Multi-laboratory evaluation of the reproducibility of polymer biodegradation assessments applying standardized and modified respirometry methods

This research evaluated the intra- and interlaboratory variability when applying OECD 301F and OECD 301B Ready Biodegradation respirometric test methods to quantify polymer biodegradation as well as the impact of method modifications including test duration, inoculum level and test substance concentration on results. This assessment synthesizes results of mineralization studies on 5 polymers of varying structural components, molecular weight, charge, and solubility, evaluated at 8 different laboratories in 4 different countries, providing significant geographic variation in inoculum source as well as lab to lab variations in test setup. Across all laboratories, intralaboratory variability was low (<= 18 % absolute difference) indicating the reproducibility of results between replicates and uniformity of test setup in each laboratory. Interlaboratory variation was also low for all 5 polymers with extent of mineralization being comparable in all OECD 301F and 301B studies even when test methods were modified. Across all studies mean mineralization was 89 +/- 5.5 % for polyethylene glycol 35,000, 85 +/- 7.4 % for polyvinyl alcohol 18-88, 44 +/- 13 % for carboxymethyl cellulose (DS 0.6), 48 +/- 4.1 % for a modified guar gum, and 88 +/- 6.2 % for microcrystalline cellulose (MCC) at study completion. Due to the lack of polymeric reference materials, MCC was evaluated and found to be a suitable reference material for polymers that biodegrade rapidly in screening studies. An additional respirometric study was conducted quantifying mineralization of the 5 polymers in river water to evaluate the relationship with OECD 301 results using activated sludge as the inoculum. A similar extent of mineralization was observed for all 5 polymers in the OECD 301 and river water studies but time to reach the maximum extent of mineralization was longer using river water as the inoculum source likely due to the lower microbial counts (10(6) CFU/L) in the test system.