that transmission by this route constitutes a signifi-
cant public health problem. Both Fox and Goldfield have recently presented
analyses of the epidemiclogic and clinical characteristics of the candidate
virus infections as well as the continuing problems with disease surveil-
lance. They have clearly outlined the difficulties inherent in studying
this question, which may account largely for the lack of conclusive evi-
dence in this area (13, 14). Only Hepatitis A virus is known to be regularly
transmitted in drinking water, a route which constitutes only a small pro-
portion of Hepatitis A transmission and which almost always results from
gross sewage contamination of a drinking water supply (15, 16). The only
viral disease consistently linked to transmission by recreational waters is
pharyngoconjunctival fever of adenoviral etiology (7, 8, 9, 17, 18).
Surprisingly, despite the widespread belief that swimming represented
a dangerous exposure during epidemics of paralytic poliomyelitis, there are
only two episodes which are cited classically as possible evidence of water-
borne transmission of poliovirus. In 1954, Little (19) described an explo-
sive outbreak of poliomyelitis in Edmonton, Alberta that occurred late in the
enterovirus season. He felt that a possible reservoir of infection was pres-
ent at Devan, a town located 20 miles upstream from the Edmonton municipal
water supply intake. Just before the outbreak of poliomyelitis in Edmonton,
an outbreak of the disease had occurred in Devan. Devan1 s primary treatment
sewage plant, which discharged its effluent into the river, had malfunctioned
42
-------
TABLE 23. FECAL COLIFORM DENSITIES AT MADISON BEACHES, 6/7/76 to 9/6/76*
- ' ' .
Number
Beach
Lake Mendota
Warner
Tenny
James Madison
Willows
Spring Harbor
Marshall
Geometric mean
6/7
710
140
50
287
600
170
229
6/14
4200
80
1670
7000
810
240
956
6/21
<10
20
1080
20
370
40
43
6/28
2500
10
50
40
150
<10
,44
7/5
220
80
3200
10
120
10
94
7/12
<10
<10
<10
20
40
10
4
of Organism/ 100ml
7/19
620
90
<10
60
70
<10
25
7/26
430
30
220
60
10
10
51
8/2
10
30
260
190
250
110
86
8/9
20
10
<10
20
-50
10
11
8/16
<10
10
<10
<10
110
20
5
8/23
20
<10
130
310
1250
300
82
8/30
140
10
<10
<10
100
480
20
Geometric
9/6
90
10
10
<10
10
10
10
mean
65
17
32
32
122
26
39
Lake Monona
___ _. _ . . <10 70 <10 120 360 38
£ Esther 270 70 <10 20 10,000 60 30 70 40 <10 10 10 70 50 36
Olbrich
Esther
Olin
South Shore
Brittingham
B . B . Clarke
180
270
100
20
160
80
17
70
230
140
640
40
30
<10
20
40
20
10
1000
20
<10
10
30
20
30
10,000
40
1500
860
70
60
10
20
30
<10
300
30
150
30
50
40
120
70
10
20
10
140
<10
40
40
<10
50
20
240 80 50 23
10 <10 <10 110 15
10 <10 360 30 35
10 <10 40 40 16
Lake Front 10 100 10 200 60 10 380 90 10 10 90 20 90 300 46
Geometric mean 74 100 13 30 252 12 85 41 11 1 13 5 51 86 48
Lake Wingra
Vilas 120 370 <10 <10 <10 20 100 120 20 <10 <10 10 60 20 13
Geometric mean-
all beaches 125 289 18 28 111 8 51 49 27 3 7 17 35 20 24
*Data obtained from the City of Madison Health Department.
-------
TABLE 24. ENTEROCOCCI DENSITIES AT MADISON BEACHES, 6/7/76 to 9/6/76
Number of
Beach
Lake Mendota
Warner
Tenny
James Madison
Willows
Spring Harbor
Marshall
Geometric mean
Lake Monona
Olbrich
Esther
Olin
South Shore
Brittingham
B.B. Clarke
Lake Front
Geometric mean
Lake Wingra
Vilas^
Geometric mean-
all beaches
6/7
770
320
390
23
360
230
238
120
420
100
80
110
40
30
91
300
150
6/14
1520
30
280
6400
600
1160
620
17
80
300
190
870
100
150
139
300
279
6/21
6/28
<10 22,400
10
19,500
<10
140
80
36
20
10
20
10
90
10
110
24
30
29
50
420
300
90
10
224
680
120
20
20
. 50
3160
490
159
100
178
7/5
2230
230
1430
10
100
<10
W, 95
40
7600
170
560
380
40
170
257
20
140
7/12
<10
20
50
20
140
100
26
90
20
30
30
<10
40
20
40
19
Organism/ 100ml
7/19
780
20
150
10
40
90
66
240
30
100
<10
60
60
1090
60
130
66
7/26
2410
50
610
40
150
70
177
240
160
140
90
230
300
600
213
190
195
8/2
70
50
440.
570
1020
280
251
30
180
190
70
10
140
90
71
250
133
'8/9 '
40
30
80
20
90
30
42
10
10
<10
<10
90
<10
10
5
170
16
8/16
<10
<10
40
<10
170
70
9
20
40
20
40
80
110
<10
25
90
17
8/23
<10
40
400
820
1150
430
137
20
50
1610
10
110
<10
30
34
120
68
B/SCT
360
<10
30
50
310
520
67
240
100
50
10
160
60
360
88
30
72
Geometric
9/6
130
60
30
<10
60
10
23
260
130
70
140
20
70
40
79
70
46
mean'
103
26
246
39
189
64
82
66
97
62
28
88
36
80
60
95
71
Data obtained from the City of Madison Health Department.
-------
TABLE 25. FECAL COLIFOBM DENSITIES AT MADISON BEACHES, 6/13/77 TO 8/29/77*
JS
Un
" i
Number of Organism/100 ml
Beach
Lake Mendota
Warner
Tenny
James Madison
Willows
Spring Harbor
Marshall
Geometric mean
Lake Monona
Olbrich
Esther
Olin
South Shore
Brittingham
B.B. Clarke
Lake Front
Geometric mean
Lake Wingra
Vllas
Geometric mean-
all beaches
6/13
80
10
<10
210
<10
80
23
<10
70
10
20
110
20
120
23
310
28
6/20
30
70
80
20
530
20
57
20
60
280
40
60
50
70
60
60
59
6/27
40
<10
640
20
120
80
41
90
20
30
<10
50
30
170
28
480
41
7/4
270
20
30
150
30
59
380
30
<10
<10
<10
50
40
11
70
25
7/11
540
<10
10
120
30
29
860
70
50
<10
<10
<10
330
19
270
28
7/18
330
110
1130
870
10,000
50
511
600
310
410
400
1000
2500
2800
802
1600
695
7/25
10
10
420
600
800
20
86
30
370
30
<10
<10
<10
60
11
120
32
8/1
40
170
10
100
120
90
65
20
<10
210
100
20
<10
20
15
100
32
8/8
350
630
1200
70
850
1000
501
1200
170
420
20
<10
8000
1370
211
140
297
8/15
<10
150
70
<10
30
30
15
80
10
350
10
30
<10
10
19
420
21
8/22
60
10
10
10
70
60
25
110
670
360
20
170
80
80
128
1310
75
8/29
20
20
210
1100
970
340
177
6600
<10
190
<10
30
60
180
46
897
101
Geometric
mean
55
25
70
71
227
66
67
125
41
83
8
18
27
123
41
279
59
Data obtained from City of Madison Health Department.
-------
TABLE 26. ENTEROCOCCI DENSITIES AT MADISON BEACHES, 6/13/77 TO 8/29/77'
Number of Organism/100
Beach 6/13
Lake Mendota
Warner
Tenny
James Madison
Willows
Spring Harbor
Marshall
Geometric mean
Lake Monona
Olbrich
Esther
Olin
South Shore
Brittingham
B.B. Clarke
Lake Front
Geometric mean
Lake Wingra
Vilas
Geometric mean-
all beaches
30
<10
10
120
10
180
20
20
40
20
<10
90
<10
200
16
160
21
6/20
30
170
430
10
150
<10
39.
, 90
50
440
10
10
10
170
44
40
41
6/27
190
<10
20
10
190
50
27
260
140
30
50
130
520
540
153
460
78
7/4
780
20
110
60
30
79
1020
30
20
20
50
50
30
51
110
64
7/11
4300
20
<10
220
260
94
1610
<10
30
<10
10
10
2510
28
490
55
7/18
2930
670
1600
670
30,000
490
1771
1940
1050
730
1430
5540
6100
18,000
2783
2570
2280
7/25
50
30
1050
270
4560
10
164
60
670
10
40
40
10
160
52
40
83
ml
8/1
150
180
100
80
50
90
100
70
<10
10
1340
10
40
10
23
50
46
8/8
1890
2200
5600
760
2230
990
1842
470
540
690
60
100
24,000
9000
809
290
1070
8/15
10
80
20'
10
160
310
45
20
30
160
90
150
10
10
39
140
45
8/22
30
10
100
30
150
50
43
60
580
40
20
180
100
540
115
6630
101
8/29 Geometric mean
50
30
1000
1390
2410
450
362
1830
180
50
150
70
100
320
177
530
261
163
41
127
94
349
88
116
208
68
61
39
73
71
304
91
264
108
Data obtained from City of Madison Health Department.
-------
at that time. A second possible waterborne episode was described by Bancroft
(20) in 1957 which involved an outbreak of poliomyelitis in a veterans low
cost housing complex. There was a clear association between the distribu-
tion of illnesses and location of flush valve toilets not provided with
vacuum breakers. There was also a consistent temporal correspondence
between the outbreak and the occurrence of extreme fluctuations of pressure
in the water mains. In neither of these studies was the suspected contami-
nated water source cultured for the presence of virus.
Another approach to the question of transmission of viruses in recrea-
tional water has been the surveillance of reported illness episodes in
swimmers using waters of defined bacteriologic quality. The first such
large scale field study compared swimmers using: two beach sites of varying
bacterial contamination on Lake Michigan, an Ohio River swimming site and a
nearby swimming pool, and two beaches on Long Island Sound (10). While close
monitoring of the bacteriologic quality of the various swimming sites was
done, illness incidence was determined by the report of the participating
families and no attempt was made to establish etiologies for the illnesses
reported. The investigators demonstrated that an appreciably higher overall
illness incidence occurred in the swimming group as compared to the Tion-
swimming group, regardless of bathing water quality. Secondly, they demon-
strated an increased incidence of illness among swimmers that correlated with
increasing levels of bacterial contamination, although the data on this
question were not conclusive. Finally, those who swam in the river site with
high coliform densities had a significantly greater incidence of gastro-
intestinal illnesses.
More recently, a study has been reported comparing illness experience
in swimmers using a relatively bacteriologically clean versus a relatively
dirty ocean beach in New York City (11). This study had some similarities to
Stevenson's in that water quality was monitored closely, but the illness
frequency and nature was defined by the report of the participants. The
study was done on the weekends to isolate single water exposure and those
who swam during the middle of the week were eliminated. Swimming was defined
as submersion of the swimmer * s head and each participant was carefully moni-
tored for this activity. This study found a significant increase in gastro-
intestinal illnesses between swimmers who had definite contact with the water
(i.e. head submerged) compared to those who did not. Differences in gastro-
intestinal illnesses between swimmers and nonswimmers were statistically
significant only at the beach which had substantial bacterial contamination.
Numerous indicator organisms were monitored through the two years of the
study reported to date, and the density of fecal coliforms and enterococci
showed the closest relationship with the frequency of gastrointestinal
symptomatology.
A number of reports have dealt with the isolation of specific entero-
viruses from different types of recreational waters in association with
human infections. The earliest report involves an outbreak of disease of
Coxsackievirus Bl etiology, an agent which had been undetected in the study
area for the previous 14 years (1). This virus was concomitantly recovered
from a city wading pool*- A similar situation was described in Berlin in
which virologlc surveillance of the city swimming pool during one summer
47
-------
resulted in the recovery of Coxsackievirus B3 from 20 percent of the water
samples. This virus was also the predominant agent recovered from clinical
specimens obtained from children primarily manifesting meningitis or
encephalitis (2). A final report dealing with swimming pools indicated that
Coxsackievirus Bl was recovered from the children's section of two Moscow
swimming pools during a two-month period with concomitantly proven human
infections with this agent during the same period(3).
A small outbreak of Coxsackievirus A16 illness in five children occurred
some days after they had been bathing in a lake(5). Coxsackievirus A16 was
isolated from a 10-liter sample of lake water and from rectal swabs of two
patients. A somewhat larger outbreak with accompanying isolation of the
etiologic agent from a lake was reported by Hawley et al. in 1973 (4). This
outbreak involved 21 identified cases of acute viral illnesses in children at
a summer camp located on the shores of Lake Champlain. Coxsackievirus B5 was
isolated from 62 percent of the clinical cases and 17 percent of sampled
asymptomatic individuals. The same virus was also recovered from the water
of the beach along the lakeshore which the camp used for its swimming acti-
vity. However, epidemiologic analysis of the outbreak fairly conclusively
indicated that the principal mode of transmission was persontoperson, in
that all but four cases were clustered in one cabin.
The lack of evidence of waterborne transmission in the Lake Champlain
outbreak is similar to the findings of a summer camp study where virologic
surveillance of the campers and environment was considerably more systematic
(21). Over the course of the summer, 37 of 681 participating campers had
virus isolates. None of the environmental samples, including those taken
from the camp swimming pool on a weekly basis, resulted in the isolation of a
virus. As in the Lake Champlain study, virus spread was shown to be primar-
ily or solely by direct transmission, in that specific viruses tended to con-
centrate in groups of boys living in the same cabins. There was no apparent
transmission of illnesses by swimming or by any other environmental
exposures.
None of the above studies, including those that reported isolations of
enteroviruses from recreational water, provide evidence that the swimming
waters were to any extent a medium of transmission. They do establish
several facts, however. It is clear that recreational waters not receiving
sewage effluent can be contaminated with enteroviruses and that the serotype
found in the water is likely to be the same one predominating in concomitant
human infections. These studies provided the basis for questions we
attempted to address in the investigations reported in this paper. If, in
fact, recreational waters can be contaminated with enteroviruses either from
the bathers using the water or from other undefined nonsewage sources, then
recreational water transmission of enteroviruses might be examined more
profitably as one of the regular transmission routes rather than the
occasional source of recognizable outbreaks.
As is characteristic of north temperate climates, enterovirus activity
in Wisconsin is highly seasonal with little circulation prior to June, a sharp
peak, of incidence in July and August, and then a decrease in Activity through
September and October (22). In most years a number of different serotypes
48
-------
circulate, although the predominance of a single serotype is not an uncommon
occurrence. If recreational water transmission is a significant factor in
this seasonal occurrence; examining the association between swimming and
documented enteroviral infections, without regard for the site of swimming
or the microbiologic characteristics of the swimming water, might be a
useful approach to this question. Recognizing that person-to-person trans-
mission is the primary mode of enterovirus spread, we realized fully that
water transmission might not be discernible by epidemiologic study even if
this mode occurred fairly commonly. Consequently, we simultaneously under-
took a prospective (cohort) study offering the advantages of more complete
and reliable exposure data and a direct determination of risk, and a case
control study to ensure a sufficient number of cases for analysis.
The major logistic aspects of the prospective study in school children
were very successful. Just under 150 children were recruited each year.
Over 90 percent of the biweekly questionnaires were returned and there was
an 80 percent response to the activity profile questionnaire distributed at
the end of the study. Because each reported illness was followed up by
phone interview to characterize the illness, we had additional confirmation
of the accuracy of reporting.
The findings in this study based on reported illnesses confirmed those
of Stevenson and Cabelli in that total illness and gastrointestinal illness
among swimmers exceeded that of nonswimmers in eight of the ten two-week
survey periods for the two summers. The average percent of total reported
illnesses and of gastrointestinal illnesses for the two summers seasons
showed a 2- to 3-fold increase in swimmers compared to nonswimmers. Equally
significant were the peak illness rates among swimmers, which coincided
closely with peak enteroviral activity in Madison as defined by isolation
studies in our pediatric clinic population. Also notable was that gastro-
intestinal illnesses were reported by swimmers in nine of the ten periods
for the two years combined, while in only two periods did nonswimmers report
illness of the gastrointestinal tract.
The prospective study revealed no consistent association between ill-
ness frequency and the type of recreational water used (i.e. swimming pool
or beach). Most surprisingly, we were unable to demonstrate any association
between the frequency of swimming either by two-week period or for the sum-
mer as a whole and the incidence of either illness or significant serologic
rise in neutralizing antibodies to the most frequently encountered Coxsack-
ievirus B or echovirus serotypes. Finally, none of the differences between
swimmers and nonswimmers in any of the analyses of the cohort study reached
statistical significance.
The major and critical difficulty encountered in the school study was
the inadequate number of nonswimmers recruited. Because of the starting
date of the grant in 1976, there was only time to recruit children agreeable
to entering the study. Under these conditions, there was an average of only
26 children.per two-week period who did not swim. In 1977, vigorous effort
was made to recruit volunteers who declared themselves to be infrequent
swimmers. Despite this, the average number of nonswimmers by two-week
period in 1977 was only 21. Computing swimming indexes for every child over
49
-------
each summer period revealed that only 17 percent of the children swam on the
average of less than one time per period. The small size of our nonswlmmer
control group in the prospective study prohibited meaningful statistical
analyses. Nevertheless, the results were all consistent with the hypothesis
that swimming activity is associated with increased risk of apparent viral
illness, both respiratory and gastrointestinal.
As anticipated, the case control study in the pediatric clinic did
supply sufficient numbers for meaningful analysis. The information gather-
ing and virologic sampling required by the study design was satisfactorily
accomplished in Clinic B. This was not the case in Clinic A where the study
was initiated in June 1976. The difficulties encountered in this clinic
limited satisfactory data collection until the beginning of August and the
first two weeks of September when Clinic B was added. This occurred just at
the beginning of peak enteroviral activity that summer. Another design error
in 1976 was the decision to collect questionnaire information on all patients
seen by the seven pediatricians at Clinic B while only two of them were
cooperating in collection of virologic samples.
These methodologic problems were corrected in the 1977 clinic study by
the placing of a full-time member of our staff in the clinic and limiting
data collection to the patients of the two cooperating pediatricians. As a
result, we obtained a much more precise picture of the various diagnostic
categories which comprised the study population and achieved a two-thirds
sampling for virus isolation of those judged clinically to have viral-like
illnesses.
The achievement of overall clinic isolation rates of 48 and 56 percent,
respectively, in 1976 and 1977 provided sufficient numbers of etiologically
defined cases. This was a central requirement in testing our hypothesis.
There are several reasons to accept the etiologic significance of the enter-
oviral isolates made, even though relatively few acute and convalescent
serum specimens were obtained from the isolate group. In 1977, 11 serum
pairs were available on patients from whom a Coxsackievirus B or echovirus
was isolated and in seven of these (63 percent) a 4-fold or greater rise in
neutralizing antibody titer to the isolated serotype was documented. The
four serum pairs not showing a significant rise all had acute serum titers
of at least 1:24 and three of these four had an initial titer of 1:96 or
above. Previous work has indicated that if the initial titer of neutralizing
antibody is above 1:32, documentation of a significant rise is seldom
achieved. This admittedly small set of paired serum specimens does suggest,
however, that most of the isolations represent a true infection. More
importantly, in ascertaining etiologic significance, is that in both years of
the study, pharyngeal specimens were positive in 85 percent of the children
with isolates. Kepfer et al. (23), have documented the importance of
pharyngeal specimen isolation in the etiologic diagnosis of enteroviral
infections compared to stool isolation alone. Finally, in the 1977 season
we obtained viral isolation specimens from 28 well control children. This
sample of controls was not representative. Nevertheless, non-polio entero-
viruses were isolated from only two of the 28 controls (7 percent), which is
very similar to the rate of enterovirus recovery found in asymptomatic
children tested in other studies.
50
-------
Statistical analyses of the Clinic B data when age adjusted revealed a
significant association between swimming and illness in both years of the
study. In 1976, statistical significance was seen for the nonsampled viral-
like illnesses and total infectious illnesses verses the well group. The
differences in swimming between those with enterovirus isolates and the well
controls were not significant, but the odds ratio of this comparison was the
highest seen in 1976. In contrast, the odds ratio in the sampled group with
no isolates versus well controls was less than one.
The larger numbers generated in 1977 allowed statistical analyses of
exclusive beach and pool swimmers, as well as total swimming, between the
various diagnostic categories. As in 1976, the age adjusted rate of swim-
ming among those with nonsampled viral-like illnesses was statistically
significant. Similarly, exclusive beach and pool swimming was signifi-
cantly greater in those with infectious disease-like illnesses. Both of
these categories obviously contained undefined numbers of enterovirus
infections. Those with known enterovirus illnesses swam significantly
more frequently in the two weeks prior to the clinic visit than did well
children, and the excess of exclusive beach swimmers in the enterovirus
group was highly significant. The only ill group in 1977 with swimming
rates the same as the well controls were those sampled with no isolate,
a finding which recapitulated the 1976 data.
The case control clinic study then consistently demonstrated an age
adjusted statistically significant association between swimming and infec-
tious disease-like illness. More to the point, the statistical association
is strongest for enteroviral disease. Since this study design cannot
provide direct evidence that the association is causal, the results must
be examined for confounding factors which might have produced either a
spurious or indirect association between swimming activity and enterovirus
illness.
In assessing the validity of our findings, several logistic aspects of
the study must be considered. One of the major concerns of a case-control
design is that exposure to the hypothesized causal factor must be determined
by retrospection and is, therefore, subject to varying degrees of Imprecision.
We do not think this represents a serious problem in our study since the
period in question encompassed only the two weeks prior to the clinic visit.
Furthermore, statistically significant associations were found only on the
question of whether or not the child had been swimming. There is no apparent
reason that accuracy of reporting this event might have differed markedly
and consistently between ill and well children. Information about the
frequency and location of swimming could be expected to be less precisely
recalled, but this information was not of central interest as will be
discussed below.
A second concern is that we did not determine the extent of exposure to
the water. Intuitively, recreational water transmission of viruses, especially
the enteroviruses, seems most likely if water is taken into the mouth and some
quantity is swallowed. This event seems most likely if there is frequent
submerging of the swimmer's head. The case-control study provides no direct
Information on this question. However the results of the activity profile
questionnaire in the school study indicate that the majority of children over
51
-------
six years of age could be expected to have had extensive water exposure.
Furthermore, there is no prior reason to suspect that degree of exposure
differed consistently between the various diagnostic categories. Finally,
the greatest difference in swimming rates was seen between the enterovirus
isolate group and the well controls below the age of three. Children this
age can be assumed to have had the lowest frequency of head submersion.
This age group, however, might be expected to put their hands in their mouths
more frequently while engaging in activities at a swimming site.
Another aspect of this study which bears on its validity is that the
results are biologically plausible. Work of others reviewed at the begin-
ning of this discussion indicates that the enteroviruses, excreted both
from the pharynx and intestinal tract, can contaminate and persist in bodies
of recreational waters not receiving sewage effluent. Secondly, in 1977 the
weekly age-adjusted swimming rates of the enterovirus isolate group did not
exceed that of the well, controls until the peak of enterovirus activity in
Madison. Lastly, if the risk of acquiring an enteroviral illness is
increased by swimming (or swimming place exposures), the effect of this
transmission route might be most apparent in the age group with the highest
incidence of enteroviral illnesses and the lowest swimming rate under normal
circumstances. Our data confirm this postulation, showing the most markedly
different age-specific swimming rates between the controls and enterovirus
isolation group in the birth to 3 year age range.
Consistency of the demonstrated association between swimming and
enteroviral illnesses is difficult to assess because there are no other
reported studies that are similar enough to ours for comparison. Neverthe-
less, both years of our study demonstrated essentially similar results. In
1976, the differences between controls and the enterovirus isolation group
were not statistically significant. This probably results, however, from
the relatively small number in the latter group, since the odds ratio (rela-
tive risk) of this comparison was higher than any other comparison in that
year.
Finally, do the data lend support to a specific association between
swimming and an increased risk of enterovirus illness? This question really
has two parts: 1) are the enteroviruses uniquely involved; 2) if swimming
activity does have a role in enterovirus transmission, is water the medium
or are viruses transmitted via person-to-person contact at the swimming site?
Clearly this study design cannot provide conclusive evidence on these ques-
tions since the extent of water contact for each swimmer is not known and
no attempts at virus recovery from the recreational waters were made. How-
ever, the data do provide some suggestions that the association was specific
for the enteroviruses and that water served as a transmission medium.
The basis for these tentative conclusions rests largely on the results
from those children who had specimens with no isolated viruses. We feel
that this group was largely devoid of undiagnosed enteroviral infections
since our enterovirus isolation rate was so high. In addition, this group's
age composition was more similar to the well controls than the isolate group,
and antibody testing of 25 serum pairs from this group revealed no Coxsackie-
virus B infections. Therefore, we feel that that the sampled group with no
isolates, who had clinical syndromes that were very similar to the enterovirus
52
-------
isolate group, provides an ill control group for valid comparison.
In neither year of the study did the group of children with apparent but
undefined viral illnesses show swimming activity that differed statistically
from the well controls. In addition, they showed no consistent seasonal
excess swimming patterns or age-specific swimming rate differences. This
implies that there was not simply a nonspecific association between swimming
and all apparent viral illnesses. Slmilary, the nonsampled viral-like ill-
ness group which undoubtedly contained some enteroviral illnesses showed a
swimming association intermediate between the isolate and no isolate groups.
The two studies of viral infections in children's camps cited above (4,21)
offered evidence against swimming water transmission of the virus infections
detected in the campers. Significantly, the epidemiologic data spoke against
other kinds of casual camp contacts, which would equal in intensity that
which occurs at poolside. In both instances, transmission was clearly related
to the family-like contact among cabinmates.
Our own work with the transmission of experimental r-hinovirus infections
(24) showed that even the intimate contact of married couples was not
sufficient to transmit this strictly respiratory pathogen unless the trans-
mitter was quite symptomatic and shedding large quantities of virus. Fox et
al. (25) had previously observed this phenomenon in his virus watch family
studies showing a markedly lower rate of transmission of respiratory
infection ( 20 versus 70 percent) if the index case in the family was
asymptomatic. In studies of intrafamilial person-to-person virus transmission
such as those of Fox's group, an attack rate in susceptibles of 50 percent is
considered high. Water containing any significant continuous level of
enteroviral contamination might be as or more efficient a transmission route
since our infectious dose study of polio type 1 vaccine shows a 50 percent
infectivity rate with about 70 TCID_0 and a 100 percent rate at 150 TCID .
The one finding in both the school and clinic studies that does not
support the association between enteroviral illness and swimming was the lack
of a swimming frequency response. In the school study, illnesses were
reported more frequently in swimmers than in nonswimmers, but no increase in
illness frequency was seen as swimming frequency increased. Likewise, in the
clinic study, well swimmers swam as frequently as did those in any of the ill
groups. This was unexpected and unexplainable, yet it was a consistent
finding in both years of both studies. If recreational water does serve as a
transmission medium, this finding suggests an all or nothing phenomenon
implying fairly constant viral contamination during peak enterovirus activity.
Finally, the prospective school study revealed no differences in illness
frequency with either type of recreational water used, pool or beach water.
The clinic study involved use of so many different swimming locations that
systematic analysis was not possible. The 1977 clinic results, however,
provide some intriguing, if inconclusive, findings. There was a statistically
significant excess of beach swimmers with enterovirus isolates as compared to
controls and" this effect was most marked among those children in the birth to
3 year old group. Secondly, the enterovirus isolate group was twice as likely
to have used the one beach on Lake Wingra than the controls. This beach is
53
-------
the most heavily used in Madison and in 1977 had the highest mean counts of
fecal coliforms and enterococcus.
We feel that the data developed in this study of enterovirus transmission
in recreational water have demonstrated the value of an epidemiologic approach
to this question. The cohort study of school children has confirmed previous
findings of increased respiratory and gastrointestinal illnesses in swimmers
versus nonswimmers. However, differences were not statistically significant
and the specific etiology of these illnesses could not be defined. The
case-control clinic study has provided the first information we are aware
of supporting the hypothesis that recreational water can play a role in
enterovirus transmission. Not only was a statistically significant associa-
tion demonstrated but the differences between various categories of ill
children provide an Internal consistency that leads validity to the statis-
tical associations.
54
-------
REFERENCES
1. McLean DM, Larke KPB, McNaughton GA, et al: Enteroviral syndromes in
Toronto, 1964, Can Med Ass J 92:658-661, 1965.
2. Liebscher S: Enteroviren im Schwimmbadwasser. Z. Gesamte Hyg
16:198-200, 1969.
3. Osherovich AM, Chasovnikova GS: Some results of a virologic investi-
gation of environmental sources. Hyg Sanit 34:424-427, 1969.
4. Hawley HB, Morin DP, Geraghty ME, et al: Coxsackievirus B epidemic at
a boys' summer camp. J Am Med Ass 226:33-36, 1973.
5. Denis FA, Blanchovin E, Poitiers F, et al: Coxsackie A16 infection
from lake water. J Am Med Ass 228:1370-1371, 1974.
6. Parrott RH, Rowe WP, Huebner RJ, et al: Outbreak of febrile pharyngitis
and conjunctivitis associated with type 3 adenoidal-pharyngeal-
conjunctival virus infection. N Engl J Med 251:1087-1090, 1954.
7. Kjellen L, Zetterberg B, Svedmyr A: An epidemic among Swedish
chj.ld.ren caused by an adenovirus type 3. Acta Paediat Scand 46:461-568,
1957.
8. Kaji M, Kimura M, Kamiya S, et al: An epidemic of pharyngo-conjunctival
fever among school children in an elementary school in Fukuoka
Prefecture. Kyushu J Med Sci 12:1-8, 1961.
9. Foy HM, Cooney MK, Hatlen JB: Adenovirus type 3 epidemic associated
with intermittent chlorination of a swimming pool. Arch Environ Hlth
17:795-802.
10. Stevenson AH: Studies of bathing water quality and health. Am J Publ
Hlth 43:529-538.
11. Cabelli VJ, Dufour AP, Levin MA, et al: The impact of pollution on
marine bathing beaches: An epidemiological study. Am Soc Limnol
Oceanogr Spec Symp, 2:424-432, 1975.
X2. Melnick J, Wenner HA: Enteroviruses. In Diagnostic Procedures for
Viral and Rickettsial Infections, 4th ed. Edited by Lennette EH,
Schmidt NJ, American Public Health Associateion, Washington DC, 1969,
pp 529-602.
55
-------
13. Fox JP: Human-associated viruses in water. In Viruses in Water.
Edited by Berg G, Bodily HL, Lennette EH et al. American Public
Health Association, Washington DC, 1976, pp 39-49.
14. Goldfield M: Epidemiologic indicators for transmission of viruses by
water. In Viruses in Water. Edited by Berg G, Bodily HI, Lennette
EH et al. American Public Health Association, Washington DC, 1976,
pp 70-85.
15. Moseley JW: Transmission of viral diseases by drinking water. In
Transmission of Viruses by the Water Route. Edited by G Berg.
John Wiley and Sons, New York, 1965, pp 5-23.
16. Craun GF, McCabe JL: Review of the causes of waterborne-disease
outbreaks. J Amer Water Works Ass 65:74-84, 1973.
17. Ormsby HL, Aitchison WS: The role of the swimming pool in the trans-
mission of pharyngeal-conjunctival fever. Can Med Ass J 73:864-866,
1955.
18. Bell JA, Rowe WP, Engler JI et al: Pharyngoconjunctival fever. J Am
Med Ass 157:1083-1092, 1955.
19. Little GM: Poliomyelitis and water supply. Can J Publ Health
45:100-102, 1954.
20. Bancroft PM, Engelhard WE, Evans CA: Poliomyelitis in Huskerville
(Lincoln) Nebraska. Studies indicating a relationship between
clinically severe infection and proximate fecal pollution of water.
J Amer Med Ass 164:836-847, 1957.
21. Paffenbarger RS, Berg G, Clarke NA, et al: Viruses and illnesses in
a boys'summer camp. Amer J Hyg 70:254-274, 1959.
22. Nelson D, Hiemstra H, Minor T and D'Alessio D: Non-polio enterovirus
activity in Wisconsin based on a 20-year experience in a diagnostic
virology laboratory. Amer J Epidemiol 109:352-361, 1979.
23. Kepter PD, Hable KA and Smith TF: Virus isolation rates during
summer from children with acute upper respiratory tract disease and
healthy children. Am J Clin Path 61:1-5, 1974.
24. D'Alessio DJ, Peterson JA, Dick £R, et al: Transmission of experimental
rhinovirus colds in volunteer married couples. J Infect Dis 133:28-36,
1976.
25. Fox JP, Cooney MK, Hall CE: The Seattle virus watch. V. Epidemiologic
observations of rhinovirus infections 1965-1969 in families with
young children. Am J Epidemiol 101:122-143, 1975.
56
-------
Appendix A
Clinic Study Questionnaire
For: Teenagers, Children & Babies
The Quisling Clinic and the Dept. of Preventive Medicine, University of Wis.
Medical School are studying the possibility of acquiring viral infections
by swimming lakes or pools. The information requested below is important to
the study. We would appreciate your cooperation in answering the following:
Child's Name
Age Sex
Address
Street " City Date
In the past 2 weeks, my child swam: (circle one)
a) 0 times b) 1-3 times c) 4-6 times d) 7-9 times e) 10 or more times
Please list beaches or pools (not home backyard wading pools) he/she
swam at during past 2 weeks and number of times:
2. Is your child ill today or has he/she been ill in the past 2 weeks?
No Yes
Reason for seeing physician today: (check one or more)
Common cold ^ Fever Vomiting, Allergy injection
Sore throat ___ Sores in mouth Stomach pain _^ Gen'l physical exam
Ear ache Persistent or diarrhea Well-child check-up
Rash Headaches Injury Other;
_ ^Vaccination __
Thank you. Please return to receptionist.
57
-------
Appendix 8
School Study Post Card Questionnaire
For the 2 weeks
beginning Hon.,
ending Sun.,
Dear Parent:
As part of the virus-swimming study that you and your child are
participating in, we would like to have you answer the following
questions and mail this card as soon as possible.
Child's Name
1. In the past 2 weeks, my child swam days.
2. Please list baaches and/or pools used:
3. Has your child been ill in the past 2 weeks? Yes No
4. Phone number during the day:
58
-------
Appendix C
Enterovirus Research Laboratory
Department of Preventive Medicine
University of Wisconsin
465 Henry Mall
Madison, Wi 53706
SWIMMER STUDY ACTIVITY PROFILE
Parents - Please help your child complete this questionnaire which will help
us determine how much contact he or she has with other children during the
summer.
Child's name
Please circle the answer that best describes you while you were in our
swimming study.
1. I have (a) one, (b) two, (c) three, (d) more than three, brothers and
sisters who live at home.
2. I am (a) the oldest, (b) the youngest, (c) neither the oldest not the
youngest, child living at home.
3. During the summer, I play with a brother or sister (a) once in awhile,
(b) fairly often, (c) most of the time.
4. During the summer, I like best of all to play with (a) my brother or
sister, (b) kids my age, (c) kids younger than me, (d) kids older than
me, (e) just myself.
5. When I do things with other kids during the summer, I prefer doing them
with (a) just one or two friends, (b) more than just one or two friends.
6. My favorite summer activities are (a) mostly indoors, (b) mostly
outdoors, (c) about equally divided between the indoors and outdoors.
7. I (a) spend up to one week at a summer camp, (b) more than a week at a
summer camp, (c) don't go to a summer camp.
8. I go swimming (a) with an organized group that meets regularly, (b)
sometimes with an organized group, (c) usually with a friend or two,
(d) mostly by myself.
9« I (a) oan't swim, (b) don't swim very well, (c) am a good swimmer.
59
-------
Page 2 - Swimmer Study Activity Profile
10. I (a) don't like to put my head under the water, (b) put my head
under the water once in awhile, (c) put my head under the water fre-
quently .
11. When I go swimming, (a) like to spend most of the time out of the
water, (b) spend most of the time in the water, (c) divide my time
equally between the water and out of the water.
Please complete the following (for the year you were in our study).
12. List your favorite summer-time activities (except for sports):
13. What sports did you play in the summer?
14. List any teams you played on:
, met times per week
, met _. _times per week
15. List any organized swimming groups you belonged to (including lessons)
, met times per week
t met times per week
16. List any other organizations that you belonged to that held regular
meetings:
r y
'. V
, met __. times per week
, met _ times per week
, met times per week
60
-------
SWIMMER STUDY ACTIVITY PROFILE
Scoring Key
QUESTION ANSWER POINTS
only child
a
b
c
d
a
b
c
a
b
c
a
b
c
d
e
a
b
a
b
c
a
b
c
0
5
10
20
30
15
5
10
2
4
6
12
6
9
3
0
3
6
0
10
5
15
20
0
12-13 0-10
16 0-15
61
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO.
EPA-600/1-80-006
3. RECIPIENT'S ACCESSION NO.
TITLE AND SUBTITLE
Epidemiologic Studies of Virus Transmission in Swimming
Waters
5. REPORT DATE
January 1980 issuing date
6. PERFORMING ORGANIZATION CODE
. AUTHOR(S)
D.J. D'Alessio, T. E. Minor, D. B. Nelson, C. I. Allen,
A.A. Tsiatis
8. PERFORMING ORGANIZATION REPORT NO.
. PERFORMING ORGANIZATION NAME AND ADDRESS
University of Wisconsin
Madison, Wisconsin 53706
10. PROGRAM ELEMENT NO.
1CC614
11. CONTRACT/GRANT NO.
R-804161
(Report #1)
12. SPONSORING AGENCY NAME AND ADDRESS
Health Effects Research Laboratory - Cinn, OH
Office of Research & Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final; 2/1/76-1/31/79
14. SPONSORING AGENCY CODE
EPA/600/10
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Retrospective and prospective epidemiologic studies were conducted to determine
if swimming activities increase the risk of acquiring enteroviral infection in child-
ren. The retrospective study consisted of a surveillance of recent swimming activities
and clinical histories in 3,774 children who visited a pediatric clinic. A highly
statistically significant increased rate of swimming activity was found among children
who had enterovirus associated illnesses as compared to the well controls.
The prospective study examined the relationship between swimming activities and
enteroviral infections in 296 elementary school children. Swimming rates for the
entire season showed no relationships to reported illnesses. This lack of a relation-
ship appeared to be the results of a failure to find enough children who were infre-
quent or nonswimmers. Nevertheless, the trend toward a decreased illness rate in
children who refrained from swimming for two weeks is consistent with the retrospective
study results.
To our knowledge, this is the first study that has found a statistically signifi-
cant association between exposure to recreational waters and an increased risk of
enteroviral disease. Various internal consistencies of the data discussed in this
report support the validity of the association and suggest that water served as the
transportation medium.
17.
KEY WORDS AND DOCUMENT ANALYSIS
a.
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COS AT I Field/Group
enterovirus
public health
water resources
Swimming and disease
Recreation
Gastroenteritis
57K
57U
06M
68D
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (ThisReport)
unclassified
21. NO. OF PAGES
74
20. SECURITY CLASS (Thispage)
unclassified
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION is OBSOLETE
62
» U.S. GOVERNMENT PRINTING OFFICE. 19SO -657-14&/5564
------- |