Prevalence of Shigella species and its drug resistance pattern in Ethiopia: a systematic review and meta-analysis

Background Shigella species are a major cause of dysentery and may attribute for death worldwide. Currently antibiotic resistance became the critical challenges for management of infectious disease. The aim was to conduct a systematic review and meta-analysis of Shigella species and its drug resistance pattern in Ethiopia. Methods A comprehensive literature search was conducted through internet searches using database of MEDLINE, PubMed, Google scholar, EMBASE, HINARI, Cochrane Library and reference lists of previous prevalence studies from January 1999 to November 2018. Results were presented in forest plot, tables and figures with 95% CI. The Cochrane Q test and I2 test statistic were used to test heterogeneity across studies. The Pooled estimate of Shigella species and its drug resistance pattern was computed by a random effects model. Results The pooled prevalence of Shigella species in Ethiopia was 6.6% (95% CI 4.7–8.8). In the subgroup analysis, the highest prevalence was observed among patients in Health facility (8.5%, 95% CI 6.2–11.5) whereas the lowest prevalence was observed in Community based studies (1.6%, 95% CI 0.8–3.4). In addition, Shigella species were highly resistant to ampicillin, amoxicillin, erythromycin and multi-drug resistant (MDR) with the pooled resistance proportions of 83.1% (95% CI 75.7–88.6), 84.1% (95% CI 75.6–90.1), 86.5% (95% CI 70.9–94.4) and 83.2% (95% CI 77.1–87.9), respectively. On the other hand, comparably low resistance pattern was reported for ciprofloxacin 8.9% (95% CI 6.0–12.8), ceftriaxone 9.3% (95% CI 3.9–20.5), and norfloxacin 8.2% (95% CI 3.8–16.6) and gentamycin 17.3% (95% CI 11.2–25.9). Subgroup analyses indicated that study years were associated with a decreasing Shigella prevalence over time (p = 0.002). Conclusion The pooled estimate showed high burden of Shigella infection and its high proportion of drug resistance pattern to ampicillin, amoxicillin and erythromycin in Ethiopia. Therefore, initiating and scale up of performing drug susceptibility test for each shigellosis case, educate the community and health care providers on appropriate use of antibiotics need to be considered and strengthened. Electronic supplementary material The online version of this article (10.1186/s12941-019-0321-1) contains supplementary material, which is available to authorized users.

(AMR) every year. The experts suggest that this figure will rise to 4.2 million in Africa and 10 million globally by 2050, if nothing is done [3][4][5].
The treatment of Shigellosis has currently become more challenging due to the emergence of drug resistant species and associated with a variety of biological, pharmacological and societal variables with the worst combinations in low and middle income countries [6][7][8]. Multidrug-resistant Shigella significantly vary from area to area of the world in relation with the practice of widespread use of antimicrobial agents [3,9].
Antibiotic resistance is a natural phenomenon that occurs whenever antibiotics are in use. However, there are human behaviors that contribute to the rapid development and spread of bacterial antibiotic resistance. Availability and use of broad spectrum antibiotic without prescriptions facilitate the development of resistance by Shigella species [10].
Different studies have been carried out in different parts of Ethiopia at different times to document the epidemiology of and drug Susceptibility pattern of Shigella species. However, there is no summarized prevalence data of this bacterial infection and its drug Susceptibility pattern at country level to help in the formulation of appropriate intervention methods. Therefore, the present study is the first of its kind and aimed to determine the pooled prevalence pooled prevalence and drug susceptibility pattern of Shigella species in Ethiopia.

Search strategy
A comprehensive literature search was conducted on the prevalence of shigella species and its antimicrobial resistance pattern among Ethiopia population. Potentially relevant studies were identified through a literature search of Medline, PubMed, Google scholar, HINARI and Cochrane Library. The search was based on the combination of the following special index search terms (medical subject headings (MeSH) and Boolean operations: "Shigella" AND "Prevalence "OR "Epidemiology" AND "Drug Resistance, Microbial" AND "Dysentery, Bacillary/ epidemiology" AND "Ethiopia" "title and abstract" from January 1st 1999 to November 20th 2018. Articles search were focused on published studies with epidemiological and/or clinical data. All records were managed in Endnote version X7 (Clarivate Analytics, Philadelphia, PA, USA). The search was carried out from April 5th, 2016 to November 20, 2018. The limit of language was English and the limit of study group was human.

Eligibility criteria
We reviewed abstracts from initial search using defined inclusion and exclusion criteria.
Inclusion criteria Studies were selected for systematic review and meta-analysis, 1, if they were conducted in Ethiopia 2, study design been cross-sectional 3, studies reported the prevalence of Shigella species and its drug resistance pattern 4, literatures published in the English language and 5. Published articles were considered.
Exclusion criteria Studies, which were not fully accessed after reading the titles and abstracts were excluded since we are unable to assess the quality of each article in the absence of their full texts.

Data extraction
The data extraction was done by three researchers (S.H, Z.Y and G.M) independently from included studies using a standardized and pretested format prepared in Microsoft Excel. The data abstraction format included first author, study design, region in the country (study site in the country), publication year, sample size, population characteristics, age group of study participants, prevalence of Shigella species, medical treatment type, and resistance pattern of Shigella species. Disagreement on data extractions between researchers were resolved through discussion and consensus.

Quality assessment
The quality was assessed using 9 point Joanna Briggs Institute (JBI) critical appraisal tools, the following criteria is established: sample frame appropriate to address the target population, study participants sampled in an appropriate way, adequate sample size, study participants sampled in an appropriate way, study subjects and the setting described in detail, data analysis conducted with sufficient coverage of the identified sample, valid methods used for the identification of the condition, the condition measured in a standard and reliable way for all participants, appropriate statistical analysis and adequate response rate. Individual studies were assigned a score that was computed using different parameters in line with the review objectives. The responses were scored 0 for "No and not reported" and 1 for "Yes". Total scores ranged between 0 and 9. Studies with medium (fulfilling 50% of quality assessment parameter) and high quality were included for analysis [11] (Additional file 1: Table S1).

Statistical analysis
The data entry and analysis were done using Comprehensive Meta-Analysis (version 3.1) software. The original articles were described using forest plot, figures and tables. Since there was heterogeneity among studies, random effect model was used to compute the pooled prevalence and antimicrobial susceptibility of Shigella species. The estimated pooled prevalence rate with 95% confidence interval (CI) was presented.

Heterogeneity and publication bias
Statistical heterogeneity was evaluated by Cochran's Q test and I 2 statistic. The I 2 provides an estimate of the percentage of the variability in effect estimates that is due to heterogeneity rather than sampling error or chance differences. So, the existence of heterogeneity was verified using Cochran's Q test (p < 0.10 indicates statistically significant heterogeneity) and I 2 test that measures level of statistical heterogeneity between studies. I 2 (values of 25%, 50% and 75% are considered to represent low, medium and high heterogeneity respectively) [12,13]. Begg intercept and Mazumdar rank correlation statistics test methods were used to statistically assess publication bias (p < 0.05 was consider as suggestive of statistically significant publication bias) [14]. As the results of the test suggested a possible existence of a significant publication bias, the final effect size was determined by applying Duval and Tweedie's Trim and Fill analysis in the random-effects model.

Identified studies
Through electronic database search, we have found a total of 372 studies. Of which, 335 were excluded based on the titles and abstracts, three article were not reported prevalence of Shigella species and the remaining five were excluded due to unavailability of full text articles. Finally, 29 studies were found to be eligible and included in the meta-analysis (Fig. 1 Do not provide prevalence rate (n=3) Studies Included in systemaƟc review and meta-analysis =29

Study characteristics
Selected articles were published from 1999 to 2018. Also, all included publications were obtained from 5 regions and 1 City administration, but no data was obtained from other regions (Afar, Benishangul-Gumuz, Gambela and Somali). The total study populations involved for the estimation of Shigella species were (8521). Of this, 7571 study participants were included in the meta-analysis of Shigella species drug resistant pattern. A total of twenty-nine studies were considered eligible for quantitative syntheses. Twentyfive studies were used to estimate the proportions of drug resistant for Shigella species (Table 2). Nine studies were used to estimate the prevalence of Shigella species in children and seven studies were used to estimate for adult population. Of the total studies, only seven reported the serogroup of Shigella (Table 1) Table 1). All studies utilized stool specimens for diagnosis of Shigella species (Table 1). Data from health facility based study: 6717 patients with diarrhea and 113 without diarrhea and from community based study: 1691 apparently healthy subjects were considered for quantitative syntheses. The prevalence of Shigella in stool samples of apparently healthy subjects and patients ranged from 0.0 to 3.1% from 1.1 to 34.6%, respectively. MacConkey, xylose lysine desoxycholate and Salmonella Shigella agar used for culture and disk diffusion method for antimicrobial susptablity (Additional file 1: Table S2).
Among the above study parameters considered during meta-regression analysis, study years were associated with a decreasing Shigella prevalence over time (p = 0.0028).   Table S4).

Discussion
Antimicrobial resistance (AMR) is widely growing public health threat worldwide, particularly in resource limited countries including Ethiopia, where infectious diseases are widespread. It is now occurring across the world [1]. Shigella species is highly antibiotic resistant among eight drugs resistance bacteria. Regarding antibiotic resistance, if nothing is done, a person will die every three second by 2050 [44].
This meta-analysis determined the pooled prevalence of shigella species in Ethiopia using 29 studies. According to the results of this meta-analysis, the pooled prevalence was 6.6% (95% CI 4. 7-8.8). This finding is in agreement with 6.6% (95% CI 3.4-9.7) Shegilla prevalence in systematic review among US military and similar populations [45].
Regional prevalence of Shigella species among regions was also calculated, hence a higher prevalence of Shigella species (8.8%) was reported in Tigray, which was nearly four times higher than the finding from Central Ethiopia (2.2%), even though the studies conducted and included in this review and meta-analysis from this region were small. The variations in prevalence estimates may be due to differences in the study populations and year of study.
Pooled prevalence of Shigella species by study year (26.9%) was reported in 1999.2003 and (4.9%) in 2014-2018. This finding is inconsistent with the study done in China [46]. This result showed a decreasing Shigella prevalence over time. The decreased in prevalence through time might be due to decrease in poverty, increase quality of life, rise of awareness on sanitation and hygiene, prevention and control strategy of communicable disease through deployed of health extension workers at community level across the country.
In this study, the pooled prevalence of shigellosis in children was (7.0%), while in adult population (2.2%). This finding confirms that Shigella cause diarrhoeal morbidity among infants and young children than adults [47] and the third leading cause of diarrhoeal deaths in children younger than 5 years [48]. The higher occurrence of Shigella in children compared to adults suggests a higher vulnerability of children to Shigella infection, this may be due to the unhygienic food handling practices, compromised sanitation, malnutrition and the ability to cause disease with a small inoculum and facilitating spread from person to person [49].
Based on the data obtained from twenty-five published articles; Shigella species showed high resistance to amoxicillin, ampicillin, and erythromycin. This finding is in line with the study done in Germany [50].
In Ethiopia the drug of choice on shigellosis treatment is norfloxacin, ciprofloxacin, and ceftriaxone for adult. This is similar with guideline USA [51]. However this review and meta-analysis showed slightly high resistance was reported on norfloxacin, ciprofloxacin, and ceftriaxone.
Furthermore, the occurrences of Shigella isolates resistant to two or more drugs were high (83.2%). This increment may be due mobile genetic units (including plasmids, gene cassettes in integrons and transposons) [1], inadequate access to effective drugs, unregulated dispensing, truncated antimicrobial therapy, medication sharing, counterfeit drugs, bacterial evolution, climate changes, lack of medical practitioner with proper training, poor-quality and unhygienic sanitary conditions [52][53][54].
In this review, included studies primarily used stool culture for Shigella identification. This estimate appears to be less sensitive method than molecular methods and may underestimate the actual occurrence of this bacteria [47].

Conclusion
The pooled estimate provides high burden of Shigella infection and its high proportion of drug resistance pattern to ampicillin, Amoxicillin and Erythromycin in Ethiopia. Therefore, initiating and scale up of performing drug susceptibility test for each shigellosis case, educate the community and health care providers on appropriate use of antibiotics need to be considered and strengthened.      SXT 66.7

Limitation
The patient data were collected from urban and rural populations of the country that took service from hospitals but urban dwellers have better access to health care setting. Therefore, the pooled estimates are more applicable to patients in urban areas than rural. Resistance pattern of shigellosis was not reported for money drugs by some studies and in most studies Shigella infection were not tested at species level.

Additional files
Additional file 1: Table S1. Study design and quality assessment of the studies included in systematic review and meta-analysis of shigella species. Table S2. Subgroup meta-analysis of Shigella species prevalence estimation in Ethiopia from 1999 to 2018. Additional file 2: Figure S1. Egger regression intercept of analysis of included studies reporting on the prevalence of Shigella species in Ethiopia. Figure S2. Begg and Mazumdar rank correlation of analysis of included studies reporting on the prevalence of Shigella species in Ethiopia.