ID Diagnostics: September 2018

A new microbroth dilution platform could facilitate routine antimicrobial susceptibility testing for anaerobic bacteria. Most microbiology labs don’t routinely perform antibiotic susceptibilities testing (AST) for anaerobic bacteria. However, anaerobes are becoming increasingly resistant to antimicrobials, including clindamycin, beta-lactam/beta-lactamase inhibitor combos, and even metronidazole, so anaerobic AST is likely to become increasingly relevant in the future.

Enter the Sensititre anaerobic MIC plate. Sensititre is ThermoFisher Scientific’s automated microbroth dilution system: essentially, you inoculate a 96-well plate that comes with loaded with serial dilutions of a bunch of antimicrobials you want to test, pop it into the machine, and then it incubates and then reads the plates for you. The authors (who declared no conflicts of interest and weren’t being funded by ThermoFisher) tested isolates from 56 invasive anaerobic infections using both the Sensititre anaerobic MIC plate and the bioMerieux’s ATB ANA test strips. Discordant results were refereed by a third test (either E-test of MIC test strip). The anaerobes were predominantly Bacteroides, but also included Fusobacterium, Finegoldia, Actinomyces, Cutibacterium (formerly Propionibacterium), and Clostridium species.

Agreement between the two AST methods was 90-100% for most bug-drug combinations; notable exceptions included gram-negative anaerobe (i.e. Bacteroides and Fusobacterium) susceptibilities to piperacillin and piperacillin-tazobactam (agreements of 70-75%), and for all anaerobes and chloramphenicol (agreements of 80-90%). Specifically, the Sensititre plate undercalled resistance to piperacillin and pip-tazo, mislabeling resistant isolates as sensitive in 13% of cases each. In every other case of disagreement, confirmatory testing supported the results of the Sensititre plate.

To summarize, the Sensititre anaerobic MIC plate, part of an automated AST system, has overall greater diagnostic accuracy than traditional methods, with the notable exception of piperacillin and piperacillin-tazobactam susceptibilities for gram-negative anaerobes. 30191340

 

Give the proletariat access to the B-D-glucan assay and they will overdiagnose fungal infections. In fellowship I first learned about B-D-glucan as a test for PJP pneumonia, because we see way more PJP than invasive fungal infection (IFI) at the county hospital, and honestly it’s a pretty good test for PJP (sensitivity 92% and specificity 86% using BAL as the gold standard; see 17426225). But as a test for IFI, particularly in a patient population with a low pretest probability of disease, B-D-glucan is less useful.

In this study, Fabre and colleagues report their experience with use of the B-D-glucan in patients at low risk for IFI (e.g. no history of hematologic stem cell or solid organ transplantation and no hematologic malignancy) at Johns Hopkins Hospital over a six-month period. Two ID physicians reviewed each chart and determined the appropriateness of the B-D-glucan order based on the patient’s clinical presentation and risk factors. Finally, the authors surveyed trainees to assess their knowledge about the test.

Forty percent of all B-D-glucan orders were performed in low-risk patients, and 96% were ordered by medicine services, most often to investigate a new respiratory process (e.g. pneumonia, pulmonary nodules, or mediastinal LAD) or new fever or leukocytosis. Half of the tests were deemed inappropriate, mostly due to lack IFI or PJP risk factors or lack of both risk factors and a consistent clinical presentation. Of the 56 positive B-D-glucan tests, 70% represented false positives, and these tests resulted in 39 days of unnecessary antifungal therapy. Inappropriate testing was estimated to cost $45,000 annually, not including the costs of inappropriate antifungal therapy and prolonged lengths of stay.

Most egregiously, the majority of trainees didn’t even know what they were testing for with the B-D-glucan assay: 56% believed it was a marker of Mucorales infection, only 63% and 54% recognized the test as a potential marker of Candida and Pneumocystis infections, and only 15% could correctly identify potential causes of a false-positive test result. And this is at one of the top internal medicine residency programs in the country! 30186888


Serum IL-8 in combination with either a lateral flow assay or PCR of BAL fluid is both highly sensitive and highly specific for invasive aspergillosis. This was prospective cohort study out of Austria that enrolled 106 adults with hematologic malignancy and suspected pulmonary infection. The patients received a bronchoscopy and serum testing: investigators measured various cytokine levels and performed an aspergillus-specific lateral flow assay and aspergillus PCR on both serum and BAL fluid. Invasive aspergillosis (IA) was graded as possible or probable using the standard consensus definitions.

Within the cohort, 11 patients had probable IA, 32 had possible IA, and 63 were classified as not having IA. When distinguishing probable IA from no IA, the most useful biomarker was serum IL-8; adding other cytokine measurements did not improve serum IL-8’s diagnostic accuracy further. When the BAL aspergillus lateral flow assay was added to serum IL-8 (e.g. a positive result being either or both tests returning positive), they had a 100% sensitivity and a specificity of 86%-94%, depending on how long the sample was left out. When the BAL aspergillus PCR was combined with serum IL-8, the sensitivity was 91% with a specificity of 97%. Serum IL-8 also showed potential to identify probable/possible IA versus no IA, with a sensitivity of 16% but a specificity of 98%.

Distinguishing possible IA from no IA remains challenging, but this study shows that serum IL-8 is probably a useful biomarker for IA and other invasive mold infections, and in combination with lateral flow or PCR assays from BAL fluid it provides near-perfect diagnostic accuracy for probable IA. This work brings us one step closer to a highly sensitive and specific biomarker panel for invasive mold infections. 29972764