Risk Factors on Healthcare-Associated Infections Among Hospitalized Tuberculosis Patients in China From 2001- 2020: A Systematic Review and Meta-Analysis

had a negative impact on HAIs among TB hospitalized patients in Chinese hospitals. These ndings provided evidence for the policy makers and hospital managers to make effective infection prevention and control measures to constrain the rising of HAIs. It is also required that more cost-effective infection prevention and control measures should be widely applied in routinely medical treatment and clinical management to healthcare-associated TB.


Introduction
As a chronic communicable disease, tuberculosis (TB) is acknowledged as a global public health issue, because it is one of the 10 top causes of disability and a leading cause of death globally (1). In 2019, the World Health Organization (WHO) reported that around 10.0 million people were infected with TB and 1.4 million died in 198 countries and territories (2). Particularly, China had the third highest burden of TB in the world and 8.4% of all global cases were detected in China (2). Since the pandemic of COVID-19 is threatening the globe, it reverses the deduction on TB incidence by having a negative impact on the Gross Domestic Product (GDP) per capita and undernutrition, which are two key determinants of TB incidence. The WHO estimated that TB cases could increase by more than 1 million per year in the period 2020-2025 because of the COVID-19 pandemic (2). This exacerbates the situation of diminishing TB.
Hospitals, as the main body to provide health care to treat TB, are also the main repository of healthcareassociated infections (HAIs). HAIs are an adverse outcome during providing health care to patients in hospitals. It affects patients' safety and extends the hospitalization, thereby patients suffering from the increasing morbidity and mortality (3,4). HAIs are also a global public health problem. In 2011, the WHO conducted a literature review and indicated that 7 and 10 out of every 100 hospitalized patients had a HAI in high-income countries and middle-income countries respectively (5,6). The HAIs prevalence in China was estimated around 3 out of every 100 hospitalized patients in 2018 (7). Although it was lower than that reported by the WHO, China has the largest population in the world. This still indicates China has a high burden attributable to HAIs.
Usually, a TB patient needs to be treated with multiple antibiotics for 6 to 12 months (3). This results in a long time exposure to antibiotics for patients. Antibiotics use in many studies has been recognised as one of the risk factors to increase HAIs incidence (8, 9). As for TB patients themselves, especially the elderly, they have a low immunity to ght against the infection (10). As a consequence, TB patients are at high risk to get HAIs. Therefore, it is essential for health professionals and managers to understand what kind of risk factors have an impact on healthcare-associated TB. Thus, it can contribute to guideline and regulation about constraining healthcare-associated TB in hospitals.
Several studies investigating the potential risk factors on healthcare-associated TB have been conducted in China However, all of them were based on a single hospital, which is sporadic. There is still scare evidence-based guideline or regulation to help health professionals and hospital managers to increase the awareness of the hazards of healthcare-associated TB and take effective infection control measures to constrain the occurrence of healthcare-associated TB. The aim of this systematic review and metaanalysis is to investigate the risk factors associated with healthcare-associated TB between patients with HAIs and those without HAIs in Chinese hospitals.

Systematic search strategy
The PICO/S (Population, Intervention, Comparison, Outcome, and Study type) tool was applied to de ne the scope of the literature. Search Chinese corresponding terms in the title, abstract and keywords included healthcare-associated infections/cross infections/hospital acquired infections/nosocomial infections, tuberculosis, risk factors/in uencing factors/ and China for the Chinese databases. The logical word is "AND". As to Medline and EMBASE, medical subject heading (MeSH) terms in key words were adopted. The MeSH terms were ("cross infection" AND "tuberculosis" AND "risk factors" AND "China").

Inclusion and exclusion criteria
The inclusion criteria were as follows: 1) risk factors analysis by using a case-control or cohort study; 2) a multi-centre study or a single-centre study; 3) study language being either English or Chinese.
The exclusion criteria were as follows: 1) conference papers or editorials/letters; 2) duplicate studies and repeated data published in different journals concurrently; 3) any study outside China; 4) only description on the prevalence or landscape of HAIs; 5) risk factors concluded without any statistical inference (e.g. according to the clinical experience).

Data abstraction
Two independent reviewers (X.L. and N.R.) screened the obtained literature by title and abstract to determine the eligibility of studies. Potential disagreement was resolved by discussion. Identi ed studies were retrieved in full text and were checked again for eligibility. In case of exclusion of the study the reason was documented. Moreover, cross-references were also considered by screening the bibliography of eligible studies as well as bibliography of cross-references.
Quality assessment of the included studies on risk factors for HAIs among TB hospitalized patients X.L. and N.R. jointly assessed the quality of the included studies for risk factors analysis associated with HAIs among TB hospitalized patients according to the criteria of JBI' (Joanna Briggs Institute) critical appraisal tools (12). This tool includes 8 questions to identify the quality of a cross-sectional study. For each question, there are 4 options to choose (Yes, Unclear, No and Not applicable). The less the number of a positive option (Yes) is, the more the uncertainty of a study is. Otherwise, the quality of a study is better. Moreover, a value was assigned to each answer in order to calculate the scores for each study, that is, 2 points for 'Yes', 1 point for 'Unclear or Not applicable', and 0 point for 'No'.

Statistical analysis
Review Manager 5.3 software was deployed to assess the risk factor odds ratio (OR) among TB hospitalized patients. Meta-analyses of OR in HAI prevalence were performed. Heterogeneity between studies was assessed based on the and I 2 statistics to select the meta-analysis model. When results had a (P<0.05) and/or I 2 > 50%, data were considered heterogeneous and the random-effects model was used; otherwise, the xed-effects model was used. P<0.05 was considered as a statistical signi cance.

Results
Characteristics of eligible studies Figure 1 shows that totally, 851 records were searched from the included databases. Eventually, 11 published articles were incorporated in the quantitative meta-analysis after reading the eligible full texts with the inclusion and exclusion criteria.
Integration of the included studies on risk factors between patients with HAIs and patients without HAIs among TB hospitalized patients Table 1 presents that in total, 11,922 TB hospitalized patients were included in the systematic review and meta-analysis. Of them, 1,133 were diagnosed as having HAIs, while the rest (110,789) were not with HAIs. All of the included studies were undertaken in tertiary general hospitals, which covered 10 regions in 9 provinces in China. Table 2 shows that the most frequent risk factors reported from the included studies were age (11/11), use of antibiotics (11/11), invasive procedure (9/11), and length of hospitalization (9/11).
Course of disease (> 5 years vs. ≤ 5 years) Negativity of the sputum smear for acid fact staining Use of anti-tuberculous drug √ √ 2 Use of glucocorticoid √ 1 Quality assessment of the included studies All of the 11 included studies were cross-sectional with a median score (10 points). All of the studies clearly de ned the inclusion for the hospitalized patients, identi ed the confounding factors which had an impact on the occurrence of HAIs among TB hospitalized patients, and measured the outcomes in a valid and reliable way. However, none of the studies described the study setting with more details and only mentioned the hospital without any details. Also, variation of the HAIs prevalence existed among the included studies. It mainly showed that some studies did not describe how the HAIs prevalence among the TB hospitalized patients was calculated. The same situation was also applied to the standard criteria for the TB patients with HAIs and the strategy to deal with the confounding factors. (See Appendix 1) Meta-analyses of all the potential risk factors

Discussion
Our systematic review and meta-analysis rst provided a comprehensive analysis of risk factors on HAIs among TB hospitalized patients in Chinese general hospitals. Our review found that older than 65 years, presence of complication, presence of diabetes mellitus, invasive procedure, longer than 15 days of hospitalization stay, secondary tuberculosis, smoking, presence of underlying disease, and use of antibiotics were the main risk factors which had a negative impact on HAIs among TB hospitalized patients in Chinese general hospitals. These ndings provided evidence for the policy makers and hospital managers to make effective infection prevention and control measures to constrain the rising of HAIs.
Our systematic review found that TB hospitalized patients older than 65 years were more susceptible to HAIs than those younger than 65 years (OR: 2.89 [2.01-4.15]). The number of aged population is rapidly rising in the globe. 1 in 6 people beyond the age of 65 years has been estimated in the world in 2050, which is much higher than 1 in 11 in 2019 (23). Elderly TB patients are recognized as the immunecompromised patients, who are at high risk of acquiring HAIs (24). This suggests that the elderly TB patients should be set as the priority and the surveillance of the elderly TB patients should be strengthened. Thus, it can result in decreasing the risk of getting HAIs among aged TB patients.
Presence of complication, diabetes mellitus, underlying disease, and secondary tuberculosis were found the main factors associated with high prevalence of healthcare-associated TB in our review. All of them mainly deteriorated the TB patients' immune systems. Consequently, TB patients were more susceptible to HAIs compared with those without above conditions. One study has con rmed that it was 2.5 times more likely to develop TB among patients with diabetes mellitus in developed countries (25). It was also found that diabetes mellitus was highly associated with multi-drug resistant TB (MDR-TB) in Asia (OR: 1.40[1.01-1.95]) in a systematic review and meta-analysis (26, 27). Presence of complication, diabetes mellitus, underlying disease, and secondary tuberculosis de nitely affect the healthcare-associated TB patients' safety and prognosis, and put a challenge to the worldwide health system infection control strategy and clinical management, which further alert the global health development without any control measure.
The hazard of acquiring HAIs among TB hospitalized patients who had invasive procedure was higher than those without invasive procedure (3.80 [2.25-6.42]) in our review. Invasive procedure was actually not frequently reported that had a negative impact on healthcare-associated TB in existing research.
However, invasive procedure, such as indwelling invasive device and surgery, indeed make the TB hospitalized patients predisposed to HAIs. A single-centre point-prevalence survey in an American hospital showed that 96.8% of hospitalized adult patients had at least one indwelling device (28). It indicated that invasive procedure was still widely adopted as a treatment among hospitalized patients, which put the patients into high exposure to HAIs.
Our review also found that smoking increased the risk of obtaining HAIs among TB hospitalized patients  (30,31). TB hospitalized patients who stay longer in hospital are more likely to have underlying disease or complications or invasive procedure and so on (32,33). As previously mentioned, they promote the probability of the exposure to HAIs for TB hospitalized patients. It is necessary for the hospitals to shorten the hospitalization stay for the TB patients in order to lower the risk of HAIs exposure.
TB hospitalized patients were more risky to get HAIs if they used antibiotics during the treatment compared with their counterparts (OR: 2.77 [2.35-3.27]). This nding is consistent with current studies (34). Use of anti-tuberculosis drug is the main way to treat TB for a long period (21). During the treatment, it is common to use antibiotics for TB patients concurrently. As a consequence, an increasing number of TB patients have been infected with multi-drug resistance. Goedele et al. also mentioned that successful TB treatment is compromised by drug resistance because of irrational use of antibiotics to treat TB (29).
This indicates that prudent and high-quality antibiotics prescription and rational use of antibiotics are essential to constrain the overuse of antibiotics, thereby reduction in the occurrence of healthcareassociated TB. Moreover, antimicrobial stewardship program has been widely recommended as one way to achieve the rational use of antibiotics, including making straightforward and strict rational use of antibiotics guidelines and regulations, improving the clinical doctors' antibiotics prescription behaviour, and increasing the awareness of antimicrobial resistance among the public (35,36).
The key to preventing the healthcare-associated TB after clearly investigating the risk factors associated with high occurrence of healthcare-associated TB is to implement effective infection prevention and control measures. Current literature has suggested that the exposure and transmission of healthcareassociated TB can be reduced with the implementation of infection prevention and control guidelines, which can increase the identi cation and isolation of the potential healthcare-associated TB patients (37). The United States, WHO, and other institutions have recommended effective infection prevention and control measures to decrease the occurrence of healthcare-associated TB, particularly a hierarchy of effective infection control measures, including administrative controls, environmental controls, and personal respiratory protection (38). Specially, the administrative controls are considered as the rst and most important component to decrease the exposure to healthcare-associated TB. It is also required that more cost-effective infection prevention and control measures should be widely applied in routinely medical treatment and clinical management to healthcare-associated TB. Hence, the exposure and transmission of the healthcare-associated TB can be reduced and prevented.
Our systematic review and meta-analysis also has some limitations. First, the included studies were mostly published in Chinese journals since they were conducted in Chinese general hospitals. It encourages the researchers to publish their ndings about risk factors for healthcare-associated TB in English journals to share the advanced knowledge. Second, we found the included studies were all conducted in a single hospital. It is required that multi-centre studies could be undertaken in future research to strengthen the current evidence base.

Conclusions
Older than 65 years, presence of complication, presence of diabetes mellitus, invasive procedure, longer than 15 days of hospitalization stay, secondary tuberculosis, smoking, presence of underlying disease, and use of antibiotics were the main risk factors which had a negative impact on HAIs among TB hospitalized patients in Chinese hospitals. These ndings provided evidence for the policy makers and hospital managers to make effective infection prevention and control measures to constrain the rising of HAIs.. It is also required that more cost-effective infection prevention and control measures should be widely applied in routinely medical treatment and clinical management to healthcare-associated TB.
Hence, the exposure and transmission of the healthcare-associated TB can be reduced and prevented.

Declarations
Ethics approval and consent to participate It is not applicable.

Consent for publication
It is not applicable.
Availability of data and materials The datasets and materials analysed during the current study are available from the corresponding author on reasonable request.

Author contributions
The systematic review and meta-analysis was conducted under the charge of H.L. X.L. and N.R. searched the databases to nd the relevant articles and assessed the quality of the included studies. X.L.
undertaken the statistical analysis. X.L. and N.R. drafted and revised the manuscript. M.Z. adjusted the tables and gures. H.L. and Z.F. reviewed the manuscript, provided the feedbacks and did the proofreading.

Funding
Our study was funded by Seed project grant of Wuhan University O ce of International Affairs.

Competing interests
The author(s) declared no potential con icts of interest with respect to the research, authorship, and/or publication of this article.