How is pulmonary TB diagnosed in Korea, and when can you lift isolation?

Published 2026-09-14 · Updated 2026-09-14 · Last checked against sources 2026-09-14

Before I looked it up

It has come down a lot these days, but Korea was once a country where tuberculosis was so rampant that there used to be a separate department called the ‘tuberculosis department’. Even now, when people who had tuberculosis before, or people like the homeless, come to the hospital and get tested, it is not uncommon for tuberculosis to be diagnosed. Our ICU has four isolation rooms (two of them negative-pressure), and a patient with suspected or diagnosed tuberculosis ends up in one of these beds. So every time a tuberculosis patient comes in, what to test and when, and when the isolation can be lifted, gets in the way, so I organised it once.

Recently, with tuberculosis patients, we use a combination tablet called ‘Tubis’, which has made taking the medication considerably easier, but still, having to give it for a long period cannot be helped…

What I found

The note I started from was a diagnosis-and-schedule note built on the 5th edition of the Korean tuberculosis guideline. Going back to the guideline itself, the skeleton held but several numbers did not. Three things first.

• Isolation comes off after at least 2 weeks of effective treatment, clinical improvement, and — if the patient was smear-positive — three consecutive negative follow-up smears (at least 8 hours apart, at least one a morning specimen) or a negative culture. My note said “two consecutive negatives”. The guideline and the national programme guideline both say three, or one negative culture.
• Xpert MTB/RIF (or Ultra) goes out with the first sputum, together with smear and culture, in every suspected case (grade IA). That is the headline change of the 5th edition — TB detection and rifampicin resistance in about 2 hours. In Korea, rifampicin resistance on Xpert is treated as MDR until susceptibility testing says otherwise, because isolated rifampicin resistance is rare here.
• The 2-month visit is the fork. Sputum smear and culture, chest film, and the conventional susceptibility result all land there. Isoniazid and rifampicin susceptible → ethambutol can be stopped. Cavity at baseline plus a positive 2-month culture → consider extending the continuation phase.

The rest of what I found, in order.

• Sputum: at least two, ideally three specimens — not “three consecutive mornings”. The first is collected on the spot; the rest at home before breakfast. The first specimen catches 83–87%, the second adds 10–12%, the third 3–5%.
• Smear sensitivity is 25–80% against culture. What it buys is speed (a positive result is reported within 24 hours) and a fast read on who is infectious. Liquid culture (MGIT) takes about 2 weeks, solid media 3–8 weeks; both are inoculated.
• Baseline before treatment (guideline Table 5): sputum smear and culture, Xpert, rapid susceptibility test, chest film, weight, symptoms and adherence, vision, CBC, liver and renal function. If indicated: HIV (risk factors present — the Korean guideline does not test everyone), hepatitis B and C, diabetes screen. Uric acid is not on the list.
• Liver, blood and renal tests after baseline are “as needed”. There is no fixed interval in Table 5. What there is instead is a threshold: ALT five times the upper limit with no symptoms, or three times with symptoms of hepatitis → stop the hepatotoxic drugs at once (IIIA).
• Vision is checked at baseline and at 1 month, then as needed — not “monthly”. The instruction is to tell the patient to report any change immediately, especially past 2 months of ethambutol.
• Monthly sputum until two consecutive negative smear-and-culture results in anyone who was culture- or smear-positive, and a last sputum at the end of treatment (IIIA). Culture still positive past 3 months → repeat susceptibility testing.
• End of treatment: chest film plus a post-TB lung disease assessment in every patient (IIIA). Dyspnoea or a significant residual lesion on imaging → pulmonary function testing at the end of treatment or within 6 months. New in the 5th edition.
• MDR-TB with fluoroquinolone susceptibility: 6-month BPaLM or 9-month MDR-END (IA), with no preference between them. The 18–20-month regimen is for patients who fit neither. MDR-END is a fixed regimen — levofloxacin, delamanid, linezolid, pyrazinamide — not an “individualised combination” as my note had it. And the definition of XDR has changed: it is fluoroquinolone plus a group A drug (bedaquiline or linezolid), not injectables.
• Where things stand in September 2026: the 5th edition (January 2024) is current — there is no 6th. The national programme guideline has a 2026 edition (January 2026) and the isolation-release criteria are unchanged from 2025. WHO Module 4 has a 2025 edition that adds a 6-month BDLLfxC regimen; the ATS/CDC/ERS/IDSA 2025 guideline conditionally recommends a 4-month HPMZ regimen for drug-susceptible TB. Neither is in the Korean 5th edition yet.

What I was reading

The Korean Guidelines for Tuberculosis, 5th edition (2024), issued by the Korea Disease Control and Prevention Agency with the Korean Academy of Tuberculosis and Respiratory Diseases — the full 430-page document. The 2025 and 2026 National Tuberculosis Management Guidelines from the same agency, for the isolation-order and work-restriction criteria. Summaries of the WHO consolidated guidelines Module 4 (2025 edition) and the ATS/CDC/ERS/IDSA 2025 update, only where they differ from the Korean guideline. Every number and threshold below was checked against the 5th-edition text; the international guidelines are cited only for the differences. What follows is what I read in September 2026.

0. How far “a country where TB was rampant” has come

Evidence: national notification statistics (Public Health Weekly Report 2026;19(16); 5th-edition preface).

• Notified TB patients: 39,557 in 2011 (50,491 on the incidence definition) → 16,264 in 2022 → 17,070 in 2025 (33.5 per 100,000), of whom 13,713 were new patients (26.9 per 100,000). Down 66% from the 2011 peak.
• Still second-highest incidence among 38 OECD countries (35.0 per 100,000). And patients aged 65 and over are 62.5% of the total (101.5 per 100,000 — 6.4 times the rate under 65); that group grew 1.3% in a year when the total fell.
• MDR in retreatment cases 5.5% versus 1.9% in new patients (2.9-fold). This is why “had TB before” means thinking about resistance from the first specimen.

1. Diagnostic criteria

Strength: 5th-edition recommendations, grades as given.

Confirmed — organism detected

• A positive culture is the most important confirmatory test. Confirmation is not complete until M. tuberculosis has been separated from non-tuberculous mycobacteria — the NTM share of positive AFB cultures in Korea is rising.
• A positive nucleic acid amplification test (TB-PCR, Xpert) counts as detection. In smear-positive sputum, sensitivity and specificity are 90–100%; in smear-negative sputum they drop (Xpert 68%, Ultra 77.5%). False positives exist, so the test is not ordered when TB is not suspected. Ultra can be falsely positive within 5 years of a previous treated episode.

Smear- and culture-negative TB — clinical diagnosis

Active TB can be diagnosed and treated without an organism, on the combination of:

• Symptoms — cough for 2–3 weeks or more without an obvious cause is the trigger to test (IIIA); fever, weight loss, night sweats.
• Imaging — upper-lobe infiltrate, cavity, miliary pattern. A chest film alone does not diagnose TB (ID); compare with previous films (IIA).
• Histology — caseating granuloma (also seen in NTM, fungi, brucellosis, syphilis).
• Response — in culture-negative disease, radiographic improvement after 1–2 months of empirical treatment supports the diagnosis.

2. The tests

Strength: 5th-edition recommendations, grades as given; performance figures from the meta-analyses the guideline cites.

Imaging

• Chest film: the first test; cannot confirm (ID).
• Chest CT: when the film cannot settle activity in a smear-negative patient, or when the differential is difficult (IIIB). What the film commonly misses: hilar and mediastinal nodes, mild infiltrates, active upper-lobe disease hidden in scar, faint miliary disease.

Microbiology — at least two, ideally three sputum specimens (IA)

• Collection: first specimen on the spot, in a ventilated sputum booth (or outdoors) with the collector in an N95; the others at home before breakfast. Two spot specimens performed no worse than the traditional set in recent studies, which is why the minimum became two.
• Smear: positive results reported within 24 hours (IIIA). Sensitivity 25–80% against culture. Ziehl-Neelsen or Kinyoun; fluorescence staining adds about 10 percentage points and is recommended for high-volume laboratories. Its job is infectiousness and the isolation decision.
• Culture: ordered with the smear (IA), on both solid and liquid media (IIA). Liquid (MGIT) about 2 weeks, solid (Löwenstein-Jensen, Ogawa) 3–8 weeks. Some isolates grow only on solid media. A positive culture triggers prompt TB-versus-NTM identification (IIIA).

Molecular — the headline of the 5th edition

5th-edition recommendation (IA): when TB is suspected, perform Xpert MTB/RIF, Xpert Ultra, or an equivalent rifampicin-resistance-detecting nucleic acid amplification test together with smear and culture. If MDR is suspected, perform Xpert or Ultra (IA).

• Xpert MTB/RIF: about 2 hours from specimen. TB plus rifampicin resistance (sensitivity 95.3%, specificity 98.8% for resistance). In a Korean tertiary-centre series of 3,359 specimens, Xpert sensitivity was 74.1% and specificity 97.5% against smear’s 38.8% and 96.7%, with half the results reported within 3.1 hours.
• Xpert Ultra: limit of detection 10¹–10² CFU/mL. Pooled sensitivity 90.9%, specificity 95.6% (Xpert 84.7% and 98.4%). The gain is in smear-negative disease (60.6% → 77.5%).
• Rifampicin resistance on Xpert = treat as MDR until susceptibility testing says otherwise — reasonable in Korea because isolated rifampicin resistance is very rare. Positive predictive value falls in low-resistance populations.
• The guideline says explicitly that Xpert is useful when a fast result is needed for infection control — deciding whether to admit to isolation.

Drug susceptibility testing

• Every patient’s first culture isolate gets a rapid (molecular) test for isoniazid and rifampicin and conventional testing (IA). Smear-positive specimens can go straight to rapid testing (IB).
• Isoniazid resistance → rapid and conventional fluoroquinolone testing (IA). Rifampicin resistance → rapid fluoroquinolone testing plus conventional testing including group A drugs (IA).
• Still culture-positive after 3 months or more of treatment → repeat susceptibility testing (IIIA).
• Turnaround: conventional (proportion method on solid media) 3–4 weeks after culture, liquid within 2 weeks — add the 3–8 weeks of culture and the conventional result usually arrives around the 2-month visit, which is where Table 5 puts it. Liquid-based testing covers mainly isoniazid and rifampicin.
• Definitions (5th edition, WHO 2021): MDR-TB = isoniazid plus rifampicin resistance; RR-TB = rifampicin resistance; pre-XDR-TB = RR/MDR plus resistance to at least one fluoroquinolone; XDR-TB = RR/MDR plus fluoroquinolone plus at least one group A drug (bedaquiline or linezolid). “Plus an injectable” is the old definition.

Tests for infection (TST, IGRA) — not for active disease (ID)

• They cannot separate active from latent infection, so they are not used to diagnose active TB (ID). Adjunctive use only in culture-negative, hard-to-diagnose disease such as extrapulmonary TB (IB).
• TST: 2 TU PPD RT 23 intradermally, read at 48–72 hours, induration 10 mm or more is positive. BCG given before age 1 is disregarded.
• IGRA: QuantiFERON or T-SPOT.TB; ESAT-6 and CFP-10 antigens, so BCG does not interfere.

Others

• Bronchoscopy with BAL when sputum is negative and suspicion is strong — smear, culture, NAAT on the lavage.
• Tissue: AFB culture plus Xpert or Ultra on lymph-node specimens (IIA), NAAT on tissue (IIA). Tissue culture takes 2–6 weeks, so Xpert goes with it.

3. The schedule after admission — 5th-edition Table 5, “baseline and follow-up tests”

Strength: guideline table (● required, ○ as needed). Recommendations graded where the text grades them.

Admission to start of treatment

• Sputum smear and culture ×2–3 plus Xpert plus rapid susceptibility — the required baseline set. The isolation decision turns on these.
• Chest film (PA and lateral).
• CBC, liver function (AST, ALT, bilirubin), renal function (BUN, creatinine) — the pre-treatment baseline (IIIA).
• Vision (IIIA) before ethambutol. Formal ophthalmology (acuity, colour) if there is an existing visual problem.
• History: liver disease, alcohol, previous anti-TB drug reactions, interacting drugs (IIIA). Abnormal LFT or a hepatitis history → hepatitis B and C serology. HIV when risk factors are present — the Korean guideline; WHO and the ATS guideline test everyone, and that is a real difference.
• Uric acid — not in Table 5. Asymptomatic hyperuricaemia on pyrazinamide has no clinical significance and urate-lowering drugs are not recommended.
• 2-week LFT — there is no 2-week time point in Table 5. Abnormal baseline or risk factors → more often, on clinical judgement.

Week 2 — the isolation decision

• Fever from TB itself usually settles within 2 weeks on standard therapy.
• Release from isolation (5th edition, chapter IX-2, IIIA): ① at least 2 weeks of effective anti-TB drugs, ② clinical improvement, ③ if smear-positive, three consecutive negative follow-up smears (at least 8 hours apart, at least one a morning specimen) or a negative culture.
• National programme guideline, 2025 and 2026 editions — criteria for lifting an admission-and-isolation order: at least 2 weeks of effective drugs and the attending physician’s judgement that home treatment is feasible. The conversion criterion is three consecutive negative smears or one negative culture (after the first negative, two more at 8–24-hour intervals). Retreatment after default or failure, no ventilated private room at home, residents of congregate settings, and foreign nationals awaiting deportation must meet both the 2-week minimum and conversion.
• Before smear conversion, a patient who has improved clinically can still be discharged to isolate at home (IIIB) if outpatient TB care is linked, there is no child under 6 or immunosuppressed person in the household, and there is a ventilated private space. Where home isolation is not feasible, and in RR/MDR/XDR-TB, admission for the infectious period is considered.
• Return to work or school (national programme guideline): 14 days or more of treatment, respiratory symptoms resolved, smear converted.

Month 1

• Sputum smear and culture (●), weight (● — dose re-adjusted to weight), vision (●), symptoms and adherence (●). Ask about GI upset, rash, arthralgia, visual change.

Month 2 ★ — end of the intensive phase

• Sputum smear and culture (●) — conversion check. Chest film (●).
• Conventional susceptibility result (●) → isoniazid and rifampicin susceptible → ethambutol can be stopped (IIB). Intensive → continuation phase (HR, or HRE until susceptibility is known).
• If the 2-month culture is positive, the guideline says check four things: adherence during the intensive phase; extensive cavitation and high bacillary load; comorbidity undermining adherence or response; resistance to first-line drugs.
• Extension: cavity at baseline plus positive 2-month culture → consider extending the continuation phase (IIB). The guideline is candid that there is no clear criterion — the 2016 ATS guideline recommended a 3-month extension on expert opinion, the 2022 WHO guideline does not recommend extension by default, and the Korean evidence is a single-centre retrospective study. Duration is individual.

Months 3–5

• Bacteriologically positive at the start: monthly sputum until two consecutive negative smear-and-culture results (IIIA). Culture positive past 3 months, or reversion after conversion → repeat rapid and conventional susceptibility testing (IIIA) and evaluate for failure.
• NTM in sputum during treatment is usually contamination or colonisation — not to be mistaken for failure.
• Vision and liver tests as needed. Past 2 months of ethambutol, the patient is told to report any visual change at once.

End of treatment (6 months; 9 if extended)

• Bacteriologically positive at the start → a final sputum at the end of treatment (IIIA). Chest film (●) for before-and-after comparison; the film alone does not judge response (ID).
• Post-TB lung disease assessment in every patient (IIIA) — symptoms, examination, chest film. Dyspnoea or a significant residual lesion → pulmonary function testing at the end of treatment or within 6 months. First appearance of this item in the Korean guideline.

4. Regimens — in brief

Strength: 5th-edition recommendations (IA / IIB); trials as cited.

Drug-susceptible TB — 2HRZE/4HR

• Two months HRZE, then four months HR (HRE until susceptibility is confirmed). Six months total; extension considered for cavity plus positive 2-month culture.
• Fixed-dose combinations have been used in Korea since 2012. The four-drug tablet Tubis® contains isoniazid 75 mg, rifampicin 150 mg, ethambutol 275 mg and pyrazinamide 400 mg; the two-drug Tubis-Two® contains isoniazid 150 mg and rifampicin 300 mg. Meta-analysis shows non-inferiority to separate tablets.

(5th-edition Table 2. Weight is measured monthly and the dose adjusted.)

• Where the international guidelines have moved (September 2026): the ATS/CDC/ERS/IDSA 2025 update conditionally recommends a 4-month HPMZ regimen (isoniazid, rifapentine, moxifloxacin, pyrazinamide for 2 months, then HPM for 2 months; Study 31/A5349) for drug-susceptible pulmonary TB at age 12 and over. It is not in the Korean 5th edition, and rifapentine access in Korea is the practical question.

MDR/RR-TB

• Fluoroquinolone-susceptible MDR: 6-month BPaLM or 9-month MDR-END (IA), chosen on treatment history, susceptibility, disease extent, adverse effects and interactions — no stated preference. The 18–20-month long regimen is for patients who fit neither (IA).
  • BPaLM = bedaquiline, pretomanid 200 mg once daily, linezolid, moxifloxacin. In TB-PRACTECAL, the 24-week BPaLM arm had a favourable outcome in 89% versus 48% for standard care.
  • MDR-END = levofloxacin, delamanid, linezolid, pyrazinamide for 9 months — a fixed regimen. In the Korean trial (Lancet 2022), treatment success was 75.0% versus 70.6% with the long regimen, meeting non-inferiority.
• This is the change the 5th edition is remembered for: 18–20 months down to 6 or 9.
• Since then: WHO Module 4 (2025 edition) adds a 6-month BDLLfxC regimen (bedaquiline, delamanid, linezolid, levofloxacin, clofazimine — the BEAT-Tuberculosis trial, NEJM 2026), including pre-XDR. Not yet in the Korean guideline.

5. Isolation — 5th edition, chapter IX-2

Strength: expert consensus (IIIA / IIIB).

• Start: isolate on suspicion, before confirmation (IIIA). Every admission is assessed for infectious TB at the door and re-assessed during the stay. Negative-pressure room, or failing that a well-ventilated single room with its own toilet and washbasin. N95 for everyone entering, door kept shut, room left empty for at least an hour after discharge.
• Release: as in the week-2 section above — 2 weeks, clinical improvement, and for smear-positive patients three consecutive negative smears or a negative culture.
• Sputum is collected in a negative-pressure room, a sputum booth, or outdoors.

6. Adverse-effect monitoring — 5th edition, chapter II-3

Strength: expert consensus (IIIA).

• Hepatotoxicity (isoniazid, rifampicin, pyrazinamide): about 20% of patients on first-line drugs have a transient, asymptomatic rise in liver enzymes that resolves on its own. ALT 5× the upper limit regardless of symptoms, or 3× with symptoms of hepatitis → stop the hepatotoxic drugs immediately (IIIA). Asymptomatic and under 5× → continue with close observation. Once ALT is at or below 2× the upper limit, reintroduce one drug at a time. Check hepatitis A/B/C, alcohol and herbal medicines. The reference enzyme is ALT, not AST.
• Isoniazid neuropathy: pyridoxine.
• Rifampicin: drug interactions. Thrombocytopenia → stop rifampicin and do not rechallenge even after recovery (IIIA).
• Pyrazinamide: arthralgia can be managed with NSAIDs while continuing; gout → stop (IIIA). Uric acid commonly rises and is clinically unimportant; urate-lowering drugs are not recommended.
• Ethambutol optic neuritis: baseline vision; stop at once for any change in acuity, colour or field. Closer monitoring in the elderly and in renal impairment. Stop ethambutol once susceptibility is confirmed.

7. Extrapulmonary specimens — 5th edition, chapter VI

Strength: guideline recommendations (IIA) and cited meta-analyses.

• Pleural fluid: lymphocyte-predominant exudate with ADA above 40 IU/L and no other cause → tuberculous pleurisy is likely (IIA; sensitivity 92%, specificity 90%). Above 70 strongly suggestive; below 40 unlikely. Also raised in rheumatoid disease, lymphoma and empyema. NAAT sensitivity in pleural fluid 20–90%.
• CSF: NAAT recommended (IIA) but sensitivity is low (Xpert 40–56%, Ultra 47–77%) — a negative does not exclude. ADA cut-off 5–10 IU/L (sensitivity 44–100%). Protein up, glucose down, lymphocytes.
• Tissue (node, bone, peritoneum): AFB culture plus NAAT plus histology. Ascitic ADA cut-off 36–40 IU/L (27 in cirrhosis).

What I take from it

1. Xpert with the first sputum, alongside smear and culture. It is the test that lets you decide isolation early.
2. The 2-month visit is the fork — conversion, conventional susceptibility, whether ethambutol stops, whether the course extends.
3. Cavity plus a positive 2-month culture → consider extending, on weak evidence.
4. Isolation comes off at 2 weeks plus improvement plus three negative smears (or a negative culture) — not two.
5. Post-TB lung disease is assessed in every patient at the end; PFT when there are symptoms or residual imaging.
6. MDR is 6-month BPaLM or 9-month MDR-END, with no preference. XDR now means fluoroquinolone plus bedaquiline or linezolid, not injectables.
7. Vision at baseline and 1 month, then as needed — not monthly. What replaces the monthly test is telling the patient to speak up the day something changes.
8. From inside a hospital — the 65-and-over group is now 62.5% of Korean TB, and it is the one group still growing. The isolation bed will keep filling from there.

Sources

Guidelines

• Korea Disease Control and Prevention Agency; Korean Academy of Tuberculosis and Respiratory Diseases. Korean Guidelines for Tuberculosis, 5th edition (2024). Chapter I (diagnosis), Table 5 p.48 (baseline and follow-up tests), Table 2 p.39 (fixed-dose combination dosing), chapter II-3 (adverse effects, post-TB lung disease), chapter III p.79 and p.103–104 (resistance definitions, short MDR regimens), chapter VI (extrapulmonary TB), chapter IX-2 p.383–384 (isolation). Official listing: https://tbzero.kdca.go.kr · https://www.korea.kr/archive/expDocView.do?docId=40809
• Korea Disease Control and Prevention Agency. 2026 National Tuberculosis Management Guidelines (January 2026) — criteria for lifting admission-and-isolation orders and work/school restrictions; unchanged from the 2025 edition, which was also checked. Distribution notice: https://www.kha.or.kr/kha_home/notice_list.do?mode=view&articleNo=46101
• World Health Organization. WHO consolidated guidelines on tuberculosis. Module 4: treatment and care (2025 edition). https://www.who.int/publications/b/77823
• Saukkonen JJ, Duarte R, Munsiff SS, et al. Updates on the Treatment of Drug-Susceptible and Drug-Resistant Tuberculosis: An Official ATS/CDC/ERS/IDSA Clinical Practice Guideline. Am J Respir Crit Care Med. 2025;211(1):15-33. doi:10.1164/rccm.202410-2096ST · PMID 40693952

Epidemiology

• Shin EJ, Lee HW, Kim JE, Lee SE, et al. Analysis of tuberculosis incidence and trends in 2025 [in Korean]. Public Health Weekly Report. 2026;19(16):693-710. doi:10.56786/PHWR.2026.19.16.1

Trials behind the regimens

• Dorman SE, Nahid P, Kurbatova EV, et al. Four-Month Rifapentine Regimens with or without Moxifloxacin for Tuberculosis. N Engl J Med. 2021;384(18):1705-1718. doi:10.1056/NEJMoa2033400 · PMID 33951360
• Nyang’wa BT, Berry C, Kazounis E, et al. A 24-Week, All-Oral Regimen for Rifampin-Resistant Tuberculosis. N Engl J Med. 2022;387(25):2331-2343. doi:10.1056/NEJMoa2117166 · PMID 36546625
• Conradie F, Bagdasaryan TR, Borisov S, et al. Bedaquiline-Pretomanid-Linezolid Regimens for Drug-Resistant Tuberculosis. N Engl J Med. 2022;387(9):810-823. doi:10.1056/NEJMoa2119430 · PMID 36053506
• Mok J, Lee M, Kim DK, et al. 9 months of delamanid, linezolid, levofloxacin, and pyrazinamide versus conventional therapy for treatment of fluoroquinolone-sensitive multidrug-resistant tuberculosis (MDR-END): a multicentre, randomised, open-label phase 2/3 non-inferiority trial in South Korea. Lancet. 2022;400(10362):1522-1530. doi:10.1016/S0140-6736(22)01883-9 · PMID 36522208
• Conradie F, Badat T, Poswa A, et al. A Pragmatic Trial of a 6-Month Strategy for Rifampicin-Resistant Tuberculosis (BEAT-Tuberculosis). N Engl J Med. 2026;394(24):2429-2439. doi:10.1056/NEJMoa2503687 · PMID 42341301