Changes to freshwater toxicity and bioaccumulation methods

Notes taken from a discussion led by Dr. Dave Mount; EPA ORD

EPA Office of Research and Development, Duluth, Minnesota
Published March 8, 2019
Change Hyalella 10‐d Hyalella 42‐d Repro Midge 10‐d Midge Life‐cycle Lumbriculus 28‐d
Food and ration changed X X X X  
Starting age/size requirements changed X   X X  
Changed guidance for low DO X X X X X
Chloride and bromide minimums in water X X      
Inclusion of sand control X X X X  
Biomass added as endpoint X X X X  
Survival‐normalized reproduction added as endpoint   X   X  
Test acceptability criteria more stringent X X X X  
Laboratory proficiency criteria X X X X  
Waterborne method discussed X X X X  
Option to change starting age   X   X  
Change organism number       X  
Test endpoints modified, some eliminated       X  
Alter guidance on sample size         X
Relax loading rate         X
Flexibility in chamber size         X
Reduced minimum replication         X

There has been a lot of work over the past decade by EPA’s Office of Research and Development Duluth, MN laboratory as well as by the USGS laboratory in Columbia, MO to improve freshwater sediment toxicity tests, especially regarding the quality and reproducibility of the ten-day and lifecycle tests with the midge, Hyalella (amphipod) and Lumbriculus (worm). The revised core methods have been peer reviewed and should be finalized in Dec. 2018. The goal is to keep the EPA and ASTM methods essentially identical. Once approved, the revised test methods will become the third edition of the test manual.

The biggest change in methods involves the food and food ration, especially for the Hyalella reproduction test and the midge lifecycle test. Improvements to the food ration have resulted in much better test organism growth and have reduced the potential influence of the sediment’s nutritional value on organism performance (organic carbon is not a good measure of the nutritional value of a sediment and there is no real way to measure or control the sediment’s nutritional value). A limited food ration can cause problems with the Hyalella test. The revisions to the method have optimized the food quality (a combination of flake fish food and either yeast-cereal leaves-trout chow (YTC) or marine diatoms) and quantity. Problems have also been observed with the midge food (tetrafin) which was blended too finely in the previous method, making it less available to the midge test organism. In the revised method, the food is sieved rather than blended and as a result, midge growth is more consistent.

The acceptable test organism age window for the ten-day Hyalella test has been tightened. Control survival is often poor if midges are less than 24-hours old so the revised method will use 3-day old test organisms. For the 10-day midge test, the revised method will specify size rather than age; the goal is to avoid midge pupation before the end of the test.

In the revised method, it will be left to the discretion of the scientist as to whether they should only aerate those test chambers with low dissolved oxygen or all test chambers. The authors of the method revisions recognize that the tests have to be flexible/adaptable for various programs (e.g. Superfund, ambient monitoring, pesticide testing, etc.). The specifics of the test method should be considered in light of the context and purpose of the study.

Regarding the use of reconstituted water in tests, the chloride concentration is very important to the US laboratory strain of Hyalella (the organism needs higher levels of chloride). A neutral substrate (clean quartz sand) can be used as a type of control for both the Hyalella and midge test, even in a water-only test. The method revision will recommend occasional inclusion of a control treatment of clean course sand; if the Hyalella or midge performance in this treatment meets test standards, the investigator can be assured that other test parameters are also acceptable. The inclusion of a sand control is also a way to frame the influence of the nutritional content of the sediments being tested; if test sediment organism performance is significantly below that in the sand control, the investigator will know that the results can be attributed to sediment toxicity. Significantly better growth in the sediment treatment when compared to the sand control suggests there is better nutritional content in the sediment being tested. With the midges, growth dynamics are important. There doesn’t appear to be a limit to midge growth with feeding; however, excessive feeding can lead to low DO conditions. As a result, there is more potential for performance differences in tests due to the nutritional content of the sediment. To address this, field sediment studies require multiple points of reference to attempt to dampen problems caused by improving food rations. In some cases, one may see growth in a field sediment that lies between the organism performance in the lab control (usually a clean natural sediment) and in a sand control; in such cases, there may be uncertainty about whether the field sediment is actually toxic, or if it might have a lower nutritional content than the lab control. In the revised Hyalella and midge test methods, biomass (survival and weight) and survival-normalized reproduction (survival and number of young) have been added as endpoints because these measures are less variable. For reproductive endpoints, normalizing the results to the number of females is necessary because the gender of the organism can’t be controlled. The revisions to the 10-day and chronic methods have tightened/increased required minimum control survival and weight acceptance criteria for Hyalella and the midge. Hyalella grow for 28-days and then the adults reproduce. If reproduction occurs earlier than 28 days, the test needs to be started with younger organisms (this would only be known after the test is underway). The midge lifecycle test uses three-day old organisms, which improves control survival. To minimize control problems, the culture and test should be at the same temperature. The midge lifecycle method now uses 10 rather than 12 organisms per test chamber.

For the Lumbriculus bioaccumulation test, more emphasis has been placed on the chemicals of interest (i.e. detection level needed), and there is more flexibility in the chamber size and number of worms used. The loading rate is not as stringent in the revised test because it does not appear to significantly impact the outcome of the test. The recommendation for five replicates/treatment has been changed; the investigator should set the number of replicates based on the level of confidence desired in the test results and data quality objectives.

Dr. Dave Mount