Global prevalence of macrolide-resistant Staphylococcus spp.: a comprehensive systematic review and meta-analysis

Abstract

Background

Staphylococcus is a genus of bacteria responsible for various infections ranging from mild skin to severe systemic diseases. Methicillin-resistant Staphylococcus aureus (MRSA) and coagulase-negative staphylococci (CoNS) are significant challenges owing to their resistance to multiple antibiotics, including macrolides, such as erythromycin, clarithromycin, and azithromycin.

Objective

This study aimed to systematically review and synthesize data on the prevalence of macrolide resistance in Staphylococcus spp., identify trends and changes in resistance patterns over time, and assess how testing methods and guidelines affect reported resistance rates.

Methods

The study conducted a systematic search of the Scopus, PubMed, Web of Science, and EMBASE databases. Studies have reported the proportion of macrolide-resistant Staphylococcus spp. Two authors independently extracted and analyzed the data using a random-effects model. Heterogeneity was assessed, and subgroup analyses were performed based on country, continent, species, AST guidelines, methods, and period.

Results

In total, 223 studies from 76 countries were included. The pooled prevalence of resistance to erythromycin, clarithromycin, and azithromycin were 57.3, 52.6, and 57.9%, respectively. Significant heterogeneity was observed across studies (I² > 95%, p < 0.001). Oceania (72%) had the highest erythromycin resistance, whereas Europe had the lowest (40.7%). Subgroup analyses revealed variations in resistance based on the species, with higher resistance in MRSA than in MSSA and CoNS than in other species. Over time, a slight decrease in erythromycin resistance has been observed (59.6% from 2015–2019 to 55% from 2020–2023).

Conclusion

This study emphasizes the high prevalence of macrolide resistance in Staphylococcus spp. and its notable regional variation. These findings highlight the necessity for standardized methodologies and global surveillance to manage macrolide resistance effectively. Controlling antibiotic resistance should prioritize enhancing public health measures and updating treatment guidelines.

Systematic review registration

https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=557756, CRD42024557756.

1. Introduction

Staphylococcus is a genus of bacteria that can cause many infections, from mild skin infections to serious systemic diseases. These infections can affect the skin, lungs, bloodstream, and medical devices and have become a significant treatment challenge, particularly for methicillin-resistant Staphylococcus aureus (MRSA) (Tong et al., 2015; Cheung et al., 2021). It is estimated that approximately 30% of people carry S. aureus on their bodies without any symptoms. In 2019, S. aureus was associated with more than 1 million deaths, with an estimated range of 816,000 to 1,470,000 deaths (Ikuta et al., 2022). In the United States, the rate of invasive MRSA infections in the black population (66.5 cases per 100,000 person-years) is more than twice that of the white population (27.7 cases per 100,000 person-years). In Australia, the incidence of Staphylococcus aureus bacteremia (SAB) is 5.8 to 20 times higher among Indigenous Australians than among non-Indigenous Australians. Similarly, in New Zealand, Māori and Pacific Island communities have significantly higher rates of SAB than those of European descent (Tong et al., 2015). In recent years, there has been a significant increase in the rate of MRSA colonization in healthy individuals, potentially contributing to the spread of MRSA in both community and hospital settings (Barcudi et al., 2020). In addition, MRSA is a pathogen resistant to multiple antibiotics, complicating infection management and leading to increased healthcare costs and adverse outcomes (Abebe and Birhanu, 2023; Lan et al., 2024; Saleem et al., 2025). Globally, the pathogen-drug combination with the most significant increase in attributable burden was MRSA. Its attributable deaths have doubled from 57,200 (range 34,100-80,300) in 1990 to 130,000 (range 113,000-146,000) in 2021(Naghavi et al., 2024).

Antibiotic resistance is a global health crisis that threatens the effectiveness of treatments for bacterial infections. Misuse and overuse of antibiotics have accelerated the development of resistance, rendering many therapies ineffective (Yadav and Kapley, 2021; Estany-Gestal et al., 2024). Macrolides, such as erythromycin, clarithromycin, and azithromycin, are widely used to treat various staphylococcal infections. However, the increasing emergence of macrolide resistance in Staphylococcus spp. has become a critical challenge in treating infections caused by these bacteria. Resistance to macrolides has been attributed to the methylation of specific targets in the 23S rRNA by methylases encoded by erm genes, particularly erm(C) and erm(A), which can be constitutive or inducible. In addition, efflux pumps, such as ABC-F proteins encoded by msr genes and major facilitator superfamily transporters encoded by mef genes, drug inactivation by phosphotransferases encoded by mph genes, and esterase encoded by ere genes, confer macrolide resistance (Leclercq, 2002; Miklasinska-Majdanik, 2021; El Mammery et al., 2023; Mahfouz et al., 2023). These mechanisms show regional variation, reflecting differences in the prevalence of resistance genes and differences in antibiotic use practices (Miklasinska-Majdanik, 2021).

Overall, antibiotic resistance reduces the effectiveness of these antibiotics and complicates the treatment of common staphylococcal infections such as skin infections, pneumonia, and bacteremia.

The global burden of macrolide-resistant staphylococci affects both public health and healthcare systems. Data indicate increasing infection rates and resistance patterns, particularly in healthcare-associated infections where S. aureus is a leading cause of morbidity and mortality (An et al., 2024). The economic impact is also profound, with resistant infections leading to longer hospital stays, more complex treatment regimens, and increased healthcare costs (Lodise and McKinnon, 2007). However, the limited number of effective treatment options for resistant infections increases the risk of adverse outcomes. This underscores the importance of developing novel therapeutic approaches and implementing stringent infection control measures (Guo et al., 2020).

Previous research on macrolide resistance in staphylococci has been limited by study design and reporting inconsistencies, making it difficult to draw robust conclusions and identify consistent trends. In addition, many studies require extensive regional analyses, limiting the generalizability of findings and their impact on global health. Furthermore, gaps in understanding the temporal trends and dynamics of resistance highlight the need for longitudinal studies and broader surveillance efforts (Leclercq, 2002; Khader et al., 2019). Hence, standardized methodologies and collaborative efforts across regions are essential to improving our understanding and managing macrolide resistance in staphylococci.

The primary objective of this study was to systematically review and analyze the available data on the prevalence of macrolide resistance in Staphylococcus spp.

The secondary objectives were to identify trends and changes in resistance patterns over time, explore heterogeneity in resistance rates across regions and populations, and assess the impact of testing methods and guidelines on reported resistance rates. By addressing these objectives, this study aimed to fill the existing knowledge gaps and provide comprehensive insights into the dynamics of macrolide resistance in Staphylococcus spp. to guide future research and clinical practice.

2. Methods

This study was conducted according to PRISMA guidelines and included a meta-analysis to increase the robustness of the results. The study was registered in the PROSPERO registry under the code CRD42024557756.

2.1. Eligibility criteria

The inclusion criteria for this meta-analysis stipulated that studies must investigate Staphylococcus spp. macrolide resistance, report resistance rates, specify sample size determination and have complete English-language articles available. Only cross-sectional studies providing antimicrobial resistance (AMR) data, mainly those reporting baseline resistance levels before any interventions, were included. Such studies offer a population-based overview of resistance rates at a specific time and are, therefore, suitable for estimating the prevalence of macrolide resistance. Studies were excluded if published in languages other than English and were review articles, case reports, and case series studies.

2.2. Information sources

A comprehensive search was conducted in several major online databases, including Scopus, PubMed, Web of Science, and EMBASE, focusing on studies published through December 2023. These databases were selected for their extensive coverage of biomedical literature, ensuring a broad scope for the systematic review.

2.3. Search strategy

The search syntax was tailored to each database according to their respective guidelines (“Staphylococcus*” OR “S. aureus” OR “S. epidermidis” OR “S. saprophyticus” OR “S. lugdunensis” OR “S. hominis” OR “S. capitis” OR “S. haemolyticus” OR “CoNS” OR “MRCoNS” OR “MRSA” OR “MSCoNS” OR “VISA” OR “VSSA”) AND (macrolide* OR azithromycin OR clarithromycin OR erythromycin OR roxithromycin OR telithromycin OR spiramycin OR fidaxomicin) AND (resistant* OR susceptible*). This rigorous methodological approach ensured comprehensive coverage of relevant research topics.

2.4. Selection process

The systematic online database search results were imported into EndNote (version 20), removing duplicate entries. Two authors (NG and EP) independently screened and analyzed the relevant publications to minimize bias. Disagreements were resolved by a third author (TN).

2.5. Data collection process

Data extracted included first author(s), publication year, country, diagnostic method, sample source, number of positive tests, and total sample size. To ensure accuracy, two authors (MM and MB) extracted the data independently, and any disagreements were resolved by consensus.

2.6. Study risk of bias assessment

The quality of the included studies was assessed using the JBI tool. Two authors (MB and TN) independently evaluated the quality, and a third author (MSH) resolved disagreements.

2.7. Synthesis methods

This comprehensive systematic review and meta-analysis aimed to determine the global prevalence of macrolide-resistant Staphylococcus species. The analysis used proportions as the primary outcome measure. The main objective was to assess the prevalence of macrolide-resistant Staphylococcus strains, while the secondary objective sought to identify sources of heterogeneity between studies. Subgroup analyses investigated potential variability in resistance rates across different demographic and methodological factors. Additionally, trends in macrolide resistance over time were examined.

A random effects model was employed to analyze the data, allowing for considering variability within and between studies. The degree of heterogeneity was estimated using the DerSimonian-Laird method for τ². Along with τ², the Q-test for heterogeneity and the I² statistic (Higgins and Thompson, 2002) were also calculated. Heterogeneity was considered present if τ² > 0, regardless of the Q-test results.

Subgroup analyses were performed across various factors to explore sources of heterogeneity, including countries, continents, antibiotic susceptibility testing (AST) guidelines, AST methods, Staphylococcus species, coagulase status, and year groups. This stratification helped identify macrolide resistance patterns and potential drivers across regions and testing protocols.

A Logit Transformation was applied to the proportions of macrolide-resistant Staphylococcus species to account for variations in the proportion data and stabilize the variance. The logit transformation—also known as the log-odds transformation—was used to ensure that the outcome variable remained within the 0 to 1 range, mainly when dealing with extreme proportions of resistance. This transformation also normalized the distribution of proportions, facilitating more accurate meta-regression modeling.

Meta-regression analysis was conducted to explore temporal trends in macrolide resistance over time. Moderator variables included country, continent, AST guidelines, and year group. This analysis aimed to identify how macrolide resistance in Staphylococcus species has evolved across different geographical regions and under varying testing conditions.

Outliers and influential studies were identified using studentized residuals and Cook’s distances. Studies with studentized residuals exceeding the 100 × (1–0.05 / (2 × k)) th percentile of a standard normal distribution were flagged as potential outliers (after applying a Bonferroni correction for α = 0.05 and for k studies in the meta-analysis). Studies with Cook’s distances greater than the median plus six times the interquartile range of Cook’s distances were considered influential and examined for their impact on the overall estimates.

Funnel plot asymmetry was assessed using rank correlation and regression tests, with the standard error of the observed results serving as the predictor. This approach was used to evaluate potential publication bias. All statistical analyses were performed using R (version 4.2.1) and the metafor package (version 3.8.1) (Cochran, 1954; Begg and Mazumdar, 1994; Higgins and Thompson, 2002; Sterne and Egger, 2005; Viechtbauer, 2010; Viechtbauer and Cheung, 2010; Kuhn et al., 2015).

3. Results

3.1. Descriptive statistics

A total of 21,273 records as results of the systematic search were collected in reference manager software (EndNote version 20), and 14,285 duplicated articles were removed. Thousand eighty-eight articles were assessed in the title abstract for this section; 990 full-text articles were evaluated and excluded. Eventually, this systematic review and meta-analysis included 207 eligible studies. The reports came from 76 countries and six continents. The reports cover the years 2015 to 2023. The screening and selection of presages are summarized in the PRISMA flowchart (Figure 1). Characteristics and references of included studies are presented in Table 1.

Figure 1.

PRISMA flow diagram of study selection: this diagram illustrates the process of study identification, screening, eligibility assessment, and inclusion for the review. From a total of 21,273 records identified through databases, 207 studies were included in the final review after exclusion based on criteria such as duplication, irrelevance, and lack of data on antibiotic-resistant isolates.

Table 1.

A summary of the included studies in the meta-analysis is provided below, highlighting the characteristics employed.

Author	Countries	AST method	AST guideline	Quality group	Species	Erythromycin	Clarithromycin	Azithromycin
Asbell et al. (2015)	United States	MM	C	L	MRSA	ND	ND	283
Abbasi et al. (2017)	Iran	DD	C	L	MRSA	30	ND	ND
Changchien et al. (2016)	China	DD	C	L	MRSA	159	ND	ND
Qin et al. (2017)	China	MIC	C	L	MRSA	109	ND	ND
Baek et al. (2016)	South Korea	AM	C	L	MRSA	338	ND	ND
Noordin et al. (2016)	Malaysia	DD	C	L	MRSA	297	ND	ND
Gitau et al. (2018)	Kenya	DD	C	L	MRSA	129	ND	ND
Coombs et al. (2020)	Australia	AM	C	L	MRSA	174	ND	ND
Shashindran et al. (2016)	ND	DD	C	L	MRSA	84	ND	ND
Horvath et al. (2020)	Hungary	MM	E	L	MRSA	122	ND	ND
Numanovic et al. (2021)	ND	DD	E	L	MRSA	9	ND	ND
Nichol et al. (2019)	Canada	MM	C	L	MRSA	ND	305	ND
Chaleshtori and Kachoie (2016)	ND	DD	C	L	MRSA	ND	ND	10
Chen Y. L. et al. (2021)	Taiwan	DD	C	L	MRSA	16	ND	ND
Khemiri et al. (2017)	Libya	DD	E	L	MRSA	30	ND	ND
Li et al. (2016)	China	MIC	C	L	MRSA	553	ND	ND
Napp et al. (2016)	United States	ND	ND	L	MRSA	ND	ND	37
Akbariyeh et al. (2017)	ND	DD	C	S	MRSA	2	ND	ND
Elzorkany et al. (2019)	India	DD	C	L	MRSA	159	ND	ND
Dormanesh et al. (2015)	Iran	DD	C	L	MRSA	32	ND	ND
Larsen et al. (2015)	Denmark	DD	E	L	MRSA	56	ND	ND
Valle et al. (2016)	Philippines	AM	C	L	MRSA	3	ND	ND
Guo et al. (2021)	China	DD	C	L	MRSA	65	ND	ND
Tekeli et al. (2016)	Turkey	AM	C	L	MRSA	131	ND	ND
Xie et al. (2016)	China	DD	C	L	MRSA	58	ND	ND
Nasirian et al. (2018)	Iran	DD	C	L	MRSA	88	ND	ND
Chauhan et al. (2021)	India	DD	C	H	MRSA	15	ND	ND
Livermore et al. (2015)	ND	MM	E	L	MRSA	123	ND	ND
Modukuru et al. (2021)	India	DD	C	H	MRSA	174	ND	ND
Ukpai et al. (2021)	Nigeria	DD	MG	L	MRSA	122	ND	ND
Pushkar et al. (2022)	India	DD	C	L	MRSA	31	ND	ND
Islam and Shamsuzzaman (2015)	Bangladesh	DD	C	L	MRSA	ND	ND	11
Preeja et al. (2021)	India	DD	C	L	MRSA	54	ND	ND
Yao et al. (2023)	China	AM	C	L	MRSA	173	ND	ND
Conceicao et al. (2021)	Portugal	DD	E	L	MRSA	92	ND	ND
Raut et al. (2017)	Nepal	DD	C	L	MRSA	40	ND	ND
Pradhan et al. (2021)	Nepal	DD	C	L	MRSA	964	ND	ND
El-Baghdady et al. (2020)	Egypt	DD	C	L	MRSA	94	ND	ND
Liang et al. (2018)	China	AM	C	L	MRSA	51	ND	ND
Fateh Amirkhiz et al. (2015)	Iran	DD	C	L	MRSA	ND	ND	30
Chen P. Y. et al. (2021)	Taiwan	MM	C	L	MRSA	233	ND	ND
Taherirad et al. (2016)	Iran	DD	C	L	MRSA	36	ND	ND
Bhattacharya et al. (2016)	India	DD	C	L	MRSA	ND	180	ND
Ukpai et al. (2021)	Nigeria	DD	C	L	MRSA	122	ND	ND
Leibler et al. (2017)	United States	AM	C	L	MRSA	13	ND	ND
Lee et al. (2020)	Taiwan	MIC	C	L	MRSA	889	ND	ND
Kong et al. (2018)	China	DD	C	L	MRSA	5	ND	ND
Petrović et al. (2016)	Serbia	DD	C	L	MRSA	27	ND	ND
de Benito et al. (2018)	Spain	DD	C	L	MRSA	45	ND	ND
Goudarzi et al. (2018)	Iran	DD	C	L	MRSA	50	ND	ND
Ouidri (2018)	Algeria	DD	C	L	MRSA	9	ND	ND
Esmaeili Benvidi et al. (2017)	Iran	DD	C	L	MRSA	59	ND	ND
Yitayeh et al. (2021)	Ethiopia	DD	C	L	S. Saprophiticus	12	ND	ND
Asbell et al. (2015)	United States	MM	C	L	MRCONS	ND	ND	120
Sheeba et al. (2021)	India	ND	C	L	CONS	182	ND	ND
Almasri et al. (2016)	Palestinian Territories	MIC	C	L	Staphylococcus Spp	131	ND	ND
Maleki et al. (2019)	Iran	DD	C	L	S. aureus	18	ND	ND
Peng et al. (2021)	China	AM	C	L	S. haemolyticus	35	ND	ND
Al-Naqshbandi et al. (2019)	Iraq	AM	ND	L	S. haemolyticus	30	ND	ND
Pfaller et al. (2020)	ND	MM	E	L	S. haemolyticus	159	ND	ND
Bensaci and Sahm (2017)	United States	MM	E	L	S. haemolyticus	406	ND	ND
Khan et al. (2017)	Qatar	AM	C	L	S. haemolyticus	19	ND	ND
Khan et al. (2017)	India	DD	C	L	S. haemolyticus	4	ND	ND
Murugesan et al. (2015)	India	DD	C	L	S. haemolyticus	9	ND	ND
Bolatchiev (2020)	Russia	DD	E	L	S. haemolyticus	19	ND	ND
Belbase et al. (2017)	Nepal	DD	C	L	S. haemolyticus	34	ND	ND
Junaidi et al. (2023)	Malaysia	DD	C	L	S. haemolyticus	53	ND	61
Zamanian et al. (2021)	Iran	DD	C	L	S. haemolyticus	1,010	ND	ND
Ackers-Johnson et al. (2021)	Uganda	DD	E	L	S. haemolyticus	14	ND	ND
Kang and Kim (2019)	South Korea	ND	ND	L	S. haemolyticus	10	ND	ND
Al-Habsi et al. (2020)	Oman	AM	C	L	S. haemolyticus	2	ND	ND
Skender et al. (2022)	India	MIC	C	L	S. haemolyticus	1	ND	ND
Solomon and Salaudeen (2021)	Nigeria	DD	C	L	S. haemolyticus	7	ND	ND
Saxena et al. (2019)	India	DD	C	L	S. haemolyticus	3	ND	ND
Xu et al. (2019)	China	DD	C	L	S. haemolyticus	2	ND	ND
Talapan et al. (2023)	Romania	MM	C	L	S. haemolyticus	835	ND	ND
Cavanagh et al. (2016)	Norway	MIC	E	L	S. haemolyticus	29	ND	ND
Guo et al. (2019)	China	DD	C	L	S. haemolyticus	184	ND	ND
Shittu et al. (2015)	Nigeria	DD	C	L	S. haemolyticus	10	ND	ND
Bishr et al. (2021)	Egypt	MM	C	L	S. haemolyticus	19	ND	18
Getaneh et al. (2021)	Ethiopia	DD	C	L	S. haemolyticus	30	ND	ND
Mutonga et al. (2019)	Kenya	AM	ND	L	S. haemolyticus	3	ND	ND
Kumar et al. (2018)	India	ND	C	L	S. haemolyticus	29	ND	ND
Belete (2020)	Ethiopia	DD	C	L	S. haemolyticus	7	ND	ND
Bhavana et al. (2019)	India	DD	C	L	S. haemolyticus	6	ND	ND
Peterside et al. (2015)	Nigeria	DD	C	L	S. haemolyticus	27	ND	ND
Al-Taweel (2020)	Iraq	DD	C	L	S. haemolyticus	9	ND	ND
Hasanvand et al. (2019)	Iran	DD	C	L	S. haemolyticus	32	ND	ND
Wangai et al. (2019)	Kenya	DD	C	L	S. haemolyticus	29	ND	ND
Lee et al. (2019)	ND	MM	C	L	S. haemolyticus	31	ND	ND
Sutter et al. (2016)	United States	DD	C	L	S. haemolyticus	24,213	ND	ND
Luo et al. (2020)	China	AM	C	L	S. haemolyticus	67	ND	25
Tang et al. (2020)	ND	MIC	C	L	S. haemolyticus	21	ND	ND
Suneel Kumar et al. (2021)	India	DD	C	L	S. haemolyticus	9	ND	ND
Rahimi (2016)	Iran	DD	C	L	S. haemolyticus	87	ND	ND
Mehreen et al. (2018)	Pakistan	DD	C	L	S. haemolyticus	49	ND	ND
McHardy et al. (2017)	United States	MM	C	L	S. haemolyticus	193	ND	ND
Asaad et al. (2016)	ND	AM	C	L	S. haemolyticus	23	ND	ND
Javidnia et al. (2015)	Iran	DD	C	L	S. haemolyticus	16	ND	ND
Rampelotto et al. (2022)	Brazil	MM	C	L	S. haemolyticus	167	ND	ND
Choi et al. (2019)	South Korea	AM	C	L	S. haemolyticus	5	ND	ND
Li et al. (2018)	China	MIC	C	L	S. haemolyticus	216	ND	ND
Bai et al. (2019)	China	AM	C	L	S. haemolyticus	134	ND	ND
Aguinagalde et al. (2015)	India	MIC	E	L	S. haemolyticus	ND	190	199
Diriba et al. (2020)	Ethiopia	DD	C	L	S. haemolyticus	30	9	ND
Shidiki et al. (2018)	Egypt	DD	ND	H	S. haemolyticus	100	ND	ND
Selim et al. (2022)	Saudi Arabia	DD	ND	L	S. haemolyticus	100	ND	ND
Sultan et al. (2015)	India	DD	C	L	S. haemolyticus	32	ND	ND
Manandhar et al. (2021)	Nepal	DD	C	L	S. haemolyticus	127	ND	ND
Mahfouz et al. (2023)	Egypt	DD	C	L	S. haemolyticus	52	51	52
Yang et al. (2017)	China	AM	ND	S	S. haemolyticus	12	12	12
Soroush et al. (2016)	Iran	DD	C	S	S. haemolyticus	68	ND	ND
Hailegiyorgis et al. (2018)	Ethiopia	DD	C	L	S. haemolyticus	4	ND	ND
Mesbah Elkammoshi et al. (2016)	Malaysia	DD	C	L	S. haemolyticus	179	ND	ND
Agarwal et al. (2016)	India	DD	C	L	S. haemolyticus	10	ND	ND
Mama et al. (2019)	Ethiopia	DD	C	L	S. haemolyticus	22	ND	ND
Gungor et al. (2021)	Turkey	MIC	E	L	S. haemolyticus	36	ND	ND
Ramakrishna et al. (2021)	India	DD	C	L	S. haemolyticus	171	ND	ND
Wang et al. (2017)	China	AM	C	L	S. haemolyticus	4	ND	ND
Salarvand et al. (2023)	Iran	DD	ND	L	S. haemolyticus	88	ND	ND
Firoozeh et al. (2020)	Iran	DD	C	L	S. haemolyticus	17	ND	ND
Liu et al. (2015)	China	MIC	C	L	S. haemolyticus	116	ND	ND
Fu et al. (2020)	China	AM	C	L	S. haemolyticus	189	ND	ND
Akpaka et al. (2017)	Germany	DD	C	L	S. haemolyticus	124	ND	ND
Svent-Kucina et al. (2016)	Slovenia	DD	C	L	S. haemolyticus	8	ND	ND
Goudarzi et al. (2020)	Iran	DD	C	L	S. haemolyticus	86	ND	ND
Fasihi et al. (2016)	Iran	DD	C	L	S. haemolyticus	94	ND	ND
Okuda et al. (2016)	Gabon	MIC	E	L	S. haemolyticus	8	ND	ND
Ahangarzadeh Rezaee et al. (2016)	Iran	ND	ND	L	S. haemolyticus	104	ND	ND
Biset et al. (2020)	Ethiopia	DD	C	L	S. haemolyticus	2	ND	ND
Olufunmiso et al. (2017)	Nigeria	DD	C	L	S. haemolyticus	122	ND	ND
Tahbaz et al. (2019)	Iran	DD	C	L	S. haemolyticus	19	ND	ND
Rukan et al. (2021)	Pakistan	DD	C	L	S. haemolyticus	68	ND	ND
Eibach et al. (2017)	Ghana	DD	E	S	S. haemolyticus	14	ND	ND
Dayie et al. (2021)	Ghana	DD	C	L	S. haemolyticus	5	ND	ND
Salah et al. (2021)	Yemen	AM	ND	L	S. haemolyticus	4	ND	ND
Weldu et al. (2020)	Ethiopia	DD	ND	L	S. haemolyticus	7	ND	ND
Wan et al. (2016)	Taiwan	MIC	C	L	S. haemolyticus	274	ND	ND
John et al. (2023)	Nigeria	DD	C	L	S. haemolyticus	6	6	ND
Duncan et al. (2016)	United States	ND	E	L	S. haemolyticus	548	ND	ND
Saini et al. (2021)	India	DD	C	L	S. haemolyticus	14	ND	ND
Sanchez et al. (2020)	Spain	ND	C	L	S. haemolyticus	81	ND	ND
Chen P. F. et al. (2021)	China	MIC	C	L	S. haemolyticus	27	ND	ND
Almohammady et al. (2020)	Egypt	DD	C	L	S. haemolyticus	15	ND	ND
Iliya et al. (2020)	Kenya	DD	C	L	S. haemolyticus	26	ND	ND
Abouelnour et al. (2019)	Egypt	DD	C	S	S. haemolyticus	107	ND	ND
Boncompain et al. (2023)	Argentina	DD	C	L	S. haemolyticus	7	ND	ND
Al-Tamimi et al. (2021)	Jordan	DD	C	L	S. haemolyticus	57	ND	ND
Ullah et al. (2022)	Pakistan	DD	C	S	S. haemolyticus	5	ND	ND
Khan et al. (2015)	Nepal	DD	C	L	S. haemolyticus	4	ND	ND
Shivappa et al. (2018)	Turkey	DD	C	L	S. haemolyticus	7	ND	ND
Muhammad et al. (2020)	Pakistan	DD	C	L	S. haemolyticus	ND	ND	14
Kahsay et al. (2018)	Ethiopia	DD	C	L	S. haemolyticus	6	ND	ND
Liang et al. (2018)	China	AM	C	L	S. haemolyticus	26	ND	ND
Zhang et al. (2015)	China	DD	C	L	S. haemolyticus	58	ND	ND
El-Kersh et al. (2016)	Saudi Arabia	AM	C	L	S. haemolyticus	5	ND	ND
Fateh Dizji et al. (2023)	Iran	DD	C	L	S. haemolyticus	45	ND	ND
Baz et al. (2021)	Egypt	DD	ND	L	S. haemolyticus	39	38	ND
Vijay and Dalela (2016)	India	DD	C	L	S. haemolyticus	14	ND	ND
AL-Salihi et al. (2023)	Iraq	DD	C	L	S. haemolyticus	6	ND	ND
Joachim et al. (2017)	Tanzania	DD	C	L	S. haemolyticus	11	ND	ND
Goes et al. (2021)	Brazil	DD	C	L	S. haemolyticus	29	ND	ND
Sapkota et al. (2019)	Nepal	DD	C	L	S. haemolyticus	16	ND	ND
Abdulmanea et al. (2023)	Saudi Arabia	AM	C	L	S. haemolyticus	30	ND	9
Adhikari et al. (2023)	Nepal	DD	C	L	S. haemolyticus	226	ND	ND
Zhou et al. (2020)	China	AM	C	L	S. haemolyticus	17	ND	ND
Kim et al. (2020)	South Korea	DD	C	L	S. haemolyticus	ND	ND	14
El-Amir et al. (2019)	Egypt	DD	C	L	S. haemolyticus	ND	ND	2
Arabestani et al. (2018)	Iran	DD	C	L	S. haemolyticus	160	ND	ND
Roden et al. (2019)	ND	ND	ND	L	S. haemolyticus	5	ND	ND
Al-Humaidan et al. (2015)	Saudi Arabia	DD	C	L	S. haemolyticus	5	ND	ND
Mansson et al. (2015)	Sweden	DD	E	L	S. haemolyticus	6	ND	ND
Garza-Gonzalez et al. (2019)	Mexico	DD	C	L	S. haemolyticus	871	ND	ND
Bhatt et al. (2016)	China	DD	C	L	S. haemolyticus	81	ND	ND
Maina et al. (2016)	Kenya	AM	C	L	S. haemolyticus	36	ND	ND
Wurster et al. (2018)	United States	ND	C	L	S. haemolyticus	107	ND	ND
Cavalcante et al. (2020)	Brazil	ND	ND	H	S. haemolyticus	9	ND	ND
Taha et al. (2019)	Sweden	DD	E	L	S. haemolyticus	506	ND	ND
Kurup and Ansari (2019)	Guyana	DD	C	L	S. haemolyticus	14	ND	ND
Kulshrestha et al. (2021)	India	DD	C	S	S. haemolyticus	25	ND	ND
Mottola et al. (2016)	Portugal	MIC	C	L	S. haemolyticus	8	ND	ND
Lenart-Boron et al. (2016)	Poland	DD	E	L	S. haemolyticus	23	ND	ND
Uyar Güleç et al. (2020)	Turkey	MIC	C	L	S. haemolyticus	45	ND	ND
Al-Qaisi and Al-Salmani (2020)	Iraq	DD	C	S	S. haemolyticus	50	ND	ND
Kpeli et al. (2016)	Ghana	ND	C	L	S. haemolyticus	15	ND	ND
Demir et al. (2020)	Turkey	DD	E	L	S. haemolyticus	30	ND	ND
Singh and Hota (2019)	India	AM	C	L	S. haemolyticus	8	ND	ND
Dilnessa and Bitew (2016)	Ethiopia	DD	C	L	S. haemolyticus	3	ND	ND
Rajkumar et al. (2017)	India	DD	C	L	S. haemolyticus	3,058	ND	ND
Hoffmann et al. (2015)	Austria	MM	E	L	S. haemolyticus	73	ND	74
Kumar and Shetty (2021)	India	DD	C	L	S. haemolyticus	43	ND	ND
Ahmad et al. (2020)	India	DD	C	L	S. haemolyticus	ND	ND	3
Juda et al. (2016)	Poland	DD	E	L	S. haemolyticus	75	ND	ND
Ibadin et al. (2017)	ND	DD	C	L	S. haemolyticus	48	ND	ND
Tsige et al. (2020)	Ethiopia	DD	ND	L	S. haemolyticus	25	ND	ND
Banawas et al. (2023)	Saudi Arabia	MM	ND	L	S. haemolyticus	27	ND	ND
Mascaro et al. (2019)	Italy	DD	E	L	S. aureus	16	ND	ND
Lennartz et al. (2019)	Germany	DD	E	L	S. aureus	42	ND	ND
Al Zebary et al. (2017)	Iraq	DD	NCCLS	L	S. aureus	10	ND	ND
Sakabe and Del Fiol Fde (2016)	Brazil	ND	ND	L	S. aureus	5	ND	ND
Lin et al. (2018)	China	DD	C	L	S. aureus	28	ND	ND
Doss et al. (2017)	Egypt	DD	C	L	S. aureus	13	7	10
Liang et al. (2023)	China	AM	C	L	MSSA	127	ND	ND
Oydanich et al. (2017)	United States	AM	C	L	S. aureus	62	ND	ND
Gajdacs et al. (2021)	Hungary	ND	E	S	Staphylococcus Spp	ND	ND	67
Mostafa et al. (2015)	Iran	DD	C	L	S. aureus	95	ND	69
Soumya et al. (2017)	India	DD	ND	S	S. epidermidis	152	ND	ND
Parastan et al. (2020)	Iran	DD	C	L	S. aureus	93	ND	ND
Farah et al. (2019)	Saudi Arabia	MIC	C	L	S. aureus	507	ND	ND
Sotoudeh Anvari et al. (2015)	Iran	DD	C	L	S. epidermidis	13	ND	ND

AST method (Multiple Method, MM), Disk Diffusion (DD), Minimum Inhibitory Concentration (MIC) method, and Automated Method (AM). Publication Bias: Risk (S), Low Risk (L), High Risk (H). AST guideline: CLSI: C, EUCAST: E.

3.2. Comprehensive overview of antibiotic resistance prevalence

Among 360 reports, the proportion of erythromycin-resistant isolates was 0.573 (95% CI: 0.556–0.590), based on 144,746 resistant isolates out of 293,411 isolates tested. The heterogeneity among reports was significant (I² = 96.09%, p = 0.001). Similarly, the proportion of clarithromycin resistance, as assessed by 30 reports involving 4,015 resistant isolates out of 8,045 tested isolates, was 0.526 (95% CI: 0.380–0.668), with significant heterogeneity between reports (I² = 98.76%, p = 0.001). In addition, the proportion of azithromycin-resistant isolates, derived from 83 reports containing 5,227 resistant isolates out of 10,553 isolates tested, was 0.579 (95% CI: 0.514–0.641), again with significant heterogeneity between reports (I² = 96.50%, p = 0.001).

3.2.1. Prevalence of erythromycin resistance

A total of 293,411 isolates from 721 studies were included in the erythromycin resistance analysis. The estimated mean proportion based on the random effects model was 0.573 (95% CI: 0.556–0.590). This result indicates that the mean proportion differed significantly from zero (z = 8.400, p < 0.001). The heterogeneity between studies was significant, as noted in the Q-test (Q(720) = 42,007.095, I² = 98.29%, p < 0.001) (Table 2). A forest plot illustrating the observed results and the random effects model estimate is shown in Figure 2. Using the trim-and-fill method, the adjusted proportion was 0.501 (95% CI: 0.483–0.518). Analysis of the studentized residuals identified several studies with values greater than 3.979, suggesting potential outliers within the model. After excluding these potential outliers, the proportion was 0.501 (95% CI: 0.483–0.518). Cook’s distance analysis also indicated that several studies were overly influential. After removing these influential studies, the proportion remained unchanged at 0.501 (95% CI: 0.483–0.518). Both the rank correlation test and the regression test suggested a potential funnel plot asymmetry (p < 0.001 and p = 0.018, respectively) (Table 3).

Table 2.

Meta-analysis statistics of worldwide antibiotic resistance in Staphylococcus spp.

Antibiotic	K (n, N)	Proportion 95%CI (LCI, HCI)	I2	P1	P2
Erythromycin	721 (144,746, 293,411)	0.573 (0.556, 0.590)	98.29%	p < 0.001	p < 0.001
Clarithromycin	30 (4,015, 8,045)	0.526 (0.380, 0.668)	98.76%	p = 0.727	p < 0.001
Azithromycin	83 (5,227, 10,553)	0.579 (0.514, 0.641)	96.50%	p = 0.017	p < 0.001

K: Number of reports, n: Number of resistant isolates, N: Number of total isolates, LCI: 95% Lower Limit Confidence Interval, HCI: 95% Higher Limit Confidence Interval, P1: p-value of difference from zero resistance rate, P2: p-value of heterogeneity between reports.

Figure 2.

Forest plot of resistance rates for macrolide antibiotics against Staphylococcus: the forest plot summarizes the resistance rates of Staphylococcus species to Azithromycin, Erythromycin, and Clarithromycin across various studies. Each dot represents an individual study’s data point, with red squares indicating pooled resistance estimates and black bars showing confidence intervals.

Table 3.

Evaluation of publication bias in meta-analysis.

Antibiotic	Egger test	Begg test	Fail and safe	Trim and Fill
Erythromycin	p < 0.001	p = 0.837	104,799	0.501 (0.483, 0.518)
Clarithromycin	p = 0.890	p = 0.432	0	0.526 (0.380, 0.668)
Azithromycin	p < 0.001	p = 0.264	473	0.519 (0.455, 0.582)

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 (NFS), 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.

3.2.2. Prevalence of clarithromycin resistance

The clarithromycin resistance analysis included Eight forty-five isolates from 30 studies. The estimated average proportion based on the random-effects model was 0.526 (95%CI, 0.380, 0.668). Therefore, the average outcome was not significantly different from zero (z = 0.349, p = 0.727). According to the Q test, the outcomes were heterogeneous (Q (29) = 2347.241, I ² = 98.76%, p < 0.001). A forest plot showing the observed outcomes and the estimate based on the random effects model is shown in Figure 2. With the fill and trim method implementation, the proportion changed to 0.526 (95%CI, 0.380, 0.668). Examination of the studentized residuals revealed that none of the studies had values greater than 3.144. 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 overly influential. Neither the rank correlation nor the regression test indicated funnel plot asymmetry (p = 0.432 and p = 0.890, respectively) (Figure 3).

Figure 3.

Funnel plots for publication bias analysis: funnel plots assessing the presence of publication bias in resistance studies for Erythromycin (left), Clarithromycin (middle), and Azithromycin (right). Symmetrical distributions indicate minimal bias, whereas asymmetries may suggest potential bias.

3.2.3. Prevalence of azithromycin resistance

The analysis of azithromycin resistance included data from 83 studies with 10,553 isolates. Using a random effects model, the estimated mean proportion was 0.579 (95% CI: 0.514, 0.641), indicating that the mean outcome differed significantly from zero (z = 2.385, p = 0.017). The heterogeneity of the outcomes was confirmed by the Q-test (Q(82) = 2342.061, I² = 96.50%, p < 0.001). After using the fill-and-trim method, the proportion was adjusted to 0.519 (95% CI: 0.455, 0.582). Analysis of the studentized residuals showed no study exceeded a value of 3.431, indicating no outliers in the model. Furthermore, Cook’s distance analysis indicated that no single study had an undue influence on the results. While the regression test revealed funnel plot asymmetry (p < 0.001), the rank correlation test did not reveal significant asymmetry (p = 0.264).

3.3. Subgroup analysis

This section provides a detailed summary of the subgroup analyses performed on antimicrobial resistance. The full dataset is available in Table 4. The analyses examined variations in resistance rates across geographic regions, antimicrobial susceptibility testing (AST) methods, time trends, and study quality.

Table 4.

Meta-analysis statistics of worldwide antibiotic resistance in staphylococcus spp. and subgroup analysis results.

Category	Subgroup	K (n, N)	Proportion 95%CI (LCI, HCI)	I2	P1	P2	P3
Erythromycin
Overall	ND	360 (144,746, 293,411)	0.573 (0.556, 0.590)	96.09%	p < 0.001	p < 0.001	NA
Countries	China	105 (11,791, 18,008)	0.731 (0.692, 0.766)	96.02%	p < 0.001	p < 0.001	p < 0.001
	Nepal	24 (3,996, 6,515)	0.620 (0.568, 0.669)	92.18%	p < 0.001	p < 0.001
	Rwanda	1 (25, 138)	0.181 (0.125, 0.254)	0.00%	p < 0.001	p > 0.999
	Iran	85 (5,568, 9,107)	0.627 (0.579, 0.671)	93.57%	p < 0.001	p < 0.001
	Kuwait	4 (4,805, 11,978)	0.421 (0.370, 0.474)	94.75%	p = 0.004	p < 0.001
	Ethiopia	35 (582, 1,493)	0.431 (0.345, 0.522)	86.72%	p = 0.137	p < 0.001
	India	79 (7,667, 14,709)	0.557 (0.514, 0.599)	94.19%	p = 0.010	p < 0.001
	Cameroon	1 (111, 201)	0.552 (0.483, 0.620)	0.00%	p = 0.139	p > 0.999
	South Korea	11 (1,155, 1786)	0.576 (0.333, 0.787)	97.89%	p = 0.550	p < 0.001
	Poland	8 (333, 873)	0.350 (0.235, 0.486)	91.39%	p = 0.031	p < 0.001
	Spain	13 (12,781, 39,342)	0.425 (0.304, 0.556)	98.27%	p = 0.262	p < 0.001
	Malaysia	8 (1,587, 2070)	0.704 (0.421, 0.886)	98.64%	p = 0.151	p < 0.001
	Kenya	8 (475, 766)	0.517 (0.323, 0.705)	94.92%	p = 0.872	p < 0.001
	United States	28 (48,494, 84,187)	0.558 (0.495, 0.618)	99.50%	p = 0.070	p < 0.001
	Australia	2 (199, 477)	0.720 (0.121, 0.979)	93.84%	p = 0.527	p < 0.001
	Hungary	1 (122, 153)	0.797 (0.726, 0.854)	0.00%	p < 0.001	p > 0.999
	Nigeria	23 (857, 1,353)	0.626 (0.523, 0.718)	89.63%	p = 0.017	p < 0.001
	Taiwan	7 (2,849, 5,223)	0.724 (0.331, 0.933)	99.71%	p = 0.257	p < 0.001
	Colombia	2 (144, 353)	0.260 (0.055, 0.681)	94.04%	p = 0.255	p < 0.001
	Libya	1 (30, 32)	0.938 (0.782, 0.984)	0.00%	p < 0.001	p > 0.999
	Switzerland	4 (127, 243)	0.536 (0.457, 0.613)	21.30%	p = 0.377	p = 0.283
	Pakistan	9 (470, 762)	0.703 (0.482, 0.857)	95.53%	p = 0.070	p < 0.001
	Eritrea	2 (14, 102)	0.159 (0.067, 0.331)	60.43%	p < 0.001	p = 0.112
	Oman	2 (12, 60)	0.194 (0.089, 0.371)	32.07%	p = 0.002	p = 0.225
	Croatia	1 (523, 542)	0.965 (0.946, 0.978)	0.00%	p < 0.001	p > 0.999
	Brazil	17 (1,013, 1740)	0.546 (0.436, 0.651)	93.59%	p = 0.412	p < 0.001
	Ghana	11 (133, 833)	0.165 (0.117, 0.226)	74.51%	p < 0.001	p < 0.001
	Denmark	2 (185, 1856)	0.182 (0.030, 0.614)	99.08%	p = 0.134	p < 0.001
	Japan	2 (216, 223)	0.966 (0.931, 0.983)	0.00%	p < 0.001	p = 0.431
	Philippines	1 (3, 108)	0.028 (0.009, 0.083)	0.00%	p < 0.001	p > 0.999
	Thailand	2 (39, 43)	0.900 (0.761, 0.962)	0.00%	p < 0.001	p = 0.431
	Palestinian Territories	5 (539, 870)	0.569 (0.452, 0.680)	90.18%	p = 0.248	p < 0.001
	Turkey	10 (1,371, 2,736)	0.626 (0.464, 0.765)	94.97%	p = 0.127	p < 0.001
	Canada	1 (521, 535)	0.974 (0.956, 0.984)	0.00%	p < 0.001	p > 0.999
	Israel	2 (274, 451)	0.358 (0.031, 0.906)	98.19%	p = 0.687	p < 0.001
	Jordan	5 (124, 170)	0.787 (0.450, 0.943)	89.75%	p = 0.089	p < 0.001
	Egypt	16 (987, 1,409)	0.788 (0.689, 0.863)	92.58%	p < 0.001	p < 0.001
	Iraq	23 (749, 1,445)	0.565 (0.470, 0.656)	88.97%	p = 0.177	p < 0.001
	Saudi Arabia	19 (1,141, 4,320)	0.610 (0.410, 0.778)	97.77%	p = 0.279	p < 0.001
	Portugal	7 (295, 590)	0.535 (0.398, 0.666)	87.32%	p = 0.622	p < 0.001
	Serbia	1 (27, 50)	0.540 (0.402, 0.672)	0.00%	p = 0.572	p > 0.999
	Algeria	2 (18, 72)	0.250 (0.164, 0.362)	0.00%	p < 0.001	p > 0.999
	South Africa	5 (208, 400)	0.530 (0.317, 0.733)	90.36%	p = 0.788	p < 0.001
	Argentina	5 (86, 181)	0.479 (0.355, 0.605)	59.85%	p = 0.742	p = 0.041
	Guyana	2 (52, 72)	0.737 (0.213, 0.967)	92.79%	p = 0.388	p < 0.001
	Mexico	7 (1,448, 4,153)	0.522 (0.408, 0.634)	95.37%	p = 0.702	p < 0.001
	France	2 (106, 227)	0.270 (0.037, 0.780)	79.93%	p = 0.389	p = 0.026
	Qatar	1 (19, 20)	0.950 (0.718, 0.993)	0.00%	p = 0.004	p > 0.999
	Russia	1 (19, 27)	0.704 (0.510, 0.844)	0.00%	p = 0.040	p > 0.999
	Vietnam	5 (313, 408)	0.763 (0.619, 0.864)	85.59%	p < 0.001	p < 0.001
	Afghanistan	1 (11, 98)	0.112 (0.063, 0.191)	0.00%	p < 0.001	p > 0.999
	Uganda	4 (182, 303)	0.621 (0.355, 0.830)	88.22%	p = 0.375	p < 0.001
	United Arab Emirates	1 (1, 3)	0.333 (0.043, 0.846)	0.00%	p = 0.571	p > 0.999
	Italy	7 (664, 1,434)	0.408 (0.301, 0.525)	93.28%	p = 0.124	p < 0.001
	Burkina Faso	1 (21, 149)	0.141 (0.094, 0.207)	0.00%	p < 0.001	p > 0.999
	Mozambique	1 (84, 236)	0.356 (0.297, 0.419)	0.00%	p < 0.001	p > 0.999
	Romania	1 (835, 1,672)	0.499 (0.475, 0.523)	0.00%	p = 0.961	p > 0.999
	Norway	2 (58, 375)	0.173 (0.068, 0.373)	92.52%	p = 0.003	p < 0.001
	Indonesia	2 (139, 211)	0.645 (0.541, 0.738)	27.49%	p = 0.007	p = 0.240
	Kazakhstan	1 (1, 5)	0.200 (0.027, 0.691)	0.00%	p = 0.215	p > 0.999
	Tanzania	3 (70, 249)	0.280 (0.133, 0.495)	87.51%	p = 0.045	p < 0.001
	United Kingdom	2 (203, 631)	0.392 (0.219, 0.596)	77.49%	p = 0.298	p = 0.035
	Tunisia	2 (21, 99)	0.215 (0.028, 0.722)	93.70%	p = 0.258	p < 0.001
	Uruguay	1 (5, 100)	0.050 (0.021, 0.115)	0.00%	p < 0.001	p > 0.999
	Germany	5 (394, 1,695)	0.283 (0.181, 0.413)	94.22%	p = 0.002	p < 0.001
	Slovenia	1 (8, 274)	0.029 (0.015, 0.057)	0.00%	p < 0.001	p > 0.999
	Gabon	1 (8, 103)	0.078 (0.039, 0.148)	0.00%	p < 0.001	p > 0.999
	Greece	2 (343, 715)	0.398 (0.217, 0.612)	88.85%	p = 0.350	p = 0.003
	Yemen	1 (4, 11)	0.364 (0.143, 0.661)	0.00%	p = 0.372	p > 0.999
	Austria	2 (146, 1,098)	0.133 (0.114, 0.154)	0.00%	p < 0.001	p > 0.999
	Gambia	1 (26, 293)	0.089 (0.061, 0.127)	0.00%	p < 0.001	p > 0.999
	Bangladesh	1 (19, 29)	0.655 (0.469, 0.803)	0.00%	p = 0.100	p > 0.999
	Niger	1 (4, 10)	0.400 (0.158, 0.703)	0.00%	p = 0.530	p > 0.999
	Bulgaria	2 (296, 870)	0.340 (0.309, 0.372)	0.00%	p < 0.001	p > 0.999
	Sweden	2 (512, 572)	0.654 (0.031, 0.991)	98.64%	p = 0.759	p < 0.001
	Myanmar (Burma)	1 (86, 153)	0.562 (0.483, 0.639)	0.00%	p = 0.126	p > 0.999
Continents	Asia	417 (44,949, 81,522)	0.638 (0.616, 0.660)	96.87%	p < 0.001	p < 0.001	p < 0.001
	Africa	119 (3,856, 8,241)	0.476 (0.423, 0.529)	93.22%	p = 0.373	p < 0.001
	ND	54 (26,002, 58,611)	0.535 (0.453, 0.616)	99.54%	p = 0.399	p < 0.001
	Europe	66 (17,977, 53,239)	0.407 (0.359, 0.456)	97.82%	p < 0.001	p < 0.001
	Americas	63 (51,763, 91,321)	0.544 (0.499, 0.588)	99.08%	p = 0.057	p < 0.001
	Oceania	2 (199, 477)	0.720 (0.121, 0.979)	93.84%	p = 0.527	p < 0.001
AST Guideline	CLSI	563 (114,948, 218,991)	0.584 (0.565, 0.604)	98.25%	p < 0.001	p < 0.001	p < 0.001
	EUCAST	67 (24,762, 66,311)	0.430 (0.382, 0.480)	98.71%	p = 0.006	p < 0.001
	Multiple Guideline	8 (778, 1,415)	0.507 (0.314, 0.697)	97.26%	p = 0.946	p < 0.001
	NCCLS	6 (398, 871)	0.353 (0.200, 0.543)	90.72%	p = 0.128	p < 0.001
	ND	74 (3,686, 5,519)	0.660 (0.601, 0.715)	92.23%	p < 0.001	p < 0.001
	BSAC	1 (65, 79)	0.823 (0.723, 0.892)	0.00%	p < 0.001	p > 0.999
	FMS	1 (4, 10)	0.400 (0.158, 0.703)	0.00%	p = 0.530	p > 0.999
	CASFM	1 (105, 215)	0.488 (0.422, 0.555)	0.00%	p = 0.733	p > 0.999
AST method	Automate	98 (9,062, 14,658)	0.660 (0.612, 0.705)	95.75%	p < 0.001	p < 0.001	p = 0.001
	Disk Diffusion	452 (72,252, 128,319)	0.557 (0.537, 0.576)	96.69%	p < 0.001	p < 0.001
	MIX	73 (34,775, 77,633)	0.566 (0.507, 0.624)	99.44%	p = 0.028	p < 0.001
	MIC	59 (25,047, 65,498)	0.568 (0.510, 0.624)	98.98%	p = 0.023	p < 0.001
Species	MRSA	212 (41,180, 58,142)	0.710 (0.679, 0.740)	97.67%	p < 0.001	p < 0.001	p < 0.001
	S. saprophyticus	2 (91, 181)	0.593 (0.320, 0.819)	74.47%	p = 0.514	p = 0.048
	Staphylococcus spp	19 (955, 1997)	0.522 (0.423, 0.619)	92.87%	p = 0.662	p < 0.001
	S. hominis	5 (125, 166)	0.751 (0.678, 0.812)	0.00%	p < 0.001	p = 0.448
	CoNS	42 (4,066, 7,352)	0.568 (0.505, 0.629)	95.30%	p = 0.034	p < 0.001
	S. lugdunensis	4 (236, 1,142)	0.313 (0.144, 0.552)	91.74%	p = 0.121	p < 0.001
	S. aureus	342 (92,286, 210,496)	0.496 (0.475, 0.516)	98.33%	p = 0.680	p < 0.001
	S. haemolyticus	8 (500, 692)	0.787 (0.544, 0.919)	94.55%	p = 0.023	p < 0.001
	S. epidermidis	41 (1953, 2,818)	0.676 (0.601, 0.744)	90.89%	p < 0.001	p < 0.001
	MSSA	37 (2,868, 9,758)	0.305 (0.221, 0.404)	98.29%	p < 0.001	p < 0.001
	MRCoNS	5 (381, 488)	0.777 (0.526, 0.916)	94.78%	p = 0.032	p < 0.001
	MSCoNS	1 (10, 69)	0.145 (0.080, 0.249)	0.00%	p < 0.001	p > 0.999
	VSSA	1 (57, 61)	0.934 (0.838, 0.975)	0.00%	p < 0.001	p > 0.999
	VISA	1 (11, 11)	0.958 (0.575, 0.997)	0.00%	p = 0.030	p > 0.999
	S. capitis	1 (27, 38)	0.711 (0.549, 0.832)	0.00%	p = 0.012	p > 0.999
Coagulase	CPS	593 (136,402, 278,468)	0.565 (0.546, 0.584)	98.49%	p < 0.001	p < 0.001	p = 0.021
	CoNS	109 (7,389, 12,946)	0.632 (0.584, 0.677)	95.26%	p < 0.001	p < 0.001
	ND	19 (955, 1997)	0.522 (0.423, 0.619)	92.87%	p = 0.662	p < 0.001
year group	2020_2023	379 (62,408, 148,526)	0.550 (0.525, 0.575)	98.32%	p < 0.001	p < 0.001	p = 0.002
	2015_2019	342 (82,338, 144,885)	0.596 (0.575, 0.616)	97.62%	p < 0.001	p < 0.001
Clarithromycin
Overall	ND	30 (4,015, 8,045)	0.526 (0.380, 0.668)	98.76%	p = 0.727	p < 0.001	NA
Countries	Canada	2 (590, 3,348)	0.179 (0.135, 0.234)	92.92%	p < 0.001	p < 0.001	p < 0.001
	Japan	4 (2,261, 2,455)	0.660 (0.249, 0.920)	97.28%	p = 0.462	p < 0.001
	Egypt	5 (105, 171)	0.590 (0.358, 0.788)	79.21%	p = 0.452	p < 0.001
	Iran	3 (28, 77)	0.388 (0.177, 0.651)	78.75%	p = 0.407	p = 0.009
	India	5 (896, 1735)	0.612 (0.438, 0.761)	97.37%	p = 0.205	p < 0.001
	Kazakhstan	1 (1, 5)	0.200 (0.027, 0.691)	0.00%	p = 0.215	p > 0.999
	Nigeria	4 (37, 56)	0.666 (0.407, 0.852)	55.45%	p = 0.205	p = 0.081
	Ethiopia	2 (17, 70)	0.244 (0.157, 0.358)	0.00%	p < 0.001	p = 0.589
	China	3 (79, 121)	0.729 (0.490, 0.883)	44.92%	p = 0.060	p = 0.163
	Pakistan	1 (1, 7)	0.143 (0.020, 0.581)	0.00%	p = 0.097	p > 0.999
Continents	Americas	2 (590, 3,348)	0.179 (0.135, 0.234)	92.92%	p < 0.001	p < 0.001	p = 0.095
	Asia	17 (3,266, 4,400)	0.580 (0.404, 0.738)	98.29%	p = 0.372	p < 0.001
	Africa	11 (159, 297)	0.529 (0.358, 0.693)	81.76%	p = 0.747	p < 0.001
AST Guideline	CLSI	24 (3,467, 7,231)	0.453 (0.291, 0.626)	98.92%	p = 0.599	p < 0.001	p = 0.115
	ND	4 (115, 137)	0.837 (0.765, 0.889)	0.00%	p < 0.001	p = 0.989
	EUCAST	2 (433, 677)	0.640 (0.601, 0.677)	8.00%	p < 0.001	p = 0.297
AST method	MIX	3 (597, 3,355)	0.192 (0.136, 0.263)	91.13%	p < 0.001	p < 0.001	p = 0.060
	Disk Diffusion	19 (707, 1,511)	0.503 (0.385, 0.620)	90.51%	p = 0.964	p < 0.001
	MIC	6 (2,696, 3,162)	0.614 (0.352, 0.824)	98.76%	p = 0.396	p < 0.001
	Automate	2 (15, 17)	0.861 (0.619, 0.959)	0.00%	p = 0.007	p = 0.927
Species	MRSA	6 (576, 2,353)	0.607 (0.269, 0.867)	98.70%	p = 0.552	p < 0.001	p = 0.582
	S. aureus	12 (2,630, 3,066)	0.632 (0.422, 0.802)	97.35%	p = 0.216	p < 0.001
	MSSA	2 (560, 2,167)	0.273 (0.154, 0.436)	98.16%	p = 0.008	p < 0.001
	S. epidermidis	3 (15, 46)	0.560 (0.113, 0.927)	82.89%	p = 0.837	p = 0.003
	CoNS	3 (199, 335)	0.320 (0.074, 0.735)	93.00%	p = 0.404	p < 0.001
	Staphylococcus spp	3 (29, 57)	0.439 (0.132, 0.802)	85.03%	p = 0.772	p = 0.001
	S. lugdunensis	1 (6, 21)	0.286 (0.134, 0.508)	0.00%	p = 0.058	p > 0.999
Coagulase	CPS	20 (3,766, 7,586)	0.581 (0.398, 0.745)	99.15%	p = 0.385	p < 0.001	p = 0.570
	CoNS	7 (220, 402)	0.392 (0.180, 0.655)	89.36%	p = 0.426	p < 0.001
	ND	3 (29, 57)	0.439 (0.132, 0.802)	85.03%	p = 0.772	p = 0.001
Year Group	2020_2023	17 (946, 3,990)	0.405 (0.281, 0.543)	96.40%	p = 0.177	p < 0.001	p = 0.032
	2015_2019	13 (3,069, 4,055)	0.674 (0.467, 0.830)	98.65%	p = 0.098	p < 0.001
Azithromycin
Overall	ND	83 (5,227, 10,553)	0.579 (0.514, 0.641)	96.50%	p = 0.017	p < 0.001	NA
Countries	United States	6 (630, 1,511)	0.452 (0.296, 0.618)	96.64%	p = 0.577	p < 0.001	p = 0.009
	Nepal	2 (94, 162)	0.554 (0.402, 0.696)	65.84%	p = 0.487	p = 0.087
	Spain	2 (170, 883)	0.348 (0.033, 0.894)	99.34%	p = 0.656	p < 0.001
	India	17 (1910, 3,360)	0.575 (0.458, 0.685)	96.84%	p = 0.207	p < 0.001
	China	8 (916, 1,137)	0.768 (0.569, 0.893)	94.57%	p = 0.011	p < 0.001
	Brazil	2 (65, 108)	0.808 (0.050, 0.997)	94.54%	p = 0.520	p < 0.001
	Egypt	6 (92, 149)	0.609 (0.417, 0.773)	57.17%	p = 0.262	p = 0.040
	Pakistan	5 (64, 75)	0.831 (0.722, 0.903)	0.00%	p < 0.001	p = 0.746
	Bangladesh	6 (88, 182)	0.500 (0.398, 0.601)	41.23%	p = 0.993	p = 0.130
	Iran	11 (469, 805)	0.563 (0.475, 0.648)	80.97%	p = 0.160	p < 0.001
	Iraq	3 (108, 150)	0.770 (0.306, 0.962)	92.89%	p = 0.243	p < 0.001
	Saudi Arabia	2 (45, 93)	0.481 (0.060, 0.930)	96.42%	p = 0.954	p < 0.001
	Malaysia	1 (61, 209)	0.292 (0.234, 0.357)	0.00%	p < 0.001	p > 0.999
	Kazakhstan	1 (1, 5)	0.200 (0.027, 0.691)	0.00%	p = 0.215	p > 0.999
	Indonesia	1 (12, 22)	0.545 (0.341, 0.735)	0.00%	p = 0.670	p > 0.999
	South Africa	1 (66, 89)	0.742 (0.641, 0.822)	0.00%	p < 0.001	p > 0.999
	Austria	2 (148, 1,098)	0.135 (0.116, 0.156)	0.00%	p < 0.001	p > 0.999
	South Korea	1 (14, 25)	0.560 (0.366, 0.737)	0.00%	p = 0.549	p > 0.999
	Australia	1 (58, 63)	0.921 (0.823, 0.967)	0.00%	p < 0.001	p > 0.999
	Hungary	2 (94, 172)	0.533 (0.209, 0.830)	95.14%	p = 0.861	p < 0.001
Continents	Americas	8 (695, 1,619)	0.493 (0.345, 0.643)	95.81%	p = 0.929	p < 0.001	p = 0.013
	Asia	58 (3,782, 6,225)	0.604 (0.540, 0.666)	94.36%	p = 0.002	p < 0.001
	Europe	6 (412, 2,153)	0.311 (0.149, 0.537)	98.29%	p = 0.098	p < 0.001
	ND	3 (122, 255)	0.466 (0.053, 0.932)	97.70%	p = 0.923	p < 0.001
	Africa	7 (158, 238)	0.641 (0.485, 0.772)	66.45%	p = 0.076	p = 0.007
	Oceania	1 (58, 63)	0.921 (0.823, 0.967)	0.00%	p < 0.001	p > 0.999
AST Guideline	CLSI	67 (4,179, 7,351)	0.590 (0.528, 0.649)	94.82%	p = 0.005	p < 0.001	p = 0.111
	ND	8 (227, 430)	0.705 (0.388, 0.900)	94.99%	p = 0.198	p < 0.001
	Multiple Guideline	1 (26, 60)	0.433 (0.315, 0.560)	0.00%	p = 0.303	p > 0.999
	EUCAST	7 (795, 2,712)	0.363 (0.163, 0.625)	99.12%	p = 0.304	p < 0.001
AST Method	MIX	9 (736, 1857)	0.668 (0.479, 0.815)	97.25%	p = 0.080	p < 0.001	p = 0.121
	Disk Diffusion	53 (2,265, 4,274)	0.553 (0.481, 0.622)	93.39%	p = 0.151	p < 0.001
	MIC	6 (1,022, 1,601)	0.663 (0.564, 0.750)	89.68%	p = 0.002	p < 0.001
	Automate	8 (875, 1,098)	0.748 (0.483, 0.904)	95.80%	p = 0.065	p < 0.001
Species	MRSA	23 (1,353, 2,733)	0.637 (0.528, 0.733)	95.45%	p = 0.014	p < 0.001	p = 0.074
	mrCoNS	1 (120, 147)	0.816 (0.745, 0.871)	0.00%	p < 0.001	p > 0.999
	S. lugdunensis	1 (21, 28)	0.750 (0.561, 0.876)	0.00%	p = 0.012	p > 0.999
	S. aureus	40 (2,907, 6,072)	0.546 (0.442, 0.645)	97.45%	p = 0.387	p < 0.001
	MSSA	3 (157, 533)	0.185 (0.067, 0.417)	94.30%	p = 0.011	p < 0.001
	S. epidermidis	5 (157, 283)	0.509 (0.338, 0.678)	81.37%	p = 0.917	p < 0.001
	CoNS	6 (407, 567)	0.767 (0.571, 0.891)	91.58%	p = 0.010	p < 0.001
	Staphylococcus spp	4 (105, 190)	0.562 (0.296, 0.797)	88.15%	p = 0.660	p < 0.001
Coagulase	CPS	66 (4,417, 9,338)	0.558 (0.485, 0.629)	96.90%	p = 0.121	p < 0.001	p = 0.312
	CoNS	13 (705, 1,025)	0.679 (0.563, 0.777)	88.70%	p = 0.003	p < 0.001
	ND	4 (105, 190)	0.562 (0.296, 0.797)	88.15%	p = 0.660	p < 0.001
Year Group	2015_2019	44 (3,537, 7,509)	0.584 (0.492, 0.671)	97.55%	p = 0.073	p < 0.001	p = 0.901
	2020_2023	39 (1,690, 3,044)	0.569 (0.483, 0.651)	93.20%	p = 0.117	p < 0.001

K: Number of reports, n: Number of resistant isolates, N: Number of total isolates, LCI: 95% Lower Limit Confidence Interval, HCI: 95% Higher Limit Confidence Interval, P1: p-value of difference from zero resistance rate, P2: p-value of heterogeneity between reports, P3: p-value of difference between groups.

3.3.1. Subgroup analysis based on countries

Subgroup analysis revealed statistically significant differences in antimicrobial resistance prevalence between countries for azithromycin, clarithromycin, and erythromycin. Austria had the lowest resistance rate for azithromycin, with a prevalence of 13.5%, while Australia had the highest resistance rate at 92.1%. Pakistan had the lowest resistance rate (14.3%) for clarithromycin, while China had the highest (72.9%). The Philippines had the lowest resistance rate of 2.8% for erythromycin, while Canada had the highest resistance rate of 97.4% (Figure 4).

Figure 4.

Global prevalence of antibiotic resistance in Staphylococcus: maps showing the worldwide prevalence of resistance to Erythromycin (A), Clarithromycin (B), and Azithromycin (C). Regions with higher resistance proportions are highlighted in warmer colors (e.g., red), while areas with lower resistance rates are shown in cooler tones (e.g., green).

3.3.2. Subgroup analysis based on continents

Subgroup analysis revealed statistically significant differences in antimicrobial resistance prevalence between continents, particularly for azithromycin and erythromycin. Europe had the lowest resistance rate for azithromycin, with a prevalence of 31.1%, while Oceania had the highest resistance rate of 92.1%. Similarly, Europe had the lowest resistance rate for erythromycin, with a prevalence of 40.7%, while Oceania had the highest resistance rate at 72% (Figure 5A).

Figure 5.

subgroup analysis results were illustrated in figures (A) compression of the prevalence of antibiotic-resistant staphylococcus isolates between continents; (B) Compression of the prevalence of antibiotic-resistant staphylococcus isolates between AST guideline; (C) Compression of the prevalence of staphylococcus isolates AST method; (D) Compression of the prevalence of antibiotic-resistant staphylococcus isolates based on species (E) Compression of the prevalence of antibiotic-resistant staphylococcus isolates based on coagulase; (F) Compression of the prevalence of staphylococcus isolates before and after 2020.

3.3.3. Subgroup analysis based on AST guideline

The subgroup analysis identified statistically significant differences in antibiotic resistance prevalence, including erythromycin, between different antimicrobial susceptibility testing (AST) guidelines. For erythromycin, the NCCLS guideline showed the lowest resistance rate with a prevalence of 35.3%, while the BSAC guideline showed the highest resistance rate at 82.3% (Figure 5B).

3.3.4. Subgroup analysis based on the AST method

Subgroup analysis revealed a statistically significant disparity in the prevalence of antibiotic resistance, including erythromycin, among the various AST methods. For erythromycin, the AST method with the lowest resistance rate was Disk Diffusion, with a prevalence of 55.7%. Conversely, the AST method, with the highest resistance rate, was automated, with a prevalence rate of 66% (Figure 5C).

3.3.5. Subgroup analysis based on species

Subgroup analysis revealed statistically significant differences in antibiotic resistance prevalence among different species, including erythromycin. For erythromycin, MSCoNS had the lowest resistance rate with a prevalence of 14.5%, while VISA had the highest resistance rate with a prevalence of 95.8% (Figure 5D).

3.3.6. Subgroup analysis based on coagulase

Subgroup analysis revealed statistically significant differences in the prevalence of antibiotic resistance, including erythromycin, among different coagulase types. For erythromycin, the coagulase type with the lowest resistance rate was ND, with a prevalence of 52.2%. In contrast, the highest resistance rate was observed for CoNS, with a prevalence of 63.2% (Figure 5E).

3.3.7. Subgroup analysis based on year-group

The subgroup analysis identified statistically significant differences in antibiotic resistance prevalence among different groups, including clarithromycin and erythromycin. For clarithromycin, the period with the lowest resistance rate was 2020–2023, with a prevalence of 40.5%, while the highest resistance rate was observed in 2015–2019, with a prevalence of 67.4%. Similarly, for erythromycin, the lowest resistance rate occurred during 2020–2023, with a prevalence of 55%, while the highest resistance rate was observed during 2015–2019, with a prevalence of 59.6% (Figure 5F).

3.4. Meta-regression

Meta-regression analysis was performed to examine the relationship between antimicrobial resistance rates and year of reporting. No statistically significant correlation was observed for erythromycin (r = −0.041, p-value = 0.007, 95% CI [−0.071, −0.011]) (Figure 6A). Similarly, the correlation was not statistically significant for clarithromycin (r = −0.123, p-value = 0.263, 95% CI [−0.339, 0.093]) (Figure 6B). These results suggest that resistance rates for azithromycin and clarithromycin remained relatively stable over the study period. In contrast, a statistically significant positive correlation was observed for azithromycin (r = 0.005, p-value = 0.929, 95% CI [−0.1, 0.11]) (Figure 6C), indicating an upward trend in erythromycin resistance rates over time.

Figure 6.

Trends in antibiotic resistance over time (2015–2023): meta-regression analysis plots the trends of resistance proportions for: (A) Erythromycin, showing a slight but statistically significant decline. (B) Clarithromycin, demonstrating a non-significant downward trend. (C) Azithromycin, with no significant trend observed. Data points represent study-specific resistance proportions over time, with bubble sizes reflecting sample size.

4. Discussion

This systematic review and meta-analysis thoroughly evaluated the prevalence and trends of macrolide resistance in Staphylococcus species, explicitly focusing on resistance to erythromycin, clarithromycin, and azithromycin. By analyzing data from 207 studies conducted in 76 countries between 2015 and 2023, our findings provide valuable insights into global patterns of macrolide resistance in Staphylococcus species. Erythromycin, the first macrolide antibiotic discovered, remains effective in treating minor skin infections caused by penicillin-resistant S. aureus strains (Washington and Wilson, 1985). This meta-analysis revealed that erythromycin was the most commonly tested macrolide in antibiotic susceptibility studies, with data from 207 studies in 76 countries. The pooled prevalence of resistance was 57.3%, with significant heterogeneity between studies (I² = 96.09%, p < 0.001). Evidence of publication bias was also detected using Egger’s test (p < 0.001), resulting in an adjusted pooled prevalence of 50.1% after Fill and Trim analysis. These variations may be due to differences in study populations, periods, sampling methods, or clinical specimen types.

Subgroup analyses revealed significant regional differences in erythromycin resistance rates. Oceania had the highest resistance rate (72%, based on two reports), while Asia contributed the most studies (417 reports) with a pooled prevalence of 63.8%. In particular, China, Iran, and India reported resistance rates of 73.1, 62.7, and 55.7%, respectively, based on 105, 85, and 79 reports. In contrast, Europe had the lowest pooled prevalence of erythromycin-resistant isolates (40.7%, 44 reports), with Spain (13 reports) and Poland (8 reports) reporting prevalence rates of 42.5 and 35%, respectively. The lower resistance rates in Europe reflect increased public awareness and effective public health interventions to curb antimicrobial resistance.

On the other hand, prevalence rates of over 90% for erythromycin-resistant isolates in countries such as Qatar, Canada, Libya, Japan, and Croatia raise significant concerns. However, because these findings are based on AST performed at a single clinical center in each country, the results cannot be generalized to the entire population in these regions. This underscores the need for comprehensive national surveillance systems to monitor antimicrobial resistance in these areas.

Subgroup analysis by species revealed a pooled prevalence of erythromycin resistance in 49.6% of S. aureus isolates (342 reports). In addition, some studies included in this meta-analysis reported erythromycin resistance rates for S. aureus in two subgroups: MSSA (methicillin-susceptible S. aureus) and MRSA. The prevalence of resistance was significantly higher in MRSA than in MSSA (71% vs. 30.5%). However, more studies have focused on MRSA than MSSA (212 vs. 37). These findings are consistent with other meta-analyses that have reported pooled prevalence rates of erythromycin-resistant S. aureus isolates (Eshetie et al., 2016; Khanal et al., 2021; Chelkeba et al., 2022; Chelkeba and Melaku, 2022; Ezeh et al., 2023; Xu et al., 2024). However, most of these studies were based on data from one African country and had fewer studies than ours. Moreover, Chelkeba et al. (2022) and Chelkeba and Melaku (2022), during two separate meta-analyses conducted in Ethiopia, reported 50 and 45% prevalence rates for erythromycin-resistant S. aureus isolates in women with bacteriuria and patients with wound infections, respectively. In a meta-analysis review, Ezeh et al. (2023) reported a prevalence rate of 47% for erythromycin-resistant S. aureus isolates in Nigeria (66 reports) up to 2022. However, data from our meta-analysis highlighted a higher prevalence of erythromycin resistance in Nigeria (23 reports) than in Ezeh et al. (2023) (62.6% vs. 47%). The observed discrepancy in prevalence rates may be due to differences in the periods and number of studies included in these two meta-analyses. Subgroup analysis by species revealed a high pooled prevalence of erythromycin resistance among CoNS isolates at 56.8% (based on 42 reports). In addition, some studies independently reported the frequency of specific CoNS species, allowing pooled prevalence rates to be calculated for each species. Among these, S. epidermidis was the most commonly studied CoNS species (41 reports), with a pooled erythromycin resistance prevalence of 67.7%. Similar to our findings, Deyno et al. (2018) also reviewed the prevalence of antimicrobial resistance among clinical isolates of CoNS in Ethiopia through 2016, reporting a 30% prevalence of erythromycin-resistant CoNS. The discrepancy between our findings and Deyno et al. (2018) may be due to differences in the periods and geographic regions covered by these two meta-analyses. Specifically, our meta-analysis included data collected between 2015 and 2023, whereas Deyno et al. (2018) focused on data up to 2016. Furthermore, our study provided a global overview of antimicrobial resistance prevalence, whereas Deyno et al. (2018) limited their analysis to Ethiopia.

In addition, five studies reported a resistance prevalence of 77.7% among [methicillin-resistant Coagulase-Negative Staphylococci (MRCoNS)], which was significantly higher than the 14.5% reported in a single survey of MSCoNS. However, due to the unequal number of studies, this comparison lacks balance, and further research is needed to make a comprehensive and accurate comparison.

Overall, the prevalence of MRCoNS was significantly lower than that of MRSA. This difference may be attributed to the lower frequency of CoNS infections than S. aureus infections, reducing antimicrobial exposure. However, CoNS have transitioned from being non-pathogenic to emerging as pathogenic strains, potentially acquiring resistance genes from S. aureus (Yu et al., 2017).

In contrast, the prevalence of erythromycin resistance decreased slightly over time, from 59.6% in 2015–2019 to 55% in 2020–2023. This decline may reflect increased national efforts to combat antimicrobial resistance and the implementation of updated treatment guidelines and surveillance systems in developed countries. Similarly, a meta-analysis by Xu et al. (2024), found no significant change in erythromycin-resistant S. aureus isolates from Cystic fibrosis patients when comparing the periods 2008–2015 and 2015–2021.

Based on AST guidelines, the subgroup analysis showed higher resistance levels in the CLSI group compared to the EUCAST group (58.4% vs. 43%). However, this finding may be influenced by more studies using CLSI guidelines (563) compared to EUCAST guidelines (67 studies). Both guidelines are widely used but differ in their breakpoints for determining resistance. For example, EUCAST defines resistance as MIC >1, whereas CLSI uses MIC ≥8. Similarly, EUCAST considers a zone diameter of <21 mm resistant, while CLSI uses a zone diameter of ≤13 mm. These differences and variations in the number of studies likely contributed to the observed differences in erythromycin resistance prevalence.

This meta-analysis found fewer studies evaluated susceptibility testing for azithromycin and clarithromycin than erythromycin. It may be due to the limited clinical use of azithromycin and clarithromycin for treating staphylococcal infections compared to erythromycin. The pooled prevalence of azithromycin resistance was similar to that of erythromycin (57.3% vs. 57.9%). However, significant heterogeneity between studies was observed (I² = 96.5%, p < 0.001), and Egger’s test indicated potential publication bias (p < 0.001). After applying fill and trim analysis, the pooled prevalence of azithromycin resistance was adjusted to 51.9%.

The highest resistance rates were reported in Oceania (92.1%, based on one report), while most studies (58 reports) were conducted in Asia, with a pooled prevalence of 60.4%. Specifically, India and Iran contributed 17 and 11 reports, respectively, with 57.5 and 56.3% resistance prevalence rates. Like erythromycin, Europe had the lowest prevalence of azithromycin resistance (31.1%, based on six studies). This low prevalence may be due to the limited number of European studies and the infrequent use of azithromycin to treat staphylococcal infections in this region. Alarmingly, high levels of azithromycin-resistant isolates were identified in Pakistan, Brazil, and China.

Subgroup analysis by species showed that S. aureus was the most commonly studied species, with a pooled resistance prevalence of 54.6% (40 reports). In addition, 23 studies reported a high prevalence of azithromycin resistance among MRSA isolates (63.7%), compared with only three studies evaluating MSSA isolates, which showed a much lower resistance prevalence of 18.5%. However, this comparison was biased due to the unequal number of studies. Subgroup analysis by the AST method showed that disc diffusion was the most commonly used method for antibiotic susceptibility testing, probably because of its accessibility and widespread acceptance. However, the highest prevalence of azithromycin resistance was associated with the automated method (74.8%, based on eight reports). Like erythromycin, the prevalence of azithromycin resistance decreased slightly over time, from 58.4% in 2015–2019 to 56.9% in 2020–2023.

Clarithromycin was the third macrolide antibiotic studied in this meta-analysis, with a pooled resistance prevalence of 52.6%; however, there was considerable heterogeneity between studies (I² = 98.76%, p < 0.001). Most of the reports (17) were from Asia, with a pooled prevalence of 58%. S. aureus was the dominant species, with a resistance prevalence of 63.2%; six studies showed a prevalence rate of 60.7% among MRSA isolates and 27.3% among MSSA isolates (two reports). In contrast to erythromycin and azithromycin, the prevalence of resistance to clarithromycin decreased significantly over different periods (67.4% from 2015 to 2019 and 40.5% from 2020 to 2023).

Clarithromycin, the third macrolide antibiotic examined in this meta-analysis, had a pooled resistance prevalence of 52.6%, although significant heterogeneity between studies was observed (I² = 98.76%, p < 0.001). Most reports (17 studies) were from Asia, with a pooled resistance prevalence of 58%. S. aureus was the predominant species, with a resistance prevalence of 63.2%. Among MRSA isolates, six studies reported a resistance prevalence of 60.7%, while MSSA isolates had a lower prevalence of 27.3% (based on two reports). In contrast to erythromycin and azithromycin, clarithromycin resistance decreased significantly over time, from 67.4% in 2015–2019 to 40.5% in 2020–2023.

This meta-analysis is the first to compare the prevalence of resistance to azithromycin and clarithromycin in Staphylococcus species. As a result, no previous meta-analyses have provided comparable global results.

A significant limitation of this study is the lack of differentiation between Staphylococcus species isolated from healthcare and community settings. This distinction is critical, as antibiotic resistance rates in healthcare settings are typically higher than in the community. Another limitation is the lack of data on resistance to newer macrolides, primarily due to the limited number of studies investigating them. This gap highlights the need for further research to provide accurate and comprehensive evidence.

5. Conclusion

This meta-analysis highlights a relatively high prevalence of macrolide resistance in S. aureus and CoNS isolates worldwide. These elevated resistance rates underscore the importance of regular epidemiologic surveillance of antimicrobial resistance and the implementation of stewardship programs. Most of the studies included in this analysis were conducted in Asia, while Europe had the lowest macrolide resistance rate. In addition, resistance to erythromycin and azithromycin remained relatively stable between 2015–2019 and 2020–2023. Nevertheless, antimicrobial susceptibility testing before treatment is recommended, and further research into the molecular and genetic mechanisms of macrolide resistance is strongly encouraged.

Funding Statement

The author(s) declare that no financial support was received for the research and/or publication of this article.

Data availability statement

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding authors.

Author contributions

TN: Data curation, Writing – original draft, Writing – review & editing. AZ: Supervision, Writing – original draft, Writing – review & editing. EP: Investigation, Project administration, Writing – original draft. MM: Investigation, Resources, Visualization, Writing – review & editing. NG: Project administration, Validation, Writing – original draft. MB: Investigation, Methodology, Project administration, Writing – original draft. MS: Formal analysis, Software, Supervision, Writing – original draft.

Conflict of interest

AZ was employed by Quality Control Department of Temad Mfg, Co.

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declare that no Generative AI was used in the creation of this manuscript.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.