Abstract
Bacterial endotoxin testing is performed for quality monitoring of dialysis fluid at many dialysis units. We measured bacterial activity using LAL assay twice a week for 3 years and three months, and investigated causes of variances in measured values. We found that sporadic increase or decrease by 0.5log occurred with a frequency of 1.79%. The actual rate of bacterial contamination at sampling confirmed by intermediate precision was only 0.45% and technical problems with the assay showed a rate of 0.36% by intra-assay precision. These findings suggested that over half of the variances in data were not caused by contamination, facilitating the development of a strategy for bacterial control. To determine whether the variance was caused by contamination, we should (1) re-confirm the data as soon as possible, (2) at the same sampling site showing the un-expected data as well as (3) at another sampling site simultaneously in order to confirm equal fluid quality.
