A clinical risk score can predict invasive mold disease in patients with hematologic malignancy. The authors sought to develop a model that could determine individualized risk of invasive mold infections in patients with leukemia and lymphoma. They built their model using prospectively collected data from a large hematology center in Italy, which comprised 2187 patients treated over a 9-year period. The primary outcome was development of proven or probable invasive mold disease within 60 days of admission. Data collected included the nature of the underlying cancer and its treatment, the patient’s comorbidities, previous infections and their complications, and 19 candidate risk factors determined a prior based on a literature review. A multivariate analysis included all variables with a univariate p<0.1 and applied backward stepwise selection until all remaining variables had p<0.05. The final model was validated using n=100 bootstrap resampling of the training data set.
The authors analyzed a total 4127 admissions to construct the risk model. In the total cohort, the prevalence of invasive mold disease was 3.3% with a median onset of 22 days from admission. Invasive mold infections occurred more frequently in patients with AML, ALL, or CLL and during admissions for salvage chemotherapy or allogeneic HSCT; infections were less often seen in lymphoma and multiple myeloma, consolidation chemotherapy, and autologous HSCT. Of the 19 prespecified risk factors, 12 were associated with invasive mold disease with p<0.01; when entered into a multivariable model, the retained variables included a prior episode of invasive fungal infection (AOR 5.17; p<0.0001). receipt of high-dose steroids (AOR 2.21; p=0.001), uncontrolled malignancy (AOR 1.96; p=0.001), high-risk chemotherapy (AOR 2.2; p=0.002), PMNs <100 cells/mm^3 (AOR 4.56; P<0.001), and total lymphocytes <50 cells/ul (AOR 3.12; p<0.001).
These data were used to provide a nice normogram of the resulting clinical risk score, which assigns probabilities of invasive mold disease from 1% to 80% across a 45-point scale. While I don’t see this scoring system entering active clinical use (these scoring systems all take too much time to calculate), this is still a paper worth reading because it points out several factors associated with invasive mold disease, including several that are readily apparent from the patient’s admission note and complete blood count and can help inform your gestalt impression of a patient’s risk for invasive mold infection. 30974130
Add tularemia to the list of diseases that can be transmitted by solid organ transplantation. In July 2017, three recipients from a single donor (1 liver and 2 kidneys) developed fever and sepsis; one died. The donor, who lived in the southwestern US, had been hospitalized with acute alcohol withdrawal, pneumonia, and multiorgan failure. Ultimately, Francisella tularensis was identified from the recipients’ blood cultures and later recovered from rabbit carcasses near the donor’s home.
I have two thoughts about this case report (OK, three thoughts, the least worthy of which is: why the heck did this case report need 28 named authors and 61 collaborators?). First, we really ought to be cautious about organs from donors who died of medically attended illnesses with an unclear and potentially infectious etiology rather than, say, trauma or other noninfectious illness. Second, this is the sort of case that, while I’ll probably never encounter it in practice, is 100% going to show up on the ID boards in a couple of years. 30730826
Might macrolides help defang cryptococcus? This is an in vitro and mouse study, but I mention it because the topic – repurposing already-licensed antimicrobials – is readily clinically applicable. One of cryptococcus’s most important virulence factors is its thick polysaccharide capsule, which helps the organism resist innate immunity (i.e. opsonization by complement). The authors of this study hypothesized that macrolides, which have numerous non-antibiotic modulatory effects, might inhibit cryptococcal capsular formation and improve host immune responses. The authors grew C. gatti and C. neoformans in the presence of clarithromycin and azithromycin and found that both agents reduced the thicknesss of the cryptococcal capsule and the amount of capsular polysaccharides. Both C. gatti and C. neoformans showed increased susceptibility to H2O2 and opsonophagocytic killing by murine neutrophils; C. gatti also proved more readily phagocytized by murine macrophages. So, perhaps macrolides might have some role as virulence-attenuating agents in severe cryptococcal infections. 30936099