To the Editor—In a recently published letter the editor, 1 Zorrilla-Vaca and Vaca-Gonzalez questioned the methodology and the results of our systematic review on nosocomial outbreaks due to contaminated drugs, especially on outbreaks due to contaminated propofol. 2 In their opinion, important articles had not been included in our review because of a poor search strategy and/or insufficient bibliographic sources, resulting in an incorrect mortality rate. Herewith, we would like to respond to their questions and remarks.
The main concern of Zorrilla-Vaca and Vaca-Gonzalez addresses our omission of an article by Bennett et al 3 in 1995, which summarizes 7 nosocomial outbreaks that could be traced to contaminated propofol. Although we were well aware of this publication at the time of our review, we decided not to include it because all of these outbreaks had previously been published by the Centers for Disease Control and Prevention (CDC) 4 and this primary publication had already been included in our work, cited as reference 113. Thus, including the article by Bennett et al would have resulted in bias due to double publication.
Secondly, Zorrilla-Vaca and Vaca-Gonzalez criticize that an editorial by Trépanier et al 5 in 2003 on nosocomial infections caused by propofol had not been adequately acknowledged in our review. Once again, this omitted publication is not a primary description of a nosocomial outbreak but rather is a summary of events published previously. It refers to the aforementioned article by Bennett et al 3 and to 3 additional outbreaks reports: Kuehnert et al 6 in 1997, McNeil et al 7 in 1999, and Henry et al 8 in 2001. Two of those articles are also included our review, cited as references 35 and 87, respectively. 6 , 8 The article by McNeil et al. 7 was not included because it only reports 1 of the 7 outbreaks that had been published by the CDC previously. 4
However, we cannot deny the likelihood that some reports of nosocomial outbreaks caused by contaminated propofol or other contaminated substances were not included in our review. No matter how complex the search algorithm for a literature search, the possibility always remains that some relevant data are lacking. In addition, the focus of our review was infections due to contaminated drugs in general (original title: “Hospital acquired infections related to contaminated substances”) rather than infections caused by propofol in particular. Thus, the key words described in the methods section of our publication did not take a specific type of substance into account. If we had specified particular substances, the number of possible substances that we would have had to investigate individually would have been too great to handle: solutions of sodium chloride, potassium, or glucose, propofol, heparin, insulin, erythrocyte concentrates, plasma albumin, all other kinds of formulas for an intravenously application, ultrasound gels, disinfection fluids, drugs used for inhalation, all kinds of substances for external use only, and many more.
Finally, Zorrilla-Vaca and Vaca-Gonzalez suggest the use of additional bibliographic sources for a more robust data search in systematic reviews on nosocomial outbreaks. Our work was based on searches of PubMed (one of the databases they recommend), the Outbreak Database, 9 and reference lists of all retrieved articles. To date, 3,200 outbreak reports have been filed in the Outbreak Database. To our knowledge, the Outbreak Database represents by far the largest collection on nosocomial outbreaks available worldwide, and it has often been used for research on various topics related to nosocomial outbreaks such as general epidemiological research, risk factor analysis, and infection control guideline preparation. 10 – 12
Nevertheless, we do agree with Zorrilla-Vaca and Vaca-Gonzalez that reviews on specific high-risk substances such as propofol should be carried out more regularly because several new outbreaks caused by contaminated propofol 13 , 14 have been published since our own systematic review in 2007. More up-to-date reviews will keep staff on the wards and infection control personnel better informed regarding the risk and the epidemiology of nosocomial infections.
Financial support: No financial support was provided for this article.
Potential conflict of interest: Both authors state that they have no conflict of interest related to this article.