Acute and Chronic
Toxicity Testing
Standard Methods
Multiple methods have been standardized
(certified) by multiple organizations
American Society for Testing and Materials (ASTM)
Organization for Economic Cooperation and
Materials (OECD) – (Europe based)
National Toxicology Program (NTP)
All above standardized protocols available from
US EPA, Federal Register and researchers that
developed the programs
Advantages of Standard Methods
Tests are uniform and comparable to previous results
within the same or other laboratories
Can be replicated (confirmed) by other laboratories
Makes it easier for decision makers to accept test
results
Logistics are simplified, developmental work already
done
Methods establish baseline from which
modifications can be made if necessary
Data generated can be combined with those from
other laboratories for use in QSAR, ERA’s
Advantages of Standard Methods (con’t)
Detailed listing of
apparatus, dilution
water, test material, test
organisms, etc
Experimental,
analytical and
documentation
procedures are detailed
Acceptability criteria
are listed
Disadvantages of Standard Methods
Often very specific hard to apply to other
situations or answer other questions
Tend to be used in inappropriate situations
(research, cause and effect evaluation)
May not be applicable to natural environment
Acute vs. Chronic Toxicity Tests
Can broadly classify toxicity tests based on length of
exposure
Acute Toxicity test
Drop dead testing
Time = 2 days (invertebrates) to 4 d. (fish)
LD50
LC50
TLm (median tolerance dose)
EC50 (effective concentration)
Lose equilibrium, sit on bottom “ecologically” dead
Not very ecologically relevent but quick, relatively cheap
(but still ~$700-1,200 per test)
Acute vs chronic toxicity testing (con’t)
Chronic toxicity testing
Growth, reproduction
More ecologically relevant data but takes longer,
more expensive
Shows effect at much lower dose
Test requires much more “baby-sitting”
Acute Testing - theory
Population of organisms has normally
distributed resistance to toxicants acute
toxicity test designed to identify mean
response
Regulations allow 5% of species to be
impacted
Most tests only use 2-3 species (up to 6)
not really enough to protect 95% of all
species!
Acute Toxicity Test Organisms
Use of test species
based on
Lab hardiness
Common
Known life cycle
Cheap
Short-lived
Normal distribution of resistance/sensitivity
100
Mean response
Frequency
Protected
5% allowable
impact
0
Resistance (log [X]
Experimental design for toxicity tests
Freg. of response (i.e death)
Percent mortality
Integration
of
Looking for
this area of
response
Log [X] Log [X]
To save money while finding area of mean response use a two step process
Step 1 – Screening test
Expose 5–10 organisms to 10x increasing [ ]
for 24-96 hours
Trying to determine range in which median
lethal concentration (LC50) will fall
100
Screening test
% Responding
0
[X] mg/L
# dead none none some all RIP all RIP
0 0 30% 100% 100%
Concen. 10-3 10-2 10-1 100 101
Step 2 – Definitive test
From previous results
low = 10-2 = 0.01 mg/L
high = 100 = 1.0 mg/L
Run test using logarithmic scale of concentrations because
organisms usually respond logarithmically to toxicants
Usually use at least 5 concentrations + control
Control – checks toxicity of dilution water, health of test organisms,
stress level of testing environment (test chambers, lighting,
temperature, etc)
If >10% of control organisms die throw out test!
Use 10 – 30 organisms randomly split up among tanks
Set up for definitive test – example 1
Treatment Division Concentration
(mg/L)
1 10-2 0.01
2 10-1.5 0.032
3 10-1 0.1
4 10-0.5 0.32
5 100 1.0
Control 0.0
Set up for definitive test – example 2
low = 101 µg/L
high = 103
Treatment Division Concentration
(µg/L)
1 103 1000
2 102.5 316
3 102 100
4 101.5 31
5 101 10
control 0
Analysis of Toxicity Tests
Based on hypothesis that resistance to toxicants is
normally distributed
Use a probit transformation to make data easier to
analyze
Based on SD so each probit has a percentage
attached to it
Mean response defined as probit = 5 so all probits
are positive easier to visualize
Can use probit analysis to calculate LC50 because
probit transformation will straighten the cumulative
distribution line
Probit Analysis
Response of organisms to toxic chemicals = normal distribution
Cannot measure normal distribution directly because effect is
cumulative, so graph as cumulative distribution
Normal distribution Cumulative distribution
# Responding
Dose Log Dose
Converting a curvilinear line to straight line
Difficult to evaluate a curved line
Conversion to a straight line would make evaluation easier
Cumulative distribution Probit transformed
100%
7
Probit Units
% Mortality
5
50
Straight line (easier to
LD50, analyze)
TLM) 3
0
Log Dose Log Dose
Note: probit forces data towards middle of
distribution good because most organisms
are “average” in their response
Relationship between normal distribution
and standard deviations
34.13%
Mean
13.6%
2.13%
-2 -1 0 1 2
Standard deviations
Difficult to deal with SD (34.13, 13.6, etc) so rename SD to probits
34.13%
Mean
13.6%
2.13%
3 4 5 6 7
Probits
Example probit analysis
Concentration Deaths %
(mg/L)
Control 0/10 0
0.3 0/10 0
1 0/10 0
3 1/10 10
10 4/10 40
30 9/10 90
100 10/10 100
Look at data should be able to tell immediately that LC50 should be between 10 and 30 mg/L
Graph fit line by eye (approximately equal number above and below line)
Uses of LC50
1. 1. Application factor
LC50 x n = ___ = allowable dose
Good if do not have better information (chronic tests)
2. Rank hazards lower LC50 = more toxic
3. Lead to chronic testing
Remember: LC50 does not provide an ecologically
meaningful result bad because trying to protect
ecosystem need more ecosystem level testing
Probit is trade-off between cost and getting
sufficient data to make a decision about the
environmental toxicity of a chemical
Chronic toxicity testing
Sublethal
Time = 7d. to 18 months
Endpoints are
growth
Reproduction
brood size (Ceriodaphnia dubia can have 2-3 broods
in seven days)
Hatching success
Analysis of chronic tests
Analysis of Variance (hypothesis testing)
Test for significant difference from control (C + 5
doses)
Regression analysis
EC20 (concentration that causes 20% reduction
relative to control)
Results of Analysis of Variance test
Respiration (gC/L/d.)
*
Community
* *
C 1 3 10 30 100
Concentration of Hg (mg/L)
Determination of EC20
Control
Response (growth)
response
10 μg
20% reduction relative to control
8 μg
Control Dose EC20 eg. 1 mg/L = discharge limit
Ecosystem Tests
(microcosms, mesocosms)
AOV design (4 reps X 3 treat., 3 rep X 4)
Time = 1 – 2 years
$106 /year
Endpoints are
Biomass
Diversity
Species richness
Etc.
All toxicity tests try to determine level of
toxicant which will or will not cause an effect
NOEC – No Observable Effect Concentration
Highest conc not signficantly different from control
LOEC – Lowest Observable Effect Concentration
Lowest test concentration that is significantly different
from control
MATC – Maximum Allowable Toxicant
Concentration
Geometric mean of NOEC and LOEC
Often called the “chronic value”
MATC
MATC = √NOEC + LOEC
Results of Analysis of Variance test
Respiration (gC/L/d.)
*
Community
* *
C 1 3 10 30 100
Concentration of Hg (mg/L)
If there is magic on earth, it is in water
Photo by R. Grippo
