|Home | About | Journals | Submit | Contact Us | Français|
Ithank my colleagues Pappas et al. (1) for their comments on my Commentary in this journal (2). I did mention that “the steroids had been mostly used for epidural injection, largely causing fungal meningitis (91%)” (2), but to lump the end results into “intrathecal injection” may be a not-too-serious oversimplification. Whereas bacterial meningitis is an unusual complication of a spinal epidural abscess (3), inadvertent entry of a needle, intended for the epidural space, into the subdural/subarachnoid space is also known (4). Introducing a fungal nidus into the epidural space can also apparently quickly result in the same meningitic process (probably enhanced by coinjection of steroids), as would a direct intrathecal injection (4, 5). Drugs, such as the steroids in this instance, carefully injected into the epidural space may migrate only in that space, but the anatomic barrier that an enlarging granulomatous process (classically disrespectful of tissue planes) has to cross to gain access to the intrathecal sac is only a few millimeters, particularly in older patients (as in this outbreak) (6).
It is welcome to now see the aggregate susceptibility results on the outbreak isolates. These results are in line with the impression obtained from testing other dematiaceous fungi in vitro.
I believe that there are animal models for black molds; references 27 to 31 in my Commentary should be noted, particularly 28, 30, and 31, which refer to central nervous system (CNS) infections in such models (2).
Pappas et al. (1) expand on the differences between animal models, in general, and human disease. While there are differences, the models may still help to illuminate our way, and those authors do not cite an example of a CNS mycosis for which preclinical studies in animal models failed to predict the efficacy of the same therapies studied in humans. The authors quote (1) the penetration into cerebrospinal fluid (CSF) as a factor in the recommendation to utilize voriconazole in treatment, but drug penetration into CSF does not, in contrast, predict the efficacy of a therapy in humans (or in animals), as my references 17, 28, 30, 31, 37, 38, and 49 to 61, I feel, substantiate (2).
Reference is made (1) to a decision analysis model used to calculate the theoretical efficacy of different treatment strategies, but the model is unpublished, and we need to see the assumptions made in generating the model. Inserting concrete, but “soft,” numbers can make a model seem mathematically very precise, down to 10ths of a percent. High percentages are quoted (1) for the side effects of the drugs (even if used at usual doses), but how many of those side effects are clinically important and costly? One may hesitate about the model, since the central point is completely unknown: in how many infections would prophylaxis or empirical therapy have prevented progression? Information that could more easily be derived, and would be illuminating, is the cost of hospitalization and care for the 678 cases thus far, including the 44 deaths. How do those expenditures compare with prophylactic drug costs, therapeutic drug monitoring, etc.?