When Rapid Blood Culture Identification Results Don’t Correlate, Part 2: Contamination

All laboratories are prone to contamination events. Blood products, analyzers, reagents, media, etc. all have the potential to be contaminated. If you are a molecular microbiologist, then you have to worry about not only bacterial, but also nucleic acid contamination. 

The Issue

The topic of my blog last month focused on discrepant results between blood culture and PCR. Traditional blood culture workflow involves correlating the Gram stain result to what grows in culture. Nowadays, many laboratories are also performing PCR on positive blood cultures. Because we know PCR is more sensitive, it may be easy for some to justify discrepancies. Let’s image that gram positive cocci in clusters were observed in the Gram stain, the PCR detected Staphylococcus and Enterococcus DNA, but only S. aureus grew in the culture. Where did the Enterococcus come from and where did it go? It was not observed in the Gram stain and it didn’t grow in cultures, so was it “real”? Possibly. It could be a contaminant or it could be real, just present in low numbers. It’s difficult to say without having to invest more effort.

When this type of situation occurs in my laboratory, three things happen. First, we review the data. For example, if the Gram stain is discrepant, then we review the Gram stain or perform an acridine orange stain (in the case of positive PCR, but negative culture). If it’s the PCR, then we would make sure that a result entry error did not occur, etc. Second, we add the comment, “clinical correlation needed”. We have found little value in going back to the blood culture bottle and trying to recover the missing organism because in most cases when we look hard enough, using selective agar and other strategies, we do find the organism from the PCR results buried among overgrowth. Therefore, our approach is to let the clinician know that they must use other clinical data to aid in their diagnosis. Third, we document all discrepant blood culture PCR results; which includes an automatic notification to the doctoral director.

Next, let’s imagine that two more blood cultures (from different patients) become positive all within a relatively short period of time from the first discrepant result noted above. gram negative bacilli are observed in one culture and the other displays gram positive bacilli. PCR detects Enterococcus DNA in both cases. What are the odds of that happening? Not good. Something strange is going on!

The Solution

A contamination investigation needs to immediately occur. The two likely sources of contamination are 1) the PCR assay or 2) the blood culture bottles. To determine whether the issue is due to amplicon or target contamination of the PCR assay, we need to identify which instruments reported the Enterococcus. Was it a single instrument or were different instruments involved? Our laboratory performs routine “swipe” tests of the environment as part of our quality control, which allows us to monitor contamination. Swipe tests may also be performed 1) after a known contamination event (i.e., spill due to cracked or leaky product) to ensure that decontamination was properly carried out, 2) to investigate increased positivity rates, or 3) follow up on unusual results, such as the scenario outlined above.

PCR may be performed on a random sampling of uninoculated bottles to determine whether the issue is due to contamination of the blood culture media. If the contamination is high density, this may be useful; however if it is low density, then all bottles you test may still be negative. If the contamination is due to bacterial DNA, then Gram stain or culture will not be useful, hence the need for PCR. It is important to note that the presence non-viable organisms and/or nucleic acids (at levels that can be detected by PCR) is a known limitation noted in the package insert of some blood culture media and PCR manufacturers. If contamination is suspected, then immediately file a report with the manufacturer. Be sure to document lot numbers and expiration dates so that they may alert other customers.
The Conclusion

Human error contributes to the majority of discordant laboratory results. However, errors in interpretation and result entry/clerical errors are only part of the problem. Contamination events only complicate matters. If the test volume is significant, then the number of discordant results should be quickly realized, especially if there truly is a contamination issue. It is important to have a process in place to help reconcile contamination events as quickly as possible as they have the potential to majorly impact operations and patient care.

References

1. https://labmedicineblog.com/2018/02/20/when-rapid-blood-culture-identification-results-dont-correlate-part-1-clinical-correlation-needed/

-Raquel Martinez, PhD, D(ABMM), was named an ASCP 40 Under Forty TOP FIVE honoree for 2017. She is one of two System Directors of Clinical and Molecular Microbiology at Geisinger Health System in Danville, Pennsylvania. Her research interests focus on infectious disease diagnostics, specifically rapid molecular technologies for the detection of bloodstream and respiratory virus infections, and antimicrobial resistance, with the overall goal to improve patient outcomes.