Prevalence of Multidrug Resistant Tuberculosis (MDR-TB) in East Africa: a Systematic Review and Meta- Analysis

Background: The rate of multidrug resistant tuberculosis is increasing at an alarming rate throughout the world. It is becoming an emerging public health problem in East Africa. Prevalence of MDR-TB among TB positive individuals in the region has not been synthesized. Determining the pooled prevalence of MDR-TB among newly diagnosed and previously treated TB cases in East Africa is the main objective of this review. Methods: In this systematic review and meta-analysis, the researcher searched six electronic databases: Google scholar, PubMed, EMBASE, Scopus, Science direct and Web of Science. Funnel plot symmetry visualization conrmed by Egger’s regression asymmetry test and Begg rank correlation methods was used to assess publication bias. The pooled prevalence estimate was calculated using Der Simonian and Laird’s random Effects model. A total of 16 articles published in East Africa from 2007 and 2019 were included in this study. STATA software (version 14, Texas, USA) was used for analysis. Results: Out of 1025 articles identied citations, a total of 16 articles published in East Africa from 2007 and 2019 were included in this meta-analysis. The pooled prevalence of MDR-TB among newly diagnosed TB cases and previously treated TB patients to be 4% (95%CI=2-5%) and 21% (95%CI: 14-28%), respectively. Living conditions, lifestyles (smoking, alcohol use, and drug abuse), previous medical history, diabetes history and HIV infection risk factors contribute to higher prevalence of MDR-TB in East Africa. Conclusion: An early diagnosis of tuberculosis and rapid detection of drug resistant Mycobacterium tuberculosis is a serious to identify who are not to the standard and to avoid spreading of resistant strains. It is also very essential to strengthen tuberculosis control and improved monitoring of chemotherapy.


Introduction
Tuberculosis (TB) disease, caused by the Mycobacterium tuberculosis (M. tuberculosis), is the main cause of death. It is killing about 2 million people each year globally and nearly half million new cases of MDR-TB emerges every year [1]. TB is ranked as the second leading cause of death from contagious disease worldwide [2]. Countries such as Russia, China and India alone accounted 47% of the global burden. Different studies carried out in many countries have shown the MDR rate is much higher in previously treated TB patients than in new TB patients [3]. The maximum TB mortality and morbidity occurs in middle and low income countries like East African countries [4]. High mortality of the TB disease was predominantly observed in human immunode ciency virus (HIV) [5]. Currently, East Africa countries such as Ethiopia, Rwanda, Kenya, Tanzania and Uganda are among the 30 high MDR-TB burden countries. According to Gobena et al. [6], Ethiopia is one of the high burden countries with regards to TB, TB/ HIV, and MDR-TB. According to the 2016 report, 89.7% of the global incident cases of MDR-TB were accounted in 30 high MDR-TB burden countries. Among the 30 high burden TB countries ranked in the world, Kenya has fth highest burden in Africa [5,6].
The emergence of MDR-TB causes a considerable risk to TB control programs worldwide [7]. Co-infection multidrug-resistant tuberculosis (MDR-TB) did not get key attention until recently in East Africa, where the TB prevalence and risk factors are highest [8]. MDR-TB, even though being nearly 100% treatable, it is one of the major community health problems that is most-frequently the root cause of death among adults [9]. One-third of the world's population is infected with TB. The presence of the drug resistant gene in M. tuberculosis is the major confront of controlling TB [10]. Multidrug-resistant (MDR) (resistance to at least both to rst-line anti-TB drugs such as isoniazid and rifampicin) TB occurred among 3.6% and 18% of new and previously treated TB cases, respectively (5.6% among all cases) [4]. The main challenge in the treatment and control of tuberculosis is a high burden of MDR-TB globally [11]. Treatment of MDR-TB requires long time with a combination of second-line anti-TB drugs; most of them are less effective, has a considerable rate of adverse effects, and more expensive than rst-line drugs [12].
Signi cant attention is not given to MDR-TB in East Africa and it has been ignored until recently in this region. Even if there are some studies about the prevalence of MDR-TB in the region, only little is known about the associated risk factors to MDR-TB in East Africa. Some studies in the region indicated that low socioeconomic status, history of anti-TB treatment, the presence of HIV infection, sex, history of diabetes mellitus, alcohol use and malnutrition were the hypothesized risk factors associated with the expansion of MDR-TB in the region [13,14]. A study done in Tanzania showed that risk factors associated with MDR-TB were previous history of treatment with rst line anti-TB (OR = 3.3, 95% CI 1.7-6.3), smoking (OR = 1.9, 95% CI 1.0-3.5), contact with TB case (OR = 2.7, 95% CI 1.4-5.1) and history of TB [15]. Similarly, a study done in Ethiopia revealed that a history of previous treatment (AOR = 21; 95% CI: 17.8-28) and

Inclusion and exclusion criteria
In this review, studies reporting on MDR in new TB cases and on MDR in previously treated TB cases were included in the nal analysis. The review included all eligible studies published in English from 2007 to 2019 in East Africa. Extraction of eligible review for this study used the following information: rst author's name; study area, year of publication, number of individuals screened for new cases of MDR-TB, number of positive for new cases of MDR-TB, the number of previously treated MDR-TB, and a number of previously treated MDR-TB who are positive. Because of any of the following reasons, studies were excluded from the analysis: duplicate publication of the same study, articles available only in abstract form, case reports, anonymous reports, studies with data unrelated to MDR-TB, studies conducted outside of East Africa and studies without adequate data.

Data extraction
For data rescue, the author developed an extraction format to extract the following information: Author(s), year of publication, study period, study area, sample size, new cases of MDR-TB and previously treated MDR-TB. The author examined the quality of the studies by using a set of criteria such as: quality of reported data and appropriateness of the method used for MDR-TB detection among newly diagnosed and previously treated TB cases.

Data analysis
The researcher used STATA statistical software (version 14, Texas, USA) to analysis the data. To assess the descriptions of original studies forest plots were used in this study. Funnel plot symmetry visualization followed by Egger's regression asymmetry test and Begg rank correlation methods were used to identify the presence of publication bias. By using Der Simonian and Laird's random effects model, point prevalence estimate of each study with a 95% con dence interval was used to estimate pooled prevalence.

Results
A total of 1025 articles were recovered from Google scholar, PubMed, EMBASE, Scopus, Science direct, and Web of Science. The process of article selection is indicated in Fig. 1. Of these, 271 articles were excluded due to duplication. Out of these, 665 research articles which don't match with this study objective were excluded after reviewing their title and abstracts. Then the remaining 89 articles were identi ed for full text review. Out of the 89 articles reviewed in full text, 16 articles were retained for nal meta-analysis based on the inclusion-exclusion criteria and quality assessment and 73 studies were removed prior to analysis for different reasons such as full text unavailability, overlapped studies, case reports and articles with lack of su cient data.

Publication bias and Heterogeneity
As indicated in Figs. 2 and 4, funnel plot symmetry visual inspection was used to determine the occurrence of publication bias and the result showed absence of publication bias. In this review, absence of publication bias was statistically con rmed by Egger's weighted regression test (bias coe cient (B) = 4.7, -2.18-5.46; P = 0.37) and (bias coe cient (B) = 3.4, (95% CI=-3.01-1.7; P = 0.54) for MDR-TB among newly diagnosed cases and among previously treated cases, respectively. In this study heterogeneity analysis for newly diagnosed and previously treated MDR-TB showed that the occurrence of high variance among studies (I 2 = 94.7%, P = 0.000) and (I 2 = 99.1%, P = 0.000), respectively, which was statistically signi cant.
[17] to 91% (95%CI: 81-101) by Okethwangu et al. [18]. The researcher found that the pooled burden of MDR-TB among newly diagnosed and previously treated TB cases varied considerably from country to country in the study region.

Discussion
Multidrug resistant tuberculosis is a major global health problem. TB continues to be a great health treat in developing countries and is compounded by the high burden of MDR-TB [31]. There is still a serious knowledge gap about multidrug resistant tuberculosis in East Africa countries. Therefore, in order to nd out the current circumstance of MDR-TB in the region, recent and evidence based studies are signi cantly required. Thus, this systematic review and meta-analysis addressed the prevalence of new MDR-TB and previously treated MDR-TB using 16 selected studies conducted from 2007 to 2019 in East Africa. In this review and meta-analysis, the prevalence of newly diagnosed MDR-TB ranged from 1% (95%CI: 0.0-2%) to 20% (95%CI: 11-29%) with a pooled prevalence of 4% (95%CI: 2-5%). The results of this metaanalysis showed that the pooled prevalence of newly diagnosed MDR-TB was somewhat higher than a previous meta-analysis report by Eshetie et al. [32] where 2% (95% CI 1-2%) of newly diagnosed TB patients have MDR-TB. According to the results of this meta-analysis, the prevalence of MDR-TB among previously treated cases ranged from 1% ((95%CI:-0.00-1.00) to 91% (95%CI: 81-101) with a pooled prevalence of 21% (95%CI: 14-28%). Similarly, the same study done in Ethiopia [32] showed that the pooled prevalence of MDR-TB among previously treated cases was 15% (95% CI 12% − 17%) which was lower than the present study. In this review, the researcher found that the level of MDR-TB were higher in East Africa than reported globally for both new and previously treated MDR-TB. This study result showed that MDR-TB estimates as almost six times higher as compared to the global average reported by WHO for both new and previously treated (21% vs 3.6%) TB positive patients [33]. According to this metaanalysis, the pooled prevalence of MDR-TB was higher than the pooled prevalence of MDR-TB with a previous meta-analysis report that was conducted in Ethiopia [23]. According to the WHO report of 2015, globally the prevalence of MDR-TB of new cases and previously treated cases of TB was 3.5% and 20.5%, respectively, while Sub-Saharan Africa countries contributing the highest proportion and these levels in recent years have remained unchanged [34,35]. The main reasons behind the appearance of MDR-TB globally are multi-factor and are related to living conditions [36], life style [15], previous medical history [37,38], diabetes history [39,40], HIV infection [41], and education level [37].
Poor and crowded living conditions of a family might make easy the spread of TB. Thus, these lower socioeconomic family groups should get the highest concern for MDR-TB prevention efforts [42]. Patients that live in a household with more than one room were ve times at lower risk of having MDR-TB than those living in a household with only one room [8]. This might be as a result of high risk of getting resistant strains from infected people in crowded places. Different lifestyles such as alcohol abuse, smoking, drugs use, etc. are the major risk factors associated with the development of MDR-TB. Cigarette smokers were more likely to be infected with MDR-TB infection and have less chance to be cured. Male cigarette smokers were more likely to be infected with MDR-TB and susceptible compared than females.
This might be due to that men are predominantly drink alcohol, smoke and consume the drug compared to women [6, [43][44][45].
The longer exposure of a patient to anti-tuberculosis drugs was also associated with the development of MDR-TB. Patients under multiple situation of anti-tuberculosis treatment might create greater antibiotic resistance with the consequent development of MDR-TB and XDR-TB cases. Previous TB disease and chemotherapy are the most important risk factors associated with MDR-TB [37,38,46]. The global diabetes mellitus (DM) epidemic creates a serious bottleneck to the TB control program [47]. Individuals with diabetes, as compared to non-diabetic controls, were two-to three-folds more likely to develop TB. Weakened immunity in diabetic patients is thought to contribute to the development of latent TB infection to active cases [39,40].
A study done in Ethiopia revealed that human immunode ciency virus (HIV) infection was identi ed as a signi cant factor associated for MDR-TB [22]. Moreover, individuals living with HIV might also be more likely to be exposed to MDR-TB because of longer hospitalization in settings with low infection control. Drug malabsorption in HIV infected people can also lead to drug resistance and has been shown to result in treatment failure [48]. In one study conducted by Al-Darraji et al. [49] revealed that there was 20% higher occurrence of MDR-TB among HIV-positive persons than those HIV-negative persons. Patients with low education level have been associated with the development of MDR-TB. [37]. A study conducted by Ronaidi et al. [50] showed that the number of MDR-TB infected patients is signi cantly higher among the lower education group compared to the higher education group. It has been reported that the higher prevalence of pulmonary TB with unsuccessful treatment outcome was observed in patients who had a lower education level [51]. However, studies conducted in Ethiopia and Sudan showed that educational status was not signi cantly associated with MDR-TB [52][53][54].