LETTER
We have read with interest the recent article by Lahanas and colleagues published in the October 2013 issue of the Journal of Clinical Microbiology (1). While this study did confirm some useful characteristics of the Alfred 60/AST detection system with a previous report by Ilki et al. from 2010 (2), we feel that some important technical aspects of this device may still have been overlooked, resulting in an underestimation of its performance characteristics.
Lahanas and colleagues concluded from their analysis that the Alfred 60/AST device was more accurate at screening negative rather than positive urinary tract infection (UTI) samples. This conclusion was based on the observation that this device failed to detect 9 of the 80 isolates, these episodes, in the author's opinion, representing “true false negatives.” We have carefully evaluated these findings and provide sound evidence to question the authenticity of these purported false-negative cases.
For one of the nine purported false-negative samples (the Pseudomonas aeruginosa isolate from sample no. MB12216254), the instrument result was not negative but was considered invalid due to inaccurate pipetting volume (unpublished data provided by Lahanas et al. in 2013). This sample should have been excluded from statistical analysis.
In addition, the 3 cases reported in Table 1 (unpublished data from authors) show a plate count of <104 CFU/ml taken from catheters (catheter specimen urine [CSU]). For CSU samples, the following details were reported in the article: (i) the Alfred 60/AST configuration settings included an incubation period of 240 min and a detection threshold of 800 CFU/ml, and (ii) for petri dishes, the interpretation criteria to consider the culture consistent with UTI included any pure or predominant uropathogen isolated (0 CFU/ml).
False-negative cases from unpublished data from 2013a
This configuration of the Alfred 60/AST is inconsistent with the adopted criteria because specimens with a count of >0 but <800 will be reported negative by Alfred 60/AST but positive with plate cultures. Since any pure or predominant uropathogen is considered to be consistent with UTI using petri dishes, the proper cutoff for the indwelling catheters would have been 5 h = 0 CFU/ml with the Alfred 60/AST device. A general purpose broth was used in the study, while a specific dedicated product is available to boost the detection of some specific organisms, such as Pseudomonas spp. or Candida spp. isolated from CSU.
Finally, the criteria used for accounting for false-positive results are questionable because, as stated by international guidelines (3, 4), the clinical interpretation is crucial for UTI investigation. We have noted that in this study the clinical interpretation was only applied to plates and not to the Alfred 60/AST results, whereas clinical interpretation should have been applied to or excluded for both methods. Based on our concerns, it is our opinion that potential users of this automated device should be correctly informed of these technical issues prior to routine use of this device and interpretation of data where this instrument was used.
ACKNOWLEDGMENT
Gianpiero Spezzotti is an employee of Sire Analytical System Srl (Udine, Italy), the company that develops and manufactures the Alfred 60/AST device.
FOOTNOTES
For the author reply, see doi:10.1128/JCM.00356-14.
- Copyright © 2014, American Society for Microbiology. All Rights Reserved.