Tuberculosis Rarely Stays in the Lungs, Seven-Year Hospital Study Finds

Tuberculosis has a reputation as a lung disease, but a sweeping seven-year study from one of Europe’s leading infectious disease hospitals is challenging that picture in a way that could change how doctors everywhere hunt for the infection. Researchers at the National Institute for Infectious Diseases Lazzaro Spallanzani in Rome analyzed the records of 2,085 adults hospitalized with tuberculosis between 2018 and 2024, and found that nearly a third of them, 632 patients or 30.3 percent, had extrapulmonary tuberculosis, meaning the bacterium had spread beyond the lungs to other organs and tissues. What surprised the team most was not how common the extrapulmonary form was, but how rarely it appeared alone. In 91.8 percent of those cases, patients also carried concurrent pulmonary disease, and only 8.2 percent had truly isolated extrapulmonary tuberculosis with no detectable lung involvement at all.

The finding carries a punchy implication that the study’s authors argue should reshape clinical practice: tuberculosis is best understood not as a disease of a single organ but as a multisystemic infection that frequently seeds multiple sites in the same patient. When Mycobacterium tuberculosis establishes itself in the body, it does not politely confine itself to the respiratory tract. It can travel through the bloodstream and lymphatic system to lymph nodes, the pleural membranes surrounding the lungs, the abdomen, bones and joints, the genitourinary tract, and, in the most feared presentations, the central nervous system. The Rome data suggest that when clinicians find tuberculosis in one of these distant sites, the lungs are almost certainly involved as well, even when a patient reports no respiratory symptoms.

The anatomical distribution of disease in the cohort was strikingly consistent with patterns reported in other low-incidence settings. Lymph nodes were the most frequently affected site, involved in 42.7 percent of extrapulmonary cases, followed by the pleura at 38.6 percent, the abdomen at 20.1 percent, and the musculoskeletal system at 15.2 percent. Lymph node tuberculosis, often presenting as painless swelling in the neck, is notoriously difficult to distinguish from lymphoma or other causes of enlarged glands without tissue sampling. Pleural tuberculosis, in which the infection inflames the membranes lining the chest cavity, can mimic heart failure or malignancy. Abdominal tuberculosis can masquerade as inflammatory bowel disease, and skeletal tuberculosis can imitate ordinary arthritis or bone tumors. Each of these mimicries contributes to delayed diagnosis, which in turn allows the infection to progress and spread further.

To understand which clinical features tracked with different patterns of disease, the researchers used multivariable logistic regression, a statistical technique that adjusts for confounding factors to isolate genuine associations. Compared with patients who had isolated extrapulmonary tuberculosis, those with concurrent pulmonary and extrapulmonary disease were significantly more likely to present with cough, dyspnea, fever, and fatigue, with all of these associations reaching statistical significance at p values below 0.05. That makes intuitive sense: active lung disease generates the classic respiratory and constitutional symptoms that prompt patients to seek care. More intriguing, however, were the associations on the other side of the ledger. Malignancy and mental disorders were linked to isolated extrapulmonary tuberculosis, suggesting that patients with cancer or psychiatric conditions may present atypically, with their tuberculosis hiding outside the lungs and escaping detection by standard respiratory screening.

HIV co-infection emerged as a particularly important modifier of disease behavior. In the Rome cohort, patients living with HIV were more likely to have abdominal and central nervous system involvement. This aligns with long-standing clinical knowledge that immunosuppression dismantles the immune barriers, particularly the CD4 T-cell responses, that normally contain tuberculosis bacilli within granulomas, the organized clusters of immune cells that wall off the infection. When those granulomas fail, the bacilli disseminate more freely, and sites that are rarely affected in immunocompetent people, such as the meninges surrounding the brain, become vulnerable. Central nervous system tuberculosis is among the most dangerous forms of the disease, carrying substantial risks of death and permanent neurological disability even with prompt treatment, which makes the identification of HIV-associated disseminated disease a clinical priority.

One of the study’s most practically useful contributions is its detailed accounting of microbiological confirmation rates across specimen types. Overall, 79.3 percent of extrapulmonary tuberculosis cases were microbiologically confirmed, a high figure that reflects the diagnostic arsenal available at a tertiary referral center, including acid-fast bacilli microscopy, nucleic acid amplification tests, and mycobacterial culture. But the yield varied substantially depending on where the sample came from. Respiratory specimens, genitourinary samples, lymph node material, and osteoarticular samples showed the highest positivity rates, while other specimen types performed considerably worse. This variation matters because extrapulmonary tuberculosis is often described as paucibacillary, meaning the burden of bacteria at affected sites is low, and every negative test in a suspicious case delays treatment. Knowing which sampling strategies pay off can help clinicians maximize their diagnostic efforts rather than settling for a clinical diagnosis of exclusion.

On the question of drug resistance, the news from Rome was largely reassuring. Most isolates were fully susceptible to first-line anti-tuberculosis drugs, and multidrug-resistant tuberculosis, defined by resistance to at least isoniazid and rifampicin, the two most powerful drugs in the standard regimen, occurred in only 2.4 percent of cases. In an era when drug-resistant tuberculosis strains circulate globally and demand long, toxic, and expensive treatment courses, a low resistance rate in a large European cohort is a meaningful indicator that transmission of resistant strains remains limited in this setting. Nevertheless, the authors emphasize that drug susceptibility testing remains essential for every confirmed case, because even a small reservoir of resistant disease can expand rapidly without systematic surveillance.

Treatment outcomes in the cohort were similarly encouraging. At twelve months, treatment success stood at 85.0 percent, and there were no differences in success rates between patients with isolated extrapulmonary disease and those with concurrent pulmonary involvement. In-hospital mortality was 2.1 percent, a low figure for a disease that remains one of the world’s leading infectious killers. These outcomes likely reflect the advantages of care at a specialized referral center, where experienced clinicians, rapid molecular diagnostics, and multidisciplinary teams combining infectious disease physicians, microbiologists, radiologists, pathologists, and surgeons can be brought to bear on complex cases. They also suggest that when extrapulmonary tuberculosis is diagnosed and treated properly, the prognosis is generally good, even for patients with widespread disease.

The public health implications of the study extend beyond the hospital ward. Because extrapulmonary tuberculosis so rarely occurs in isolation, the authors argue that every patient found to have tuberculosis at any site should undergo systematic bidirectional assessment, meaning that clinicians should look for pulmonary disease in patients presenting with extrapulmonary findings, and vice versa. This matters for infection control as well as individual care, since patients with concurrent pulmonary involvement may be capable of transmitting the infection through respiratory droplets even when their presenting complaint is a swollen lymph node or abdominal pain. Missing the pulmonary component means missing the opportunity to isolate the patient, trace their contacts, and interrupt transmission chains. In low-incidence countries like Italy, where tuberculosis elimination is the stated goal, each missed case carries a disproportionate epidemiological cost.

The study, published in BMC Infectious Diseases and funded by the Italian Ministry of Health through its Ricerca Corrente program for research institutes, arrives at a moment when global tuberculosis control remains stubbornly off track. The World Health Organization continues to record millions of new cases annually, and the proportion of extrapulmonary disease has been rising in low-incidence settings as pulmonary cases decline faster. What the Rome experience demonstrates is that the old diagnostic habit of thinking of tuberculosis in compartments, pulmonary here and extrapulmonary there, no longer matches the biological reality of the infection. The bacterium behaves as a systemic pathogen, and the diagnostic response must be equally systemic: aggressive sampling of all suspected sites, use of the highest-yield specimen types, bidirectional screening of lungs and other organs, and heightened vigilance in patients with HIV, malignancy, or mental health conditions whose disease is most likely to hide in plain sight.

Subject of Research: Epidemiology, clinical presentation, and diagnostic yield of extrapulmonary tuberculosis in a low-incidence setting

Article Title: Extrapulmonary tuberculosis: a 7-year retrospective study at a tertiary referral hospital in Rome

Article References: Extrapulmonary tuberculosis: a 7-year retrospective study at a tertiary referral hospital in Rome. (n.d.). https://doi.org/10.1186/s12879-026-14320-4

Image Credits: AI Generated

DOI: 10.1186/s12879-026-14320-4

Keywords: extrapulmonary tuberculosis, tuberculosis, Mycobacterium tuberculosis, HIV co-infection, diagnostic yield, drug resistance, multidrug-resistant tuberculosis, low-incidence setting, lymph node tuberculosis, pleural tuberculosis, public health, retrospective study