The statement by Ms Josephson that the 24-hour specimen results in the 48-hour group do not contribute to the determination of safety at 48 hours is incorrect. We were unwilling to make that assumption. In fact, some burettes in the 48-hour group were contaminated at 24 hours but not at 48 hours. Therefore, contamination of the burette at 24 hours clearly could contribute to the determination of safety in the 48-hour group. Since any contamination might lead to bacteremia, we felt that all contamination was an appropriate end point used to determine safety. Furthermore, the risk of contamination of the patient may be time dependent. To compare comparable time intervals, successive 24-hour cultures in the 24-hour group were compared with successive 24-hour cultures in the 48-hour group. It just so happens that the study design allowed one group (the 48-hour one) to have their burette remain the same.