You don't need to be signed in to read BMJ Blogs, but you can register here to receive updates about other BMJ products and services via our site.


Primary Care Corner with Geoffrey Modest MD: Another Concerning LTBI IGRA Article

17 Oct, 16 | by EBM

By Dr. Geoffrey Modest

Given the recent article of the USPSTF LTBI (latent TB infection) recommendations ( ) I decided to bring up another study which questions the role of  IGRAs (interferon-γ release assays) –see  King TC. Am J Respir Care 2015; 192(3): 367. I initially decided that the prior articles and blogs were pretty concerning about the stability of IGRAs, but this study does add another wrinkle: it was an industry-sponsored study, still had pretty striking conclusions that 1 in 5 people with positive IGRAs would revert to negative, yet then concluded that T-SPOT.TB (one of the IGRAs) is “an accurate and reliable way to screen healthcare workers”.


  • 20,095 pairs of T-SPOT.TBresults from 16,412 health care workers (HCWs) at 19 geographically diverse US hospitals (urban, suburban, rural), looking at serial T-SPOT.TB  testing, from 2010 to 2014, where the second test result was at least 150 days after the first one.
  • 77% female, median age 41 and no other demographics included
  • Their analysis excluded those with borderline tests from either the initial or subsequent test, approx 2.3% of the results
  • Conversion was defined as a negative result turning positive; reversion was the opposite


  • The mean conversion rate was 0.8%, but 2.3% in hospitals that checked only high risk HCWs, and 0.6% in hospitals that checked all HCWs
  • The mean reversion rate was 17.6%, but 13.9% in hospitals that checked only high risk HCWs, and 20.7% in those that checked all HCWs
  • The baseline test positivity rate was 2.3%, but varied from 0-27.4%; and was 8.4% in hospitals that checked only high risk
  • There was a strong correlation between observed conversion rates by T-SPOT and reported TB incidence of the state where the hospital was located
  • 23% of those with borderline results were positive on retesting


  • Of course, one of the primary issues with LTBI screening is that there is no gold standard.
  • This article does cite 5 studies which challenge the IGRA’s stability and reliability.
  • The researchers think their test was better (again, remember that this is an industry-sponsored study), and had lower reversion rates than other studies because: they had higher and more rigorous cutpoints than prior T-SPOT tests (though, without a gold standard, hard really to know the effect of this on the sensitivity of the test), they have lower likelihood of sample processing errors than QuantiFERON IGRAs, but they do note that all IGRAs, which measure an immune response to TB, are subject to “biological variability”. It is also unknown in this study whether the patients had been treated or not, which other studies have suggested might be part of the reason for reversions (though the authors did not comment on this).
  • One really interesting note is that this industry-sponsored study was able to state on the one hand that though >20% of the HCWs who had a positive IGRA actually reverted to negative about 1 year later, on the other hand they concluded that the IGRA is an accurate/reliable test. It seem to me as a casual observer that the possibility of 20% false positives, which they acknowledge as a possibility, is a tad high. And any positive can lead to unnecessary but potentially toxic drugs, potential stigma, etc. So, as mentioned in prior blogs, I am sticking with the PPD, except in unusual circumstances (e.g., the patient really cannot get the PPD read in 48-72 hours).


Primary Care Corner with Geoffrey Modest MD: USPSTF LTBI Screening Recommendations

11 Oct, 16 | by EBM

By Dr. Geoffrey Modest

The USPSTF just came out with their recommendations for screening for LTBI (latent TB infection) in populations at increased risk, giving it a “B” Grade (a recommended service, with “high certainty that the net benefit is moderate, or there is moderate certainty that the net benefit is moderate to substantial”  (see ).



  • The Natl Health and Nutrition Examination Survey in 2011-12 found a national prevalence of LTBI of7% of the US population and 20.5% in those foreign-born
  • 5-10% of those with LTBI progress to active TB during their lifetime


  • Benefits of screening
    • No eligible studies were identified showing that targeted screening for LTBI in primary care settings in asymptomatic adults improves quality of life or reduces active TB/transmission/mortality
    • Accuracy of tests:
      • Pooled estimates for sensitivity/specificity of TST (tuberculin skin testing)depends on degree of induration
        • 5 mm: sensitivity 0.79, specificity 0.30-0.97
        • 10 mm: sensitivity 0.79, specificity 0.97
        • 15 mm: sensitivity 0.52, specificity 0.99
      • Pooled estimates for sensitivity/specificity of IGRAs (interferon-γ release assays), for different tests
        • T-SPOT: sensitivity 0.90, specificity 0.95
        • QuantiFERON-Gold: sensitivity 0.77, specificity 0.98
        • QuantiFERON-Gold In-Tube: sensitivity 0.80, specificity 0.97
      • No studies were identified that evaluated sequential screeningstrategies of using both TST and IGRAs in asymptomatic patients
    • Benefits of treatment: does CDC-recommended regimes improve quality of life/reduce progression or transmission of TB?
      • INH for 6 months va placebo leads to a relative risk reduction of 0.35 (1.4% in placebo group over 5 yrs, vs 0.5% with INH)
      • Rifampin for 4 months is equivalent to INH for 9 months
      • Once-weekly rifapentine plus INHx3 months (directly-observed therapy) vs INH for 9 months were also statistically equivalent
    • Harms of screening: no eligible studies
    • Harms of treatment:
      • INH hepatitis
        • Hepatitis incidence rate vs placebo: INH had RR=4.59 for 24 weeks (Number-Needed-To-Harm, NNH=279) and RR=6.21 for 52 weeks
        • Mortality rate vs placebo: INH had RR=2.35 (NNH= 6947, 0.14/1000 persons treated)
      • INH discontinuance rate from adverse events: 1.8% vs 1.2% on placebo (mostly GI distress)
      • Rifampin (in 3 studies comparing Rifampin to INH): for INH in the 3 studies, hepatotoxicity in 5.2%, 3.7% and 11.4%; for Rifampin: 0%, 0.7%, and 4.4%: ie, pooled INH RR of 3.29 vs Rifampin
      • Also <1/2 the rate of discontinuations with Rifampin vs INH


  • The USPSTF recommendation does appropriately highlight the important issue of checking TB status. There are increasing numbers of high risk people in this country, especially in foreign-born. And I have seen several untreated people develop active TB late in life, after being here for decades.
  • I do have concerns about the reliability and stability of the IGRAs, as noted in prior blogs (see, which also goes into more detail on identifying who are high-risk of LTBI. But most importantly, it reviews a couple of studies finding quite remarkable inconsistency of IGRAs: many people who were initially positive but subsequently negative on repeat testing within a couple of months and without treatment, i.e. high and unexplained “reversion” rates
  • In terms of treatment regimens, the CDC-recommended ones overall are the INH for either 6 or 9 months (though the 9 month is preferred given somewhat better results) or the rifampin for 4 months (reviewed in Over the past several years, I have been prescribing only the rifampin one (unless there are difficult drug-drug interactions), without a problem. easier/shorter regimen and better tolerated
  • Though I realize the issues of anecdotal medicine, I will still bring up one of my long-term patients who reinforced 2 of the above points to me: he is a 60 yo Haitian man I have known for several decades with really bad uncontrolled diabetes, hypertension, HIV, and now on dialysis (of those, his HIV has been consistently well controlled, with CD4 in the 250-300 range and viral load consistently suppressed). He worked for decades as a health aide in a nursing home and had repeatedly negative annual PPDs (about as good as one can get for accuracy of PPD screening). He was being evaluated by a transplant physician who insisted that he go to ID/HIV clinic to be cleared for surgery. They did a QuantiFERON-Gold test which came back positive for TB. He was started on INH, but within 2 months developed a nearly fatal case of INH hepatitis leading to a many week stay in an ICU. He recovered, but I decided to check the QuantiFERON-Gold again, which turned out to be negative……
  • So, my bottom line: I still rely on TST screening for TB, based on that prior blog and reinforced by my anecdotal case, though I understand the appeal of the  IGRAs (I had had assumed I would be using IGRAs in my patient population, which is largely foreign-born and has had a high rate of prior BCG vaccination, since it made sense mechanistically that it would be more accurate than PPDs and there is no need to come back in 48-72 hours for a reading, no need for booster PPDs in those who had not had one for a long time, etc.) And, I do find the rifampin based treatment of LTBI to be easy and well-tolerated, though with forewarning about red urine….

For the CDC recommendations on PPD screening, see )

EBM blog homepage

Evidence-Based Medicine blog

Analysis and discussion of developments in Evidence-Based Medicine Visit site

Creative Comms logo

Latest from Evidence-Based Medicine

Latest from EBM