Reply to “Not Over Yet: Fungal Infections following Methyl Prednisolone Injections Smoulder On”

REPLY

Molecular diagnosis of invasive fungal infections (IFI) is challenging, mainly due to the difficulties of efficient primary sample preparation and the lack of a gold-standard confirmation because of the low culturability from infection sites for many fungal species, such as Aspergillus fumigatus.

Inefficient nucleic acid extraction from primary samples exists in two aspects: first, the low circulating levels of pathogens or naked DNA for most IFIs, and second, the relatively large amount of human DNA compared to target fungal DNA present in the extracts may either inhibit downstream detection or cross-react with fungal primers/probes to cause false-positive results. Concerns of inefficient DNA extraction have been raised from Barnes and White in their commentary (1) on our recent study of fungal DNA detection in the blood of patients who received contaminated methylprednisolone injections (2). We were equally concerned about this issue, which is why we adopted the highly efficient MolYsis technology, composed of selective lysis of blood cells and pathogen enrichment steps (3), in combination with FastPrep mechanical lysis, to ensure a highly efficient DNA extraction. We were able to detect a lower limit of 200 fg of genomic targets from cell disruption using this approach.

It is not uncommon for broad-spectrum fungal PCR to be vulnerable to trace contamination from common assay sources, such as tubes, enzymes, and reagent. Therefore, a late-coming positive result (late threshold cycle [C_T] value) is often expected from negative controls in such detection. We agree with the authors of this commentary that investigation into the false positivity would be beneficial to ensure the specificity of the assay. In fact, we did test samples from both healthy volunteers and the EDTA blood collection tube controls, and even these samples generated panfungal CT values of >35. However, these high-C_T-value samples failed to amplify or generate a clean single band on agarose gel in internal transcribed spacer (ITS) PCR, possibly due to the bigger amplicon sizes (250 to ∼300 bp) relative to that (∼80 bp) produced by 28S panfungal real-time PCR, and thus sequencing was not performed.

The lack of clinical information is concerning, but our accessibility to such data was very much limited under the special situation where data collection was based on initiatives of physicians at local clinics. Ethical considerations precluded more involved clinical evaluations.

Finally, we greatly appreciate the sincere comments and suggestions from our colleagues. It is gratifying to see more fungal diagnosis experts and their teams being geared up for more extensive and in-depth studies to advance molecular diagnostics to benefit medical mycology.