When Rapid Blood Culture Identification Results Don’t Correlate, Part 1: Clinical Correlation Needed

More and more laboratories perform rapid (i.e., multiplex PCR) blood culture identification. For the most part, it has been a wonderful addition to the laboratory workflow, not to mention the added benefits of provider satisfaction and improved patient care. Because the PCR only provides the organism identification (sometimes only to the family-level, i.e.; Enterobacteriaceae), laboratories must continue to culture the positive blood for definitive identification and/or antimicrobial susceptibility results. So what do you do when the results don’t correlate?

The Issue

From time to time, the PCR result is not going to correlate with the direct Gram stain or with the culture results. Although this is an issue one would fully anticipate, what do you do when this happens? Do you take some sort of action to arbitrate? Do you report the results as is?

First of all, the PCR assays do not detect all organisms. They only detect the most common bloodstream pathogens. Therefore, one should fully expect to observe cases in which the Gram stain would be positive, but the PCR results would be negative (scenario 1).  This is not a surprise.

Additionally, one should also assume that the PCR will occasionally detect organisms that were present at the lower limit of detection of the Gram stain. An example of this would be that the Gram stain is positive for one morphology (i.e.; Gram-positive cocci), but the PCR is positive for two organisms (i.e.; Staphylococcus and a Proteus species). Most of these cases tend to correlate with culture. In other words, although the second organism was not originally observed in the Gram stain, it was detected via PCR and then it also subsequently grew in culture (scenario 2).

Another type of discordant result laboratories sometimes experience is when the organism detected via PCR does not grow in culture for whatever reason. Similar to scenario 2 stated above, except that the culture is also negative for the second organism (scenario 3). Perhaps the patient was treated with antibiotics and the organism is no longer viable for culture? Perhaps a sampling or processing error was to blame?

The Solution

Depending on the scenario and how much work you want to do, you can either repeat testing or try an alternative method. Take scenario 2 for example. If the PCR detects two organisms and the Gram stain is only positive for one, then review of the original Gram stain is warranted. It is possible that the Gram-negative was somehow missed. Our eyes tend to go to the darker, more obvious structures. Perhaps the Gram-negative organism was faintly stained and it was overlooked? It is also possible that the Gram-positive is present in much lower numbers and only Gram-negative organism was originally observed. If the Gram stain result remains the same after review (only one organism observed), then there is nothing much left to do except to wait for the culture. That being said, an alternative method, such as acridine orange can be utilized in this type of scenario (two different cell morphologies). Acridine orange is a fluorescent stain that improves organism detection, as it is more sensitive than the Gram stain (1, 2).

If only the Proteus is growing (and the Staphylococcus isn’t from scenario 2) and we normally subculture positive blood to blood, chocolate, and MacConkey agars, then perhaps including an additional media that inhibits Gram-negative growth would be beneficial.

Scenario 3 can be a little more difficult to solve because you can’t make a non-viable organism grow. It just is what it is. [Spoiler alert: in next month’s blog I plan to write about when you should change your thinking from true-positive to false-positive.]

Regardless of why the result is discrepant, our laboratory appends a comment to the discordant result which says, “Clinical correlation needed.” This lets the clinician know that the results are abnormal and that they must use other relevant information to make a definitive diagnosis. In addition to the comment, we also make sure the discrepancy is notified to laboratory technical leadership (i.e.; Doctoral Director, Technical Lead/Specialist). This allows us to keep track of discrepancies as they may become important to know about in the future (see next month’s blog).

The Conclusion

In terms of organism detection, nucleic assays (i.e., NAATs) can provide superior sensitivity over antigen and culture-based methods of organism detection (i.e., sensitivity = PCR > culture > Gram). From the laboratory perspective, other potential benefits of utilizing nucleic acid detection methodologies include decreased TAT, simplified workflows, and reduced hands-on time. In terms of patient care, many have noted improved outcomes due to increased sensitivity and decreased time to result.

Although advances in technology can significantly improve analytical performance, they can also add complexity to the post-analytical process. Making sense of the results can sometimes lead to confusion. It is important to know the product’s limitations and what your risk(s) is. This should already be known and included in your Individualized Quality Control Plan (IQCP). Lastly, guiding the clinician to proper result interpretation is also important to maintain valuable patient care.

  1. Mirrett, S., Lauer, B.A., Miller, G.A., Reller, L.B. 1981. Comparison of Acridine Orange, Methylene Blue, and Gram Stains for Blood Cultures. J. Clin. Microbiol. 15(4): 562-566.
  2. Lauer, B.A., Reller, L.B., and Mirrett, S. 1981. Comparison of Acridine Orange and Gram Stains for Detection of Microorganisms in Cerebrospinal Fluid and Other Clinical Specimens. J. Clin. Microbiol. 14(2): 201-205.


Martinez Headshot-small 2017

-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.

One thought on “When Rapid Blood Culture Identification Results Don’t Correlate, Part 1: Clinical Correlation Needed”

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s