Delayed ethanol analysis of breath specimens: long-term field experience with commercial silica gel tubes and breathalyzer collection.

Delayed ethanol analysis was performed on breath specimens collected with commercial silica gel tubes using multiple Breathalyzer instruments. Eleven hundred and nine results were obtained from an ethanol testing program over a five-year period. Only 2.5% of the specimens had apparent collection errors. For the valid specimens, the most frequent result was 0.11 g/210 L and the mean result was 0.14 g/210 L. For 642 specimens, delayed results were compared with direct results. Direct results were greater than delayed results for 55%, less than for 27%, and equal to for 18% of the pairs. When fixed tolerance limits of +/- 0.03 were used, 81% of the direct results were confirmed. The confirmation percentage was best in the critical range of direct results, 0.05 to 0.15 g/210 L. The collection tubes showed no substantial variability in retaining ethanol during storage and releasing ethanol for analysis.