Systematic review of latent tuberculosis infection research to inform programmatic management in Ireland

The World Health Organisation (WHO) End Tuberculosis (TB) Strategy and the WHO Framework Towards Tuberculosis Elimination in Low Incidence Countries state that latent tuberculosis infection (LTBI) screening and treatment in selected high-risk groups is a priority action to eliminate TB. The European Centre for Disease Prevention and Control (ECDC) advises that this should be done through high-quality programmatic management, which they describe as having six key components. The research aim was to systematically review the literature to identify what is known about the epidemiology of LTBI and the uptake and completion of LTBI screening and treatment in Ireland to inform the programmatic management of LTBI nationally. A systematic literature review was performed according to a review protocol and reported in adherence with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement. Twenty-eight studies were eligible for inclusion and described LTBI screening or treatment performed in one of five contexts, pre-biologic or other immunosuppression screening, people living with HIV, TB case contacts, other vulnerable populations, or healthcare workers. The risk of bias across studies with regard to prevalence of LTBI was generally high. One study reported a complete cascade of LTBI care from screening initiation to treatment completion. This systematic review has described what published research there is on the epidemiology and cascade of LTBI care in Ireland and identified knowledge gaps. A strategy for addressing these knowledge gaps has been proposed.


Introduction
Reactivation of latent tuberculosis (TB) infection (LTBI) is a significant challenge for global TB elimination efforts. It is estimated that 23% of the world's population and 13.7% of Europe's population have LTBI [1]. The World Health Organisation's (WHO) End TB Strategy and Framework Towards Tuberculosis Elimination in Low-Incidence Countries state that LTBI screening and treatment in selected high-risk groups are priority actions to eliminate TB [2,3]. The European Centre for Disease Prevention and Control (ECDC) advises that this should be done through highquality programmatic management, which they describe as having six key components (Table 1) [4].
Risk groups with a high prevalence of LTBI or a high risk of TB reactivation should be prioritised for LTBI screening and treatment [2][3][4]. For some cohorts, whether programmatic LTBI screening and treatment occurs depends on the country-specific epidemiology of LTBI and the resources available for screening and treatment (Table 2) [3,4]. As well as identifying cohorts who should be screened and treated for LTBI, it is important to know whether programmatic LTBI management in these cohorts is feasible by having prior knowledge of the uptake and completion of LTBI screening and treatment (known as the cascade of care) and having considered its cost and cost-effectiveness [4].
Many countries with a low incidence of TB are establishing programmatic LTBI management to achieve TB elimination after researching the prevalence of LTBI in different cohorts and the feasibility of programmatic LTBI management. In the United Kingdom (UK), they have identified that immigrants from countries with a very high incidence of TB contribute significantly to the case burden nationally [5,6]. They have demonstrated a high prevalence of LTBI among these immigrant cohorts and demonstrated that the rate of TB reactivation over time was significant, suggesting that TB reactivation, as opposed to primary TB infection, explained the high TB incidence in this cohort [7,8]. Furthermore, they have researched the feasibility, acceptability and cost effectiveness of different screening strategies among high-risk immigrant cohorts [9][10][11][12]. Public Health England has established a national LTBI testing and treatment program for immigrants from countries with a high incidence of TB informed by their research on the prevalence of LTBI and feasibility of programmatic screening in this cohort [13]. This was a key action of their national collaborative strategy for TB [14]. Evidently, LTBI epidemiological and cascade of care research informed and enabled Public Health England to establish programmatic LTBI management in a target risk cohort.
The aim of this systematic review was to identify what is known about the epidemiology and cascade of care of LTBI in Ireland to inform its programmatic management nationally.

Methods
A systematic literature review was performed according to a review protocol and reported in adherence with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement (Appendix 1) [15]. The protocol for this systematic review was registered with the Open Science Framework (https:// doi. org/ 10. 17605/ OSF. IO/ 8ED29) and is available in Appendix 2.
Studies eligible for inclusion were those that described any group of patients who were screened or treated for LTBI in Ireland and reported using any one or a combination of chest radiography, tuberculin skin resting (TST) or interferon-gamma release assay (IGRA) testing to screen for LTBI. Studies had to report at least one of the following outcomes (chosen because they describe the cascade of LTBI care): the proportion of people screened out of the target population, the prevalence of a positive screening test in the target population, the proportion of those diagnosed with LTBI who were offered treatment, the proportion of Identification of groups at risk of having LTBI or an increased risk of progressing to active TB. These target groups should be prioritized for LTBI screening and treatment 2 Definition of diagnostic approach for LTBI detection, including both the selection of diagnostic test(s) and the diagnostic algorithm most appropriate for each target group 3 Provision of LTBI treatment using treatment regimens that are effective and promote adherence and completion by different target groups 4 Implementation of patient-centred strategies for service delivery 5 Effective health education and communication with target groups and health care providers 6 Programme monitoring and evaluation those diagnosed with LTBI who started treatment for LTBI, the proportion of those diagnosed with LTBI who completed treatment out and the cost of performing screening or treatment of LTBI cases identified. Clinical audits, randomized controlled trials, diagnostic accuracy studies, retrospective cohort reviews and prospective cohort reviews published between the 1st of January 2000 and the 31st of December 2019 (inclusive) were eligible for inclusion. Studies published in languages other than English were not eligible for inclusion. Studies where it was not possible to extract data on patients screened in Ireland alone were excluded.
A search of MEDLINE (via OVID), Embase, Web of Science, Google Scholar and published abstracts from national conferences in Ireland was conducted (search strategy is described in Appendix 2, date of last search: 14th of May 2020). The references of included studies were also searched. The literature search and data extraction were each conducted independently by two reviewers, and any disagreements relating to study eligibility or data extraction were resolved by discussion and mutual agreement. For the prevalence of a positive screening test, the risk of bias was assessed using a tool designed for TB prevalence studies that was derived from on an existing tool for prevalence studies (Appendix 3) [16].

Search results
The results of the search are described in Fig. 1. Fifty-two articles were identified for full-text review from the search of the indexed literature, Google Scholar, conference abstract searches, and the references of included articles. In total, 28 studies were identified as meeting the review inclusion criteria.
The reporting of the cascade of care was generally very incomplete (Table 4). Only six studies described the proportion of the target population that completed screening. Provider recommendation, patient acceptance and patient completion of LTBI treatment were reported in 10, 12 and eight studies, respectively. One of 28 studies reported the complete cascade of care from screening initiation to treatment completion [17]. One study reported an estimate of the cost of LTBI screening and treatment [26].

Risk of bias assessment
The overall risk of bias in assessing the prevalence of LTBI in the included studies was high ( Table 5, Appendix 3). Convenience sampling occurred in 25 of 26 studies. In 12 of 26 studies, there was a lack of a description of the patient exclusion criteria or how TB disease was identified. In 19 of 26 studies, the response rate, or the proportion of the target population who were screened, was not reported. Overall, the risk of bias was low (score [6][7][8] in two studies, moderate (score [3][4][5] in seven studies and high (score 0-2) in 17 studies.
In the 11 studies evaluating LTBI screening in immunosuppressed patients, the risk of bias was high in six of 11 studies, moderate in four of 11 studies and low in one of 11 studies. In the two studies evaluating LTBI screening in people living with HIV, the risk of bias was high in one study and moderate in the other. In the eight studies evaluating LTBI screening in recent TB contacts, the risk was high in six of eight studies, moderate in one of eight studies and low in one of eight studies. In the two studies evaluating LTBI screening in asylum seekers, the risk of bias was moderate in one study and low in the other. All four studies which evaluated LTBI screening in health care workers had a high risk of bias. Aside from the risk of bias within studies, there is a risk of reporting bias across

Latent TB infection screening and treatment outcomes in patients undergoing immunosuppression
Gnanasekaran et al. [17] was the only study that reported the proportion of the target population screened (95% of the target cohort) ( Table 6). The median prevalence of a positive IGRA across all studies in this cohort was 7% (interquartile range (IQR) 7-8%). When considering only the studies where the risk of bias was moderate-low, the prevalence of a positive IGRA was 7% (IQR 5-7%). The median prevalence of a positive TST across all studies in this cohort was 17% (13-26%) and when considering only studies where the risk of bias was moderate-low, the median prevalence of a positive TST was 17% (IQR 15-18%). In all seven studies where the proportion of patients who were offered and accepted LTBI treatment was reported [17-20, 26, 27, 46], all patients were offered and accepted treatment. The median proportion of patients completing treatment was 100% (IQR 90-100%), with all patients in three studies [17,19,26] completing treatment, and 61% of patients in one study [23] completing treatment. Jordan et al. [26] reported the cost of treating four patients with LTBI diagnosed using an IGRA as €1652 and 21 patients diagnosed using a TST as €6174, although the methodology used to make these cost estimates is unclear. Gnanasekaran et al. [17] Retrospective Secondary centre Pre-biologic screening in patients with IA (n = 39) O'Flynn et al. [18] Retrospective Tertiary centre Pre-biologic screening in patients with IA (n = 70) Awan et al. [19] Prospective Secondary centre Pre-biologic screening in patients with IA (n = 25) O'Flynn et al. [20] Prospective Tertiary centre Pre-biologic screening in patients with IA (n = 109) Hurley et al. [21] Retrospective Tertiary centre Pre-organ transplantation screening (n = 101) Safwat et al. [22] Retrospective Secondary centre Pre-biologic screening (43% of cohort) (n = 78) Haroon et al. [23] Prospective Tertiary centre Pre-biologic screening in patients with IA (n = 132) O'Flynn [24] Unclear Tertiary centre Pre-biologic screening in patients with IA (n = 188) Martin et al. [25] Prospective Tertiary centre Pre-biologic screening in patients with IA (n = 150) Jordan et al. [26] Retrospective Tertiary centre Pre-biologic screening in patients with IA (n = 63) Kelly et al. [27] Retrospective Tertiary centre Pre-biologic screening in patients with psoriasis (n = 101) Ní Cheallaigh et al. [28] Prospective Tertiary centre People living with HIV (n = 256) Ali et al. [29] Retrospective Tertiary centre PLWHIV (n = 331), Occupational screening of new entrant health care workers (n = 2410) Higgins et al. [30] Retrospective PHD TB outbreak (6 cases) in the community (n = 268) Glynn et al. [31] Prospective PHD Contact tracing of 39 sporadic TB cases (n = 701) O'Donovan et al. [32] Prospective PHD TB outbreak (6 cases) in university students (n = 71) O'Meara et al. [33] Retrospective PHD TB outbreak (3 cases) in a primary school (n = 307) O'Sullivan et al. [34] Retrospective PHD TB outbreak (13 cases) in a secondary school (n = 1200) Bambury et al. [35] Retrospective PHD Contact tracing of 118 sporadic TB cases (n = 1082) Gaensbaeur et al. [36] Prospective PHD TB outbreak in two creches (n = 268) Hennessy [37] Retrospective PHD Children tuberculin skin tested before BCG (1854) Tam et al. [38] Retrospective Tertiary centre TB-related referrals to a specialist paediatric clinic (n = 13) Millar et al. [39] Retrospective PHD Asylum seekers attending communicable disease screening in Cork and Kerry (n = 4780) Doyle et al. [40] Retrospective PHD Asylum seekers undergoing communicable disease screening 1998-2003 (n = 236) Smyth et al. [41] Retrospective Unknown Health care workers with significant exposure to infectious TB (n = 41) Kelly et al. [42] Retrospective Tertiary centre Occupational screening of overseas health care workers (n = 505) Power et al. [43] Retrospective Tertiary centre Overseas nursing recruits from India (n = 54) Arya et al. [44] Retrospective Tertiary centre Health care workers with a positive TST referred to the TB clinic (n = 243)

Latent TB infection screening and treatment outcomes in people living with HIV
Ni Cheallaigh et al. [28] reported the proportion of people living with HIV who had a positive test when screened using an IGRA as 18% when T-SPOT was used and 11% when QuantiFERON was used (sample sizes 256 and 247, respectively). When a TST was used among PLWHIV, the proportion of patients diagnosed with LTBI was 10% in the study by Ni Cheallaigh et al. [28] and 11% in the study by Ali et al. [29] (sample sizes 93 and 331 respectively). However, the risk of bias in the study by Ali et al. [29] was high. Ni Cheallaigh et al. [28] reported that all patients who were diagnosed with LTBI were offered treatment. No study reported on the proportion of patients completing LTBI treatment in this cohort. These studies did report on the outcome of LTBI screening, including the prevalence of a positive screening test among the screened population. Therefore, they could not be assessed using the selected risk of bias tool Study Total risk of bias score

Latent TB infection screening and treatment outcomes in recent TB contacts or prior to BCG vaccination
Three studies reported on the proportion of the target sample screened as part of contact tracing with 97%, 83% and 79% respectively being screened (  [36]. The prevalence of a positive TST in these studies was 9% and 7% respectively. One study reported on the proportion of recent TB contacts diagnosed with LTBI who were offered treatment, 61% [31]. Two studies reported on the proportion of TB case contacts accepting treatment as 31% and 67% [30,35]. Two studies reported on the proportion of patients completing treatment as 33% and 77% [30,35]. Two studies described the outcome of LTBI screening prior to BCG vaccination, one of which reported 13 cases of LTBI being offered treatment, 10 of whom accepted and completed treatment [38]. Millar et al. [39] reported on the proportion of the target sample screened in asylum seekers (28%) where screening was voluntary. Doyle et al. [40] reported that of 334 TSTs placed in a cohort of asylum seekers, only 236 were read. In this study, when screened using TST, 5/236 (2%) of those read were positive. Of these five patients, three were started on treatment. It is unclear if the remaining patients were not offered or declined LTBI treatment, and it is unclear how many completed treatment.

Latent TB infection screening and treatment outcomes in health care workers
Five studies reported on LTBI screening in this cohort (Table 8) [29,[42][43][44]46]. Two studies reported on the prevalence of LTBI in health care workers screened using a TST [29,46]. In a cohort of new entrant health care workers, 32% had LTBI [29] while in health care workers with significant exposure to infectious TB, the prevalence was 56% [46]. Kelly et al. [42] reported that of new entrant health care workers from overseas were screened using a TST or an IGRA, 17% had a positive test result of which 85% were offered LTBI treatment [43]. Only 26% accepted treatment, all of whom completed treatment. Arya et al. [44] reported of 243 health care workers with a positive TST referred to a TB clinic, only 59% accepted LTBI treatment, but it is not reported how many were offered LTBI treatment. Of these, 62% completed treatment [44].

Discussion
This research presents a comprehensive review of studies describing LTBI prevalence and screening and treatment outcomes in Ireland and highlights the significant knowledge gaps. The findings demonstrate that there are few studies that are reliably informative as to the prevalence of LTBI across all risk cohorts in Ireland. Studies were all performed on a local or regional level. When considering only the studies where the risk of bias was moderate-low, the prevalence of a positive IGRA among immunosuppressed patients was 7% (IQR 5-7%). There is no published research describing the prevalence of LTBI in people from countries with a high incidence of TB, people who are homeless, people in prisons and people who use intravenous drugs in Ireland, and for asylum seekers, there were only two studies describing the prevalence of LTBI, both of which had a moderate or high risk of bias. Regarding health care workers, only two studies, both performed in the same centre and both with a high risk of bias, were informative as to the prevalence of LTBI. Despite these cohorts having an increased risk of TB in other low-incidence countries [47][48][49], it is unclear if the incidence of TB in these cohorts in Ireland is high because TB cases are not described according to these characteristics in recent national surveillance reports. However, a 2015 report describing risk factors for TB cases notified in 2013 reported that approximately 20 to 25% of cases had "high endemicity residence", approximately 30% had "high endemicity origin" and approximately 10% had "substance abuse" [50]. Additionally, significant TB outbreaks have been reported in the Irish prison system within the previous decade [51]. Studies assessing the prevalence of LTBI and risk of TB reactivation in people from countries with a high incidence of TB, people who use drugs and prison inmates should be a future research priority in Ireland.
Research describing the cascade of LTBI care in Ireland was limited. Among immunosuppressed patients, treatment acceptance and completion appeared to be generally high, although the number of patients with LTBI described in these studies was small. Among TB case contacts, provider recommendation of treatment was reported as 61% in one study [31], treatment acceptance was reported as 31% and 67% [30] [30], and treatment completion was reported as 33% and 77% [31,35]. Among health care workers, two studies reported that the acceptance of LTBI treatment was generally low. There was insufficient information in the literature to describe the cascades of care in other cohorts and provide insight into where it should be improved. There were no studies which described the cost-effectiveness of LTBI screening and treatment, which are important if LTBI is to be managed programmatically at scale. A 2015 report describing risk factors for TB cases notified in 2013 reported that approximately 10%, 5% and 5% of TB cases occurred in TB case contacts, people with immunosuppressive illnesses and people on immunosuppressive medications, respectively [50]. Studies evaluating the cascade of LTBI care in PLWHIV should be prioritised, and further studies evaluating the cascade of LTBI care in patients on immunosuppressive treatments and TB case contacts should be encouraged. These studies would have utility when defining the diagnostic algorithm most appropriate for each target group in Ireland, which is key for effective programmatic management [4].
The strengths of this review are its rigorous methodology and that it is the first comprehensive review of TB research in Ireland, which establishes with certainty that scope and degree of research are needed. A weakness of this systematic review was that the research question, while was intentionally broad, could have been more explicitly defined at inception using the PICO model. With regard to abstract publications, the authors were not contacted to search for any further results. However, most abstracts included in this review described single-centre studies with a small sample size obtained using convenience sampling, limiting their utility when assessing the prevalence of LTBI. A limitation of this research was that there were few studies which were reliably informative as to the prevalence of LTBI because the risk of sampling bias was high across almost all studies. Therefore, the limited prevalence estimates reported in this review should be interpreted with caution.
Intensified research and innovation is a strategic pillar of the WHO End TB strategy, which should be adapted at a country level with global collaboration [2]. Studies meeting the identified research needs must be performed. The WHO describes the components of an enabling environment for high-quality research, which has relevance for LTBI research in Ireland [52,53]. These components include having a national TB research network. This could enhance collaboration between researchers, health care providers and patients and coordinate local and national TB research activities to align with national TB programme priorities [52]. The WHO recommends the formation of a country-specific TB research agenda and strategic plan to guide country-specific actions [52]. Other low TB incidence countries have advanced national LTBI research in cohorts they have identified as at risk, such as in Canada and England, where LTBI research priorities have been outlined in TB elimination strategies [14,54]. In Canada, the Public Health Agency have funded studies in Inuit people [55][56][57] and in the UK, Public Health England [8,11], the National Institute of Health Research [8,10,11,58]and the Medical Research Council [9,10] have funded LTBI research in people from countries with a high incidence of TB. TB research networks must not only contribute to local and national TB elimination efforts but also global TB elimination efforts through international collaboration [2]. Other European countries such as the Netherlands are prime examples of how countries with a low incidence of TB can be global leaders in transnational collaboration for TB research by funding and developing in their institutions TB researchers and research programmes that are guided by a national TB research agenda and the WHO Global TB Research Agenda [59]. The WHO advises that an enabling environment for TB research should have sufficient local researchers with the necessary profiles in TB research and incentives to retain them in employment and that there should be specialized training on TB for new researchers [52]. Although there are many researchers involved in other aspects of TB in Ireland, such as host-pathogen response, drug development and TB diagnostics [60,61], such is the scale of the identified LTBI epidemiological and cascade of care research needs that to meet them, dedicated TB research positions should be created within research institutions and form part of a TB-network. A high-quality research network with a well-defined research plan and strategy and the opportunity for international collaboration could attract new researchers to this field in Ireland and contribute to achieving TB elimination. The research needs identified in this systematic review would be best met by inclusion in a TB research agenda and strategic plan and delivered through a TB-network that develops local, national and international TB research.

Conclusion
This systematic review has described what published research there is on the epidemiology and cascade of LTBI care in Ireland and identified knowledge gaps. A strategy for addressing these knowledge gaps has been proposed. Study characteristics 18 For each study, present characteristics for which data were extracted (e.g., study size, PICOS, followup period) and provide the citations Page 6 and Table 3 Risk of bias within studies 19 Present data on risk of bias of each study and, if available, any outcome level assessment (see item 12) Page 8 and Table 4 Results of individual studies End TB Strategy states that the identification and management of LTBI in groups of people at high risk of reactivation is an essential part of TB elimination in low-incidence countries [18]. The End TB Strategy also suggests that epidemiological research should be conducted to determine the burden of LTBI in various geographical settings and risk groups and as a basis for nationally and locally tailored interventions, including integrated community-based approaches [18]. The Health Protection Surveillance Centre Guidelines for the Prevention and Control of TB 2010 guidelines outline which groups should be prioritized for LTBI screening in Ireland and offer guidance as to the diagnostic approach for certain target groups (Table 9) [151]. However, these guidelines do not discuss strategies for service delivery and programmatic monitoring and evaluation. In this systematic review, we aim to determine what evidence exists to describe the epidemiology of LTBI in the Republic of Ireland. Knowledge of regional LTBI epidemiology is crucial to improve the programmatic management of LTBI.

Aim
We aim to describe the epidemiology of LTBI in the Republic of Ireland including its prevalence, screening outcomes, and treatment outcomes.

Objectives
To determine:

Included studies
We will include all studies describing patients who were screened for LTBI in the Republic of Ireland. Studies must be published in English. Conference abstracts will be included for review. All studies published from 01 Jan. 2000 to 31 Dec. 2020 will be included. Studies must use any one or a combination of chest X-ray (CXR), tuberculin skin tests (TST) or interferon-gamma release assay (IGRA) to screen for LTBI. Clinical audits randomized controlled trials, diagnostic accuracy studies, retrospective cohort reviews, and prospective cohort reviews will be included. We plan to exclude any studies where it was not possible to ascertain data on only patients screened in the Republic of Ireland.

Outcomes
The outcomes chosen are screening test used, the proportion of people screened out of target population, the prevalence of a positive screening test in the target population, proportion of those diagnosed with LTBI who are offered treatment, the proportion of those diagnosed with LTBI who started treatment for LTBI, the proportion of those diagnosed with LTBI who completed treatment and the cost of performing screening and/or treatment of LTBI cases identified.

Search methods
We will search MEDLINE (via OVID), Embase, Web of Science and Google Scholar. The references of the included studies will also be searched.   The reason why LTBI screening was performed The proportion of people screened out of target sample population The proportion of people screened out of the total group of people targeted as defined by the authors The proportion of patients screened with a positive test The proportion of all patients screened who have a positive chest X-ray, TST, or IGRA The proportion of patients with a positive test offered LTBI treatment The proportion of patients offered LTBI treatment who accepted treatment The proportion of patients on treatment for LTBI who completed treatment The cost of screening for LTBI and or treatment A random sample of the target population was used 2 A national survey or multisite random sample of the target population was used Quality of selection method 0 There were no exclusion criteria stated or a risk factor for LTBI was an exclusion criterion 1 Exclusion criteria were stated and a risk factor for LTBI was not an exclusion criteria 2 The means of identification of TB was stated Response rate 0 Not reported 1 The proportion of the sample screened is reported and is under 65% 2 The proportion of the sample screened is reported and is 65% or above Quality of prevalence assessment 0 TST cut-off at 10 mm was not present/Indeterminate IGRA results were not stated 1 TST cut-off at 10 mm was present/ Indeterminate IGRA results were stated 2 TST cut-off at 5 or 15 mm was present as well/ Indeterminate IGRA results constituted < 10% The risk of bias relating to the other outcomes in the cascade of care was not formally assessed with a risk of bias tool because most of the items in risk of bias tools for non-randomized studies, such as the ROBINS-I tool [355] or the Newcastle Ottawa Scale [356], are not applicable to the primarily descriptive noninterventional studies that describe the cascade of TB care, and this limitation has always been encountered in other systematic reviews of TB cascades of care [291].