Recognition errors in the quantification of micro-organisms by fluorescence microscopy

An interlaboratory comparison of fluorescence microscopic counting of micro-organisms was carried out to assess counting errors in the measurement of micro-organisms in bioaerosols generated during handling of municipal waste.

Series of 27 replicate samples were collected in the work environment with a modified field exposure chamber. The preparation methods of three Scandinavian laboratories were compared. Four microscopists from these laboratories performed the counts which were also compared. Duplicate counting of identical microscopic fields allowed the assessment of recognition errors.

The field exposure chamber collected replicate samples with a relative standard deviation of 5% when particles < or = 15 microm aerodynamic diameter were collected. Storage time of 40-200 days had no significant influence on the total micro-organism count. Differences between preparation methods were from 2 to 35% for bacteria, and from 15 to 35% for fungal spores when samples were analysed in Oslo; the results for fungal spore counts were significantly different (P < 0.01). These differences were not confirmed when samples were analysed in Umeå, Copenhagen and Oslo using those laboratories methods. These results can be explained by less efficient redispersion of aggregates when the Umeå and Copenhagen methods were recreated in Oslo yielding a greater number of innumerable aggregates. Differences between microscopists were minor for fungal spores (2-12%) but substantial for bacteria (4-53%). A major source of error was the recognition of bacteria which had a relative standard deviation (rsd) of 37% although a lower size limit of 0.75 microm was adopted for counting of bacteria. Fungal spores were recognised with much better precision (rsd 9%).

Recognition errors of bacteria may be substantial and more specific fluorochromes are needed for fluorescence microscopic counting of micro-organisms.