Thread by @FilaMentor

Next question!

1/
This is the first of 2 tweets! #GalactoMagic

40M with no PMHx presents to ED with 3 days of fever, cough, & SOB. Vitals: T 39C, HR 120, BP 130/80, O2 93% on 4L. Exam notable for R basilar rhonchi & crackles. CXR shows RLL opacity. He is started on ceftriaxone

2/

4 days later, he is still having intermittent fevers with Tmax 38.1C, O2 94% on 2L. He does not have dysuria or hematuria. The team orders a CT chest, which reveals a dense RLL consolidation. In addition to sputum cultures, which of the following tests should be ordered?

1/10
Here, the patient has ongoing fevers and hypoxia, with a CT that shows a RLL consolidation. In addition to sputum cultures, the patient should also have blood cultures obtained to complete the fever work up. There is no dysuria or hematuria to prompt the ordering of an UA

2/
With respect to the consolidation, the patient could still have a dense bacterial PNA. The patient has no PMHx and does not have any known risk factors for IFI. Your pretest probability for a fungal infection is very low. As such, there is no indication to send serum BDG or GM

3/
Why is your pretest probability low? Why aren’t BDG and GM indicated? Because the host's risks for a fungal infection is very low, and as of right now, the clinical presentation is not consistent with a fungal infection

4/
Fungi can cause infections depending on epi & host factors. Given challenges of culture-based diagnostics, serum biomarkers such as beta-d-glucan (BDG) & galactomannan (GM) can aid in the diagnosis. Performance of these biomarkers depend on organism, host factors & prevalence

5/
No diagnostic test is perfect. In order to use tests to our advantage, we need to know disease prevalence to calculate PPV & NPV. The disease prevalence, as well as clinical presentation drives pretest probability, which influences post-test likelihood of disease

6/
To help determine pretest probability, you can use this table with fungal infections & patient characteristics

7/
Once you get a feel for pre-test probability (probability of disease based on prevalence according to host risks & clinical presentation) of the fungal infection for your patient, you can see how this value impacts the post-test probabilities of serum BDG or GM

8/
You see this study & find that BDG>80 has a sens 64% & spec 92% for fungal infections. This helps determine the PPV/NPV at different prevalence rates

pubmed.ncbi.nlm.nih.gov/16080087

Based on above, @kierenamarr calculated PPV/NPV using above sens/spec with a range of prev values

9/
Graphs show that PPV can be heavily influenced by prev of disease in a population

At BDG 80, if prev of disease is 50%, PPV is 89%
➡️ A positive test is more likely to be truly positive when disease is probable

If prev of disease is 20%, PPV is 68%

pubmed.ncbi.nlm.nih.gov/16511777/

10/
It’s important to know pretest probability of a fungal infection (knowing pt characteristics & presentation) & test characteristics (analytic performance) of biomarkers

This site is helpful in understanding how disease prevalence & PPV/NPV are related
online.stat.psu.edu/stat507/lesson…