Tuberculosis

TL;DR. Tuberculosis is a slow bacterial infection of the lungs, spread by breathing the air of someone who has it. Its distinguishing trick is patience: the bacterium survives inside the very immune cells sent to destroy it, gets walled into a small nodule called a granuloma, and can sit there for decades doing nothing. Most infected people never become ill. When immunity weakens, from HIV, malnutrition, diabetes, age, or certain drugs, the walls break down and the disease begins. TB has probably killed more humans than any other infection in history, and it is still killing over a million people a year, almost entirely among the world's poor, despite being curable with antibiotics that cost a few dollars.

Key takeaways

• In 2024 there were an estimated 10.7 million cases and 1.23 million deaths. TB is again the world's deadliest single infectious disease.
• Infection is not disease. A large share of infected people carry latent TB and never develop illness; the lifetime risk of progression is roughly 5 to 10 percent, and far higher with HIV.
• It spreads by airborne particles that stay suspended, so ventilation matters more than surface cleaning, and prolonged shared indoor air is the main risk.
• Treatment works and takes months. Standard drug-sensitive TB is cured by four drugs over six months. Stopping early is how resistance is created.
• Drug-resistant TB affected about 390,000 people in 2024, and only about 42 percent accessed treatment. Regimens have improved dramatically, from two years of injections to six months of tablets.
• The first randomised controlled trial in medical history was a TB trial, in 1948.

What it is

In short: A slow bacterium with a waxy wall, and the crucial distinction is between latent infection, which is not contagious, and active disease, which is.

Tuberculosis is caused by Mycobacterium tuberculosis, a slow-growing bacterium with an unusual waxy cell wall of mycolic acids. That wall makes it resistant to drying, to many disinfectants, and to most antibiotics, and it means the organism divides roughly once every 15 to 20 hours rather than every 20 minutes like E. coli. Everything slow about TB, including the months of treatment, follows from that.

Three states, and the distinction governs everything:

About 85 percent of active disease is pulmonary. Extrapulmonary TB affects lymph nodes, the spine (Pott's disease), the membranes around the brain (TB meningitis, the most lethal form), the pericardium, kidneys, and, when spread through the bloodstream, everywhere at once (miliary TB, named for the millet-seed appearance of countless tiny lesions on a chest X-ray). Extrapulmonary TB is generally not contagious.

The history

In short: Koch found the bacterium in 1882, and the first randomised controlled trial in medical history was a TB trial in 1948.

TB is old. DNA from M. tuberculosis complex organisms has been recovered from Egyptian mummies and from pre-Columbian remains in South America, and the spinal deformity of Pott's disease appears in skeletons thousands of years old.

In Europe it became epidemic with industrialisation and urban crowding. Called consumption or phthisis for the wasting it produced, and the White Plague for the pallor of its victims, it caused something like a quarter of all deaths in nineteenth-century Europe. It killed Keats, Chopin, the Brontë sisters, Chekhov, and Orwell, and it acquired a romantic literary association with sensitivity and genius that reads grotesquely now.

24 March 1882: Robert Koch announced to the Berlin Physiological Society that he had identified the bacterium causing tuberculosis, demonstrating it with the criteria that became Koch's postulates. World TB Day is still held on that date.

What followed, in order:

• Sanatoria (from the 1850s): rest, fresh air, and sunlight in mountain institutions. They isolated infectious patients from cities, which probably helped the population more than the individual.
• BCG vaccine (1921), developed by Calmette and Guérin from a weakened bovine strain. It protects children well against severe disseminated TB and meningitis and protects adults against pulmonary TB unreliably. It remains the most-administered vaccine in the world and the only licensed TB vaccine.
• Streptomycin (1943 to 1944), the first effective antibiotic, and the subject of the 1948 Medical Research Council trial, the first properly randomised controlled trial in medicine. It also demonstrated, within months, that single-drug therapy produces resistance, which is why TB has been treated with combinations ever since.
• Isoniazid (1952), rifampicin (1960s), and the modern short-course regimen.
• DOTS (1990s): the WHO strategy built around directly observed therapy, standardised regimens, and drug supply, which drove large falls in mortality.
• HIV (1980s onward) reversed decades of progress in sub-Saharan Africa, because HIV and TB amplify each other.
• Bedaquiline (2012), the first new TB drug class in over 40 years, followed by pretomanid and the short all-oral regimens for resistant disease.

What actually goes wrong

In short: The bacterium survives inside the very cell sent to kill it, gets walled into a granuloma, and waits for immunity to weaken.

Someone with active pulmonary TB coughs, and expels particles containing a handful of bacilli. The particles are small enough to stay suspended in room air for hours and small enough to be inhaled deep into the alveoli. The infectious dose may be as low as a few organisms.

Alveolar macrophages engulf the bacteria, and this is where it gets interesting. Normally a macrophage fuses the swallowed microbe with a lysosome full of digestive enzymes. M. tuberculosis blocks that fusion and survives inside the macrophage, using the cell that was supposed to kill it as a shelter and a taxi.

The immune system's answer is containment rather than elimination. T cells arrive and organise a granuloma: a ball of infected macrophages surrounded by other macrophages, lymphocytes, and a fibrous rim. Inside, oxygen is low, the pH is acid, and the centre becomes a cheese-like necrotic mass (caseation, from the Latin for cheese). The bacteria persist there in a slow, dormant state.

This is latent infection, and it is a stalemate. The bacteria cannot spread; the immune system cannot finish them. It can last a lifetime.

Reactivation happens when the containment fails. The granuloma liquefies, erodes into an airway, and spills bacteria into the lung, forming a cavity: a hole in the lung, rich in oxygen, where the organism multiplies rapidly and from which it is coughed out to the next person. Anything that weakens cell-mediated immunity increases the risk:

What it does to the body

Pulmonary TB: a cough lasting more than two or three weeks, at first dry then productive, sometimes with blood (haemoptysis) when a cavity erodes a vessel. Fever, drenching night sweats, loss of appetite, and progressive weight loss, the "consumption" that named it. Lungs are progressively destroyed and replaced by fibrosis, so survivors often have permanent breathlessness and bronchiectasis even after cure.

TB meningitis: headache, fever, confusion, cranial nerve palsies, developing over weeks rather than the hours of bacterial meningitis. It kills or disables a large share of those affected, especially young children, and it is the reason BCG vaccination of infants persists.

Spinal TB: destroys vertebral bodies, causing collapse, angular deformity, and spinal cord compression.

Miliary TB: seeded everywhere through the blood, with fever and multi-organ involvement, often in the very young, the very old, or the immunosuppressed.

Is it deadly?

• 1.23 million deaths in 2024, from about 10.7 million cases. TB has reclaimed its position as the leading cause of death from a single infectious agent.
• Untreated active pulmonary TB kills roughly half of those who have it, over a period of years. Around a quarter recover spontaneously and the rest become chronic sources of infection.
• Treated drug-sensitive TB is cured in over 85 percent of cases.
• Drug-resistant TB used to have cure rates near 50 percent with two years of toxic treatment; modern six-month all-oral regimens achieve roughly 85 to 90 percent in trials.
• TB with HIV remains the deadliest combination: TB is a leading cause of death among people living with HIV.

Is it contagious?

Yes, through the air, and only from people with active disease in the lungs or airway.

Key points that are widely misunderstood:

• Latent TB is not contagious. Someone with a positive skin or blood test and no active disease cannot infect anyone.
• It requires prolonged shared air, typically hours in an enclosed space. It is not caught in passing on a street, and it does not spread on doorknobs, cutlery, or bedding.
• Ventilation and ultraviolet light are effective, because the particles must stay suspended to transmit. Open windows and air changes matter more than disinfectant.
• Effective treatment rapidly reduces infectiousness, typically within two weeks for drug-sensitive disease, which is why prompt diagnosis protects households.
• Not everyone exposed is infected, and most infected people never become ill.

Who gets it

In short: A disease of poverty in the most direct sense, tracking crowding, undernutrition, HIV, and diabetes, with two-thirds of cases in eight countries.

TB is a disease of poverty in the most direct sense: it tracks crowding, undernutrition, indoor air pollution, and lack of access to care.

Geography. About two-thirds of cases occur in eight countries: India, Indonesia, China, the Philippines, Pakistan, Nigeria, Bangladesh, and the Democratic Republic of the Congo. India alone accounts for roughly a quarter of the world's cases.

In low-incidence countries, TB is concentrated among migrants from high-incidence regions (usually reactivation of infection acquired years earlier, not new transmission), people experiencing homelessness, people in prisons, people who use drugs, and people with HIV.

Prisons deserve specific mention as amplifiers: crowded, poorly ventilated, with high rates of HIV and interrupted treatment, they produce and export drug-resistant TB in several regions.

Sex. Notified cases are consistently higher in adult men, by roughly 2 to 1, a gap that reflects both real differences in exposure and behaviour and under-diagnosis in women.

Drug-resistant TB is concentrated in the countries of the former Soviet Union, India, China, and South Africa. About 390,000 people developed multidrug-resistant or rifampicin-resistant TB in 2024, and fewer than half received treatment.

Treatment, and how it works

In short: Four drugs for six months, because any single drug reliably selects resistant survivors and a subpopulation of dormant bacteria takes months to kill.

Drug-sensitive TB

The standard regimen is six months in two phases:

• Isoniazid blocks mycolic acid synthesis, destroying the cell wall's defining feature.
• Rifampicin blocks bacterial RNA polymerase, halting transcription. It is the backbone drug; resistance to it defines multidrug-resistant TB.
• Pyrazinamide works specifically in the acidic environment inside granulomas, which is why adding it shortened treatment from nine months to six.
• Ethambutol blocks cell wall assembly and is included largely to protect the others against resistance until sensitivities are known.

Why four drugs. Any large bacterial population contains a few spontaneously resistant mutants. Resistance to one drug arises at roughly 1 in 10^6 to 10^8 organisms; a cavity can contain 10^8 or more. Resistance to two drugs simultaneously requires the product of those probabilities, which is effectively impossible. Monotherapy therefore selects resistance reliably, and combination therapy prevents it. This is the same logic as HIV treatment and as combination chemotherapy in cancer.

Why six months. Most bacteria die in the first weeks. A small subpopulation of non-replicating persisters, tolerant rather than genetically resistant, survives far longer, and relapse follows if treatment stops before they are eliminated.

Adherence support is central rather than paternalistic: directly observed therapy, video-observed therapy, fixed-dose combination tablets, transport and food support, and tracing of contacts. Interrupted treatment is the main manufacturing process for resistant TB.

A four-month regimen using high-dose rifapentine and moxifloxacin is now an option for selected patients, the first shortening of drug-sensitive treatment in decades.

Drug-resistant TB

MDR-TB is resistance to at least rifampicin and isoniazid. Pre-XDR and XDR-TB add resistance to fluoroquinolones and other key drugs.

Treatment was, until recently, 18 to 24 months including daily painful injections that caused deafness in a substantial fraction of patients, with cure rates around 50 percent. The BPaLM regimen (bedaquiline, pretomanid, linezolid, moxifloxacin) is six months, all oral, and achieves far higher cure rates. Bedaquiline blocks the bacterium's ATP synthase, starving it of energy, a target no previous TB drug used. This is one of the clearest recent wins in global health.

Latent TB

Treating latent infection prevents future disease and is the main tool in low-incidence countries. Options include three months of weekly rifapentine plus isoniazid (3HP), four months of rifampicin, or six to nine months of isoniazid. Screening and treating latent TB is mandatory before starting TNF blockers or transplant immunosuppression.

Diagnosis, which is half the problem

Sputum smear microscopy is cheap and misses many cases. Xpert MTB/RIF, a cartridge-based molecular test, detects TB DNA and rifampicin resistance in under two hours and has transformed diagnosis where it is available. Culture remains the reference standard and takes weeks. AI-read chest X-ray is now WHO-endorsed for triage, which matters where radiologists are scarce. Despite all this, a large share of the world's TB cases are never diagnosed or reported, and finding them is the central bottleneck in TB control.

What treatment costs

Drug-induced hepatitis is the main reason regimens get interrupted, and it requires proper monitoring rather than avoidance of the drugs.

What the person can do

In short: Get a cough lasting three weeks investigated, complete the full course, and let household contacts be traced and tested.

• Get a persistent cough investigated, especially a cough lasting more than two or three weeks with weight loss, night sweats, or fever.
• Complete the full course. This is the single most important action, for the patient and for everyone else. Feeling better after two weeks is expected and is not cure.
• Get tested for HIV, since the two diseases travel together and treating HIV dramatically reduces TB risk.
• Let contacts be traced and tested. Household members and close workplace contacts should be screened for both infection and disease.
• Ventilate the home during the infectious period, and follow isolation advice for the first weeks of treatment.
• Nutrition matters. Undernutrition is both a cause and a consequence, and nutritional support during treatment improves outcomes; a large Indian trial (RATIONS) found that food supplementation for household contacts substantially reduced TB incidence among them.
• Stop smoking, which roughly doubles TB risk and worsens outcomes.

Living with it

TB carries heavy stigma nearly everywhere, associated with poverty, HIV, and contagion. People lose jobs, housing, and marriages over a TB diagnosis, and fear of that loss delays presentation, which increases transmission. In several high-burden countries the direct and indirect costs of an episode of TB (lost income, travel to clinics, food) exceed a household's annual income, and "catastrophic costs" from a curable disease are a formal WHO indicator for this reason.

Survivors are frequently left with permanent lung damage, and post-TB lung disease is only now being recognised as a distinct long-term condition needing follow-up.

What's next

• A better vaccine. The M72/AS01E candidate showed roughly 50 percent efficacy at preventing progression from latent infection to disease in a phase 2b trial and is now in phase 3. A vaccine of even moderate efficacy in adults would be the largest single advance possible against TB.
• Shorter regimens, including trials aiming at treating drug-sensitive TB in two months.
• Better diagnostics, particularly tests that do not require sputum (tongue swabs, urine antigen, breath) and that work in children and people with HIV, in whom current tests perform worst.
• Treating undernutrition as TB control, following the RATIONS results.
• Sustained funding, which is the structural risk: TB programme financing has repeatedly fallen short of targets, and WHO warns that recent gains are fragile.

Sources and notes

Case and death figures are from the WHO Global Tuberculosis Report 2025 (2024 data): approximately 10.7 million cases, 1.23 million deaths, 390,000 people developing MDR/RR-TB with about 42 percent accessing treatment. Koch's announcement: 24 March 1882. The 1948 MRC streptomycin trial is generally regarded as the first randomised controlled trial in medicine. Natural history of untreated TB (roughly 50 percent mortality): Tiemersma et al., PLoS ONE, 2011. BPaLM regimen: TB-PRACTECAL trial, NEJM, 2022. M72/AS01E: Tait et al., NEJM, 2019. RATIONS nutrition trial: Bhargava et al., The Lancet, 2023. Latent infection prevalence estimates have been revised downward in recent modelling and are deliberately not quoted as a single number here.

Open questions. How long latent infection genuinely persists, and what fraction of "latent" infections are actually cleared, are being actively revised. Whether BCG revaccination of adolescents helps is under trial. The best regimen duration for drug-sensitive TB may still be shorter than six months.

Next: the virus that made tuberculosis worse, and the most successful treatment story in modern medicine. 👉