Nontuberculous mycobacteria (NTM) are environmental organisms. They live in soil, in natural water and in engineered water systems, where biofilms and tolerance of disinfectants allow them to persist (Falkinham, 2009). Most people inhale or swallow them without consequence. Clinical disease is usually confined to one organ: slowly progressive lung disease in people with bronchiectasis or emphysema, cervical lymphadenitis in young children, or skin and soft tissue infection after injury or a procedure. Management of those syndromes is set out in international guidance (Daley et al., 2020).

Disseminated disease belongs to another category. When an NTM species is cultured from blood or bone marrow, or from several anatomically separate sites, the organism has crossed defences that ordinarily contain it, and the clinician should assume a defect in cell-mediated immunity until one has been looked for. Before effective antiretroviral therapy, disseminated Mycobacterium avium complex infection was a defining illness of advanced human immunodeficiency virus (HIV) infection and its relationship to the CD4 count was well described (Horsburgh, 1991). An adult with the same syndrome and a negative HIV test presents a harder problem.

That problem now has a recognised answer in a substantial proportion of cases. Adults, most of them of East or South East Asian origin, have been described with disseminated NTM and other intracellular infections caused by neutralising autoantibodies against interferon-gamma, a condition sometimes called adult-onset immunodeficiency (Browne et al., 2012). Other possibilities include late-presenting inborn errors of the interleukin-12 and interferon-gamma circuit, idiopathic CD4 lymphocytopenia, haematological malignancy and the effects of immunosuppressive drugs. This article proposes an ordered pathway for working through them. It is a proposal drawn from published series and guidelines, and it has not been validated in practice.

Control of mycobacteria depends on a conversation between infected macrophages and T cells. Macrophages produce interleukin-12, which drives T cells and natural killer cells to secrete interferon-gamma, which in turn activates the macrophage to restrict intracellular growth. Genetic lesions at almost every node of this circuit, in the receptors for interleukin-12 and interferon-gamma, in STAT1, in IRF8 and in the NEMO regulator, cause Mendelian susceptibility to mycobacterial disease, in which weakly virulent environmental mycobacteria and Bacille Calmette-Guerin cause severe infection in otherwise healthy people (Casanova and Abel, 2002; Bustamante et al., 2014). The classification of these and other inborn errors of immunity is maintained by an expert committee and is the reference point for genetic work-up (Tangye et al., 2022).

The same circuit can be interrupted without a mutation. High-titre neutralising autoantibodies to interferon-gamma block signalling and produce a phenocopy of the genetic defects in adults who were previously well (Browne and Holland, 2010). The large comparative study that defined the syndrome described patients in Thailand and Taiwan with disseminated NTM, salmonellosis, penicilliosis and varicella zoster, almost all of them with these autoantibodies and none with an identified genetic cause (Browne et al., 2012). Susceptibility is associated with particular HLA class II alleles, which fits the geographical concentration of reported cases (Chi et al., 2013). Follow-up of such patients has shown a chronic, relapsing course in which the autoantibody titre tends to persist (Chi et al., 2016).

Two further diagnoses complete the main differential. Idiopathic CD4 lymphocytopenia, defined as persistently low CD4 T cells without HIV infection or another cause, was characterised in a national survey when such patients came to attention during the HIV epidemic, and it predisposes to the same opportunistic organisms (Smith et al., 1993). Iatrogenic immunosuppression is probably now the commonest explanation of all: tumour necrosis factor antagonists in particular carry a recognised risk of mycobacterial disease (Winthrop et al., 2009), and corticosteroids, calcineurin inhibitors, cytotoxic chemotherapy and transplantation all contribute.

The pathway has seven steps. Figure 1 shows how they follow one another, and Table 1 gives the purpose of each step and the result that moves the patient on. The order is deliberate: inexpensive and widely available tests come before specialised ones, and the tests that change immediate management come before those that explain the mechanism.

Few hospitals can perform every test. The pathway is arranged so that each step is useful in isolation and so that the first four can be completed in any centre with a diagnostic laboratory and flow cytometry. Steps five and six depend on referral, and the practical requirement is a named referral route and an agreed sample protocol rather than local capability. In settings where tuberculosis is common, the first consideration is often whether the organism is Mycobacterium tuberculosis at all, since the therapeutic consequences differ entirely and misidentification is a recognised hazard. Where resources are limited, the greatest return comes from the early steps: confirming dissemination, getting the species right and excluding HIV infection and drug-induced immunosuppression.

This article proposes a sequence and does not evaluate it. Three questions would have to be answered before it could be recommended without qualification. The first is yield: in consecutive adults presenting with disseminated NTM infection, how often each step produces the explanation, and whether the order chosen here finds it sooner than an unstructured approach. The second is timing, since a pathway that reaches the right answer after many months of untreated infection has limited value; the interval from presentation to specific diagnosis would be the natural measure. The third is transferability, as the relative frequency of the underlying causes differs by region and the pathway may need reordering where the autoantibody syndrome is rare.

Two cautions apply to the pathway itself. A negative evaluation does not exclude an immune defect, and patients in whom no cause is found should be followed rather than discharged, because new phenotypes continue to be described. And the pathway addresses the cause of susceptibility, not the treatment of the infection, which follows the relevant guidance and the identified species.

Disseminated nontuberculous mycobacterial infection in an adult is a signal that the interleukin-12 and interferon-gamma axis or the T cell compartment has failed, and the cause can usually be identified if it is sought in an orderly way. The seven-step pathway proposed here moves from confirming dissemination and identifying the species, through the common acquired causes, to the autoantibody syndrome and the inborn errors, and it ends with a plan agreed between infection and immunology services. Its value in practice remains to be measured, and the measurements described above are within reach of centres that see these patients.

To be declared by the author(s) before publication.

To be declared by the author(s) before publication.