<p>Publication bias using Egger’s test.</p>

The I2 statistic and Galbraith plot confirmed heterogeneity. A univariate meta-regression, sensitivity, and subgroup analyses were conducted to identify the source of heterogeneity. Egger’s test and funnel plot were used to check publication bias.</p> <p>Results</p><p>Eight studies comprising 1,518 MDR-TB patients, with 656 new and 862 previously treated cases, were included.

Among those, 52 Pre-XDR and 13 XDR-TB cases were identified. According to the Freeman-Tukey transformation, the overall pooled prevalence of Pre-XDR-TB was 3% (95% CI: 2%-5%, I2 = 43.69%) and XDR-TB was 1% (95% CI: 0%-4%, I2 = 54.07%). In the subgroup analysis, the prevalence of Pre-XDR-TB across studies from multiple regions of Ethiopia was 7% (95% CI: 3%-13%), and in Addis Ababa, it was 3% (95% CI: 2%-4%).

Single studies in the Amhara, Tigray, and Oromia regions reported Pre-XDR-TB prevalence of 6%, 5%, and 3%, respectively. The pooled prevalence of XDR-TB in Addis Ababa was 1% (95% CI: 0%-2%). A single study conducted in Oromia reported a relatively higher prevalence of XDR-TB; 10% (95% CI: 3%-26%).

Regarding the study year, the XDR-TB prevalence was 3% (95% CI: 0%-11%) in 2005–2018 and 1% (95% CI: 0%-2%) in 2019–2024. A statistically significant publication bias was observed in XDR-TB prevalence (p = 0.024).</p> <p>Conclusion</p><p>There is a significant prevalence of Pre-XDR and XDR-TB cases among MDR-TB patients in Ethiopia. It highlights the need for strengthened surveillance in all regions and expanded drug susceptibility testing of local strains for better management of MDR-TB patients.</p></div>