Thread by @TylerLarsenMD

1/ Ok #medtwitter, here goes my first attempt at a #tweetorial, inspired by a recent question on wards from a learner I didn't know the full answer to:

“How good is a tuberculosis (TB) 'rule-out'?”

To try to answer this question we'll first start with a case.

2/ A 62yo female w/ recent renal transplant and remote hx of pulm TB s/p 1y DOT presents with fever. 4 wks PTA was hospitalized for 2 wks of cough, unintentional 20lbs weight loss & large LUL cavitary lesion on CT.

3/ Extensive work-up including bronchoscopy only reveals +human metapneumovirus. 4 AFB sputum smears and 3 MTB PCRs (including BAL) are negative. AFB cultures are NGTD. Patient is d/c on empiric posaconazole. 1 wk later she returns to ED with temp 102F. Cough is now resolved.

4/ Would you place the patient on airborne isolation and repeat at TB "rule-out"?

5/ To answer how "good", or sensitive, a TB "rule-out" is we first need to understand the sensitivity of the variety of diagnostic tests used to confirm or exclude pulmonary TB. The 3 most common in my practice setting are AFB sputum smear microscopy, MTB PCR, and AFB culture

6/ Before we discuss further we’ll pause for another poll.

Which of the following diagnostic tests is considered the gold standard for diagnosis of active pulmonary TB?

7/ The "normative process" of evaluation for active pulm TB in the US involves collecting 3 AFB sputum samples ~8 hours apart, with 1 being an early AM sample. These should be sent for smear microscopy and then AFB culture. doi.org/10.1093/cid/ci…

8/ Reported sensitivity of a single AFB sputum smear varies widely but is estimated at 50-80%. A second sample may increase the sensitivity by 5-11% but a third sample may only add another 2-3%.

The @IDSAInfo reports the overall sensitivity of three AFB smears at ~70%.

9/ Part of this is that smear results are highly variable. In general >10,000 bacilli are needed to be detected on LED microscopy and thus adequate sample volume of 10-15ml is required. Yields are higher in induced or early AM, or BAL samples, though these data are mixed

10/ Smears have notably lower sens in immunocompromised pts, particularly HIV given overall lower burden of disease. Given smears are relatively inexpensive and quick to obtain they are still recommended in the assessment for active disease.

11/ Thus three negative smears are reassuring that the patient PROBABLY doesn’t have active TB or if they do, they probably aren't very infectious. I think the best way to think about smear results is summarized in this table from @CDC. cdc.gov/tb/education/c…

12/ Interestingly, PCR/NAAT methods have better sensitivity of 66-96% but many guidelines note that a negative PCR, while helpful does not sufficiently exclude TB if the clinical suspicion is high enough.

13/ AFB culture is still considered the gold standard in diagnosis as it is the most sensitive ~90-95% (only need ~10 bacteria!) and specific ~99%. The major limitation is that AFB cultures grow slowly and test may take 4-6 weeks to result

14/ IGRAs are the diagnostic test of choice for latent TB infection (LTBI) and carry a sensitivity of 60-80% in active disease (About as good as 3 smears!). It is unclear how long it takes for an IGRA to turn positive but some studies guess between 4 and 12 weeks. PMID 21148232

15/ So, back to our case. Would you isolate? As with everything in medicine, it depends. Given a high clinical suspicion in the setting of immunocompromise, the patient was isolated until repeat BAL MTB PCR negative. Ultimately BAL grew A. fumigatus.

16/ So what did I learn?
* AFB smears are less a "Rule-out" and more a risk-stratification of infectiousness.
* If high clinical suspicion for TB, get the juiciest sputum you can (induced, AM, or BAL) and send for PCR if able
* AFB Culture remains the gold standard for diagnosis