Microbiology

Tuberculosis Diagnosis for NEET PG: Smear, CBNAAT, Culture, IGRA and Mantoux Compared

Reflex · 20 Nov 2026 · 12 min read

Last updated: 20 Nov 2026

Lungs with an upper-lobe TB cavity linked to an inset of three acid-fast bacilli, tuberculosis diagnosis
On this page

Tuberculosis diagnosis is a sequence of tests, and each one answers a different question. A smear shows acid-fast bacilli, a nucleic acid amplification test (NAAT) such as CBNAAT names the organism and screens for rifampicin resistance, culture confirms it and allows susceptibility testing, and IGRA and the Mantoux test only show that the immune system has met TB.

This post compares them within India's NTEP pathway. Learn each test as a pattern: what it detects, how fast it reports, and what a negative result can and cannot exclude.

Key Takeaways
  • NTEP's upfront test for presumptive TB is a NAAT (CBNAAT or Truenat), which also reports rifampicin resistance. A positive smear, culture or NAAT/LPA each counts as microbiological confirmation.
  • ZN smear is quick but only 22-80% sensitive. NTEP grades it from 3+ (more than 10 AFB per field) to scanty (1-9 per 100 fields).
  • In adults with presumptive pulmonary TB, Xpert MTB/RIF has pooled sensitivity of 85% and specificity of 98% (Cochrane 2019), but only 67% in smear-negative, culture-positive TB. Rifampicin-resistance sensitivity is 96%.
  • Culture is the reference standard. It detects 10-100 CFU/mL but takes about 10 days in liquid systems and 18-24 days on LJ medium.
  • IGRA and Mantoux show TB infection, not active disease, and NTEP's Standards of TB Care do not recommend them for diagnosing active TB.
  • A negative smear or a negative CBNAAT never excludes TB on its own.

Tuberculosis Diagnosis in India: NAAT First, Then Confirmation

NTEP aims for microbiological confirmation in every patient, with rapid molecular tests used upfront wherever possible (NTEP principles of TB diagnosis). A NAAT, either CBNAAT or Truenat, is offered to all notified new patients. Key populations, people with extrapulmonary TB, people living with HIV (PLHIV) and children (after X-ray screening) go directly for NAAT.

A positive smear, culture or NAAT/line probe assay (LPA) each counts as microbiological confirmation. When the upfront NAAT detects TB, the second sample goes for LPA and liquid culture, and the NAAT also serves as the rifampicin susceptibility test.

WHO's tuberculosis fact sheet likewise makes rapid tests the initial tests for everyone with TB signs and symptoms, and these tests can give initial results within 48 hours of sample collection. WHO Module 3 on diagnosis (second edition, 3 November 2026) adds a near point-of-care test class, tongue swabs for people unable to produce sputum, and sputum pooling for settings with exceptional resource constraints.

Sample Collection and Probable TB

A good sputum sample is 2-5 mL of recently discharged bronchial material with minimal oral or nasopharyngeal material, collected in a sterile container after rinsing the mouth with clean water (NTEP sample guidance). The Standards of TB Care in India ask for two sputum samples, one early morning, and allow "probable TB" to be labelled and treated when confirmation is not possible but suspicion is strong.

Tuberculosis Diagnosis Tests Compared at a Glance

Each figure below is sourced in the section on that test.

TestWhat it detectsKey numbersLimit or role
ZN smearAcid-fast bacilli on a stained slideNeeds about 1,000-10,000 CFU/mL; sensitivity 22-80%Cannot separate M. tuberculosis from other mycobacteria; a negative smear does not exclude TB
CBNAAT (Xpert MTB/RIF)M. tuberculosis complex DNA and rpoB mutationsAdults with presumptive pulmonary TB: sensitivity 85%, specificity 98%; result in about 2 hoursNTEP upfront test; 67% sensitive if smear-negative, culture-positive
TruenatM. tuberculosis DNA (real-time micro-PCR on chips)Pooled figure not quotedNamed with CBNAAT as an upfront NAAT; MTB-RIF chip tests rifampicin resistance
Culture, LJ (solid)Growth of M. tuberculosis complexDetects 10-100 CFU/mL; colonies in 18-24 daysReference standard; species identification and susceptibility testing
Culture, liquid (MGIT)Growth in brothAbout 10 days (automated liquid systems)More sensitive than LJ in one meta-analysis
IGRAInterferon-gamma response to ESAT-6 and CFP-10T-SPOT.TB in osteoarticular TB: sensitivity 0.84, specificity 0.78Infection only; no BCG cross-reaction
Mantoux (TST)Skin reaction to tuberculin PPDCut-off depends on the patient groupBCG and non-tuberculous mycobacteria (NTM) give false positives; infection only

Memory aid (written for this post, not a published mnemonic): smear sees, NAAT names, culture clarifies, IGRA infers (as Mantoux does).

ZN Stain and Smear Microscopy

The Ziehl-Neelsen (ZN) stain uses carbol fuchsin with heat and phenol, then a methylene blue counterstain. Carbol fuchsin forms stable complexes with the mycolic acids of the lipid-rich mycobacterial wall, and these resist destaining with mineral acids, so the bacilli stay red on a blue background (NTEP ZN microscopy page). The Gram staining post shows how wall chemistry decides what a stain reveals, and ZN applies the same logic. A 2016 review by Caulfield and Wengenack adds that auramine-rhodamine fluorescence is more sensitive and faster to read, and that WHO has endorsed phasing out conventional ZN microscopy in its favour.

Smear needs roughly 1,000-10,000 CFU/mL for reliable detection, against 10-100 CFU/mL for culture. Overall sensitivity is 22-80%, depending on bacterial burden, stain and technician experience (same review). The positive predictive value for mycobacteria is above 95%, but a positive smear shows acid-fast organisms, not necessarily M. tuberculosis.

NTEP Smear Grading

NTEP's smear grading page, which cites the 2005 RNTCP laboratory module, sets this scale.

GradeAFB seenFields to examine
3+More than 10 per oil-immersion field20
2+1-10 per field50
1+10-99 per 100 fields100
Scanty1-9 per 100 fields (record the exact number)100
NegativeNone in 100 fields100

To remember it, note that fewer bacilli mean more fields to read: the counts run 20, 50, then 100 for every lower grade.

CBNAAT and Truenat in Tuberculosis Diagnosis

How CBNAAT Works

CBNAAT (Xpert MTB/RIF) is a fully automated cartridge-based real-time PCR test. According to NTEP, bacilli from sputum are lysed by sonication, DNA is amplified and fluorescent molecular-beacon probes detect mutations, giving a result in about 2 hours. NTEP recommends it upfront for PLHIV, children, presumptive DR-TB and privately notified patients.

The target is the rpoB gene, which encodes the RNA polymerase beta subunit. The probes cover an 81-base-pair region where mutations account for about 96% of rifampicin resistance (Caulfield review). NTEP's CBNAAT interpretation page describes a probe that drops out completely, or a Ct difference above 4 between the five probes, as signalling resistance.

Accuracy in the 2019 Cochrane Review

The Cochrane review by Horne and colleagues pooled Xpert MTB/RIF against culture in adults with presumptive pulmonary TB:

  • Unselected adults: sensitivity 85% (95% credible interval 82-88) and specificity 98% (97-98), from 70 studies and 37,237 participants.
  • By smear status: sensitivity 98% in smear-positive patients and 67% (62-72) in smear-negative, culture-positive patients.
  • By HIV status: sensitivity 88% in HIV-negative and 81% in HIV-positive patients.
  • Rifampicin resistance: sensitivity 96% (94-97) and specificity 98% (98-99), from 48 studies and 8,020 participants, with high-certainty evidence.

These 2019 pooled values are reference figures, not the latest estimates. The review covered adults and respiratory specimens only, so do not assume they apply to children or extrapulmonary specimens.

Xpert Ultra and the Trace Result

Ultra adds two multicopy targets, IS1081 and IS6110, and uses the single-copy rpoB gene only to detect rifampicin resistance. Its detection limit is 15.6 CFU/mL against 115 CFU/mL for the previous Xpert, but the new feature and the trace category reduced specificity, according to a 2022 study of trace calls.

In the Cochrane review's single head-to-head study (moderate-certainty evidence), Ultra had sensitivity of 88% against 83% for Xpert, and specificity of 96% against 98%. Accuracy for rifampicin resistance was similar (sensitivity 95%, specificity 98% for both).

A trace result means a very small amount of M. tuberculosis DNA was detected through IS6110 and IS1081 while rpoB was not, so it gives no rifampicin result, and further investigations are needed to confirm or exclude resistance. As the same study summarises WHO's position, a trace counts as bacteriological confirmation in PLHIV, children, extrapulmonary TB and adults with no HIV risk and no anti-TB treatment in the last 5 years. In HIV-negative (or not-at-risk) symptomatic adults whose treatment was completed less than 5 years ago, a trace may reflect non-viable bacilli, so clinical judgment decides.

Truenat

NTEP names CBNAAT and Truenat together as the upfront NAAT. According to NTEP's Truenat page, Truenat is real-time PCR built into micro-PCR chips, with a separate MTB-RIF chip as the reflex test for rifampicin resistance. Treat it as a NAAT with the same role as CBNAAT; the accuracy figures in this post belong to Xpert.

Culture, LPA and Drug Susceptibility Testing

Culture is the reference standard for NAAT accuracy, and the Caulfield review stresses that it should accompany smear and NAAT testing for species identification and susceptibility testing.

Speed is the weakness. Colonies appear in about 18-24 days on solid LJ medium, automated liquid systems average about 10 days, and cultures are usually held for 6-8 weeks before a negative is reported. A meta-analysis cited in the review found MGIT analytical sensitivity of 81.5% and specificity of 99.6%, against 67% sensitivity for LJ. The culture media post covers LJ medium alongside other media.

Line Probe Assay Under NTEP

The NTEP LPA page places LPA in the drug-resistance cascade:

  • If NAAT detects rifampicin resistance, offer both first-line (FL) and second-line (SL) LPA.
  • If NAAT does not detect rifampicin resistance, offer FL-LPA for isoniazid resistance. Do SL-LPA only if isoniazid resistance is found, to check fluoroquinolone resistance.
  • Direct LPA is done on smear-positive specimens. Indirect LPA is done on culture isolates when the smear is negative.

The antitubercular drugs post explains the drugs behind these resistance labels.

IGRA and Mantoux Test: Detecting TB Infection, Not Disease

Both tests measure immune sensitisation to TB, not active disease: the Standards of TB Care in India state that IGRA and the tuberculin skin test (TST) are not recommended for diagnosing active TB, and WHO uses them to identify TB infection and decide who benefits from preventive treatment (fact sheet above).

FeatureMantoux (TST)IGRA
What is measuredSkin reaction to tuberculin purified protein derivative (PPD)Interferon-gamma released by T cells stimulated with ESAT-6 and CFP-10
FormatIntradermal injection; induration read at 48-72 hoursBlood test: QuantiFERON-TB Gold In-Tube (amount of interferon-gamma) or T-SPOT.TB (count of responding cells)
BCG and NTMFalse positives after BCG vaccination or NTM infectionNo cross-reaction with BCG strains or most NTM
Active versus latent TBCannot separateCannot separate

The CDC tuberculin skin test page supports the injection route, the 48-72 hour reading and the BCG and NTM false positives. The IGRA details come from a 2022 meta-analysis. Across seven T-SPOT.TB studies in osteoarticular TB it found sensitivity of 0.84 and specificity of 0.78, with a positive likelihood ratio of 3.9 and a negative likelihood ratio of 0.21. One included study reported inconclusive results, and the authors call the test a suboptimal discriminator.

Mantoux cut-offs differ between populations, so tie the cut-off to the patient group in the question. The NTEP paediatric TB guideline draft (Release 1, March 2022) reports 10 mm or more as positive and says 5 mm may be taken as the cut-off in HIV co-infection, using 2 TU of PPD RT23. It adds that the original RT23 lot is currently unavailable, so these cut-offs may not transfer to other PPD products. WHO's tests for TB infection module (September 2022) says IGRAs and newer antigen-based skin tests (TBSTs) use M. tuberculosis complex-specific antigens, a significant advancement over TST.

Other Tests: LF-LAM and Non-Respiratory Samples

LF-LAM is a urine point-of-care test for people living with HIV. Under WHO's 2019 policy update, inpatients qualify (strong recommendation) with TB signs or symptoms, advanced HIV disease, serious illness, or CD4 below 200 cells/mm3. For outpatients WHO only suggests it (a weaker recommendation), with signs or symptoms, serious illness, or CD4 below 100 irrespective of symptoms, and recommends against it in asymptomatic outpatients with unknown CD4 or CD4 of 100 or more.

For extrapulmonary TB, NTEP's Standard 3 sends specimens from the presumed site for microscopy, culture, DST, CBNAAT or histopathology. Standard 6 asks for respiratory specimens from children with presumptive intrathoracic TB, preferably tested by CBNAAT, smear or culture. WHO's April 2025 update recommends concurrent respiratory and non-respiratory testing in people of all ages living with HIV and in children without HIV or with unknown HIV status.

Common Mistakes When Reading Tuberculosis Diagnosis Results

  • Ruling out TB on a negative smear. Smear sensitivity is 22-80%, and Xpert detected only 67% of smear-negative, culture-positive cases.
  • Ruling out TB on a negative CBNAAT. Pooled Xpert sensitivity is 85% overall and 81% in HIV-positive patients, and NTEP's Standard 5 allows probable TB to be treated when suspicion stays strong.
  • Calling a positive smear M. tuberculosis. It shows acid-fast organisms; NAAT or culture names the organism.
  • Using Mantoux or IGRA to diagnose active TB. Neither separates latent from active TB, and NTEP's Standards of TB Care do not recommend them for diagnosing active TB.
  • Treating Ultra as simply better than Xpert. In the single head-to-head study it was more sensitive but less specific, and a trace result carries no rifampicin result.

Choosing a Tuberculosis Diagnosis Test in a Clinical Stem

When a question gives you a patient and a list of tests, work through four steps:

  1. Infection or disease? Mantoux and IGRA answer infection. Smear, NAAT and culture answer active disease.
  2. What does the stem need? A fast organism-plus-rifampicin answer points to CBNAAT. Species identification and full susceptibility testing point to culture and LPA.
  3. Who is the patient? A child or a person with HIV points to CBNAAT first, with non-respiratory samples tested alongside. An inpatient with HIV and advanced disease brings in urine LF-LAM.
  4. Does the result fit the suspicion? A negative smear or CBNAAT cannot exclude TB under strong suspicion, and culture takes weeks.

Tuberculosis Diagnosis: Putting the Tests Together

Under NTEP, tuberculosis diagnosis starts with a NAAT that names the organism and screens for rifampicin resistance, and CBNAAT reports in about two hours. Smear is quick but misses many cases, culture is the slow reference standard, and Mantoux and IGRA speak only to infection. Practise the comparison with the microbiology bacteriology practice questions, and redraw the comparison table from memory until each test's detection limit, speed and role come easily.

Practise this on Reflex

Turn what you just read into recall with 14 years of tagged PYQs.

FAQ

Frequently asked questions

The questions aspirants ask most about this topic.

Under NTEP the upfront test is a nucleic acid amplification test (NAAT), either CBNAAT or Truenat, offered to all notified new patients. It detects M. tuberculosis and screens for rifampicin resistance, and CBNAAT gives a result in about 2 hours. A positive smear, culture or NAAT/LPA each counts as microbiological confirmation.

No. In the 2019 Cochrane review of adults with presumptive pulmonary TB, pooled Xpert MTB/RIF sensitivity was 85% overall, 81% in HIV-positive patients and 67% in smear-negative, culture-positive patients, with specificity of 98%. A negative result with strong clinical suspicion is not an exclusion, and culture remains the reference standard.

3+ is more than 10 AFB per oil-immersion field (20 fields examined), 2+ is 1-10 per field (50 fields), 1+ is 10-99 per 100 fields, and scanty is 1-9 per 100 fields with the exact number recorded. Negative means no AFB in 100 fields. The fewer the bacilli, the more fields are read.

No. Both show TB infection and cannot separate latent from active TB, and the Standards of TB Care in India do not recommend them for diagnosing active TB. IGRA uses ESAT-6 and CFP-10, so it does not cross-react with BCG strains or most non-tuberculous mycobacteria, whereas Mantoux can be falsely positive after BCG vaccination or infection with those mycobacteria.

Colonies appear in about 18-24 days on solid LJ medium, and automated liquid systems average about 10 days. Cultures are usually held for 6-8 weeks before a negative is reported. In return, culture detects as few as 10-100 CFU/mL and allows species identification and susceptibility testing.

Stay ahead in your preparation

Get expert tips, exam updates, and high-yield insights delivered straight to your inbox.

No spam. Unsubscribe anytime.

More in Microbiology

View all →