Tuberculosis is a neglected tropical disease caused primarily by Mycobacterium tuberculosis, the most epidemiologically significant species within the genus Mycobacterium, and it affects individuals of all ages but disproportionately impacts adults in low- and middle-income countries due to socioeconomic vulnerabilities that heighten exposure and hinder access to care [¹,²]. Transmission occurs mainly through the inhalation of bacilli released into the air when individuals with active pulmonary tuberculosis cough, speak, or sneeze, while ingestion of milk contaminated with M. bovis represents a less common gastrointestinal transmission route [³,¹]. Infectivity persists as long as bacilli are present in sputum, typically declining after approximately 15 days of appropriate treatment. However, ongoing risks persist in communities burdened by overcrowding, limited ventilation, and poor living conditions, all of which amplify the potential for transmission. Furthermore, the disease’s lengthy and complex treatment regimen often results in low adherence, which compromises therapeutic success, prolongs infectiousness, and perpetuates transmission cycles, particularly among populations experiencing social inequality and restricted access to healthcare services.

Tuberculosis remains a significant public health problem. In Brazil, more than 85,000 new cases were reported in 2024, representing a significant increase over previous years, even relative to the pre-pandemic period, when annual totals did not exceed 80,000 cases (including new and retreatment cases). Among new cases of confirmed pulmonary tuberculosis, 36.2% were closed due to abandonment – a proportion 2.6 times the WHO-established 5% threshold – during the 2020 – 2022 period. This increase appears to be related to disruptions in health services and systems caused by the COVID-19 pandemic [⁴,⁵,⁶,⁷].

Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2), responsible for COVID-19, has infected more than 35.5 million Brazilians since 2020, causing an infectious respiratory disease transmitted through small liquid particles and often characterized by a severe inflammatory response that disrupts immune function [⁸,⁹]. Both pulmonary tuberculosis and COVID-19 affect the respiratory system, potentially causing permanent alterations that compromise lung parenchyma and pulmonary function [¹⁰,¹¹,¹²]. During the COVID-19 pandemic, priority public health care was concentrated on SARS-CoV-2 containment, which significantly affected the monitoring and treatment of other diseases, including TB – reflected in reductions in case notifications and likely increases in undiagnosed patients [¹⁰,¹³,¹⁴,¹⁵,¹⁶].

People experiencing homelessness (PEH) form a heterogeneous and vulnerable population surviving in conditions of extreme material deprivation – including lack of stable shelter, basic sanitary conditions, and economic resources – which renders them susceptible to infectious diseases, treatment barriers, and adverse health outcomes. During the SARS-CoV-2 pandemic, the homeless population grew in several settings at a rate disproportionate to general population growth [¹⁷,¹⁸].

Despite growing literature on TB and COVID-19, evidence specifically addressing PEH remains fragmented, heterogeneous, and methodologically limited. Most available studies employ divergent operational definitions of homelessness, and PEH-specific denominators are often not reported separately from broader vulnerable-population data. Understanding the TB burden and care disruptions experienced by PEH during the COVID-19 era – and distinguishing direct from indirect evidence – is essential for informing targeted public health responses.

The objective of this narrative review was to describe TB-related outcomes and care disruptions among people experiencing homelessness (PEH) during the COVID-19 era (2020 – 2025) using primary observational evidence.

This narrative review was designed to synthesize primary observational evidence on TB-related outcomes among PEH during the COVID-19 era. A narrative approach was selected given the substantial heterogeneity of the available evidence across study designs, population definitions, and outcome measures, which precluded a formal meta-analytic synthesis. No formal protocol was pre-registered; this methodological limitation is acknowledged.

Primary studies were prioritized to anchor the main findings. Secondary reviews (Laycock et al., 2021; Zhu et al., 2023) identified during the search were excluded from the primary evidence set and are used solely to contextualize background information, mechanisms, or service-delivery dynamics [¹⁹,²⁰]. These secondary sources are clearly distinguished from primary study findings throughout the manuscript.

The seven included primary studies were conducted across multiple countries and settings, including Brazil, Italy, France, and Kazakhstan. Study designs included retrospective cohorts (Gabdullina et al., Campana et al., Di Gennaro et al.)[²²,²³,²⁴], a case-control study (Alsaidi et al.)[²¹], prospective observational analyses (Di Gennaro et al.)[²³], and descriptive surveillance analyses (Dias et al., Zuim et al., Vidal et al.)[²⁶,²⁵,²⁷]. Publication years ranged from 2021 to 2025. summarises key characteristics, including the operational definition of homelessness used in each study and the degree to which PEH-specific data are separable from broader populations.

The risk-of-bias assessment revealed substantial variability in methodological quality. Three studies with robust designs and appropriate adjustment for confounding (Di Gennaro et al., Gabdullina et al., Campana et al.)[²²,²³,²⁴] were classified as low risk, and the study by Alsaidi et al. was classified as moderate risk due to limited comparability or partial representativeness of PEH [²¹]. The remaining studies were at high risk of bias due to indirect populations, descriptive designs, or a lack of control for confounding. Full ratings are provided in Supplementary Table S2.

In interpreting these findings, readers should note that the included studies employ heterogeneous operational definitions of homelessness – ranging from formally street-dwelling adults to shelter residents, shared dormitory occupants, and administrative or socially vulnerable classifications. These differences limit direct cross-study comparability, and findings from studies with indirect or partial PEH definitions are explicitly labeled as such throughout this section.

Evidence hierarchy note: The following results are presented according to an explicit evidence hierarchy. PEH-specific quantitative studies drive the primary findings (labeled ’Direct PEH evidence’). Descriptive and surveillance analyses provide supporting context (labeled ’Supporting evidence’). Qualitative and contextual literature are used to explain mechanisms and barriers (labeled ’Contextual/indirect evidence’).

Direct PEH evidence – Campana et al. (2025; São Paulo, Brazil; retrospective cohort; low risk of bias): In São Paulo, between 2017 and 2022, more than 39,000 TB cases were reported. Among PEH specifically, the proportion of unfavorable outcomes remained persistently high and statistically stable across the pre-pandemic and pandemic periods (57.8% pre-pandemic vs. 59.5% pandemic; no significant period-associated risk). Dropout rates were approximately 50%, and TB mortality exceeded 8% throughout. In contrast, unfavorable outcomes in the general population increased significantly during the pandemic period (from 21.8% pre-pandemic to 28.9% during the pandemic; HR 1.45; 95% CI: 1.37 – 1.55). These PEH-specific figures represent the most methodologically robust estimates of pandemic-era changes in TB outcomes in this population available in the reviewed literature [²²].

Direct PEH evidence – Gabdullina et al. (2023; Almaty, Kazakhstan; retrospective cohort; low risk of bias): A retrospective study of TB treatment outcomes among socially vulnerable populations from 2018 to 2021 found that unfavorable outcomes increased markedly during the pandemic period compared with the pre-pandemic period (20% vs. 11%). Mortality rose (9% vs. 6%). After adjustment for age, sex, HIV status, and alcohol use, homelessness was independently associated with unfavorable outcomes (aRR approximately 2.7 – 2.9). These PEH-specific adjusted estimates are among the strongest quantitative evidence available for homelessness as a risk factor for poor TB outcomes [²⁴].

Direct PEH evidence – Zuim, Almeida & Soeiro (2025; Rio de Janeiro, Brazil; descriptive surveillance; moderate risk of bias): A descriptive analysis of Rio de Janeiro surveillance data from 2016 to 2021 documented 2,965 TB cases among the homeless population – 1,708 (57.6%) new cases and 1,257 (42.4%) retreatment cases. Treatment interruption rates were high: 50.3% in new cases and 61.1% in retreatment cases. Proportions of deaths were 10.1% and 6.5%, respectively. The authors noted that COVID-19 worsened pre-existing inequalities and TB indicators in this population [²⁷].

Direct PEH evidence – Di Gennaro et al. (2022; Bari, Italy; prospective cohort; low risk of bias): Among 206 TB-diagnosed patients admitted from 2013 to 2021, 17 (8.3%) were identified as experiencing homelessness. Overall treatment success was 57%; TB-attributable mortality was 1% (n=4). Social vulnerability – including homelessness and prolonged street residence – was associated with longer hospitalization duration, particularly among patients under 65 years of age (43 vs. 24 days). This study provides direct PEH-specific clinical outcome data within a broader cohort, though the PEH subgroup is small [²³].

Supporting evidence – Dias et al. (2023; Brazil; descriptive-ecological; high risk of bias): A national descriptive analysis of pulmonary TB in Brazil from 2019 to 2021 estimated that PEH accounted for 4.2% of all pulmonary TB diagnoses nationally, with the Southern Region showing the highest proportional representation. Structural barriers – including underreporting, insufficient primary care coverage, and limited community health agent outreach – were identified as persistent drivers of diagnostic delay and treatment dropout. These findings provide national surveillance context but do not enable PEH-specific outcome calculations independent of broader population data [²⁶].

Contextual/indirect evidence – Alsaidi et al. (2021; Marseille, France; case-control; moderate risk of bias): In a case-control study of 64 participants from a mixed vulnerable population – including immigrants and individuals experiencing homelessness, supported by a social assistance organization – conducted during March – May 2020, 23.4% were hospitalized. Risk factors for COVID-19 infection included shared dormitory residence, poor social distancing adherence, and prior TB history or multiple comorbidities. This study provides contextual evidence regarding COVID-19 risk in mixed vulnerable populations that included PEH, but PEH-specific outcomes are not separable from the broader study population [²¹].

Contextual/indirect evidence – Vidal, Coelho & Merhy (2024; Rio de Janeiro, Brazil; qualitative field study; CASP appraisal): A qualitative study conducted between May and December 2021, from the perspective of healthcare professionals linked to street clinics, documented structural barriers to TB care for PEH: fragmented service delivery, difficulties in maintaining continuity of care, and reliance on mobile outreach units as a compensatory strategy. These findings are used to explain mechanisms and service-delivery context rather than to quantify outcomes [²⁵].

Supporting evidence – Laycock et al. (2021; narrative review; repositioned as contextual reference): As a secondary review, this work is not included in the primary evidence set. It is noted here for contextual purposes: focusing on TB in adolescents and young adults aged 10 – 24 years from 2016 to 2021, the review documented that TB progression risk increases with puberty-related immunological changes, and that homeless and marginalized youth face elevated TB risk due to congregate living and healthcare access barriers. This provides background to understand age-specific TB vulnerability [¹⁹].

Contextual reference – Zhu et al. (2023; systematic review; repositioned as contextual reference): As a secondary review, this work is not included in the primary evidence set. For context: this systematic review of 21 articles documented that mass testing in shelter settings identified high proportions of asymptomatic COVID-19 cases (up to 86% in one study) and that shelter users had higher SARS-CoV-2 prevalence than unsheltered PEH. Prolonged shared environments were the main transmission factor identified for TB in this setting [²⁸,²⁰].

This narrative review synthesized primary observational evidence from seven studies to describe TB-related outcomes and care disruptions among PEH during the COVID-19 era. The evidence hierarchy applied throughout this synthesis distinguishes between PEH-specific quantitative studies – which anchor the primary conclusions – and descriptive, qualitative, or contextual sources, which are used to explain mechanisms and service-delivery dynamics. Collectively, findings indicate a severe and disproportionate TB burden in this population, characterized by persistently high rates of treatment interruption, prolonged hospitalizations, and elevated mortality [²⁹]. These patterns predate the COVID-19 era; available evidence suggests that the pandemic in PEH did not uniformly worsen them, though disruptions in the broader population are well documented.

Several limitations should be considered when interpreting these findings. First, the heterogeneity of operational definitions of homelessness across included studies prevents direct cross-study comparisons. Studies range from those explicitly focusing on street-dwelling adults to those using administrative, sheltered, or mixed vulnerable-population classifications; the latter are clearly characterized as indirect evidence throughout this review, but their inclusion introduces imprecision into the overall synthesis of evidence.

Second, a significant proportion of included studies were at moderate to high risk of bias, particularly due to the use of administrative population definitions, limited control for confounding, and small PEH subgroup sizes. Findings from high-risk studies, while consistent with the overall pattern, should be interpreted with caution.

Third, formal quality appraisal tools were applied only to primary studies; the two secondary reviews were repositioned as contextual references and were not formally appraised. This is methodologically appropriate but limits the ability to assess the quality of the contextual literature drawn upon in the Discussion.

Fourth, the restriction to six databases and three languages may have introduced selection bias, potentially missing relevant evidence published in other languages or in grey literature. No protocol was pre-registered, which is acknowledged as a methodological limitation consistent with the narrative design.

Finally, the predominantly descriptive nature of the available primary evidence precludes rigorous causal inference about the pandemic’s effects on PEH TB outcomes. What can be concluded is that unfavorable TB outcomes remain unacceptably high among PEH across the reviewed period, and that pandemic-era disruptions likely affected an already critical situation – the magnitude and direction of any marginal pandemic effect in PEH specifically cannot be robustly estimated from this evidence base.

This narrative review described TB-related outcomes and care disruptions among PEH during the COVID-19 era (2020 – 2025). The most methodologically robust evidence confirms that PEH are disproportionately affected by unfavorable TB outcomes – including high rates of treatment interruption, prolonged hospitalization, and elevated mortality – driven by structural vulnerabilities operating across individual, social, and programmatic dimensions.

Pandemic-era disruptions were associated with increased unfavorable TB outcomes in the general population in multiple settings. Among PEH, outcomes remained persistently high across pre-pandemic and pandemic periods in the most robust study available, suggesting a pre-existing ceiling of vulnerability; in another high-quality study from a different setting, pandemic-era worsening was documented. This inter-study heterogeneity underscores the importance of context-specific, PEH-targeted surveillance and service delivery.

Evidence on TB – COVID coinfection in PEH specifically remains sparse, heterogeneous, and indirect. Future research should prioritize well-designed prospective studies with explicit PEH definitions, separable PEH-specific outcome denominators, and appropriate comparator designs to better characterize the epidemiological dynamics of TB and TB – COVID coinfection in this population.