What it is:An infection with Mycobacterium tuberculosis, spread by aerosol. About a quarter of the world's population carries it, and roughly 13 million people in the US have latent infection right now. Most of them will never get sick from it.
The core problem:TB has two completely different clinical states that need completely different treatment. Latent TB infection (LTBI)is a positive test with no disease, no symptoms, and no ability to spread it. Active TBis symptomatic, contagious, and needs multi-drug therapy. Confusing which one you're treating is the single biggest way to get a TB question wrong.
What you do about it:LTBI gets 1-3 drugs for 3-4 months to prevent progression. Active disease gets 4 drugs for a minimum of 6 months, ideally with directly observed therapy (DOT), because undertreating active TB is exactly how you manufacture drug resistance.
Think of TB as a numbers game your immune system usually wins. About 90% of primary infectionsgo no further than a positive test and a walled-off granuloma (the Ghon complex). Only 5-10%of immunocompetent people with LTBI ever progress to active disease, and nearly half of those who do, do it within 2 years. Everything about TB risk stratification and treatment duration comes back to that containment-versus-breakthrough framework.
Three patterns, one organism. Knowing which bucket a patient is in tells you whether they're contagious, whether they need isolation, and what regimen applies.
| Pattern | What's happening | Contagious? | Test/imaging |
|---|---|---|---|
| LTBI | Immune system contained the organism in a granuloma. No active replication, no symptoms. | No | Positive TST/IGRA, normal or old healed changes on CXR |
| Primary active disease | Initial infection progresses instead of being contained (more common in kids, called progressive primary TB, mimics bacterial pneumonia) | Yes | Positive test, active infiltrates, positive sputum |
| Reactivation disease | Dormant organisms in an old granuloma escape containment, usually years later. Most US adult cases work this way. | Yes | Apical/upper lobe cavitary disease typically |
| Miliary TB | A massive inoculum seeds the bloodstream, producing widespread, tiny granulomas across multiple organs at once | Yes, often severely ill | Diffuse "millet seed" pattern on imaging |
HIV is the most important risk factor for progressing from LTBI to active disease, especially in the 25-44 age range. An HIV-positive person with TB infection is over 100 times more likelyto develop active disease than someone who is HIV-negative. Any patient with new TB diagnosis should be tested for HIV, and vice versa.
M. tuberculosisspreads person to person through aerosolized droplets from coughing, so close, prolonged contact drives most transmission. Once inhaled, macrophages in the alveoli engulf the organism, but M. tuberculosishas a waxy, lipid-rich cell wall that resists killing and lets it survive and slowly replicate inside the very cells trying to destroy it.
Over several weeks, cell-mediated immunity organizes macrophages, lymphocytes, and epithelioid cells into a granuloma, walling the organism off in a caseating, necrotic center. When this calcifies on chest imaging, it's called a Ghon complex. The bacteria inside aren't dead, they're just dormant and metabolically slowed, which is exactly why LTBI treatment needs weeks to months rather than a short antibiotic course: you're trying to sterilize slow-dividing organisms hiding inside an immune structure.
If the granuloma breaks down, whether from immunosuppression, malnutrition, aging, or just bad luck, bacteria resume active replication and the disease becomes clinically apparent and infectious. This is reactivation, and it's why immunosuppression (HIV, chronic steroids, TNF-alpha inhibitors, transplant medications) is such a dominant risk factor. It isn't giving people new infections, it's unmasking old ones.
Active TB has a much larger, faster-replicating, more heterogeneous bacterial population, including some organisms that are naturally, spontaneously drug-resistant to any single agent. Using one drug alone just selects for the resistant subpopulation and you've created a resistant infection. LTBI, by contrast, is a small, slow, contained population, so 1-2 drugs sterilize it without the same resistance-amplification risk.
Classic active TB is a slow burn: weight loss, fatigue, a productive cough, low-grade fever, and drenching night sweats that build gradually rather than hitting all at once. Frank hemoptysis is a late finding, it usually shows up after cavitary disease has already developed, though it can appear earlier.
| Category | Findings |
|---|---|
| Constitutional | Weight loss, fatigue, low-grade fever, night sweats |
| Pulmonary | Productive cough (weeks to months), hemoptysis (late) |
| Exam | Thin, recently weight-losing appearance; dullness to percussion, rales, increased vocal fremitus - but a normallung exam is common even with significant radiographic disease |
| Labs | Mild-moderate WBC elevation with lymphocyte predominance, thrombocytosis, mild-moderate anemia |
| Extrapulmonary | Depends on organ involved, but the common thread is slow, progressive organ decline plus low-grade fever and constitutional symptoms - not the dramatic acute-infection picture students expect |
HIV-positive patientsare less likely to have a positive skin test, cavitary lesions on imaging, or fever, and are more likely to have extrapulmonary disease or progressive primary disease. Older adultsare far less likely to show positive skin tests, fever, night sweats, sputum production, or hemoptysis, so TB in the elderly gets misread as "just another respiratory illness." Childrenoften present as what looks like ordinary bacterial pneumonia (progressive primary TB). In all three groups, the textbook picture is the exception, not the rule.
Diagnosis happens in two separate steps that get confused constantly: screening for infection(TST or IGRA, tells you if someone was ever infected) and confirming active disease(imaging plus microbiology, tells you if they're sick right now). A positive skin test alone never diagnoses active TB.
The tuberculin skin test (TST), done by the Mantoux method, is an intracutaneous injection of purified protein derivative (PPD) containing five tuberculin units. It's read 48-72 hours laterby measuring the diameter of induration (the raised, firm area, not any surrounding redness). A patient who tests negative and then flips positive on retest one week later is showing a booster effect, not a new infection.
The interferon-gamma release assay (IGRA)measures the immune system's IFN-gamma response to TB-specific antigens in a blood sample. It's preferred over TSTfor patients age 5 and older who are likely infected, at low-to-moderate risk of progression, have had BCG vaccination (which causes false-positive TSTs), or are unlikely to return for a 48-72 hour TST reading. Results report both qualitatively (positive/negative/indeterminate/borderline) and quantitatively.
| Induration cutoff | Positive in these patients |
|---|---|
| ≥5 mm | HIV-positive, recent close contact of an active case, fibrotic changes on CXR consistent with prior TB, organ transplant or other immunosuppression (equivalent of ≥15 mg/day prednisone for ≥1 month, TNF-alpha antagonists) |
| ≥10 mm | Recent immigrants (within 5 years) from high-prevalence countries, injection drug users, residents/employees of prisons, nursing homes, hospitals, homeless shelters; mycobacteriology lab staff; silicosis, diabetes, chronic renal failure, certain malignancies, gastrectomy, jejunoileal bypass; children under 5 or those exposed to high-risk adults |
| ≥15 mm | Anyone with no known risk factors |
The threshold isn't arbitrary, it's tuned to pretest probability. In a high-risk patient (HIV, close contact), even a small reaction is meaningful because the prior probability of true infection is already high, so you accept more false positives to avoid missing true infections. In a no-risk-factor patient, you need a much bigger reaction before you believe it, because most small reactions in that group are noise (cross-reactivity, prior BCG). Same logic students already know from likelihood ratios in other diagnostic contexts.
The point of LTBI treatment is prevention, stopping the 5-10% who would otherwise reactivate. Three CDC-endorsed regimens exist, and they're considered roughly equivalent in efficacy.
| Regimen | Duration/frequency | Adult dose |
|---|---|---|
| Isoniazid + rifapentine (3HP) | Once weekly x 3 months | INH 15 mg/kg + rifapentine weight-banded (300-900 mg) |
| Isoniazid + rifampin | Daily x 3 months | INH 5 mg/kg + rifampin 10 mg/kg |
| Rifampin alone | Daily x 4 months | 10 mg/kg |
The CDC now considers 12 weeks of once-weekly isoniazid/rifapentine an equal alternativeto 9 months of daily isoniazid monotherapy, for otherwise healthy patients age 12 and older who have a predictive factor for progression: recent exposure to a contagious case, a recent test conversion from negative to positive, or radiographic evidence of healed pulmonary TB. Shorter, once-weekly regimens dramatically improve completion rates, and completion is what actually prevents reactivation, not the drug on paper.
Isoniazid depletes vitamin B6, which can cause peripheral neuropathy or CNS effects. Give pyridoxine 10-50 mg dailyto pregnant patients, people who misuse alcohol, and those with poor nutritional status. It isn't routine for every otherwise healthy patient on isoniazid.
Goals of therapy:identify the case quickly, start treatment fast, eradicate the organism, get the patient to a noninfectious state so isolation can end, prevent resistance from developing, keep the patient adherent, and cure them, generally in a minimum of 6 months. Active patients should be isolated while infectious, and DOT (directly observed therapy by a healthcare worker) is the standard of care because it's the single most effective adherence tool available.
For drug-susceptible pulmonary TB: isoniazid + rifampin + pyrazinamide + ethambutol daily for 2 months (the intensive phase), then isoniazid + rifampin for 4 more months (the continuation phase), for 6 months total. Ethambutol exists purely as coverage until susceptibility results confirm the organism is sensitive to isoniazid, rifampin, and pyrazinamide, at which point it can be dropped.
| Regimen | Intensive phase (8 weeks) | Continuation phase | Use case |
|---|---|---|---|
| 1 (preferred) | INH/RIF/PZA/EMB daily | INH/RIF daily x 18 weeks | Standard for newly diagnosed pulmonary TB |
| 2 | INH/RIF/PZA/EMB daily | INH/RIF 3x/week x 18 weeks | When daily DOT in continuation phase is hard to arrange |
| 3 | INH/RIF/PZA/EMB 3x/week | INH/RIF 3x/week x 18 weeks | Use cautiously; avoid in HIV or cavitary disease, missed doses risk failure/relapse/resistance |
| 4 | INH/RIF/PZA/EMB daily x 2 wks, then 2x/week x 6 wks | INH/RIF 2x/week x 18 weeks | Never use twice-weekly dosing in HIV or smear/cavitary-positive patients - a missed dose here behaves like once-weekly, which is inferior |
Extend treatment to 9 months(and at least 6 months from the point of culture conversion) in patients who are slow to respond, remain culture-positive at 2 months, have cavitary lesions on chest imaging, or are HIV-positive. This is a frequently tested exception to the "TB is always 6 months" reflex.
Pyridoxine 25-50 mg/day is given alongside isoniazid in active-disease regimens to anyone at risk of neuropathy: pregnant or breastfeeding patients, HIV, diabetes, alcohol use disorder, malnutrition, chronic renal failure, or advanced age. Bump it to 100 mg/day if neuropathy actually develops.
| Drug | Daily adult dose | Intermittent (DOT) dosing |
|---|---|---|
| Core four (active disease) | ||
| Isoniazid (INH) | 5 mg/kg | 15 mg/kg (1x, 2x, or 3x weekly) |
| Rifampin (RIF) | 10 mg/kg | 10 mg/kg (2x or 3x weekly) |
| Pyrazinamide (PZA) | 1000 mg (40-55 kg) to 2000 mg (76-90 kg), weight-banded | Higher weight-banded doses 2-3x weekly |
| Ethambutol (EMB) | 800 mg (40-55 kg) to 1600 mg (76-90 kg), weight-banded | Higher weight-banded doses 2-3x weekly |
| Rifamycin alternatives | ||
| Rifabutin | 5 mg/kg | Preferred rifamycin with PI-based HIV therapy - much weaker CYP3A4 inducer than rifampin |
| Rifapentine | Used weekly in 3HP LTBI regimen (weight-banded 300-900 mg) | Not approved for use in children |
All weight-banded TB doses use idealbody weight, not actual. Also, current guideline authors specifically push back on capping doses arbitrarily. Older max-dose tables weren't based on prospective data in larger patients, so clinical judgment (and serum drug concentration monitoring when needed) should guide dosing at the extremes rather than a hard ceiling.
Blocks mycolic acid synthesis in the mycobacterial cell wall, the same waxy coat that lets TB survive inside macrophages. Highly bactericidal against actively dividing organisms, which is why it's in essentially every regimen.
Hepatotoxicity:transaminase elevation is common and often transient, but stop the drug if AST/ALT exceed 5x the upper limit of normal (or 3x with symptoms), or if total bilirubin exceeds 3 mg/dL. Peripheral neuropathy:isoniazid depletes pyridoxine (B6) by competing for the same enzyme cofactor, so patients at risk (pregnancy, alcohol use, poor nutrition, diabetes, renal failure, HIV, advanced age) get prophylactic B6.
Isoniazid is also a classic CYP2C19/CYP3A4 inhibitor, so it raises levels of phenytoin and carbamazepine among other interacting drugs. Acetylator status (fast vs slow) affects metabolism and toxicity risk, though this rarely changes empiric dosing.
Inhibits bacterial DNA-dependent RNA polymerase, blocking transcription. Bactericidal against both actively dividing and semi-dormant organisms, which is part of why it's essential in every phase of therapy, not just the intensive phase.
Rifampin is a potent inducer of CYP3A4 and other CYP enzymes, plus P-glycoprotein. It slashes levels of hormonal contraceptives, warfarin, many HIV antiretrovirals, azole antifungals, and methadone, among others. Always run an interaction check when starting rifampin on a patient with other chronic medications; this single property drives most of the special-population complexity in TB care (see the HIV coinfection section).
Harmless but important counseling point: it turns urine, sweat, tears, and other body fluids orange-red, and can permanently stain soft contact lenses.
Converted to its active form, pyrazinoic acid, most efficiently in the acidic environment inside macrophages and caseating granulomas, which is exactly where dormant/semi-dormant organisms hide. That's why it's only needed for the first 2 months, its job is knocking down that specific population early.
Adverse effects:hepatotoxicity (additive with isoniazid and rifampin), hyperuricemia (usually asymptomatic, occasionally triggers gout), and arthralgias.
Inhibits arabinosyl transferase, disrupting cell wall arabinogalactan synthesis. Bacteriostatic, not bactericidal. Its real job in the standard regimen is preventing resistance amplificationduring the intensive phase while susceptibility results are pending; once susceptibility to INH, RIF, and PZA is confirmed, it can be dropped.
Dose-dependent retrobulbar optic neuritis causing decreased visual acuity and red-green color discrimination loss. Get a baseline visual acuity and color vision testbefore starting, and counsel patients to report any visual change immediately. Use with real caution in young children who can't reliably report visual symptoms.
Suspect drug resistance in patients who've had prior TB treatment, come from geographic areas with high resistance prevalence (South Africa, Mexico, Southeast Asia, the Baltic countries, former Soviet states), are homeless, institutionalized, inject drugs, or are HIV-positive, remain AFB smear-positive after 2 months of therapy, remain culture-positive after 2-4 months, fail therapy or relapse after retreatment, or have known exposure to an MDR-TB case.
When resistance is suspected or confirmed, add at least two active agents the patient hasn't received before. There's no single standard MDR-TB regimen, it has to be built around the resistance pattern. Never add a single drug to a failing regimenand never treat with monotherapy, either one just breeds resistance to the new drug too. MDR-TB treatment runs 18-24 months, versus 6 for drug-susceptible disease.
| Drug | Pregnancy status |
|---|---|
| Isoniazid, ethambutol | Relatively safe; standard components of the pregnancy regimen |
| Rifampin | Used despite rare reports of birth defects (limb reduction, CNS lesions) - benefit outweighs risk in active disease |
| Pyrazinamide | Not extensively studied, anecdotal data suggests it's likely safe |
| Ethionamide | Avoid- linked to premature delivery, congenital deformities, Down syndrome |
| Cycloserine, fluoroquinolones | Avoidin pregnancy and while nursing |
The standard pregnancy regimen is isoniazid + rifampin + ethambutol for 9 months, and B-vitamin supplementation matters more here than usual. Counsel patients with active TB against becoming pregnant during treatment given the risk to both mother and fetus.
Isoniazid and rifampin need no dose adjustmentin renal failure, they're primarily hepatically cleared. Pyrazinamide and ethambutol switch from daily to three-times-weekly dosingonce CrCl drops below 30 mL/min or the patient is on hemodialysis, with doses given after the dialysis session on dialysis days. Several second-line agents (levofloxacin, cycloserine, streptomycin, capreomycin, kanamycin, amikacin) also require frequency reduction; serum drug concentration monitoring is useful to confirm adequate absorption without toxic accumulation.
Isoniazid, pyrazinamide, ethionamide, cycloserine, and linezolid all penetrate CSF well, which matters because rifampin and ethambutol don't cross as reliably. TB meningitis is treated for 9-12 monthsinstead of the standard 6. Bone TB typically runs 9 months, sometimes with surgical debridement. Soft-tissue extrapulmonary disease can usually follow the conventional regimen and duration.
Regimens mirror the adult approach with pediatric weight-based dosing, though some clinicians still extend to 9 months. Progressive primary TB in children looks like ordinary bacterial pneumonia clinically, so a high index of suspicion in a child with risk factors (exposure to an adult case) matters.
Treating both diseases at once is dominated by one problem: rifamycins are potent CYP3A4 inducers, and most antiretrovirals are CYP3A4 substrates. Get this wrong and you either undertreat the HIV or overexpose the patient to rifabutin toxicity.
| ART backbone | Rifamycin pairing | Bottom line |
|---|---|---|
| Efavirenz-based | Rifampin-based TB regimen | Preferred combination- modest, well-characterized decrease in efavirenz levels, low discontinuation, excellent antiviral activity. Avoid efavirenz in the first trimester of pregnancy. |
| Protease inhibitor-based | Rifabutin-based TB regimen | Preferred for patients who can't take efavirenz. Rifabutin concentrations rise markedly with PIs, so the rifabutin dose must be reduced - and re-adjusted back up if the PI is later stopped. |
| Nevirapine-based | Rifampin-based TB regimen | Alternative only. Watch for hepatotoxicity when nevirapine is combined with isoniazid, rifampin, and pyrazinamide together. |
| Raltegravir/dolutegravir-based | Rifampin-based TB regimen | Alternative at increased integrase inhibitor doses; raltegravir option requires baseline viral load under 100,000 copies/mL. |
Rifampin also causes roughly a 50% decrease in zidovudine concentrationsand can worsen anemia in regimens containing it, which is one more reason zidovudine-containing backbones are lower on the preference list when a rifamycin is on board.
| Parameter | When | Watching for |
|---|---|---|
| AFB sputum smear | Every 1-2 weeks until 2 consecutive negatives | Declining bacterial burden, progress toward noninfectious status |
| Sputum culture | Monthly until negative (typically 2-3 months) | Treatment response; if still positive after 2 months, repeat susceptibility testing and check serum drug levels |
| BUN, SCr, AST/ALT, CBC | Baseline, then periodically based on risk (advanced age, alcohol use, pregnancy) | Hepatotoxicity, renal function for dose adjustment |
| Transaminases / bilirubin | Any time symptoms suggest hepatotoxicity, or per schedule in high-risk patients | AST/ALT >5x ULN or bilirubin >3 mg/dL - stop the offending drug(s) |
| Visual acuity / color vision | Baseline and periodically on ethambutol | Optic neuritis |
| Adherence | Every contact | DOT is the most effective safeguard; nonadherence is the single biggest driver of treatment failure and resistance |