Global and regional prevalence of metallo-β-lactamases in carbapenem-resistant Acinetobacter baumannii: a systematic review and meta-analysis
Our latest research paper was published in the Journal of Antimicrobial Chemotherapy (British Society for Antimicrobial Chemotherapy) on 5 February 2026.
Global and regional prevalence of metallo-β-lactamases in carbapenem-resistant Acinetobacter baumannii: a systematic review and meta-analysis
by
Stamatis Karakonstantis, Renatos Nikolaos Tziolos, Maria Paneta, and Evangelos I Kritsotakis
Abstract
Background: Emergence and spread of carbapenem-resistant Acinetobacter baumannii (CRAB) producing metallo-β-lactamases (MBLs), especially New Delhi MBLs (NDMs), are concerning due to resistance to last-resort antibiotics.
Methods: PubMed and Scopus were queried for studies published between 2020 and 2024 reporting MBL prevalence in CRAB isolates. Risk of bias was assessed across the population, setting and measurement domains. Binomial-Normal mixed-effects models were applied to estimate regional and country-specific weighted MBL and NDM prevalence proportions in CRAB. Subgroup and meta-regression analyses were performed to investigate heterogeneity.
Results: Two hundred thirty-three studies reporting 58 676 CRAB isolates were analysed. The global MBL prevalence in CRAB was 5.3% [95% confidence interval CI), 3.3%–8.2%]. Regional variability explained a substantial portion of heterogeneity in meta-regression (R2 = 35%). MBL prevalence in CRAB was highest in the Eastern Mediterranean (42.1%; 95% CI, 28.7%–56.8%) and African (36.1%; 95% CI, 12.2%–69.8%) regions, moderately high in South-East Asia (17.9%; 95% CI, 9.6%–30.9%), and low (<1%) in Europe, the Americas and the Western Pacific region. MBL prevalence in CRAB was higher in studies conducted during 2020–2024 than during 2012–2019 (7.2% versus 4.2%; adjusted odds ratio 2.3, P = 0.024). NDM was the dominant MBL in CRAB, with a global prevalence of 1.7% (95% CI, 1.1%–2.7%) in 218 studies.
Conclusions: Although its global prevalence is low, MBL-producing CRAB is common in specific regions and countries, threatening the utility of new antibiotics. Sustained surveillance, rigorous infection control and antimicrobial stewardship are required to preserve the activity of last-resort antimicrobials.
Keywords: Acinetobacter baumannii, Carbapenemases, Metallo-β-lactamases, Global epidemiology, Drug resistance, Prevalence.
Access the full paper via:
Meta-analysis
Global, regional and country-specific prevalence proportions of MBL-type and NDM-type carbapenemases among CRAB isolates were estimated as weighted averages from individual study data using a mixed-effects logistic regression model.24,25 The model assumes a binomial distribution for the number of carbapenemase-producing CRAB isolates in each study and a normally distributed random effect (log odds ratio) across studies. Maximum likelihood estimation and the Laplace approximation for integration were applied.24,25 A pooled prevalence estimate was obtained as a weighted mean. Weighting is implicit: studies with lower variability contribute more to the model's likelihood function and exert greater influence on the pooled estimate. The Binomial-Normal mixed-effects model correctly manages proportions near 0% or 100% (when variance tends to zero and classic inverse variance weights are problematic) and maintains the confidence limits for the pooled proportion within the zero to one range. The resulting confidence interval contains highly probable values for the population-averaged (pooled) prevalence proportion. Whenever computational problems arose, pooled prevalence proportions were estimated using a random-effects inverse-variance model after applying the Freeman-Tukey double-arcsine transformation.26
Caterpillar forest plots were constructed to illustrate the distribution of prevalence proportions across studies, with a 95% confidence interval (CI) for each study calculated using Wilson's score method. Higgins-Thompson's I² statistic was used as a summary index of variability across the studies that cannot be attributed to sampling error. Because I² is usually high and may not be informative for prevalence data,27 the between-study variance (τ2) and its respective 95% prediction interval (PI) were also reported. The PI describes the expected range of MBL prevalence in CRAB isolates in new studies or settings.28 The geographic distributions of MBL and NDM proportions in CRAB were illustrated with choropleth maps based on the latest country boundaries published by the World Bank.
Multivariable meta-regression analysis using the Binomial-Normal mixed-effects model was employed to examine sources of heterogeneity. Adjusted odds ratios and their 95% CІs were calculated to summarize the strength and direction of associations between study-level covariates and the prevalence of MBL carbapenemase production in CRAB. A covariate-specific R2 statistic was calculated as the portion of between-study variance reduced after including that covariate in the multivariable model. Moreover, pooled estimates of MBL carbapenemase production prevalence were obtained for each covariate level using subgroup analysis. A priori selected covariates were examined, including the year when data collection ended (examined both as a dichotomous indicator before and after 2020, and as a continuous variable using a restricted cubic spline29), geographical region [according to the World Health Organization (WHO) classification], multicounty setting (yes/no), multicentre setting (yes/no) and risk of bias (classified as either low or moderate-to-high).
To assess publication bias, we examined asymmetry in Doi plots of normal-quantiles (z-scores) against logit-transformed MBL and NDM prevalence proportions in CRAB.30 In addition, we calculated the LFK index (values close to zero indicate symmetry).30
STATA (Version 19; StatCorp, College Station, TX, USA) was used for all analyses. A value of P < 0.05 was considered statistically significant.
Figure 4. Global distribution of the country-specific estimates of the prevalence of MBL- and NDM-producing carbapenem-resistant A. baumannii clinical isolates.
Figure 3. Time trend in the annual global prevalence of MBL- and NDM-producing CRAB isolates. The bubbles represent data points from individual studies; their sizes are proportional to sample sizes. A restricted cubic spline was used to model the independent effect of the year of study conduct on the proportions of MBL CRABs (solid blue line) and NDM CRABs in particular (solid red line) in a multivariable Binomial-Normal mixed-effects model, adjusting for WHO region, multicounty setting, multicentre setting and risk of bias classification. The bluish-grey and violet shaded areas represent the corresponding pointwise 95% confidence intervals. MBL, metallo-β-lactamase; NDM, New Delhi MBL; CRAB, carbapenem-resistant A. baumannii.
