Impact of Mycobacterium tuberculosis complex lineages as a determinant of disease phenotypes from an immigrant rich moderate tuberculosis burden country

Tuberculosis (TB) caused by Mycobacterium tuberculosis complex (MTBC) remains as a major public health challenge despite of implementing several control programs [1]. World Health Organization (WHO) estimated in 2016 that, there were 10.6 million newly diagnosed TB cases including 490,000 multidrug resistant cases and 1.4 million deaths [1]. The MTBC primarily infects the lungs, but virtually can affect any site in the body. Extrapulmonary TB (EPTB) is a difficult form of TB to diagnose and treat and results in high mortality and morbidity. Recent national level report from Saudi Arabia, showed 25.6% of annually reported new TB cases were EPTB [2]. Interestingly, the annual EPTB incidence rate in Saudi Arabia is comparatively higher than developed countries, where EPTB incidences are even increasing steadily [3, 4].

In EPTB, lymph nodes, meninges, kidney, spine and joints/bones are mostly affected. However, TB in pericardium, peritoneum, pleura, liver, intestine, skin, genitourinary tracts, spleen, ileum, cecum, eyes, breast, penis, are also occasionally reported [5]. EPTB is mostly observed as a key indicator of immunosuppression. To date, bacterial determinants of pulmonary TB (PTB) or EPTB have not been clearly estimated, although large attention is directed to explore the host and pathogen determinants. There have been controversies in correlating specific lineages with morbidity and mortality due to TB infection. Several previous studies revealed conflicting evidences on the association of MTBC phylogenetic lineages and site of infection [6, 7]. However, such analysis was scarce from the Middle Eastern countries including Saudi Arabia. Saudi Arabia has a highly diverse population structure including 10.4 million immigrant workers from around the world. In addition, annually the country receives 10 million pilgrims to the Islamic holy cities. This population diversity was reflected highly on the spectrum of MTBC lineages in the country, with the presence of almost all defined lineages including indigenous strains [8‐10]. Although, a recent study reported the lineage spectrum of MTBC in PTB and EPTB cases in the country, no detailed systematic analysis was carried out to date [8]. Therefore, for the first time in the country a multicenter study on a large cohort of pulmonary and extrapulmonary TB cases has been carried out to analyze the impact of MTBC lineages towards disease phenotypes.

During the study period, 1003 extrapulmonary isolates and 1089 pulmonary isolates were enrolled. Matching the age group of patients against the PTB and EPTB incidence showed significant variations with statistical associations. The age group below 14 years showed a statistically significant predominance of EPTB (P value < 0.0001).On the other hand, PTB incidence was more significant among the age group 15–29 and 49–59. Overall, 76.6% of EPTB cases were found with an age below 45 years. Among, EPTB patients 83.3% were Saudi nationals. Demographics and other patient characteristics have been summarized in Table 1 and Additional file 1. Distribution major lineages among the study population based on their nationality is depicted in Additional file 2.

Relationships between MTBC phylogenetic lineages and clinical site of TB have been analyzed systematically in a cohort of Saudi Arabian patients for the first time. Ahead of the previous studies which reported a primary data on the site of infection and mycobacterial lineages in the country, the current study congregates with a detailed outcome from a large population of diverse patient cohort of 2092 cases [8]. However, previous international studies also showed inconsistent findings on the association of certain phylogenetic lineages with particular infection sites and disease presentation of tuberculosis [6, 7, 12, 13]. The current findings showed positive association of MTBC lineages (Delhi/CAS, EAI, Uganda-I and M.bovis) with site of infections such as lymph nodes, gastrointestinal and central nervous systems respectively.

Demographical findings of the cohort showed a domination of Saudi nationals. This finding is inversely proportional to the national tuberculosis data where Non-Saudis were reported with more TB incidences [2]. However, the finding is non-conclusive and the expected reason for this variation is that, opted study centers were four referral hospitals including a military facility which provide services largely to citizens only. The gender of study subjects showed predominance of male, which corroborates with the findings of many recent studies which showed similar trend [2, 8, 14]. In addition, localization of TB manifestations strongly depends on several clinical features. Previously younger age is one of the most common associated with extrapulmonary infections [15‐17]. The current findings also corroborate with previous studies and showed statistically significant predominance of younger age towards EPTB.

HIV positivity in the cohort was very much limited (0.33%), although 58.5% of the total cases only underwent the testing. This finding is in concordance with the low rate of annually reported HIV positive cases among the local population (1.5 cases per 100,000 for Saudis) in the country [18]. HIV testing data was inconsistently recorded (46% not tested) in the study, while untreated HIV infection is well recognized as an important factor in determining an extrapulmonary clinical phenotype, therefore its influence on clinical phenotype could not be analyzed [17].

The site of infections followed very detailed classification and reported as highly diverse with several rare sites. Lymph node was the most commonly infected sites followed by the gastrointestinal, osteoarticular and central nervous systems. This finding shows similar trend reported in a recent Saudi Arabian and international studies [3, 14, 19]. The current cohort of EPTB cases revealed several rare infection sites such as bone marrow, breast, tongue, testis, kidney, prostate glands, spleen, pancreas, duodenum, cecum and jejunum as observed in available literature [20, 21].

The major advantage of our data compared to previously published studies with limited sample size and sample diversity was the presence of all defined lineages. Phylogenetic diversity of pulmonary and extrapulmonary isolates showed the presence of all defined MTBC lineages (Lineage 1–7) in the country. Interestingly, for the first time in the country, presence of M. africanum lineages West African I and II were noticed. M.africanum has not been detected in any of the previous nationwide studies [8‐10]. Saudi Arabia annually receives approximately 8–10 millions of pilgrims (from 184 countries) from TB endemic regions, in addition to its migrant workers population of 10.4 million from around the world. This massive influx of foreign nationals solely induces a higher impact on the population structure of MTBC in the country [10, 22]. The increased possibilities of TB transmission and exportation in mass gathering particularly during Hajj were highly projected [23]. Therefore, the higher diversity of strain lineages in the study could be well explained.

Comparative analysis on the strain diversity between PTB and EPTB was a key objective of the study. The findings showed the presence of all the lineages among both groups though the proportion varied. Ancestral lineages were found relatively high among EPTB cases, which are in concordance with previously published studies from other part of the world including TB endemic regions [24, 25]. On the other hand, “Modern” strains were more common among PTB. Predominance of “Modern” strains among the pulmonary cases in the country was previously well documented, corroborating with other global regions [6, 8, 13].

Detailed statistical analysis to find an organ or site specific preponderance of MTBC lineages showed significant association of ancestral lineages, Delhi/CAS and M.bovis to lymphadenopathy and gastrointestinal TB. Although, previous studies showed predominant isolation of Delhi/CAS and M.bovis from lymph nodes and gastrointestinal sites only few of them were statistically significant [24, 26, 27]. Interestingly, association of lineage EAI with central nervous system TB which mainly included tuberculous meningitis (TBM) was significant. Although previous studies showed the mycobacterial genotypes could play a major role in disease severity, emergence of drug resistance, host response and transmissibility in TBM cases, findings of statistical significance between certain lineages and CNS disease sites are highly scarce [28, 29]. Association of Uganda-I lineage to gastro intestinal sites of infection was another significant finding. The Uganda-I lineage has not been studied in detail as a cause of EPTB and therefore very much limited information is only available in literature on its association with any clinical manifestation [30].

The influence of genomic diversity on pathobiological properties such as transmissibility, virulence, immune responses, and clinical manifestations were established in several previous studies [6, 31, 32]. In addition, lineage specific difference in mycobacterial transcriptomic responses, cytokine induction patterns in animal or cellular infection models also were proved in recent studies [33‐35]. One of the most intensively studied lineage Beijing, showed several potential mechanisms as evidences of associations towards clinical manifestations [36]. In animal models less protective Th1 response and high virulence has been documented when infected with Beijing strains [35]. Other proposed mechanisms in Beijing strains, which increase the pathobiological adaptations and phenotypic stability are DosR up- regulation and production of a phenolic glycolipid by an intact pks15–1 [37, 38]. However, the real mechanisms behind all such associations particularly on less frequently encountered lineages are still not well established.

The study has few limitations mostly related to sampling and data analysis. Although, the study population largely consisted of Saudi nationals it cannot be considered as a nationwide population based analysis. The cohort was a selective group of citizens and mostly from the central region of the country. The expatriate population was not well represented mainly due to their eligibility of treatment in study centers although they represent 33% of total population of the country. Analysis of confounding factors behind the EPTB in the current cohorts was limited and this was not included as a major objective of the study. In addition, 5.1% of the total isolates in the study could not be assigned into “defined” lineage even after combining the MIRU and spoligotyping techniques, and no alternative technique was utilized to define their phylogenetic nature.