Prevalence of Brucella melitensis and Brucella abortus aminoglycoside-resistant isolates: a systematic review and meta-analysis

Abstract

Introduction

Aminoglycosides are vital antibiotics for treating Brucella infections, because they interfere with bacterial protein production and are often combined with other antibiotics. They are cost-effective, have fewer side effects, and can penetrate biofilms. The prevalence of brucellosis has increased in recent years, increasing the need for effective treatments. In addition, the emergence of multidrug-resistant Brucella strains has highlighted the need for an updated and comprehensive understanding of aminoglycoside resistance. This systematic review aimed to provide a comprehensive overview of the global prevalence of aminoglycoside resistance in B. melitensis and B. abortus.

Methods

A systematic search of online databases was conducted and eligible studies met certain criteria and were published in English. Quality assessment was performed using the JBI Checklist. A random-effects model was fitted to the data, and meta-regression, subgroup, and outlier/influential analyses were performed. The analysis was performed using R and the metafor package.

Results

The results of this systematic review and meta-analysis suggested that the average prevalence rates of streptomycin, gentamicin, and amikacin resistance were 0.027 (95% confidence interval [CI], 0.015–0.049), 0.023 (95% CI, 0.017–0.032), and 0.008 (95% CI, 0.002–0.039), respectively. The prevalence of streptomycin resistance was higher in the unidentified Brucella group than in the B. abortus and B. melitensis groups (0.234, 0.046, and 0.017, respectively; p < 0.02). The prevalence of gentamicin resistance increased over time (r = 0.064; 95% CI, 0.018 to 0.111; p = 0.007). The prevalence of resistance did not correlate with the quality score for any antibiotic. Funnel plots showed a potential asymmetry for streptomycin and gentamicin. These results suggest a low prevalence of antibiotic resistance in the studied populations.

Conclusion

The prevalence of aminoglycoside resistance in B. melitensis and B. abortus was low. However, gentamicin resistance has increased in recent years. This review provides a comprehensive and updated understanding of aminoglycoside resistance in B. melitensis and B. abortus.

Supplementary Information

The online version contains supplementary material available at 10.1007/s42770-023-01233-6.

Introduction

Brucellosis is a zoonotic illness that affects both animal and human health. The genus Brucella contains a fastidious, tiny Gram-negative Coccobacillus bacterium that is the cause of this illness [1]. Direct contact with infected animals, consumption of infected dairy products, or inhalation of infected aerosols during the treatment of the isolate in the laboratory are ways in which the disease may contract. This disease has a significant negative economic impact on the livestock industry and seriously threatens veterinary and public health worldwide. A variety of mammals can be infected with Brucella species. One million people are expected to contract mammalian brucellosis annually, with 40% of cases developing chronic illnesses. Human brucellosis is caused by the clinical strains of Brucella melitensis (B. melitensis) and Brucella abortus (B. abortus).

Antibiotic resistance (AR) is a serious public health concern worldwide. Antibiotic-resistant bacterial infections (ARBs) are associated with increased mortality, hospitalization requirements, prolonged hospitalization, and increased healthcare expenses [2]. This bacterium is located intracellularly, which makes its eradication challenging. In addition, many antibiotics lose their ability to efficiently target Brucella-infected cells because they cannot endure for long enough to have a therapeutic impact. Therefore, the current brucellosis treatments require long-acting antimicrobial medications. Doxycycline, aminoglycoside (amikacin, gentamicin, kanamycin, neomycin, streptomycin, and tobramycin), rifampicin, tetracyclines, and co-trimoxazole (trimethoprim plus sulfamethoxazole), which are typically used in dual or triple regimens, are the medications most frequently suggested for treating human brucellosis. While the combination of doxycycline with aminoglycosides or rifampicin is commonly utilized, and antibiotic therapy is practically effective, there is a notable challenge as relapse rates range between 5 and 10%, often resulting in the incomplete eradication of the disease [3]. Antimicrobial resistance (AMR) refers to a condition in which bacteria, parasites, viruses, and fungi can survive and proliferate in the presence of antibiotics that are often effective in stopping them. Various resistance mechanisms such as altered antimicrobial targets, enzymatic hydrolysis/degradation, efflux, and impermeability can cause AMR. The selection pressure caused by the correct or incorrect use of antimicrobial agents results in resistance, which is mediated by a variety of resistance genes [4]. Infectious diseases are among the leading causes of death in humans and animals, worldwide. The World Health Organization (WHO) and Centers for Disease Control and Prevention (CDC) have raised serious concerns about the ongoing increase in multidrug-resistant microorganisms. Drug resistance is a major health concern worldwide. Resistance to drugs used to treat brucellosis is a significant issue [5]. Therapy failure is common, and many cases of brucellosis relapse following treatment have been reported, with recurrence rates ranging from 5 to 15% in moderate cases. Recently, multidrug-resistant Brucella strains have emerged in brucellosis endemic areas worldwide. Antibiotic resistance in zoonotic infections is a problem, because it reduces the options for treating illnesses in both humans and animals. Multidrug resistance has increased globally, and is now viewed as a public health hazard [6].

Although aminoglycosides are currently recommended for malt fever treatment, there is a notable absence of comprehensive global data on aminoglycoside resistance in Brucella isolates. This systematic review aimed to fill this critical gap by evaluating aminoglycoside resistance in brucellosis treatment and offering precise recommendations.

Bacterial resistance to antibiotics poses a heightened risk for ineffective treatment and increased mortality [7]. Although aminoglycosides are currently recommended for the treatment of malt fever, there is a notable absence of comprehensive global data on aminoglycoside resistance among these isolates. This systematic review aimed to assess aminoglycoside resistance in brucellosis treatment and to provide clear treatment recommendations.

Methods

Eligibility criteria

The eligibility criteria for incorporating articles in the meta-analysis included studies that investigated B. melitensis and B. abortus, reported the proportion of resistance to aminoglycoside antibiotics, determined the sample size, used the standard antimicrobial susceptibility test (AST) guidelines, and interpreted the AST results based on standard guideline breakpoints, and full-text articles published in English. The exclusion criteria were languages other than English, case reports, single-arm studies, cohort studies, and pharmacokinetic studies.

Search strategy

All included studies were identified through a systematic search of online databases, including MEDLINE, Scopus, and Embase, through 18/08/2022. The search was performed using relevant keywords, including B. melitensis and B. abortus, antimicrobial susceptibility test, aminoglycosides gentamicin, tobramycin, amikacin, plazomicin, streptomycin, and neomycin, as well as mesh terms. The protocol for conducting this meta-analysis was registered in PROSPERO with registration code CRD42023392158. The search syntax is available in Supplementary File 1.

Selection process

The results of the systematic online database search were imported into EndNote (version 20) and duplicates were removed. Two authors (F-G and S-MK) independently searched and analyzed relevant publications to prevent bias. A third author (M-S) investigated the disparities.

Data collection process

Extracted data included first author(s), publication year, country, AST technique (Disk Diffusion, E-test, Agar Dilution, Broth Dilution), AST guideline (CLSI, Non-CLSI), source of isolates (human specimens, animal samples), and species (not identified, B. melitensis, B. abortus); The article lacks differentiation between isolates of B. melitensis and B. abortus, presenting them collectively as Brucella. Moreover, it does not encompass other species beyond B. melitensis and B. abortus, specimens (Other, NA, blood cultures, joint fluid), number of resistant isolates, and the total number of isolates, to avoid errors in data extraction, two authors (F-G and S-MK) independently extracted the necessary data and agreed to the discrepant data.

Study risk of bias assessment

Owing to the inclusion of cross-sectional studies, the JBI checklist was used to evaluate the quality of the included articles [8]. Two authors (F-G and S-MK) independently evaluated the data quality. A third author (M-S) investigated the disparities.

Synthesis methods

The primary objective of this study was to determine the prevalence of quinolone-resistant Brucella strains. The number of antibiotic-resistant isolates for each aminoglycoside and total number of investigated isolates (sample size) were used to calculate the pooled prevalence of each antibiotic.

Statistics

The analysis was performed using proportions as outcome measures. The data were fitted to a random-effects model. The amount of heterogeneity (i.e., τ²) was estimated using the DerSimonian-Laird estimator [9]. In addition to estimating τ², a Q-test was used to assess heterogeneity [10], and I² statistics [11] is reported. If any amount of heterogeneity is detected (i.e., τ² > 0, regardless of the Q-test results), a prediction interval for the true outcomes is also provided in [12]. Meta-regression analysis was used to analyze the publication years of the included articles, quality assessment score, and subgroup analysis to investigate differences in prevalence between AST techniques (disk diffusion, E-test, agar dilution, broth dilution), AST guidelines (CLSI, non-CLSI), source of isolates (human and animal samples), species (not identified, B. melitensis, B. abortus), specimens (other, NA, blood cultures, joint fluid), and continents.

Studentized residuals and Cook’s distances were used to examine whether the studies were outliers and/or influential in the context of the model [13]. Studies with a studentized residual larger than the 100 × (1 − 0.05/(2 × k))th percentile of a standard normal distribution were considered potential outliers (i.e., using a Bonferroni correction with two-sided α = 0.05, for studies included in the meta-analysis). Studies with Cook’s distance greater than the median and six times Cook’s distance interquartile range were considered influential, rank correlation test [14], and regression tests [15]. The standard error of the observed outcomes was used as a predictor to check funnel plot asymmetry. The analysis used R (version 4.2.1) [16] and the metafor package (version 3.8.1) [17].

Results

Descriptive statistics

A total of 1360 records as results of the systematic search were collected in reference manager software (EndNote version 20), and 856 duplicated articles were removed. A total of 504 articles were assessed in terms of the title and abstract of the section, and 83 full-text articles were evaluated. Eventually, this systematic review and meta-analysis included 48 eligible studies [1, 5, 6, 18–60]. The screening and selection of presagers are summarized in the PRISMA flow chart (Fig. 1). The reports were from 18 nations in Asia, Europe, and South America. The specimens included blood cultures from the patients, cerebrospinal fluid, and other laboratory methods for bacterial culture. Most studies have used CLSI as AST Guidelines and agar dilution, broth dilution, E-Test, and disk diffusion as AST methods to detect antibiotic resistance. The reports covered the years 1983 to 2020. The information extracted from the articles included in the meta-analysis is summarized in Supplementary Table 1.

Fig. 1.

PRISMA flow diagram of the included articles

Prevalence of streptomycin resistance

Forty studies were included in the analysis. The observed prevalence of resistance ranged from 0.001 to 0.994, with most estimates being negative (98%). The estimated average prevalence based on the random-effects model was μ = 0.027 (95% confidence interval [CI], 0.015 to 0.049). Therefore, the average outcome differed significantly from 0 (z = − 11.334, p < 0.001). A forest plot showing the observed outcomes and the estimate based on the random-effects model is shown in Fig. 2. According to the Q-test, the true outcomes appear to be were heterogeneous (Q(39) = 152.162, p < 0.001, τ² = 2.608, I² = 74.369%). Subgroup analysis was performed according to the significant heterogeneity between reports, and the differences among AST methods, AST guidelines, countries, and continents were insignificant. Among the Brucella species groups, the non-identified Brucella species had a higher prevalence of resistance than B. abortus and B. melitensis (0.234, 0.046, and 0.017, respectively, p < 0.02). An examination of the studentized residuals revealed that one study [61] had a value larger than ± 3.227, and may be a potential outlier in the context of this model. According to Cook’s distance method, one study [61] can be considered overly influential. Figure 3 depicts the funnel plot, while Table 1 provides a concise summary of the sensitivity analysis results. The estimated average proportion based on the random-effects model after removing potential outlier reports was μ = 0.023 (95% CI, 0.013–0.041).

Fig. 2.

A Forest plot showing the overall proportion of Brucella streptomycin resistance, also subgroup difference in AST method, Brucella species, specimens, and AST guideline estimate of the random-effects model, B subgroup analysis of the proportion of Brucella streptomycin resistance in continents, C subgroup analysis of the proportion of Brucella streptomycin resistance in countries

Fig. 3.

Funnel plot of reported percentages of resistance and the result of trim-and-fill, which is used to simulate the symmetrical state of the funnel plot and estimate the overall proportion of resistance in the symmetrical state. A The overall proportion of ciprofloxacin resistance in a symmetrical state was $\widehat{\mu }$= 0.057 (95% CI, 0.033 to 0.096). B The overall proportion of gentamycin resistance in the symmetrical state was $\widehat{\mu }$ = 0.013 (95% CI, 0.008 to 0.02), C overall proportion of amikacin resistance in the symmetrical state was $\widehat{\mu }=0.014$ (95% CI, $0.007$ to $0.029$)

Table 1.

Evaluation of publication bias in meta-analysis

ANTIBIOTIC	EGGER TEST	BEGG'S TEST	FAIL AND SAFE	TRIM AND FILL
Streptomycin	< 0.001	0.155	6134	0.057 (95% CI, 0.033–0.096)
Gentamycin	< 0.001	< 0.001	6358	0.034 (95% CI, 0.024–0.049)
Amikacin	0.215	0.333	36	0.008 (95% CI, 0.002–0.039)

This table provides a comprehensive assessment of potential publication bias in the meta-analysis using a range of statistical techniques. Included are statistics generated from Egger’s method, Begg’s method, the Fail-Safe N, and the Trim-and-Fill method. These methods are applied to investigate the presence of bias and its impact on the meta-analysis results, ensuring the robustness and reliability of the findings

Prevalence of gentamycin resistance

A total of k = 38 studies were included in the analysis. The observed prevalence of resistance ranged from 0.002 to 0.073, with most estimates being negative (100%). Based on the random-effects model, the estimated average prevalence was μ = 0.023 (95% CI, 0.017–0.032). Therefore, the average outcome differed significantly from 0 (z = − 21.740, p < 0.001). A forest plot showing the observed outcomes and the estimate based on the random-effects model is shown in Fig. 4. According to the Q-test, there was no significant heterogeneity in true outcomes (Q(37) = 31.967, p = 0.704, τ² = 0.000, I² = 0.000%). Several studies [38, 42, 45, 47] had relatively large weights compared to the rest of the studies (i.e., “weight” ≥ 3/k, so a weight at least three times as large as having equal weights across studies). Examination of the studentized residuals revealed that none of the studies had a value greater than ± 3.213. Hence, there was no indication of outliers in the context of this model. Based on the basis of Cook’s distance, two studies [38, 47] can be considered overly influential. The estimated average proportion based on the random-effects model after removing the potential outlier reports was μ = 0.013 (95% CI, 0.008 to 0.02).

Fig. 4.

Forest plot showing the overall proportion of Brucella gentamycin resistance and the estimate of the random-effects model

Prevalence of amikacin resistance

Three studies were included in this analysis. The observed prevalence of resistance ranged from 0.003 to 0.033, with most estimates being negative (100%). The estimated average prevalence based on the random-effects model was μ ^= 0.008 (95% CI, 0.002–0.039). Therefore, the average outcome differed significantly from 0 (z = − 5.870, p < 0.001). A forest plot showing the observed outcomes and the estimate based on the random-effects model is shown in Fig. 5. According to the Q-test, there was no significant heterogeneity in true outcomes (Q(2) = 1.551, p = 0.461, τ² = 0.000, I² = 0.000%). Examination of the studentized residuals revealed that none of the studies had a value greater than ± 2.394. Hence, there was no indication of outliers in the context of this model. According to Cook’s distance, none of the studies could be considered to be overly influential. The estimated average proportion based on the random-effects model after removing the potential outlier reports was μ = 0.004 (95% CI, 0.001–0.027).

Fig. 5.

Forest plot showing the overall proportion of Brucella amikacin resistance and the estimate of the random-effects model

Discussion

Brucellosis is the most common zoonotic disease in the world. Annually, there are over 500,000 new cases, although they are unevenly distributed worldwide. Annual occurrence rates range from 0.3 cases per million in the UK and most of the US to over 1 case per 1000 in endemic locations, where the disease causes a significant and increasing health burden.

Treatment is prescribed to decrease symptom duration, prevent relapse, and avoid consequences such as arthritis, sacroiliac joint inflammation, spondylitis, encephalitis, endocarditis, and abortion [62]. The resistance of Brucella spp. to antibiotics has reached alarming levels worldwide, greatly affecting their treatment efficacy. Local surveillance networks are required to select the appropriate eradication regimen for each region. ABR is associated with a high mortality risk and increased economic costs, with pathogens implicated as the main cause of increased mortality.

This meta-analysis provides a comprehensive overview of universal aminoglycoside resistance in humans with Brucellosis caused by Brucella spp. A meta-analysis by Shahrabi et al. investigated the prevalence of tetracycline resistance in B. melitensis and B. abortus strains. The estimated rates of resistance to tetracycline and doxycycline were 0.017 (95% CI, 0.009–0.035) and 0.017 (95% CI, 0.011–0.026), respectively, based on 51 studies conducted from 1983 to 2020. Both drugs exhibited an increasing trend in resistance over time (tetracycline: r = 0.077, p = 0.012; doxycycline: r = 0.059, p = 0.026) [63]. In addition, in an investigation conducted by Beig et al., this study systematically reviewed and employed a meta-analysis to assess the prevalence of fluoroquinolone resistance in B. melitensis and B. abortus isolates. The resistance rates to ofloxacin, sparfloxacin, fleroxacin, pefloxacin, and lomefloxacin were 2%, 1.6%, and 4.6%. These findings underscore the necessity for targeted treatment strategies and heightened surveillance in response to the evolving fluoroquinolone resistance [64].

The abuse of antibiotics increases ecological pressure on local communities and adds to the burden of antibiotic resistance, especially in the Syrian Arab Republic, Egypt, Saudi Arabia, Turkey, Iran, and Brazil. To our knowledge, this is the first study to analyze how aminoglycosides resist Brucella infections. These studies revealed that the administration of amikacin to the patients was effective. We also observed that the rates of resistance to streptomycin were higher than those to gentamycin. Although the prevalence of amikacin- and streptomycin-resistant Brucella has not changed over the years, that of gentamycin resistance has increased, due to the overuse and misuse of antibiotics, particularly gentamycin. Health authorities should pay additional attention to the evolution of streptomycin resistance, which is a serious global health issue.

Conclusion

In conclusion, aminoglycosides remain vital for Brucella treatment because of their efficacy, low cost, and minimal side effects. Despite the increasing prevalence of brucellosis and multidrug-resistant strains, this study revealed a generally low resistance to streptomycin, gentamicin, and amikacin in B. melitensis and B. abortus. Notably, although streptomycin resistance was higher in the unidentified Brucella groups, an upward trend in gentamicin resistance was observed over the years. This concise review provides valuable insights into the optimization of aminoglycoside treatment strategies for brucellosis management.

Limitation

The following are the limitations encountered while conducting this study: in determining Brucella’s antibiotic resistance in the first step, there is no specific guideline for determining antibiotic resistance in Brucella, and we had to follow the defined guidelines and breakpoints established for Haemophilus influenzae. Additionally, no standardized laboratory method has been established to evaluate antibiotic resistance in Brucella spp. As a result, we included all the techniques used in different methods worldwide in our analysis.

Supplementary Information

Below is the link to the electronic supplementary material.

Data Availability

All relevant data utilized in this study can be found in the supplementary file.