According to estimates from the World Health Organization, 10 million people became ill with tuberculosis (TB) in 2017, and 1.3 million people died as a result1. This shows that TB continues to be an important, unresolved public health issue that causes an unacceptable level of mortality, especially considering that it is a treatable, preventable disease.

Since the pulmonary form is the most common, and its epidemiological surveillance has been the priority for the control of disease transmission, there have been limited studies and publications to date on extrapulmonary forms (EP-TP). Evidence regarding the risk factors of EP-TB is scarce, as is knowledge on TB among immigrants and its differential features in this population group. In Spain, the Program for Tuberculosis Research (PII-TB), created by the Spanish Society for Pneumology and Thoracic Surgery (SEPAR), has undertaken several studies to identify the distinctive features of cases among immigrants in our country2.

Several publications in Spain, the United States and the United Kingdom have shown that the clinical form of TB is changing 3–10, with a year-on-year increase in the rate of EP-TB, mostly in the immigrant population. EP-TB proportion now reaches almost 50% of all TB cases in some ethnic groups8–10. Diagnosing the forms of TB-EP requires a high initial clinical suspicion and can be complicated due to the difficulty in obtaining microbiological samples. The symptoms and signs can be nonspecific and many times, the chest-x-ray can be normal and the sputum smear negative. For this reason, the delay in diagnosis is frequent and can condition an increase in morbidity and mortality.

The objective of this study was to describe sociodemographic, epidemiological, microbiological and clinical variables associated with the various forms of TB and the factors associated with EP-TB risk in Spain, such as belonging to specific foreign-born population groups.

The PII-TB database contained 5058 registered cases of TB corresponding to the study period, of which 908 (17.9%) had EP-TB. The percentage of EP-TB was higher among the foreign-born population (22.1%) than among native Spaniards (15.9%; p<0.001). We also observed differences according to country of origin, with EP-TB forms being more prevalent in immigrants from India-Pakistan (42.8%), other Asian countries (25.3%), Maghreb countries (20.2%), other African countries (30.1%) and Latin America (16.8%). The proportion tended to be lower in individuals from Eastern Europe (10.8%) and Western developed countries (9.8%) (Table 1).

Sociodemographic, epidemiological, microbiological and clinical features of pulmonary TB and EP-TB cases are described in Tables 2 and 3. Our descriptive analysis showed a higher proportion of EP-TB than pulmonary TB in women, immigrants, patients older than50 years, patients with a delay in diagnosis of over 51 days (median days from onset of symptoms to initiation of treatment12), and people co-infected with HIV. In addition, most patients with EP-TB showed a negative sputum smear test and culture, and lower resistance to some antituberculosis drugs. We found no statistically significant differences in the proportion of pulmonary and EP-TB among cases of multi-drug-resistant TB (MDR-TB). A higher proportion of EP-TB cases were treated with 4 drugs rather than 3, and EP-TB cases also showed more disfavourable disease progression and had higher risk of dying as a result of their disease.

Our multi-variate analysis (Table 4) showed that EP-TB was associated with being older than 50 years, being an immigrant, HIV co-infection, a diagnostic delay of >51 days, a negative smear test, negative culture, and poorer clinical progression toward advancing disease.

Our analysis of the cases registered in SEPAR's PII-TB database during this 8-year study period showed the associated factors with EP-TB.

Industrialized countries with a low TB incidence have experienced increase in the rates of EP-TB, partly because of high immigration from low-income countries with high TB incidence8,10,12,13. The incidence of EP-TB is generally lower than pulmonary TB. According to the WHO's 2016 Global Tuberculosis Report, EP-TB is estimated to account for 15% of the total number of new TB cases worldwide1. However, the proportion of EP-TB cases varies between regions, ranging from 8% in the Western Pacific to 23% in the Eastern Mediterranean. Reports from the European Union have shown marked differences between countries, ranging from 4% in Hungary to 48% in the United Kingdom14,15.

Our study shows that extrapulmonary forms are proportionally more common among the immigrant population, which is consistent with other epidemiological studies from several regions and countries in Europe and in the United States7,11,12,14-16.

Traditionally, risk factors associated with the development of EP-TB include immunodeficiency (such as HIV co-infection17,18), very old age, being female, and being non-white7,18,19. However, little is known about other clinical, epidemiological or even microbiological factors. As shown by several studies20-22, there are also some open questions regarding the association between the phylogenetic lineage of strains from different zones and disease presentation.

There seems to be a direct link with host-related factors, and there are unquestionable classic co-morbidities such as immunocompromising disease and HIV17,23. However, there have to be probably more factors, given that rates of HIV co-infection have become less and less common, both in receiving countries and countries of origin21 (e.g. HIV co-infection was less than 4% in our study).

Analysing the geographical origin of the cases in our study, we found that the highest proportion of EP-TB was among Asian-born patients, particularly those born in South Asia (India-Pakistan). This is consistent with studies from the United Kingdom and Germany11,12. In addition, studies from other Asian regions, such as Saudi Arabia, also reported high incidence of EP-TB24, even though the prevalence of traditional risk factors (such as HIV or alcohol and drug abuse) is decreasing in these regions. This forces us to consider whether other factors could be involved, such as metabolic syndromes or genetic polymorphisms that could interfere immune response. For instance, in this population that has been studied, it has been found that the incidence of consanguinity is up to 67% which may be related to diseases of the immune system23,24.

Regarding gender and age, we found that the incidence of EP-TB was higher among women and people aged >50 years. Similar rates have been found in numerous studies worldwide, including the United States, Germany, Denmark, Benin and Taiwan7,12,16,18,25. These studies found that women were more prone to develop EP-TB than men, particularly in their post-reproductive period16,25. Neither our study nor any of the others has shown a link to bacterial load, virulence or strain resistance, which suggests that there are other specific risk factors that make women and individuals age >50 years more vulnerable to disease, such as hormonal factors.

We observed very low proportion of resistance to one or multiple drugs, and also few cases with MDR-TB strains, both in pulmonary TB and in EP-TB. We did not find statistically significant differences in resistance between these pulmonary and EP-TB forms. This also suggests that the form of disease presentation is more strongly associated with host-related factors than with intrinsic features of Mycobacterium tuberculosis, in terms of resistance or virulence.

In addition, from a clinical point of view, patients with EP-TB are much more difficult to diagnose. These cases frequently require invasive techniques and procedures for microbiological confirmation, resulting in significant diagnosis delay. They also have a worse clinical outcome (statistically significant) with a higher frequency of disease progression and greater risk of TB-associated death 26,27.

The main limitation of this study is the missing data for some important variables, as the high percentage of patients in which the result of HIV or radiological information had not been collected. Other limitation of this study is the lack of information about the exact location of EP-TB, because this information is not collected in our database.

Nevertheless, we believe that the high number of cases collected prospectively by investigators from all over Spain can be considered the main strength of this study.

The results of our study show a high frequency of EP-TB among several groups, such as immigrants, individuals with HIV, and older people.

Therefore, it is in these groups with high risk of developing extrapulmonary forms, where programs should focus on the research of active cases and in the treatment of latent infection in order to prevent its development.

There are other risk factors clearly related to EP-TB that highlight the importance of increasing clinical suspicion and improving the microbiological and/or histological diagnosis of cases in order to introduce better and earlier treatment regimens, despite of their low infectivity, EP-TB contributes significantly to the total TB load, so increased attention is needed in order to achieve the control and the elimination of the disease.

LL, TR and JAC conceived of the study, participated in its design, literature search, writing original draft, supervision, project Administration and Funding Acquisition.

The principal authors (LL, TR, JPM, JMG, AO, JA, MC) have taken part in the data collection, data interpretation and Writing – Review & Editing.

MC performed the statistical analysis. The Working Group of Integrated Program of Tuberculosis Research have taken part in the collection of the cases.

The study was funded by a SEPAR research grant (058/2014).

The authors declare that they have no financial or no-financial competing interests.