United States
Environmental Protection
Agency
Health Effects
Research Laboratory
Research Triangle Park NC 27711
Research and Development
EPA-600/S1-84-006 July 1984
&EPA Project Summary
Retrospective Epidemiological
Study of Disease Associated
with Wastewater Utilization
Hillel I. Shuval, Badri Fattal, and Yochanan Wax
A retrospective* epidemiological
study was carried out on the association
between enteric disease incidence and
wastewater utilization in 79 kibbutzim
(cooperative agricultural settlements)
in Israel having a population of 32,672.
Medical records on disease incidence
were collected directly from the patients'
files at each kibbutz clinic. These
records provided information on 24 de-
fined diseases, laboratory confirmed or
clinically diagnosed, half of which were
enteric diseases which might possibly
be transmitted by wastewater, while
the other half were non-enteric diseases
which served as controls.
Environmental data such as the
location of the sewage irrigated tracts,
wind direction and source of water
supply were collected for each kibbutz.
The kibbutzim included in the study
were divided into four categories:
I. Thirty irrigating with wastewater
effluent in addition to irrigating
with freshwater. Population:
13,531.
II. Twenty-eight not utilizing efflu-
ent for any purpose, which served
as a control group. Population:
11,096.
III. Ten using wastewater effluent as
feed-water for fish ponds. Popu-
lation: 5,005.
IV. Eleven which practiced two con-
secutive years of effluent utiliza-
tion and another two consecutive
years in which no effluent was
utilized for any purpose or vice
versa. Population: 3,040.
"Although some authorities prefer the name
historical prospective for a study of this design, the
term retrospective has been used for purposes of
simplicity
The risk of enteric disease associated
with various forms of effluent utilization
was estimated for different age groups
using both the logistic and log linear
regression models which controlled for
various confounding factors including:
length of irrigation period, geographic
region, source of water supply, source
and volume of wastewater, source of
clean irrigation water, and population
size.
The high degree of intra- and inter-
kibbutz variations of disease rate in the
68 kibbutzim in categories l-lll resulted
in conflicting findings and anomalies
which did not provide a basis for
drawing conclusions. However, no
evidence of excess risk associated with
effluent irrigation was found.
In the switch category (IV) a clearly
significant excess risk of total enteric
disease ranging from 32-112 percent
was found in the 0-4 age group during
those periods in which wastewater was
used for irrigation as compared to the
non-wastewater irrigation periods,
regardless of the direction of the
switch. Excess risk was not seen in
other age groups. A small but significant
excess risk of about 10 percent in
effluent irrigating years was also found
on an annual basis in each age group, as
estimated by the log linear model only.
None of the findings provided direct
support for the hypothesis that the main
pathway of pathogens transmission
was aerosolized effluent resulting from
sprinkler irrigation.
Data from 68 kibbutzim, which were
based on kibbutz clinic files, were
compared with cases of salmonellosis,
shigellosis and hepatitis reported to the
Ministry of Health, and no serious
-------
discrepancy was found between the
two sources. A small, significant
excess risk of total enteric disease was
found during effluent irrigation periods
but the present study did not confirm
the extreme differences reported in a
previous study of effluent irrigating
kibbutzim.
This Project Summary was developed
by EPA's Health Effects Research
Laboratory. Research Triangle Park, NC,
to announce key findings of the research
project that is fully documented in a
separate report of the same title (see
Project Report ordering information at
back).
Introduction
Land application of wastewater serves
both as an effective water pollution
control measure and as a means of
conserving water through recycling in
areas suffering from shortages of water
resources.
In the United States, the U.S. Environ-
mental Protection Agency (EPA) has been
assigned the task of evaluating the
technological feasibility and potential
health effects of land application of
wastewater as a major water pollution
control strategy. In Israel, wastewater re-
use in agriculture is part of the national
water resources management policy in-
tended to help close the gap between the
limited amounts of renewable fresh
water resources and the growing water
demand associated with increasing
populations. This policy has resulted in
the establishment of some 250 wastewa-
ter utilization projects, 90 percent of
which practice sprinkler irrigation with
oxidation pond effluent. By the year
2000, it has been estimated that Israel
will be utilizing some 300-400 million
cubic meters of effluent per year,
primarily for agricultural purposes. The
goal is that 80 percent of all municipal
wastewater will be recycled in this
manner.
The most prominent form of wastewa-
ter reuse in agriculture is by sprinkler
irrigation. This results in the aerosoliza-
tion of 0.1-1 percent of the effluent
applied to the land. In the United States,
sprinkler irrigation is also one of the most
feasible modes of land application of
wastewater.
There has been concern that the
extensive use of sprinkler irrigation with
effluent may present a potential health
hazard for field workers and the popula-
tion adjacent to the fields. Humans may be
infected by aerosols containing patho-
genic bacteria or viruses by inhalation of
particles of 0.5-5.0 nm which can enter
the respiratory system. The larger
droplets, however, can be trapped in the
upper respiratory system including the
nose and larynx, and from there they may
reach the digestive tract. Other modes of
transport of pathogens from areas
irrigated with wastewater are by direct
contact with irrigation workers, contami-
nated crops and fomites.
In 1975 investigators reported that
pathogenic bacteria may be transported
in aerosols from sewage to humans. A
later survey emphasized the potential risk
to sewage workers of enteric diseases
from occupational exposure. Earlier
studies also showed that microorganisms
were isolated from aerosols at a distance
of 1200 m downwind from a sewage
treatment plant. However, there is little
evidence of the degree of health risk from
aerosolized sewage.
A recent review of health effect studies
of wastewater aerosols is contained in
the proceedings of the EPA Symposium on
Wastewater Aerosols and Disease. The
studies reviewed did not provide clear-
cut evidence as to the health risks, if any,
associated with wastewater treatment or
reuse.
The Environmental Health Laboratory
of the Hebrew University-Hadassah
Medical School has succeeded in isolating
coliforms at a distance of 350m, salmonella
at a distance of 60m and enteroviruses at
100m downwind from the wastewater
irrigated field.
One of the first retrospective epidemic-
logical studies on the possible health
risks associated with sprinkler irrigation
of wastewater was carried out in Israel. In
this study, the data suggested that in 77
kibbutzim (cooperative agricultural settle-
ments) practicing mainly sprinkler irriga-
tion with undisinfected oxidation pond
effluent, the incidence of typhoid fever,
salmonellosis, shigellosis and infectious
hepatitis was two to four times higher
than in 130 control kibbutzim not
practicing any form of effluent irrigation.
Then investigators concluded that these
findings "provide some epidemiological
evidence for increased risk of enteric
communicable diseases" among users
of partially treated, non-disinfected
wastewater for irrigation.
The researchers themselves pointed
out that from this study they could not
provide definite proof that the suggested
added health risk in effluent-utilizing
kibbutzim was associated with the
dispersion of pathogenic microorganisms
by aerosols. This conclusion was drawn
because a number of pathways of
infection other than aerosols existed,
such as direct personal contact via
clothing or skin, of sewage irrigation and
fish pond workers, or exposure to the
crops, and fish grown in ponds using
effluent.
This preliminary survey also had
serious methodological constraints in that
investigators based their findings solely
on official communicable disease reports
which were submitted to the Ministry of
Health as required by law. These require-
ments do not include the listing of all
relevant diseases, indicating serious
reporting deficiencies. The comprehen-
siveness of kibbutz reporting to the
Ministry of Health is not known. Other
methodological deficiencies existed also.
For example, the irrigation period, as
defined in that study, was arbitrarily set
as April to November, rather than the
actual irrigation period which varies from
3 to 12 months per year. The effluent
irrigation status of the kibbutzim were
determined according to the Ministry of
Agriculture report of 1975 and assumed
that the status was correct for the entire
study period, while the disease data
covered a period of from 2 to 13 years. In a
check of the Ministry of Agriculture
reports for available years 1971, 1975
and 1978, it was found, for example, that
in the above-mentioned seven-year
reporting period, there was an increase of
about 35 percent in the number of
kubbutzim irrigating with wastewater,
while about 10 percent of the kibbutzim
ceased wastewater irrigation. Thus about
45 percent of kibbutzum changed their
wastewater irrigation status during the
study period. Detailed data on the
changes in category during the 13 years
of the study are not available, but it can be
assumed from the available data that a
similar trend of changes in status may
have occurred.
In addition, a field check on the
accuracy of the Ministry of Agriculture
reports on wastewater irrigation status
indicated that some 16 percent were
incorrectly categorized. Moreover, in that
study, communicable disease data for
different diseases were based on varying
periods of time—from two years for
influenza, up to 13 years for typhoid
fever—leading to difficulties in interpre-
ting the epidemiological implications due
to such factors as disease cycles and
epidemic periods. For the above reasons,
conclusive findings should not be based
on that study.
The World Bank monograph on appro-
priate technology for water supply and
sanitation states that the "Data on the
health effects of night soil or sewage
-------
upon sanitation or agricultural workers
are inconclusive, although the risk is self-
evident." This statement illustrates the
ambiguous situation that exists in this
area and which may lead to overly
restrictive policies concerning wastewa-
ter reuse - especially in those areas
suffering from water shortages. What
are required are carefully designed
studies that can provide accurate data to
assess the real nature of the risks involved
in sewage reuse and can serve as a basis
for developing a sound public health
policy.
In light of all the above problems, new
research based on primary medical and
environmental data collected at each
kibbutz was carried out by this project to
confirm or rebut the findings of the initial
retrospective study of wastewater irriga-
tion in kibbutzim.
The full report covers a study of the
health records during the years 1974-77.
These were taken directly from the clinic
files of 32,672 persons in 79 kibbutzim in
the following categories:
I. Thirty practicing sprinkler irriga-
tion with wastewater and exposed
to wastewater aerosols.
II. Twenty-eight not exposed to waste-
water aerosols, nor practicing any
form of wastewater irrigation or
utilization. This category served as
the control group for categories I
and III.
III. Ten not exposed to wastewater
aerosols by utilizing wastewater
as feed water for fish ponds.
IV. Eleven that switched from non-
wastewater use to wastewater
use or vice versa, thus serving as
controls for themselves.
In most cases, the method of wastewa-
ter treatment prior to utilization was
oxidation ponds, with 5-10 day detention
periods. Coliform concentrations in efflu-
ent were very high and usually ranged
from 106~7/100 ml while enteric virus
concentrations in effluent ranged from
102-103 PFU/100 ml. High concentra-
tions of Salmonella organisms are
commonly detected in municipal waste-
water in Israel. These concentrations of
enteric microorganisms were consider-
ably higher than those at comparable land
application sites in the United States.
Thus, this project in Israel represents
what may be considered a "worst case"
of population exposure to wastewater
land application practice in general, and
aerosolized pathogens of sewage origin
in particular. If little or no health risks can
be detected under such extreme condi-
tions of environmental exposure to
pathogens, the risk associated with land
application projects using significantly
higher quality effluent may be negligible
or nonexistent. If, however, some health
risks are detected under the extreme
conditions of this study, it will provide
some indication of the type and degree of
additional prophylactic measures which
should be introduced to prevent or reduce
the risks to an acceptable level.
Conclusions
1. Based on simplified analyses, no
excess of total enteric disease was
observed in effluent irrigating (category I)
kibbutzim (collective agricultural settle-
ments) over non-effluent utilizing kibbut-
zim (category II) during the irrigation
period or on an annual basis. An excess of
total enteric disease rates was observed
in kibbutzim utilizing effluent as feed
water for fish ponds (category III) over
both categories I and II during the
irrigation period and on an annual basis.
Due to the numerous and complex
confounding factors not controlled for in
a simple comparison of crude rates, no
attempt at drawing conclusions from
these observations was made.
It was concluded that a regression type
analysis by the logistic and log linear
models was required to control for con-
founding factors in order to estimate the
risk of enteric disease, if any, associated
with the various wastewater utilization
practices in the kibbutzim included in this
study.
2. Not all distributional assumptions
which underlie both the logistic and log
linear models could be examined by the
data. Those that were examined were
found to be reasonably satisfied.
The log linear model succeeded in
explaining about 50 percent of the mtra-
and inter-kibbutz variation of the enteric
disease rates (R2 = .50) based on episodes
culled from kibbutz clinic files.
However, much of this apparently
good fit is due to an artifact associated
with the high correlation between enteric
and control diseases, which might reflect
differences in quality of reporting in the
kibbutzim. Thus the fit, due to the mam
variables under study, is considerably
less. The fit for the logistic model was
about the same as that of the log linear,
after considering the effect of the quality
of the reporting.
The above conclusions concerning the
basic assumptions and fit of the two models
apply to the analysis of the risk of enteric
disease for the 68 kibbutzim in which a
comparison was made between 30
effluent irrigation kibbutzim (category I)
as against the 28 control kibbutzim, in
which only non-effluent was used for
irrigation (category II) and the 10 kibbutzim
using effluent as feed water for fish
ponds (category III).
The two models showed a far better fit
for the data in the switch category (IV)
design (11 kibbutzim that irrigated two
years with non-effluent and switched to
two years of effluent irrigation or vice
versa). The coefficient of determination
for the fit of the log linear model was very
high (R2 = .94), with an equivalently good
fit for the logistic model.
It was concluded that, despite the
numerous co-vanates entered into both
models for the analysis of effluent vs. non-
effluent utilization categories (the 68
kibbutzim), insufficient control of inter-
kibbutz variations was obtained. On the
other hand, these very same kibbutz-
specific confounding factors were almost
completely controlled for in the switch
category. This was due to its block-type
design with each community serving as
its own control. Thus, the switch category
provided a far superior design on which
the conclusions of this study are mainly
based
3. Both the logistic and log linear
models used to estimate the risk of
becoming ill with any of the 12 enteric
diseases (total enteric) in association with
various wastewater utilization practices
in the study of the 68 kibbutzim in
categories l-lll produced results which
were often highly conflicting and full of
anomalies, to the extent that is was not
possible to draw clear and significant
conclusions from most portions of the
analysis of the data set. However, on an
annual basis it could not be concluded
from the study of the 68 kibbutzim that
there was evidence of an excess risk of
enteric disease in effluent irrigating
kibbutzim (category I) over non-effluent
irrigating kibbutzim (category II).
The results did suggest that there
might be an excess risk of enteric disease
in effluent irrigating kibbutzim (category I)
over non-effluent irrigating kibbutzim
(category II) during the "period of
exposure" to effluent (i.e., the actual
months of irrigation as recorded for each
individual kibbutz) with a deficit during
the non-irrigation period, and no excess
risk at all on an annual basis. There was
also evidence pointing to an excess risk of
enteric disease in both seasons and on an
annual basis in kibbutzim utilizing
effluent as feed water for fish ponds
(category III) over category II.
Caution should be used in evaluating
these possible trends, particularly in
relation to fish ponds, since this study did
-------
not include a control group of non-
effluent utilizing fish ponds.
4. Within category I, no correlation
was found between crude enteric disease
rates and the aerosol exposure index
used, which included the following
parameters: distance; wind direction and
velocity; size of effluent irrigated field;
and length of irrigation period.
5. The results of the study for the 68
kibbutzim, based on crude rates of
specific diseases (shigellosis, salmonel-
losis and infectious hepatitis) culled from
both the kibbutz clinic files and from the
Ministry of Health records of reported
cases of communicable disease, did not
confirm the findings of a previous study,
which indicated an excess of 2.0-4.3
times for the above specific enteric
diseases in effluent irrigating vs. non-
effluent irrigating kibbutzim. Based on
kibbutz clinic files, the shigellosis rates
were found to be about 26 percent
greater in effluent irrigation kibbutzim
during the irrigation season as compared
to non-effluent irrigating kibbutzim, but
no excess for salmonellosis or infectious
hepatitis was found, nor was there any
excess risk of total enteric disease on an
annual basis.
6. In the switch category (IV), the
analysis by both the logistic and log linear
models showed a significant increase in
risk of enteric disease in the irrigation
period during the effluent irrigating years
over the non-effluent irrigating years for
the 0-4 age group, regardless of whether
the switch was from non-effluent to
effluent or vice versa. The estimated
excess in enteric disease ranged from 70
percent to 112 percent by the odds ratio
(O.R.) as estimated by the logistic model
and was about 30 percent according to
the relative risk (R.R.) as estimated by the
log linear model.
There was also a significant excess risk
of enteric disease in the other age groups
during effluent irrigation periods, how-
ever, without the same degree of uniform-
ity as found in 0-4. The estimated excess
ranged from 7-25 percent by O.R. and
R.R.
On an annual basis there was a signifi-
cant excess risk of about 10 percent in
eachage group separately and all ages by
the log linear model, which was not found
in the analysis by the logistic model. The
findings by the logistic model analysis do
not necessarily negate those of the log
linear model. Thus, it was concluded that
a small but significant excess risk of
enteric disease on an annual basis did
exist during the effluent irrigation years
in the switch category.
The finding of a significant excess risk
of enteric disease in the irrigation
season during the effluent irrigation
years before the switch to non-effluent
irrigation supports the suggestion that
the risk associated with exposure to
effluent is long term and not only
associated with the introduction of
wastewater irrigation into a previously
unexposed community.
7. The findings in the switch category
of an excess risk of enteric disease during
the effluent irrigation months, but with
little or no excess on an annual basis,
can be explained, in part, by the results of
the analysis of the annual cycles of crude
enteric disease rates in the study of the
68 kibbutzim.
The monthly crude rates for effluent
irrigating kibbutzim (category I) were
observed to increase rapidly shortly after
initiation of irrigation. Monthly enteric
disease rates for May-July were higher in
I compared to II, dropping to a lower level
for most of the remaining months of the
year. This held for the 0-4 and the all ages
groups. The annual crude rate of enteric
disease in both I and II were the same.
The control diseases showed the known
inverse cyclical pattern peaking in
December and January, with lows during
the summer months. No differences in
the annual cycle for control diseases
between I and II were observed.
The annual cycle of enteric disease
suggests that, in category I, there may
indeed be increased transmission of
enteric disease on the initiation of
wastewater irrigation in the spring, but
that, in the unique kibbutz situation with
intimate communal living including a
common dining hall, there are continuous
multiple exposures to enteric pathogens,
which result in a certain annual satura-
tion level of enteric disease, regardless of
wastewater utilization practices.
The annual appearance of lower
enteric disease rates late in the irrigation
period and after it, during the winter in
category I as compared with category II,
suggests that some form of short-term
immunity may have developed in a
portion of the exposed population.
8. Despite the evidence that there was
an excess risk of total enteric disease in
kibbutzim during the period of exposure
to effluent (i.e., the effluent irrigation
period), the annual excess risk for all ages
was nil or quite small, thus indicating
that the overall risk of enteric disease
resulting from effluent irrigation in
kibbutzim in Israel is not a public health
problem in general, but might become
one during epidemic periods. Many of
these findings are related to the unique
kibbutz lifestyle. Whether or not a similar
degree of exposure of typical U.S. families
in private homes in a non-communal
situation, to enteric disease pathogens
associated with wastewater effluent
irrigation of similar quality would result in
an excess risk of enteric disease on a
seasonal or an annual basis cannot be
determined from this study.
9. Extrapolation of these findings to
situations of wastewater utilization in
agriculture and/or land application in the
United States must take into considera-
tion the fact that, in general, the effluent
used for irrigation at kibbutzim in Israel
was only minimally treated in oxidation
ponds with detention periods of 5-10
days. It was generally of poor microbial
quality, having coliform concentrations of
about 106/100 ml and enteric virus
concentrations ranging from 102-103
PFU/ 100 ml. Whether or not the same
degree of risk of enteric disease associ-
ated with wastewater utilization found in
this study would exist if better treated ef-
fluents, typical of those from convention-
al biological treatment plants or their
equivalent, had been used was not eval-
uated.
10. Caution must be used in the
interpretation and application of all of the
above conclusions in light of the many
methodological problems associated with
this retrospective study.
Recommendations
1. In light of the finding that a
significant risk of enteric disease during
irrigation periods was associated with
irrigation with effluent from short
detention period oxidation ponds having
low efficiency in the removal of enteric
microorganisms in kibbutzim in Israel, it
is recommended that a higher degree of
effluent treatment be provided as a
prudent public health measure. Even
though the excess risk on an annual basis
is nil or small, the higher level of
treatment can be justified as a safety
factor particularly to reduce possible
increased disease transmission during
epidemic periods.
2. Despite the numerous methodologi-
cal problems revealed in estimating the
risk of enteric disease associated with the
use of similar quality effluent as feed
water in fish ponds, the evidence points
to the presence of an excess risk in this
group of kibbutzim in Israel, which
justifies taking precautionary measures.
It is recommended that a higher level of
treatment, assuring more efficient re-
moval of enteric microorganisms, be
provided prior to discharge of effluent to
fish ponds or that effluent use be avoided.
4
-------
3. In Israel and other countries suffer-
ing from serious water shortages, where
effluent reuse and/or land application
has become a matter of national water
planning policy, society should provide
long-term health monitoring programs to
achieve an unambiguous evaluation of
the risk to the public and to determine the
type of preventative action required
where significant excess risk is revealed.
This applies in particular to areas where
previously unexposed populations will be
involved in effluent irrigation and/or land
application programs for the first time.
4. Due to the serious methodological
problems revealed in this retrospective
study, it is recommended that the pro-
spective investigations be conducted to
gather more conclusive data concerning
potential health risks associated with
land application of wastewater.
5. It is recommended that special
emphasis be given to improving the
comprehensiveness and consistency of
medical recording and that all sources be
utilized, including the nurses' and
doctors' daily logs. Investigation of
outbreaks and special studies on the
highly susceptible groups of volunteers
should be included. Of critical importance
is the sampling of blood and stools in
children, effluent workers and volunteers
so as to determine infectivity rates and
immunity in relation to effluent exposure.
These cannot be readily detected by
recording cases of clinical disease.
6. Further research should study the
apparent association between effluent
utilization and the excess of control
diseases such as streptococcal throat
infections observed in this study.
This study was carried out at the
Environmental Health Laboratory of the
School of Public Health and Community
Medicine, Hebrew University - Hadassah
Medical School, Jerusalem, Israel under
U.S. Environmental Protection Agency
Grant No. R-805174. The study was
initiated in October 1977 and completed
March 1, 1982.
Project Officer Comment - The main
reason for support of this research was to
follow up on the report in Science, 194
(1976) that there was a marked health
hazard from wastewater spray irrigation.
The authors could not confirm that earlier
finding, and, in fact, discovered serious
deficiencies in the data used in that study.
They also imply that the data base
provided by the kibbutzim is inadequate
for a conclusive study and that caution
should be exercised in interpreting the
report due to the study's many methodo-
logical problems. Neither the numerator
nor denominator could be reliably deter-
mined to calculate illness rates and
ratios. A complex statistical analysis was
necessary as an attempt to control for the
inadequacies of the data as well as to
adjust for differenes in influential factors.
A clear understanding of the methods
used for data analysis is necessary in
order to interpret the results and evaluate
the findings of the study.
H. I. Shuval. B. Fattal, and Y. Wax are with the Hebrew University—Hadassah
Medical School, Jerusalem, Israel 9 WOO.
Walter Jakubowski is the EPA Project Officer (see below).
The complete report, entitled "Retrospective Epidemiological Study of Disease
Associated with Wastewater Utilization." (Order No. PB 84-190 727; Cost:
$14.50, subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
. S. GOVERNMENT PRINTING OFFICE: )984/759-102/10623
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300
------- |