Prevalence and Predictors of Pulmonary Tuberculosis among Prison Inmates in Sub-Saharan Africa: A Systematic Review and Meta-Analysis

Introduction Prisoners in Sub-Saharan Africa (SSA) are at a high risk of tuberculosis (TB) infection due to overcrowding and poor ventilation. Consequently, TB is a leading cause of morbidity and mortality in prison, and many inmates face a number of barriers to TB control and had limited information in the region. Thus, the aim of this systematic review and meta-analysis was to estimate the overall pooled prevalence of pulmonary TB and predictors among prison inmates in SSA. Methods From 2006 to 2019, a systematic review and meta-analysis was conducted using various databases, including PubMed, Embase, Web of Science, and Scopus. The data were extracted in Microsoft Excel using a standardized data extraction format, and the analysis was carried out with STATA version 14. To detect heterogeneity across studies, the I2 and the Cochrane Q test statistics were computed. To determine the overall prevalence of TB and predictors among prison populations, a random effect meta-analysis model was used. Results Of the 3,479 retrieved articles, 37studies comprising 72,844 inmates met the inclusion criteria. The pooled prevalence of pulmonary TB among prison inmates in SSA was 7.74% (95% CI: 6.46-8.47). In the subgroup analysis, the highest prevalence was found in the Democratic Republic Congo (DRC) (19.72%) followed by Zambia (11.68%) and then Ethiopia (9.22%). TB/HIV coinfection (OR 4.99 (95% CI: 2.60-9.58)), Body mass index (BMI < 18.5) (OR 3.62 (95% CI: 2.65-6.49)), incarceration (OR 4.52 (95% CI: 2.31-5.68)), and previous TB exposure (OR 2.43 (95% CI: 1.61-3.56)) had higher odds of pulmonary TB among inmates. Conclusion The prevalence of pulmonary TB among SSA prison inmates was found to be high as compared to total population. TB/HIV coinfection, BMI, incarceration duration, and TB exposure were all predictors with pulmonary tuberculosis in prison inmates. As a result, emphasizing early screening for prisoners at risk of pulmonary TB is an important point to achieving global TB commitments in resource-limited settings.


Background
Tuberculosis (TB) is an airborne infectious disease mainly caused by Mycobacterium tuberculosis (MTB), which is still a major public health issue affecting person of all ages [1]. Adults in their most productive years are primarily affected, which accounted 88% [2,3]. Globally, an estimated 10 million people developed TB disease in 2019, and there are 1.2 million TB deaths in HIV-negative people and an extra 208,000 among people living with HIV [4]. The major derivers of TB remain undernutrition, poverty, tobacco smoking, people infected with HIV, prisoners, household air pollutions, diabetes, and other comorbidities that impair the immune system which are all the risk of contracting TB [2][3][4]. TB affects 30% of the world's population and is the leading cause of death from a single infectious pathogen, accounting for 1.3 million deaths each year [5]. In 2020, the 30 high TB burden countries accounted for 86% of new TB cases, with 1.5 million people dying from the disease [6]. Low-and middle-income countries, including sub-Saharan Africa (SSA), account for 94% of all TB infections and deaths [2,7]. In 2017, more than 10 million new cases of tuberculosis were reported worldwide, with one-third of these cases going unreported due to gaps in the healthcare system [7][8][9]. Although the World Health Organization (WHO) End Tuberculosis Global Strategy sets patronizing targets for 2020-2035, it aims to detect an estimated 90% of TB cases and reduce TB deaths by 95% in 2035 as compared to 2015, particularly among TB key populations who are most at risk of TB infection but have limited access to quality TB healthcare service [10][11][12][13]. To that end, WHO has prioritized the most vulnerable TB patients, including the poor, refugees, HIV-positive people, and prisoners, who are TB key populations [14]. Prison inmates are thought to be reservoirs for MTB transmission within their walls and in the community as a whole, but they are ignored for a variety of known reasons, including a lack of quality TB diagnostic services, overcrowding or stressful surroundings, a lack of perioding active case screening, comorbid illness, poor nutrition, and poor wall ventilation, particularly in SSA prisons [15][16][17]. Facts suggest that the risk of developing tuberculosis in prison is 6-30 times higher than in the general population, but 200 times higher in SSA, particularly in overcrowded prisons [13,15]. In 2016, for example, a review of 24 SSA countries' prisons revealed TB prevalence ranging from 0.4 to 16.3% [18]. Another study published in the same year found that regional variation in TB prevalence was 5.3% in Southern and East Africa and 2.9% in Central and West Africa [18]. In SSA prisons, overcrowding and poor ventilation are severe, with data showing that 86% of countries with data had prison occupancy rates above 100%, increasing the risk of airborne TB infection [19]. Despite its burden, there is no aggregated data on pulmonary TB prevalence and its predictors in SSA prison inmates. As a result, the aim of this systematic review and metaanalysis was to precisely estimate the overall pooled prevalence and predictors of pulmonary tuberculosis among prison inmates in SSA.

Search Strategy and Study Identification. From 2006 to
October 24, 2020, we systematically searched for articles published in international databases and electronic engines such as PubMed/MEDLINE, Cochrane Library, Google Scholar, Embase, Scopus, and Web of Science. In addition, we searched the reference lists of the included studies for relevant articles. The grey literatures were found by commu-nicating and searching the digital/repository library, the institution's official website, colleagues, researchers, and other scholars. To search relevant articles for this study, the following MeSH terms and keywords were used: "prevalence"/"magnitude"/"proportion", "tuberculosis", "pulmonary tuberculosis", "PTB", "associated factors"/ "predictors"/"risk factors", "prison inmates"/"prisoners", "Sub-Saharan Africa", and "SSA". The MeSH terms were used separately and/or in combination using Boolean operators like "OR" or "AND" (see Supplementary file1). The literature search for the aforementioned databases was done from June 01, 2020, to October 24, 2020. The review protocol was entered into the International Prospective Register of Systematic Reviews (PROSPERO), which is maintained by the University of York Centre for Reviews and Dissemination (CRD4202016764). Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines were followed in this study [20,21].

Eligibility Criteria
2.2.1. Inclusion Criteria. The contents of each included article were independently reviewed by the two investigators (HB and TB). When the specimen for laboratory processing is sputum, the prevalence of pulmonary TB was reported, and quality assurance measures for laboratory diagnostic methods (volume of sputum, acid-fast bacilli (AFB), fluorescence microscopy (FM), culture, and GeneXpert MTB/RIF), diagnostic technique of standard operating procedures, and methodological quality of each original study were carefully examined.
Study area: all novel or primary studies conducted in SSA prisons were taken into account.
Publication condition: from January 1, 2006, to October 24, 2020, articles published in peer-reviewed journals were included.
Study design: observational study designs with original articles contain data reporting the prevalence or number of confirmed TB cases and predictors of pulmonary TB among SSA prison inmates.
Language: only English language publications were considered.

Exclusion Criteria.
After reviewing the abstracts and full texts of the papers, the two independent reviewers (HB and TB) extracted data with care. We excluded incomplete data, full-text inaccessible articles, studies conducted outside of SSA after contacting the primary authors via email or phone, articles that did not report the point prevalence, reviewed articles, and studies conducted in the general population.
2.3. Data extraction. The three authors (HB, TB, and ZH) extracted the necessary data using a standardized data extraction format in an excel spreadsheet. The following information was extracted from the studies: first author, country of study, diagnostic methods, specimen type, study design, year of publication, sample size, area of prisoners, and tuberculosis prevalence. When the three authors disagreed, a fourth author (FW) was consulted, and disagreements were resolved through consensus and discussion.

2
Tuberculosis Research and Treatment

Quality Assessment Tool.
To assess the quality of the studies, the two authors (HB and TB) used the Newcastle-Ottawa quality assessment tool, which was modified for nonrandomized and cross-sectional studies [22,23]. Quality assessment tool was modified for nonrandomized and crosssectional studies. There are three indicators in the tool. The first section was given five stars and assessed the quality of the study's methodology; the investigators also evaluated the quality of laboratory diagnostic procedures. The second section graded the studies' comparability out of three stars. The third section graded the quality of the original articles based on their statistical analyses, which were graded out of two stars. Studies with medium (50% of quality assessment criteria was met) and high quality (6 out of 10 scales) were included in the analysis.
2.5. Statistical Data Analysis. The necessary data was extracted using Microsoft Excel and analyzed using STATA/ SE version 14. Tables were used to summarize the included articles, and the forest plot was used to estimate the study effect size and confidence interval. Using the binomial distribution formula, the researchers calculated the standard error of TB prevalence for each original article. Using the Cochrane Q statistics and I 2 test, we examined heterogeneity in the reported prevalence of pulmonary TB [24]. Heterogeneity is classified as high with >75%, substantial with 50-75%, moderate with greater than 25 and less than 50%, and low with 25% for I 2 [25]. To determine the pooled effect of the original studies, a random effects model with Der Simonian and Laird's statistical methods was used. To identify potential sources of heterogeneity, a univariate metaregression analysis was performed using the publication year, sample size, and confirmed TB. Furthermore, subgroup analysis was performed in these studies using variables to reduce the random heterogeneity between the estimates of the original studies [26]. Potential publication bias was evaluated subjectively using a funnel plot and objectively using Egger's weighted correlation and Begg's regression cutoff tests at a 5% significance level; a p value of 0.05 indicates the presence of publication bias [27]. Another subjective assessment of publication bias is a funnel plot, which is plotted by effect size per study versus standard error of effect size. Each dot represents a single study, and symmetric dots with an inverted funnel shape indicate that there is no publication bias. If the random effects model detects publication bias, the estimate is determined using the trim and fill analysis.

Explanation
3.3. The Quality Score of the Original Studies. All of the 37 studies were undergoing the quality assessment of methodological, and methods of laboratory diagnostic criteria were critically justified by using the Newcastle-Ottawa quality assessment tool [22]. Regarding the sampling technique, almost all of the included studies used active case finding approach ≥ two or less duration of cough during the study period. Presumptive symptom-based tuberculosis screening using standardized tuberculosis screening protocols was used. All the included studies used different laboratory diagnostic techniques to confirm pulmonary tuberculosis in the prisons (Table 2).
3.4. Meta-Analysis (Heterogeneity, Publication Bias, and Sensitivity Analysis). The overall pooled prevalence of pulmonary TB among SSA prison inmates in the random effects model was 7.74% (95% CI, 6.70-8.79, I 2 = 97:3%) (Figure 2). The pooled point incidence rate in this meta-analysis from the thirty-seven studies per 100,000 prison inmates was estimated to be 10,700 per 100,000 prison inmates (95% CI: 10,340-10,952). The results of the included studies' sensitivity analyses revealed that no study significantly influenced the results, with effect sizes ranging from 2.82 to 3.07. Different factors potentially associated with heterogeneity, such as sample size, publication year, and confirmed pulmonary TB, were checked by advanced metaregression model using univariate metaregression to identify the possible source of heterogeneity, and all potential variables of TB prevalence were not statistically significant (Table 3).  (Table 4).

Subgroup
In this review, we observed that no publication bias was statistically significant between the primary studies by  (Figure 3).

Predictors of Pulmonary Tuberculosis among Prison
Inmates. Only seven of the 37 studies had data that could be analyzed to determine the associations [37,40,43,45,56,65,66]. Body mass index (BMI), TB/HIV coinfection, a history of TB, and incarceration, which prison inmates live in the prison ≥ 6 months, were identified as potential predictors for pulmonary TB among prison inmates in SSA countries ( Figure 4). BMI of 18.5 kg/m 2 was found to be significantly associated with pulmonary tuberculosis in prison inmates adjusted odds ratio (AOR): 3.62 (95% CI: 2.31-5.68). Prison inmates who have a body mass index of <18.5 kg/m 2 were 3.62 times more likely to have pulmonary TB as compared to those prisoners who have a BMI of ≥18.5 kg/m 2 (Figure 4(a)). When compared to their counterparts, inmates with HIV were 4.99 times more likely to be infected with pulmonary tuberculosis, AOR: 4.99 (95% CI: 2.60-9.58) (Figure 4(b)). Similarly, a meta-analysis of this study found that prison inmates with a history of TB were more likely to contract pulmonary TB (Figure 4(c)). The pooled estimate of the seven studies also revealed that long duration of incarceration (≥6 months) was strongly related to pulmonary tuberculosis among prison inmates, AOR: 4.52 (95% CI: 2.57, 7.98) (Figure 4(d)).

Discussion
Tuberculosis remains a major public health threat worldwide, particularly in the SSA region. The current metaanalysis found a high prevalence of pulmonary TB among SSA prison inmates. TB/HIV coinfection, BMI, long duration of incarceration, and previous TB exposure were predictors of pulmonary TB infection among prison inmates. The current meta-analysis estimated that the overall pooled prevalence of TB among prison inmates using the 37 studies was 7.74%. This prevalence is similar with the study conducted in Malaysia 7.7% [9], Southwest Iran 7.9% [33], and SSA (6.4%-8.8%) [30,38,41,56,63] as previously reported studies among prison inmates. However, the finding of this meta-analysis was higher than studies in South Thailand (2.1%) [74], Peru (2.51%) [75], Asian countries (4.5%) [76], Brazil [77], and SSA (2%-3.6%) [37,55,63] as reported in a previous systematic review conducted from prison inmates. Moreover, the finding of our study was lower than a study conducted in Brazil diagnosed with smear, GeneXpert, and culture 12%-12.9% [78,79], Nepal sputum smear and GeneXpert test 10% [31], and SSA 10%-13.7% [44,61] as conducted in previous reports among prison inmates. The possible explanations for the observed variations might be attributable to the difference in the geographical variation, overcrowding, method of diagnosis, and number of prisoners in a cell with poor ventilation. In this systematic review and meta-analysis, the pooled estimated incidence of pulmonary TB among prisoners was 10,700/100,000, which is higher than WHO target for end TB strategy incidence in 2019, which is approximately130/100,000 [8]. Despite this, high incidence of TB in prisons could be possibly the fact that they are a forgotten population in the case of early screening, diagnosis, and treatment of TB in a high-risk population like prison inmates.    The subgroup analysis of this meta-analysis also showed that the prevalence of pulmonary TB was higher in prisoners incarcerated in DRC 19.72%, Zambia 11.68%, and Ethiopia 9.22% as compared to other countries in SSA. This difference in the prevalence of pulmonary TB within countries in the prisons might be the difference in the diagnostic technique, screening method, overcrowded, and sociocultural and socioeconomic variations of the study participants. Concerning to study year, the prevalence of pulmonary TB was significantly higher than in those studies conducted after 2015 as compared to those studied before 2015. This difference might be hypothesized that due to implementation variations in strategy targets and milestones for the underlying in the end TB strategy in SSA. This might be integrated, patient-centered care and prevention for high-risk populations of TB, bold policies and supportive systems, highlevel political commitment and shortage of enhanced resources, and intensified research and innovations to end TB epidemic in high-risk groups and new diagnostics tools, drugs, and vaccines in the SSA region were needed [14].
This meta-analysis examined that the predictors of BMI, presence of TB/HIV coinfection, previous TB history, and duration of incarceration were meta-analyzable and were significantly associated with the prevalence of pulmonary TB among prison inmates in SSA. Prison inmates who are underweight were 3.62 more likely to be infected by pulmonary TB than those prison inmates who had a BMI of ≥18.5 kg/m 2 . This finding is similar with a study reported by Valença et al. and Casha and Scarci [79,80]. The possible explanation for this is that a low BMI in body construct might be in some way predisposed to TB reactivation in the lung due to congenital apical lung bullae and biome-chanical pleural stress that increases in apical pleural stress in low anteroposterior diameter chests on coughing. Underweight impairs the immune system; these might alter cytokine synthesis, reduce antigen response, and diminish function of natural killer cells, dendritic cells, and macrophages, thus might affect TB incidence. The presence of previous TB/HIV coinfection of the prison inmates was significantly associated with the expansion of pulmonary TB. Prisoners with previous TB/HIV coinfection were 4.99 times more likely to have active pulmonary TB than prisoners who had not TB/HIV infection before. This is because people living with HIV are more likely infected by Mycobacteria tuberculosis where HIV weakens the immune system due to the depletion of CD 4 T cells and HIV upregulates M. tuberculosis entry into receptors on macrophages [1,81] that weakens the body to fight the bacteria and exerts immense burden in healthcare systems in resource liming settings like SSA. Similarly, the results of this metaanalysis showed that prisoners that have not previously TB history have protective associations with enhancement of pulmonary TB. Prison inmates who have presence of previous TB history/contact were 2.43 times more likely to develop pulmonary TB as compared to its counterparts. The finding of this study is consistent with a study conducted in South India which is 3.64 times more infected [82]. Long duration of imprisonment of prisoners was 4.52 times more likely infected by M. tuberculosis as compared to those prisoners whose duration of incarceration is short [50,56,82]. This might be because prisoners are not closed system, due to the number of people entering, departing, and reentering and poor environmental situation possibly increasing the transmission probability [83].

Limitation of the Study
Despite the fact that the authors conducted a comprehensive search using various databases for articles on the prevalence and predictors of pulmonary TB among prison inmates in SSA, this systematic review and meta-analysis failed to include reports published in languages other than English. We found studies in only eleven SSA countries, with other countries underrepresented due to the small number of studies included in prisons. Furthermore, studies and important data published in peer-reviewed journals between 2006 and 2019 may have gone overlooked.

Conclusion
In this systematic review and meta-analysis, the pooled prevalence of pulmonary TB among prisoners in SSA was considerably high which needs special attention to attain the end TB strategy. Underweight, previous TB/HIV coinfection, long duration of incarceration, and previous TB exposure/contact history are predictors of pulmonary TB infection among prisoners in SSA. Therefore, establishing systems for providing adequate nutrition and balanced diet, early screening, diagnosis of long imprisonment, treatment of TB/HIV coinfection, and contact history should be implemented to address this problem and succeed the control and end TB strategy in the SSA.

AFB:
Acid-fast bacilli AOR: Adjusted odds ratio BMI: Body mass index CI: Confidence interval C-S: Cross-sectional FM: Fluorescent microscopy HIV: Human immunodeficiency virus MTB: Mycobacterium tuberculosis OR: Odds ratio PRISAM: Preferred Reporting Items for Systematic Review and Meta-Analysis SSA: Sub-Saharan Africa TB: Tuberculosis WHO: World Health Organization.

Data Availability
The literature analyzed in the current study is available from the online data sources by using the reference listed or available from the corresponding author on a reasonable request.