Global trends and molecular epidemiology of Azithromycin-resistant Salmonella: A systematic review and meta-analysis Authors: Getachew Mengistu, Mekuanint Geta, Anwar Nuru, Baye Gelaw, Feleke Moges Published: 2026 ABSTRACT AbstractIntroductionSalmonella infections pose a significant public health challenge, affecting both social and economic dimensions across various countries. Different animal species serve as reservoirs for Salmonella and contribute to drug resistance in human isolates. The improper use of antimicrobials in livestock, companion animals, and humans has led to the rise of multidrug-resistant (MDR), extensively drug-resistant (XDR), and azithromycin-resistant Salmonella serovars globally. This systematic review and meta-analysis aim to evaluate heterogeneity, estimate pooled proportions, and outline the molecular epidemiology of these resistant Salmonella serovars worldwide.MethodsWe analyzed data from 15,514 Salmonella isolates, focusing on their genetic and phenotypic characteristics. Our analysis included meta-analyses and frequency analyses based on the studies' geographical locations, the types of isolates, and potential biases. We employed a random-effects model, specifically the DerSimonian-Laird method, to evaluate heterogeneity and estimate pooled proportions. The result is presented as a proportion and confidence interval after the inverse of the Freeman-Tukey transformation; otherwise, all the calculations were made using the Freeman–Tukey double arcsine transformation.ResultsThe pooled prevalence of azithromycin resistance among Salmonella isolates was 9.12% (95% CI: 6.53–12.07). Among azithromycin-resistant isolates, multidrug-resistant (MDR) and extensively drug-resistant (XDR) serovars accounted for 31.10% (95% CI: 18.82–44.86) and 19.30% (95% CI: 6.90–35.75), respectively. Resistance mechanisms involved multiple genetic factors, including efflux pumps (mefA, acrB gene mutation), enzymatic phosphorylation (mphA), and class 1 integrons gene cassettes. Notably, resistance rates varied by region and serovar, with high prevalence in Southern Asia and Eastern Europe.ConclusionAzithromycin resistance in Salmonella is a growing global health concern, complicated by the presence of MDR and XDR strains and diverse resistance mechanisms. Urgent, coordinated efforts encompassing enhanced surveillance, antimicrobial stewardship, and research are critical to controlling resistance and preserving azithromycin’s efficacy in treating invasive salmonellosis through a One Health approach. Additionally, addressing antimicrobial use in the agriculture and veterinary sectors is critical to safeguarding public health.RegistrationWe conducted a systematic review and meta-analysis after registering the protocol with PROSPERO (CRD420251047985), following the MOOSE (Meta-Analysis of Observational Studies in Epidemiology) guidelines.Keywords: Salmonella, Azithromycin resistance, MDR, XDR, Molecular epidemiology, Genetic mechanisms, Meta-analysis Introduction Salmonella serovars remain among the leading causes of foodborne and enteric bacterial diseases worldwide, imposing substantial public health and socioeconomic burdens, particularly in low- and middle-income countries [1], [2]. These pathogens are maintained in diverse animal reservoirs, including poultry, livestock, reptiles, and wild birds, and are transmitted to humans primarily through contaminated food and water[1], [3]. Several point of evidence indicate that the widespread use of antimicrobials in human medicine, livestock production, and agriculture has accelerated the emergence and dissemination of antimicrobial resistance (AMR) determinants within Salmonella populations, thereby facilitating the spread of resistant strains through the food chain and the environment [4], [5], [6], [7], [8]. Consequently, salmonellosis has become increasingly difficult to control and treat, highlighting the growing importance of understanding resistance mechanisms and transmission dynamics. Human infections caused by Salmonella range from self-limiting gastroenteritis to severe systemic diseases such as enteric fever and invasive bloodstream infections [9]. Typhoidal serovars, principally Salmonella enterica serovars Typhi and Paratyphi, are estimated to cause approximately 21.7 million infections and more than 217,000 deaths annually [10], [11], [12], [13], whereas non-typhoidal Salmonella (NTS), comprising over 2,000 serovars, are responsible for nearly 93.8 million cases and 155,000 deaths worldwide each year, most of which are foodborne [2], [10], [11]. In recent decades, the epidemiology of salmonellosis has been increasingly shaped by the emergence of multidrug-resistant (MDR) and extensively drug-resistant (XDR) lineages, as well as shifts in the geographic distribution of dominant serovars[14]. Antimicrobial resistance has become one of the most significant threats to the effective treatment of Salmonella infections[15]. Resistance to traditional first-line agents, including ampicillin, chloramphenicol, and trimethoprim-sulfamethoxazole, became widespread during the 1990s [16], [17], followed by increasing resistance to fluoroquinolones in the early 2000s [18], [19]. These developments led to greater reliance on third-generation cephalosporins and azithromycin as key therapeutic options for both typhoidal and invasive non-typhoidal salmonellosis [20], [21], [22], [23], [24]. However, the emergence of XDR Salmonella strains, particularly those reported from South Asia, has severely restricted available treatment options [25]. XDR S. Typhi strains resistant to chloramphenicol, ampicillin, cotrimoxazole, streptomycin, fluoroquinolones, and third-generation cephalosporins have rendered azithromycin one of the few remaining effective oral therapies. The growing dependence on azithromycin has been accompanied by increasing reports of azithromycin-resistant Salmonella worldwide. Clinical treatment failures associated with azithromycin resistance have been documented, raising concerns about the future effectiveness of this critical antimicrobial. Laboratory detection of azithromycin resistance has therefore become increasingly important for clinical management and surveillance, with susceptibility commonly assessed using minimum inhibitory concentration (MIC) (<16 μg/mL) breakpoints and disk diffusion(>13 mm measured against a 15-μg disk) criteria established by international standards [26], [27], [28]. Several molecular mechanisms contribute to azithromycin resistance in Salmonella. These include target-site modification mediated by methylase genes such as erm[29], [30], [31], enzymatic inactivation through phosphotransferases and esterases encoded by genes including mph and ere [32], [33], active drug efflux mediated by transport systems such as mef and msr[31], and chromosomal mutations affecting ribosomal targets or efflux pump components such as rrl and rpl [34]. Among these mechanisms, the macrolide phosphotransferase gene mphA has emerged as a major determinant of high-level azithromycin resistance and has been associated with markedly elevated MIC values (MIC ≥ 128 μg/mL) in diverse Salmonella serovars [35], [36]. In addition, mutations in the efflux pump gene acrB have been increasingly recognized as important contributors to azithromycin resistance in both S. Typhi and S. Paratyphi A, demonstrating the growing complexity of resistance evolution[21], [22], [37]. Despite increasing reports of azithromycin-resistant and XDR Salmonella, current evidence remains fragmented across geographical regions, serovars, and study designs. Most investigations have focused on specific outbreaks, individual countries, or selected determinants of resistance, limiting a comprehensive understanding of the global distribution of resistance mechanisms, the prevalence of key genetic markers, and the evolutionary relationship between azithromycin resistance and emerging XDR lineages. Furthermore, the rapid expansion of whole-genome sequencing studies has generated a large body of molecular data that has not been systematically synthesized to identify global patterns and emerging trends. Therefore, a comprehensive meta-analysis is needed to consolidate existing evidence on the prevalence, geographical distribution, and molecular determinants of azithromycin resistance in Salmonella. By integrating phenotypic and genomic data from studies conducted worldwide, this study aims to characterize the global molecular epidemiology of azithromycin-resistant and emerging extensively drug-resistant Salmonella, identify dominant resistance genes and mutations, assess temporal and geographic trends, and provide evidence to support antimicrobial stewardship, surveillance programs, and future treatment strategies. Methods Study design and data sources We performed a systematic review and meta-analysis of published studies on global trends and the molecular epidemiology of azithromycin-resistant, emerging XDR Salmonella. Azithromycin resistance in Salmonella was defined as a zone of inhibition of 13 mm or less using a 15-μg disk or a MIC of 16 μg/ml or more, as standardized by the Clinical Laboratory Standards Institute (CLSI) [38]. Before extracting data, we registered our research protocol with PROSPERO (International Prospective Register of Systematic Reviews) under registration number CRD420251047985. This was done in accordance with the MOOSE (Meta-Analysis of Observational Studies in Epidemiology) guidelines. The study followed the procedures outlined in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [39], [40]. The study designs included a review comprising laboratory-based cross-sectional studies, comparative cross-sectional studies, and randomized clinical trials. We searched published studies using various databases: MEDLINE/PubMed, Google Scholar, Cochrane Library, Embase, and ScienceDirect. A systematic search strategy was implemented to avoid duplication, with search terms such as “Azithromycin resistant AND Salmonella AND (‘systematic review’ OR ‘meta-analysis’).” Multiple search strings incorporating Medical Subject Headings (MeSH), title/abstract (TIAB), and text words (TW) were used to identify primary studies. The search string that allowed us to identify most studies was “Azithromycin AND Salmonella”. We also examined the reference lists of global studies on Salmonella to identify additional reports. The final search was conducted on December 30, 2024. Study selection We excluded studies with unrelated titles or abstracts and those that did not report primary outcomes of interest. Studies that conducted culture-based, molecular-based, or both methods of Salmonella isolation and assessed azithromycin resistance were included. The eligibility criteria for inclusion were: (i) published in English, (ii) reported on azithromycin resistance, (iii) isolated Salmonella from humans or food samples, and (iv) conducted anywhere in the world. We also excluded letters, communications, correspondence, reports with unclear data, and studies that did not include results on azithromycin antibiotic susceptibility. Most major biomedical and scientific databases (PubMed, Embase, Web of Science, etc.) index primarily English-language journals; as a result, we only include English-language literature. Most major biomedical and scientific databases, including PubMed, Embase, Web of Science, etc., primarily index journals published in English. As a result, we included only English-language literature in our review. We also excluded non-English literature to avoid potential biases arising from translating texts from other languages. Data extraction The following variables were extracted: (i) study identifier: first author, year of publication, year of study, country, target population (humans, animals, and food samples), and sampling population (healthcare/nosocomial, community); (ii) methods/design: study design (cross-sectional, comparative cross-sectional, randomized controlled trial), sample size, sampling method, sample type, and drug susceptibility testing (DST) method; (iii) results: number of isolates, number of isolates subjected to DST, number of isolates by serovar (Typhi, Paratyphi, Typhimurium, and other non-Typhoidal Salmonella [NTS] strains), numbers of MDR serovars, numbers of XDR serovars, number of azithromycin-resistant serovars, method of gene detection (PCR, sequencing, whole genome sequencing [WGS]), numbers of isolates subjected to gene detection, and number of isolates containing specific azithromycin/macrolide resistance genes (mefA, mefB, mphA, mphB, mphD, ereA, ereB, ermA, ermB, ermC, ermV, erm42, msrA, class 1 integron [intl], R717Q mutation in acrB gene, R717L mutation in acrB gene, mutation in acrR gene, and mutation in rplV gene). The data extraction processes were customized to the specific characteristics of each study, encompassing calculations, exclusions, and assumptions. Prevalence data for Salmonella isolates that did not display azithromycin resistance were excluded, and duplicate entries were removed. Authors were approached to obtain additional online or supplementary information. Data extraction was performed twice; we assessed the overall prevalence of azithromycin resistance by examining mutations in the azithromycin resistance gene in S. Typhi, S. Paratyphi, S. Typhimurium, and other NTS serovars. Subgroup analyses were performed by target population and geographical region. Multiple cross-checks were conducted to ensure consistent data extraction and to resolve discrepancies. Assessment of quality and risk of bias The validity of the studies was determined using criteria from Reddy et al.[41], which excluded those likely to yield invalid results or non-comparable outcomes. Each study’s methodological quality was evaluated with the Joanna Briggs Institute (JBI) critical appraisal tool, ensuring compliance with the PRISMA checklist [42]. To identify potential biases, we considered: (i) target population (humans vs. animals/food samples), (ii) sample population (healthcare/nosocomial vs. community), (iii) sample size (n ≤ 384 vs. larger; calculated as n = Z2PQ/d2, with P = 0.5), (iv) sampling method (probability vs. non-probability), and (v) study design (cross-sectional, comparative cross-sectional, randomized control trial). Low-risk studies were defined as those with a sample size over 384, using prospective sampling, and having a comparative cross-sectional or randomized control trial design[41]. We acknowledged variability in microbiological methods and culture interpretation across different resource settings. Thus, we included studies regardless of media types and sample transport methods, even if they lacked detailed contaminant reporting. Heterogeneity was addressed using established inclusion criteria. The review aimed to assess the global trend and molecular epidemiology of azithromycin-resistant Salmonella serovars, necessitating a diverse representation from various hospital communities and animal communities, rather than focusing narrowly on high-risk patients[43]. Data processing and analysis Data extraction, entry, transformation, and analysis were performed using Microsoft Office Excel 2013, R, IBM SPSS (version 28), and Stata (Version 18). The azithromycin resistance levels in Salmonella were categorized as rare (<0.1%), very low (0.1%–1%), low (>1%–10%), moderate (10%–20%), high (20%–50%), very high (50%–70%), and extremely high (>70%)[44]. We estimated the pooled proportion of azithromycin-resistant Salmonella using the DerSimonian-Laird method for random-effects meta-analysis. A double arcsine transformation normalized the data, and the results were back-transformed for reporting[45]. Potential publication bias and heterogeneity were assessed using the Cochrane Q test and I2 statistics, with significance set at p < 0.10 for the Q test and thresholds of 25%, 50%, and 75% for low, medium, and high heterogeneity, respectively[46]. Begg’s rank correlation and Egger weighted regression tests were employed to test for publication bias, with p < 0.05 indicating significant bias[47]. A sensitivity analysis was performed by excluding one study at a time to determine its impact on the pooled effect size and heterogeneity. The effect estimate was considered not to significantly influence the overall result if it remained within the 95% confidence interval after exclusion[47]. Certainty of evidence evaluation This study assessed and summarized the certainty of the pooled estimates regarding the primary outcome, specifically the proportion of azithromycin-resistant Salmonella serovars, and the secondary outcome, the proportion of MDR and XDR Salmonella serovars. This evaluation was conducted using the Grading of Recommendations, Assessment, Development, and Evaluations (GRADE) methodology, utilizing the GRADEpro software [48]. Result Search and selection of studies The flow diagram in Figure 1 shows the study selection process. No systematic reviews or meta-analyses exist on the global trend and molecular epidemiology of azithromycin-resistant and emerging XDR Salmonella. Past reviews covered typhoid fever[49], fluoroquinolone resistance [50], streptomycin-resistant Salmonella[51], and general drug resistance in humans[52] and animals[53], but none focused on azithromycin resistance. Out of 5,977 papers screened, 57 articles were included in this review. Flowchart of literature search and inclusion/exclusion process. Characteristics of included studies The characteristics of the included studies are summarized in Table 1, Table 2. A total of 57 studies are included in this systematic review and meta-analysis. Most were laboratory-based, prospective cross-sectional studies, with two comparative cross-sectional studies [36], [54] and one randomized controlled trial [27]. Many were conducted in India (n=20)[24], [28], [55], [56], [57], [58], [59], [60], [61], [62], [63], [64], [65], [66], [67], [68], [69], [70], [71], [72], followed by Nepal (n=5)[73], [74], [75], [76], [77], and four each from Bangladesh[21], [37], [78], [79], China[35], [80], [81], [82], [83], and Pakistan[84], [85], [86], [87], [88]. From 2011 to 2021, these studies mainly originated from southern Asia (34/57; 59.65%), suggesting a higher prevalence of azithromycin-resistant Salmonella in the region. Of the 57 studies, 48 focused on humans, four on food, four on animals, and one on both food and humans. Most were in healthcare settings (48/57) and involved participants of all ages (43/57), mainly using blood samples (37/57). The Kirby-Bauer disk diffusion method was used in 31 studies. A total of 15,514 Salmonella isolates were included in the review of global azithromycin resistance trends and the emergence of MDR/XDR serovars (Figure 2). Table 1Characteristics of included studies by author, year, and countryAuthor, ReferenceYear of PublicationCountryTotal number of isolatesAZM Resistance (%)MDR (%)XDR (%)Ali Shah et al. [84]2020Pakistan812(2.47)54(66.67)38(46.91)Anvarinejad, et al.[132]2016Iran195(26.32)5(26.32)-Aslam, et al.[85]2021Pakistan502(4)35(70)27(54)Aziz, et al. [86]2018Pakistan13745(32.85)--Behl, et al.[55]2017India1062(1.89)--Besharati, et al. [133]2020Iran9749(50.52)34(35.05)0(0)Bhetwal, et al. [73]2017Nepal2458(3.27)--Carey, et al[24]2021India667(10.61)7(10.61)7(10.61)Choudhary, et al.[56]2013India32281(25.16)--Dave, et al.[134]2015England1940(0)--Dong, et al.[80]2020China14715(10.2)134(91.16)-Elkenany, et al.[113]2018Egypt264(15.38)26(100)-Fida, et al[87]2021Pakistan521(1.92)48(92.31)37(71.15)Garg, et al.[57]2013India424(9.52)--Gebremedhin, et al[135]2021Ethiopia208(40)8(40)2(10)Gupta, et al.[58]2013India3020(0)--Harb, et al.[136]2017Iraq3322(66.67)28(84.85)17(51.52)Harichandran,[59]2017India1180(0)0(0)0(0)Hassing, et al[106]2014Netherlands35457(16.1)--Hong, et al.[123]2021Taiwan3510(28.57)--Hooda, et al.[21]2019Bangladesh108213(1.2)13(1.2)13(1.2)Hussain, et al.[88]2019Pakistan2927(2.4)200(68.49)115(39.38)Jain, et al.[60]2020India9925(25.25)37(37.37)-Jain, et al[61]2013India34425(7.27)9(2.62)-Joshi, et al.[74]2018Nepal1591(0.63)--Katiyar, et al.[62]2020India1331(0.75)--Khadka, et al.[75]2021Nepal900(0)--Khan, et al.[63]2019India10018(18)5(5)-Khanal, et al.[76]2017Nepal880(0.00)--Khatun, et al.[78]2018Bangladesh4317(1.62)122(28.31)-Kokare, et al.[64]2021India1550(0)18(11.61)-Kuijpers, et al.[137]2017Cambodia2540(0)--Lin, et al.[81]2015China8220(24.39)--Makkar, et al[65]2018India78049(6.28)18(2.31)-Misra, et al[66]2016India1000(0)--Müller, et al.[107]2011Germany1250(0)--Nair, et al.[36]2016UK6839(1.32)7(1.02)-Nelson, et al. [77]2020Nepal1578(5.1)--Octavia, et al.[54]2021Singapore243(12.5)3(12.5)-Okanda, et al[79]2018Bangladesh240(0)7(29.17)-Parry, et al.[27]2015Vietnam1700215(12.65)510(30)-Patel, et al. [67]2017India4710(21.28)11(23.4)-Patil, et al [68]2019India2510(0)--Prabhurajan, et al.[69]2021India178(47.06)--Rai, et al. [28]2012India10736(33.64)4(3.74)-Reddy, et al[105]2011England1388(5.8)36(26.09)-Sajib, et al[37]2021Bangladesh3025104(3.44)--Shang, et al[138]2018South Korea12322(17.89)53(43.09)-Sharma, et al.[70]2019India60263(10.47)--Tîrziu, et al. [139]2020Romania5145(88.24)--Vala, et al[71]2016India6139(63.93)--Veltman, et al.[140]2021Australia3070(0)--Vlieghe, et al.[141]2012Cambodia7218(25)18(25)-Waghamare, et al.[72]2018India719(12.68)--Wang, et al[35]2017China55931(5.55)473(84.62)-Wong, et al. [82]2014China5462(0.37)12(2.2)12(2.2)Yasin, et al.[83]2021China18937(19.58)175(92.59)- Characteristics of included studies by author, year, and country Table 2Characteristics of analyzed studies.Study characteristicsNumber of studiesStudy characteristicsNumber of studiesCountryAustralia1Sampling methodConsecutive/Random1Eastern Africa1Purposive2Eastern Asia7Random18Eastern Europe1Consecutive36Northern Africa1Age groupAll43Northern Europe3Adult2Southeastern Asia4Child5Southern Asia34NA7Western Asia3Sample typeBlood37Western Europe2Blood/feces2Target populationAnimal4Blood/feces/pus1Food4Feces6Food/human1Food8Human48Rectal swabs1Sample populationCommunity8Unspecified2Community/health care1Method of antimicrobial susceptibility testingMIC determination methods8Health care48Disk-diffusion & MIC determination18Study designCross-sectional54Disk-diffusion31Comparative cross-sectional2Randomized controlled trials1Total57 Characteristics of analyzed studies. Geographical distribution of azithromycin-resistant Salmonella serovars. The distribution of studies examining azithromycin-resistant Salmonella, categorized by year, country, and region, with no studies from southern or western Africa are demonstrated. Data availability is plotted by year on the x-axis and by country on the y-axis, grouped by region. The size of each point represents the number of studies conducted for each country in a given year. Burden and heterogeneity of azithromycin resistance in Salmonella serovars This meta-analysis and systematic review identified 15,514 Salmonella isolates, including 14,580 from humans, food samples (n = 597), animal samples (n = 240), and undetermined samples (n = 97). The prevalence rates of azithromycin resistance, MDR, and XDR Salmonella serovars were 7. 45%, 25.48. 48%, and 11. 00%, respectively. Notably, azithromycin resistance was highest in undifferentiated food and human samples (50. 5%), despite the small number of isolates (Table 3). Most azithromycin-resistant isolates were from Southern Asia (573 isolates, 5.90%), while Eastern Europe had a high percentage of resistance (45 isolates, 88.24%) (Table 4). Table 3Characteristics of included studies by sample population and Salmonella isolateSample populationNumber of Salmonella isolatesAzithromycin resistant (percentage)Total isolates tested for MDRMDR Isolates (Percentage)Total isolate tested for XDRXDR Isolates (Percentage)Animal24043(17.92)14361(42.65)2010(10.0)Food59773(12.23)00(0)00(0.0)Food/human9749(50.5)9734(35.05)970(0)Human145809 90(6.79)79381989(25.05)2320266(11.46)Total155141155(7.45)81782165(25.48)2437268(11.00) Characteristics of included studies by sample population and Salmonella isolate Table 4Characteristics of included studies by region and Salmonella isolateSample populationNumber of Salmonella isolatesAzithromycin resistant (percentage)Total isolates tested for MDRMDR Isolates (Percentage)Total isolate tested for XDRXDR Isolates (Percentage)Australia3070(0.00)----Eastern Africa208(40.00)208(40.00)202(10.00)Eastern Asia1681137(8.15)1564847(54.15)54612(2.20)Eastern Europe5145(88.24)----Northern Africa264(15.38)---Northern Europe101517(1.67)82143(5.24)--Southeastern Asia2050236(11.51)1796531(29.56)--Southern Asia9736575(5.90)3828588(15.36)1741237(13.61)Western Asia14976(51.01)14967(44.97)13017(13.08)Western Europe47957(11.90)----Total155141155(7.44)81782084(25.48)2437268(11.00) Characteristics of included studies by region and Salmonella isolate Details on azithromycin resistance, MDR, and XDR Salmonella serovars are presented in Table 5. A total of 57 articles were analyzed across 15, 514 isolates, including S. Typhi (8, 955), S. Paratyphi (2, 188), and S. Typhimurium (1, 283). Notably, a significant percentage of azithromycin-resistant Salmonella serovars were identified in S. Choleraesuis (45.95%) and S. Kentucky (33.33%), although fewer isolates were tested. The predominant resistant isolates were S. Typhi (607, 6.77%) and S. Paratyphi (157, 7.17%). Multidrug- resistant S. Typhi serovars accounted for 1, 134 cases (25. 35%), while XDR serovars were identified in 235 cases (15. 69%). Table 5Characteristics of included studies by Salmonella isolateSalmonella SerovarsNumber of Salmonella isolatesAzithromycin resistant (percentage)Total isolates tested for MDRMDR Isolates (Percentage)Total isolate tested for XDRXDR Isolates (Percentage)S. Albany3510(28.57)----S. Choleraesuis3717(45.95)374(10.81)--S. Enteritidis161(6.25)70(0.0)--S. Houtenae195(26.32)195(26.32)--S. Kentucky62(33.33)----S. Paratyphi2188157(7.17)87533(3.77)2262(0.88)S. Typhi8955607(6.77)44731134(25.35)1498235(15.69)S. Typhimurium128349(3.81)1268619(48.81)55812(2.15)Other Salmonella serovars ⁎2975307(10.32)1499289(19.28)15519(12.25)Total155141155(7.44)81782084(25.48)2437268(11.00)⁎S. Amsterdam, S. Infantis, S. Labadi, S. Larochelle, S. London, S. Takoradi, S. Tamale, and unspecified Salmonella serovars Characteristics of included studies by Salmonella isolate S. Amsterdam, S. Infantis, S. Labadi, S. Larochelle, S. London, S. Takoradi, S. Tamale, and unspecified Salmonella serovars All calculations used Freeman–Tukey double arcsine-transformed values for the forest plot, but data were presented as proportions with confidence intervals after back-transformation. Despite high heterogeneity (I2 > 90%), the azithromycin-resistant Salmonella proportion was 9.12% (95% CI 6.53–12.07) (Figure 3). The Galbraith plot also showed high heterogeneity (Figure 4). Since isolates came from large samples, random fluctuations were deemed insignificant. Sensitivity analysis (figure not shown) revealed that excluding studies with bias risk due to incomplete AST reporting didn’t affect heterogeneity or pooled resistance prevalence, supporting the decision to include all studies. Forest plot illustrating the proportion of azithromycin-resistant Salmonella serovars. AZM – Azithromycin. Galbraith plot of the proportions of azithromycin-resistant Salmonella serovars among studies. The proportion and heterogeneity of MDR and XDR Salmonella serovars among isolates tested for azithromycin resistance In this meta-analysis, data from 30 studies were analyzed to assess the prevalence of MDR Salmonella among those that provided azithromycin test results. The review evaluated 8,178 Salmonella isolates for MDR, identifying 2,084 as MDR. Fig. 5 illustrates the number of isolates screened for MDR and those that developed MDR. There was a high level of heterogeneity among the included studies (I2 = 99.35, p < 0.001). The overall pooled proportion of MDR Salmonella was 31.10% (95% CI: 18.82–44.86). Forest plots illustrating the proportion of MDR among Salmonella isolates. MDR; Multidrug resistant. Additionally, this meta-analysis investigated the presence of XDR Salmonella isolates among those screened for azithromycin resistance. It included 11 studies reporting the prevalence of azithromycin resistance, with 2,437 Salmonella isolates screened for XDR, of which 268 were identified as XDR. Details of the studies reporting the XDR results are presented in Fig. 6. High heterogeneity was also observed among the included studies (I2 = 98.50, p < 0.001). The pooled proportion of XDR Salmonella serovars tested for azithromycin resistance was 19.30% (95% CI: 6.90–35.75). Forest plots illustrating the prevalence of XDR among Salmonella isolates. XDR: Extensively Drug Resistant. Molecular epidemiology of azithromycin resistance in Salmonella serovars This systematic review and meta-analysis evaluated 57 eligible articles reporting azithromycin resistance in Salmonella, of which 19 provided detailed genotypic data on the molecular determinants underpinning resistance across diverse serovars. The investigated genetic loci encompassed macrolide efflux systems, enzymatic inactivators, ribosomal methyltransferases, integron-associated elements, and ribosomal target-site mutations, with prevalence data stratified by serovar and source. Among efflux-associated determinants (Table 6), the mefA gene was identified in 9 of 565 screened isolates: a single S. Typhi isolate (out of 323 tested) and 8 of 157 unspecified serovars, the latter predominantly recovered from food and meat samples, suggesting a potential foodborne reservoir for this resistance determinant. In contrast, mefB yielded only a single positive isolate, which belonged to the non-typhoidal group, with no carriage observed among typhoidal strains. The msrA gene demonstrated a broader distribution, being present in 15 S. Typhimurium and 3 S. Typhi isolates. Mutations within the acrB gene—specifically R717Q and R717L substitutions—emerged as the most prevalent chromosomal alterations, with R717Q detected in 52 of 79 S. Typhi and 5 of 9 S. Paratyphi isolates, while R717L was observed in 4 of 49 S. Typhi and 1 of 10 S. Paratyphi isolates. In contrast, alterations in the acrR repressor and rplV ribosomal protein-coding genes were exceptionally rare, identified in only one and two S. Typhi isolates, respectively. Table 6Proportion of efflux pump coding genes and mutations among Salmonella serovars tested for azithromycin resistanceAuthor, Year [Reference]CountrySourceSerovarsTotal IsolatesmefAmefBmsrAR717Q (acrB)R717L (acrB)rplVacrRDong et al., 2020 [80]ChinaHumanS.Typhimurium150-0----Nelson et al., 2020 [77]NepalFood/MeatUndifferentiated Salmonella spp.1578------Parry et al., 2015 [27]VietnamHumanS. Typhi400------Parry et al., 2015 [27]VietnamHumanS. Paratyphi390------Sharma et al., 2019 [70]IndiaHumanS. Typhi2460-0--21Wang et al., 2017 [35]ChinaHumanS. Typhimurium310-15----Yasin et al., 2021 [83]ChinaHumanS. Typhi371-3----Khan et al., 2020 [63]IndiaHumanS. Typhi10-0-----Khan et al., 2020 [63]IndiaHumanS. Paratyphi8-0-----Nair et al., 2016 [36]UKHumanUndifferentiated Salmonella spp.683-1-----Carey et al., 2021 [24]IndiaHumanS. Typhi16---70--Hooda et al., 2019 [21]BangladeshHumanS. Typhi12---120--Katiyar et al., 2020 [62]IndiaHumanS. Typhi2---10--Octavia et al., 2021 [54]SingaporeHumanS. Typhi15---21--Sajib et al., 2021 [37]BangladeshHumanS. Typhi34---303--Sajib et al., 2021 [37]BangladeshHumanS. Paratyphi9---50--Hooda et al., 2019 [21]BangladeshHumanS. Paratyphi1---01--Note: (–) denotes not tested or data not reported. Proportion of efflux pump coding genes and mutations among Salmonella serovars tested for azithromycin resistance Note: (–) denotes not tested or data not reported. Regarding enzymatic inactivation (Table 7), macrolide phosphotransferases exhibited a serovar-dependent distribution. The mphA gene was the most frequently detected, present in 20 of 46 S. Typhimurium, 6 of 683 undifferentiated serovars, and a single S. Paratyphi isolate, while being completely absent among 348 S. Typhi isolates, suggesting serovar-specific exclusion warranting further investigation. The mphB gene was detected in only 2 of 683 unspecified isolates and was not identified in any S. Typhi isolates, while mphD yielded uniformly negative results across all tested S. Typhi strains. Erythromycin esterases, encoded by ereA and ereB, were similarly conspicuous by their absence; screening of typhoidal and S. Typhimurium populations returned no positive results, indicating that esterase-mediated degradation does not contribute to resistance in these clinically relevant serovars. Table 7Proportion of macrolide inactivation enzymes coding genes among Salmonella serovars tested for azithromycin resistanceAuthor, ReferenceYearCountrySource of sampleSerovarsTotal isolate subjectedmphA genemphB genemphD geneereA geneereB geneDong et al.[80]2020ChinaHumanS. Typhimurium1550-0-Khan, et al.[63]2020IndiaHumanS. Typhi1000-Khan, et al.[63]2020IndiaHumanS. Paratyphi810-Nair, et al.[36]2016UKHumanSalmonella Spp68362-Octavia, et al.[54]2021SingaporeHumanS. Typhi1500000Parry, et al.[27]2015VietnamHumanS. Typhi400--00Parry, et al.[27]2015VietnamHumanS. Parayphi390--00Sharma, et al.[70]2019IndiaHumanS. Typhi2460--0-Wang, et al[35]2017ChinaHumanS. Typhimurium31150-0-Yasin, et al.[83]2021ChinaHumanS. Typhi3700-0-Note: (–) denotes not tested or data not reported. Proportion of macrolide inactivation enzymes coding genes among Salmonella serovars tested for azithromycin resistance Note: (–) denotes not tested or data not reported. Ribosomal RNA methyltransferases encoded by erm family genes constitute another well-characterized macrolide resistance strategy, but their role in Salmonella remains circumscribed (Table 8). In this analysis, ermA, ermB, ermC, and ermV were universally absent across all tested typhoidal and S. Typhimurium populations. However, erm42, which confers high-level macrolide resistance via 23S rRNA methylation, was identified in 6 of 12 S. Albany isolates, underscoring a serovar-restricted yet clinically significant mechanism that merits heightened surveillance in this lineage. Beyond serovar-specific determinants, class 1 integrons were widely distributed across multiple serovars, with carriage documented in 6 of 7 S. Enteritidis, 5 of 6 S. Kentucky, 3 of 5 S. Typhimurium, and variably in S. Takoradi, S. Tamale, S. Labadi, S. Infantis, and S. Larochelle (Table 9). This pervasive integron association is particularly concerning given its co-localization with multidrug resistance cassettes, which may facilitate co-selection and persistence of azithromycin resistance even in the absence of direct macrolide selective pressure. Collectively, these findings reveal that azithromycin resistance in Salmonella is driven not by a single dominant mechanism but by a heterogeneous arsenal of genetic determinants, with mphA-mediated phosphorylation and acrB chromosomal mutations representing the principal contributors. Table 8Proportion of RNA methylase coding genes among Salmonella serovars tested for azithromycin resistanceAuthor, ReferenceYearCountrySource of sampleSerovarsTotal isolate subjectedermA geneermB geneermC geneermV geneerm42 geneDong, et al.[80]2020ChinaHumanS. Typhimurium150----Octavia, et al.[54]2021SingaporeHumanS. Typhi15000--Parry, et al.[27]2015VietnamHumanS. Typhi4000-00Parry, et al.[27]2015VietnamHumanS. Parayphi3900-00Sharma, et al.[70]2019IndiaHumanS. Typhi2460-0--Wang, et al[35]2017ChinaHumanS. Typhimurium310----Yasin, et al.[35]2021ChinaHumanS. Typhi370----Hong, et al. [123]2021TaiwanHumanS. Albany12----6Note: (–) denotes not tested or data not reported. Proportion of RNA methylase coding genes among Salmonella serovars tested for azithromycin resistance Note: (–) denotes not tested or data not reported. Table 9Proportion of class 1 integron-associated resistance cassettes genes among Salmonella serovars tested for azithromycin resistance.GeneSerovarsTotal isolate subjectednumber of isolates having the geneSource of the sampleCountryYearAuthor, ReferenceClass 1 integronS.Kentucky65ChickenEgypt2018Elkenany, et al.[113]S.Takoradi21ChickenEgypt2018Elkenany, et al.[113]S.Typhimurium53ChickenEgypt2018Elkenany, et al.[113]S.Tamale21ChickenEgypt2018Elkenany, et al.[113]S.Labadi22ChickenEgypt2018Elkenany, et al.[113]S.Enteritidis76ChickenEgypt2018Elkenany, et al.[113]S.Infantis11ChickenEgypt2018Elkenany, et al.[113]S.Larochelle11ChickenEgypt2018Elkenany, et al.[113] Proportion of class 1 integron-associated resistance cassettes genes among Salmonella serovars tested for azithromycin resistance. GRADE assessment results for certainty of evidence The methodological qualities and risks of bias in the included studies are documented in (Supplement 1). The evidence summarized in Table 10 was assessed using the GRADE framework, showing low certainty about the proportions of azithromycin-resistant and MDR Salmonella serovars. This low certainty is due to high inconsistency (I2 = 96.28% for azithromycin resistance and I2 = 99.68% for MDR, indicating high heterogeneity) and notable publication bias, evidenced by gaps on the left side of the funnel plots. Conversely, the evidence regarding XDR Salmonella serovars was rated as very low certainty because of high inconsistency (I2 = 97.86%), serious risk of bias, and significant publication bias, with no comparative or randomized studies included in the meta-analysis for XDR Salmonella. These findings suggest that Salmonella has developed resistance to azithromycin, and MDR serovars are present, but the overall certainty is low, warranting cautious interpretation. The certainty for XDR Salmonella serovars remains very low. To improve these conclusions, more high-quality, large-scale randomized trials are needed. Table 10GRADE evidence summary for the proportion of Azithromycin resistance, MDR, and XDR Salmonella serovars.a, d, iGRADE evidence profile: the proportion of azithromycin-resistant Salmonella serovarsNo studiesStudy designQuality assessmentSummary of findingsRisk of biasInconsistencyIndirectnessImprecisionPublication biasTotalPositivePrevalence (95%CI) aQuality (overall uncertainty)Importance1Randomized controlled trialsNo seriousNot assessed (single study)Serious bNo seriousNot assessed (single study)17002150.13% [0.11–0.14%]ModerateCritical2Comparative cross-sectionalSerious cSerious dNo seriousNo seriousNo serious707120.05(-0.05-0.15)Low54Cross-sectionalSerious cVery serious eNo seriousNo seriousVery serious f131079280.08(0.07-0.09)Very low57Overall1551411550.08(0.07-0.09)⊕⊕◯◯ LowGRADE evidence profile: the proportion of MDR Salmonella serovars1Randomized controlled trialsNo seriousNot assess (single study)Moderate bNo seriousNot assess (single study)17005100.30% [0.28–0.32%]ModerateCritical2Comparative cross-sectionalSerious cSerious gNo seriousNo seriousNo serious707100.05(-0.05-0.15Low27Cross-sectionalSerious cVery Serious hNo seriousNo seriousVery Serious f577115640.37(0.28-0.46Very low30Overall817820840.35(0.28-0.42⊕◯◯◯ LowGRADE evidence profile: the proportion of XDR Salmonella serovars11Cross-sectionalSerious cVery Serious iNo SeriousNo SeriousSerious j24372680.20 [0.15, 0.24]⊕◯◯◯ Very LowCriticalaThe proportion in the column may differ from the result obtained by dividing the number of positive cases by the total number of isolates because the result is derived from the forest plot, which was calculated after double arcsine transformation and considers the weight of each study when calculating the proportion.bThe control group receives a combination of ofloxacin and azithromycin, while azithromycin is administered only to the treatment group.cThere are some studies that the population is purposely selected (comparative cross-sectional or Cross-sectional study)dModerate heterogeneity: I2 = 63.52% (well above the 75% threshold for high inconsistency).eHigh heterogeneity: I2 = 96.28% (well above the 75% threshold for high inconsistency).fMissing points on the left of the funnel plot - suggests small study effects/publication biasgModerate heterogeneity: I2 = 65.43% (well above the 75% threshold for high inconsistency).hHigh heterogeneity: I2 = 99.68% (well above the 75% threshold for high inconsistency).iHigh heterogeneity: I2 = 97.86% (well above the 75% threshold for high inconsistency).jFew missing points on the left of the funnel plot - suggests small study effects/publication bias GRADE evidence summary for the proportion of Azithromycin resistance, MDR, and XDR Salmonella serovars.a, d, i The proportion in the column may differ from the result obtained by dividing the number of positive cases by the total number of isolates because the result is derived from the forest plot, which was calculated after double arcsine transformation and considers the weight of each study when calculating the proportion. The control group receives a combination of ofloxacin and azithromycin, while azithromycin is administered only to the treatment group. There are some studies that the population is purposely selected (comparative cross-sectional or Cross-sectional study) Moderate heterogeneity: I2 = 63.52% (well above the 75% threshold for high inconsistency). High heterogeneity: I2 = 96.28% (well above the 75% threshold for high inconsistency). Missing points on the left of the funnel plot - suggests small study effects/publication bias Moderate heterogeneity: I2 = 65.43% (well above the 75% threshold for high inconsistency). High heterogeneity: I2 = 99.68% (well above the 75% threshold for high inconsistency). High heterogeneity: I2 = 97.86% (well above the 75% threshold for high inconsistency). Few missing points on the left of the funnel plot - suggests small study effects/publication bias Discussion Our systematic review and meta-analysis, encompassing 57 articles and 15,514 Salmonella isolates, provides insights into the trends and molecular epidemiology of azithromycin resistance and the global rise of XDR Salmonella. We noted a significant emergence of azithromycin resistance, possibly influenced by antimicrobial use in treating salmonellosis in humans and animals [52]. A notable outbreak of XDR S. Typhi began in Hyderabad, Pakistan, in November 2016, primarily caused by the H58 strain[89], with increasing cases reported in South and Southeast Asia[52]. In 2011, XDR S. Typhi showed sensitivity to azithromycin and meropenem, and the WHO recommended ceftriaxone or azithromycin for treatment due to rising fluoroquinolone resistance[90]. The prevalence of resistance and the shrinking pipeline of new antimicrobials limit the use of azithromycin as an oral treatment option for salmonellosis [37]. Our study indicates a pooled azithromycin resistance rate of 9.12% (95% CI: 6.53-12.07) across Salmonella serovars, which is higher than the 4.3% reported in a 25-year analysis conducted in India [91]. This difference is likely due to variations in study scope and study timelines. In contrast, the resistance rate identified in this meta-analysis is lower than the 16.1% reported in a global review of non-invasive Salmonella [92]. The observed differences may be attributed to variations in serovars; specifically, the previous meta-analysis focused exclusively on non-invasive Salmonella serovar isolates, whereas our current meta-analysis examined all Salmonella serovars across various host types. This systematic review and meta-analysis emphasize the prevalence of MDR and XDR Salmonella, particularly azithromycin-resistant strains worldwide. About one-third (31.10%; 95% CI: 18.82%- 44.86%) of Salmonella isolates tested showed MDR. The prevalence of MDR Salmonella is comparable to previous reports (32.5%)[91], with significant increases noted since the late 1980s [88], [93], [94]. This increase is clearly visible in the Malawi study, which shows a rise from 6.8% between 1998 and 2010 to 97% in 2014 [95]. This trend is evident in Asia, where the impact of MDR S. Typhi has been recorded in various outbreaks, particularly in countries such as Pakistan, Vietnam, Nepal, and India [94], [96], [97], [98]. Furthermore, the rapid dissemination of the H58 S. Typhi haplotype has been linked to several MDR epidemics in sub-Saharan Africa and South and Southeast Asia [20], [99], [100]. Rising rates of MDR Salmonella have shifted treatment practices toward fluoroquinolones [93], but resistance to these drugs has since emerged [21], [88], [101], [102], prompting the World Health Organization (WHO) to recommend ceftriaxone or azithromycin [90]. Subsequently, resistance to nalidixic acid and third-generation cephalosporins has emerged[93], [103], leading to the classification of “extensively drug-resistant” (XDR) Salmonella, defined as resistant to multiple antibiotic classes, including chloramphenicol, ampicillin, cotrimoxazole, streptomycin, fluoroquinolones, and third-generation cephalosporins. In this systematic review and meta-analysis, the prevalence of XDR Salmonella was reported as 19.30% (95% CI: 6.90-35.75). The prevalence rate of XDR reported in this review is significantly higher than the previously documented rate of 2.6% [91]. This discrepancy may be attributed to the previous study concluding its search by May 10, 2019, whereas this review was conducted until December 30, 2024. Additionally, the earlier study examined all Salmonella Typhi isolates with any antibiotic test results, whereas this meta-analysis focused specifically on studies that included azithromycin test results. This review underscores the urgent need to monitor and address the rising antibiotic resistance trends in Salmonella infections. This review discusses the genetic determinants of azithromycin resistance in Salmonella, particularly in serovars of S. Typhi, S. Paratyphi, S. Typhimurium, and S. Enteritidis, highlighting their role in human and animal diseases as indicated in various literature sources[50], [91], [94], [96], [97], [98]. Azithromycin is the last oral treatment option for patients with uncomplicated XDR typhoid; however, resistance is increasing, especially in developing countries such as South Asia[37], Africa[104], and India[85], as well as in developed countries like England[105], the Netherlands[106], and Germany[107]. Resistance mechanisms include efflux pumps and modifying enzymes [108], [109], [110], [111], with studies from India 22 and Bangladesh [112] showing that genes such as mefA, mefB, mphA, mphB, ermA, and class 1 integrons contribute to azithromycin resistance. Based on this review, screening of 1,502 Salmonella isolates found that mefA[27], [70], [83], msrA[70], [83], acrR[70], rplV[70], and the R717Q/L mutation in the acrB gene[21], [24], [37], [54], [62] significantly correlate with resistance in S. Typhi. Additionally, S. Paratyphi exhibits resistance linked to the mphA[63] gene and the R717Q/L mutation[37]. In S. Typhimurium, key contributions to resistance are from mefA[35], [80], msrA[35], [80], mphA[35], [80], and class 1 integrons[113]. Macrolides inhibit protein synthesis by binding to the bacterial ribosome[114], and resistance can occur through efflux pumps[115], enzymatic drug modifications[116], ribosomal mutations[116], and methylation of rRNA[117], largely mediated by erm genes[118], [119]. All erm genes encode methyltransferases that target the N6 position of a nucleotide in 23S rRNA[120]. They differ in their ability to mono- or di-methylate this nucleotide. erm mono-methyltransferases confer low to moderate resistance to Macrolide-Lincosamide-Streptogramin B (MLSB) antibiotics (Type I phenotype) and are often found in macrolide producers with additional resistance mechanisms[121]. In contrast, erm di-methyltransferases confer a Type II phenotype, providing high resistance to MLSB antibiotics and newer ketolide drugs, commonly seen in pathogenic bacteria[119], [122]. In our study, six out of twelve S. Albany isolates carried the erm42 gene, enhancing their resistance to macrolides like azithromycin[123]. Antimicrobial resistance arises through various mechanisms, including chromosomal mutations and mobile genetic elements such as plasmids and integrons [124], [125], [126]. Class 1 integrons, which are clinically significant and commonly associated with plasmids [111], play a crucial role in the horizontal transfer of antibiotic resistance genes [111], [127], [128]. In our findings, 76.9% (20 out of 26) of Salmonella isolates tested positive for the int1 gene [113], suggesting a link to multidrug resistance [129]. Multiple Salmonella serovars, including S. Kentucky, S. Enteritidis, and S. Infantis, contained the int1 gene, according to this review [113]. We identified that 56 of 79 S. Typhi and 6 of 10 S. Paratyphi isolates carry the acrB-L mutation, conferring resistance to azithromycin. Phylogenetic analysis indicated the independent emergence of the acrB-R717 mutation in both serovars, likely driven by selection pressure from azithromycin use in endemic countries[37]. Recent studies have confirmed this mutation’s association with high azithromycin MIC in S. Typhi isolates from Bangladesh [20], [24], Pakistan[21], Singapore[54], and Nepal[23], where the drug can be purchased over the counter without a prescription [37], [130], [131]. Limitations of the review This review includes all articles published online from various data sources but does not cover unpublished data. Conclusions Emerging azithromycin resistance (9.12%) among Salmonella isolates poses challenges to current and future treatment options, as azithromycin is the only oral therapy available for salmonellosis. The rise of MDR (31.10%) and XDR (19.30%) Salmonella serovars poses significant risks, with azithromycin becoming one of the few effective treatments. Resistance rates vary by region and serovar, with notably high rates in Southern Asia and Eastern Europe, and significant rates in serovars such as S. Choleraesuis and S. Kentucky. The molecular mechanisms of azithromycin resistance are diverse, involving multiple genetic determinants, including efflux pump genes (e.g., acrB mutations), macrolide inactivation enzyme genes (e.g., mphA), and integron-associated elements, highlighting a complex resistance landscape. Over the past two decades, studies have documented a high local prevalence of resistant Salmonella, raising concerns about the potential for widespread transmission of resistance across serovars globally. Recommendations Combating azithromycin-resistant Salmonella demands a coordinated global strategy. Enhanced surveillance networks must integrate human clinical data with animal, food, and environmental monitoring to capture emerging resistance trends, all framed within a robust One Health approach to effectively disrupt multifaceted transmission pathways. Stringent antimicrobial stewardship programs in clinical settings, alongside regulatory oversight limiting agricultural antibiotic overuse, are essential to minimize selective pressure. Sustained international funding is crucial to support molecular studies and clinical trials to decipher resistance mechanisms and validate treatment protocols. Concurrently, the deployment of rapid point-of-care diagnostics must be prioritized to enable timely targeted therapy and reduce empirical misuse. Intensified awareness campaigns targeting healthcare professionals and the general public, with a strong emphasis on infection prevention and control, are critical to preserve the long-term efficacy of this vital antibiotic class. CRediT authorship contribution statement Getachew Mengistu: Writing – review & editing, Writing – original draft, Visualization, Validation, Software, Methodology, Formal analysis, Data curation, Conceptualization. Mekuanint Geta: Writing – review & editing, Validation, Methodology, Data curation. Anwar Nuru: Writing – review & editing, Visualization, Validation, Supervision, Methodology, Conceptualization. Baye Gelaw: Writing – review & editing, Visualization, Validation, Supervision, Methodology, Conceptualization. Feleke Moges: Writing – review & editing, Visualization, Validation, Supervision, Methodology, Data curation, Conceptualization. Consent for publication Not applicable Ethics approval and consent to participate Not applicable Funding The authors received no funding for this systematic review and meta-analysis. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.