PCR really is quite magical. You take a few drops of blood from a febrile patient and within a few hours, you can have an answer of the type of bacterium causing your “presumed noninfectious SIRS,” or, perhaps, the cause of febrile neutropenia. It may even be higher yield than good old fashioned blood cultures (x2, of course). Unfortunately, today’s article on PCR is, yet again, more “look what I can do” rather than “look at the improved outcomes!”
At a single tertiary care center for pediatrics, 150 MagicPlex PCR studies from 89 patients aged 2-12 years young were performed for various reasons (suspected sepsis, 55%, febrile neutropenia 25%, localized infection 12%, FUO 6%, undefined 3%). None of these patients were tested in the ED, and 149 of the 150 tests had antibiotics given before testing. 34% of patients were heme/onc patients, 17% had a stem cell transplant, 17% were ICU players, and 21% had no chronic disease. Concurrent blood cultures were drawn.
There were 60 blood stream infections identified, 5 of which were by blood culture only, 32 of which were by PCR only. For this, the authors claim a 143% increase in yield. All is not roses though, as this increase comes at the behest of 12 additional contaminants by PCR. They report their contaminant numbers are consistent with prior findings – about 90% sensitivity, 70% specific. Just under 10% of these tests led to a false positive.
Specific for what is a better question. Who is the ideal candidate for this testing? Ill appearing ICU players without a known reason? Sure. On immunosuppressants? Asplenics? transplant patients? I guess its reasonable if you dont have a source. Even in AIDS patients, pneumonia is typically common causes.
They tout, “As a matter of fact, in 24 cases out of 60, MP allowed a quick start of an appropriate, goal-directed antibiotic therapy, modifying the previous empiric regimen.” Really? An explanation of sorts would have been nice here, but they fail to provide a one liner about patient history that may have clued them in. Oh, and for every two saves, you’ll have one contaminant that you’ll jump on, bill for critical care, and start the patient on a path of antibiotics and aggressive care that they didnt need. Or set you up for premature closure. That, and magicplex missed the one Neisseria meningitidis case.
My fear is that **everyone** will get it – I mean, we can barely get providers to stop sending blood cultures on well appearing cellulitis. What happens to the pyelo without comorbidities that was on the border of getting sent home, but now gets MagicPlex, and someone decides that, despite having a transient bacteremia that was sensitive to the prescribed Levaquin, now this patient who likely could have been fine as an outpatient because they had a pan-sensitive bug now gets admitted, without measurable benefit. We still have not quite figured out when yield would be highest with magicplex. Ideally, it would be high yield in the patients who are quite ill – the UA without nitrite/leuks with 5-10 wbcs, or UA/CXR/cUS negative patient, or the febrile dialysis patient with a question of a line infection. Or the asplenic or transplant patient. But if their low yield in these situations, who cares?
As previously documented here, there is still no improvement in mortality, despite magical PCR testing. I fear that this will be the next widely overused test, that has potential to result in MUCH higher costs for the system from overdiagnosis, with currently NO added benefit.
PS – previous MagicPlex in the ED performed awful.