Pooled estimate of anticoagulation control with warfarin therapy among adult patients in Ethiopia: a systematic review and meta-analysis Authors: Firafan Shuma Teka, Ayana Korsa, Bekan Gudata Gindaba, Gemechu Gelana Ararame, Jebesa Desisa Hordofa, Lalesa Tadesse Chuko, Nega Abebe Meshesha, Birbirsa Sefera Senbeta Published: 2026 ABSTRACT BackgroundWarfarin is widely used to prevent thromboembolic events; however, achieving optimal anticoagulation control remains challenging. In Africa, Time in Therapeutic Range (TTR) is generally low, and no nationally representative pooled estimate exists for Ethiopia.ObjectiveTo estimate adequate anticoagulation control (TTR ≥ 65%) among adult patients receiving warfarin therapy in Ethiopia.MethodsA systematic review and meta-analysis was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines and registered in PROSPERO (CRD420251271413). Databases including PubMed/MEDLINE, Scopus, Embase, Web of Science, African Journals Online, and Google Scholar were searched from inception to December 10, 2025. Observational studies conducted in Ethiopia among adult warfarin users reporting TTR were included, while studies lacking relevant outcomes or involving non-adult populations were excluded. Study quality was assessed using the Joanna Briggs Institute tool. A random-effects model was applied to estimate pooled anticoagulation control, and heterogeneity was evaluated using the I² statistic.ResultsSix studies comprising a total of 2,022 participants were included in this systematic review and meta-analysis. All studies were retrospective cross-sectional and conducted among adult patients receiving warfarin therapy in hospital settings across Ethiopia. The pooled estimate of adequate anticoagulation control (TTR ≥ 65%) was 34.30% (95% CI: 28.16–40.44), with substantial heterogeneity (I² = 79.09%). The mean TTR ranged from 29% to 47.24% across the included studies. Subgroup analysis showed relatively higher anticoagulation control in Addis Ababa (39.29%) compared to other regions (Amhara and Oromia) (22.19%).ConclusionAnticoagulation control among Ethiopian patients on warfarin is suboptimal. Strengthening anticoagulation management services and establishing dedicated clinics are recommended.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12959-026-00865-6. Introduction Oral anticoagulants are essential for the management of cardiovascular diseases and the prevention of thromboembolic events [1–3]. In low- and middle-income countries, warfarin remains the most widely used oral anticoagulant for patients with mechanical heart valves, non-valvular atrial fibrillation, and venous thromboembolism (VTE) due to its affordability and availability [4, 5]. The international normalized ratio (INR) and time in therapeutic range (TTR) are commonly used indicators to monitor the effectiveness and safety of warfarin therapy [6]. TTR represents the proportion of time a patient’s INR remains within the target therapeutic range (typically 2.0–3.0 for most indications) and is a key measure of anticoagulation quality [7]. Maintaining INR within the recommended range is essential for achieving optimal therapeutic outcomes and minimizing the risk of bleeding and thromboembolic complications [8, 9]. Despite its clinical importance, achieving optimal anticoagulation control remains a global challenge. Evidence shows that only a limited proportion of patients achieve adequate TTR, with suboptimal control associated with increased risks of adverse clinical outcomes [7, 10]. In addition, anticoagulation quality may vary over time, with poorer control often observed during early phases of therapy [11]. Effective warfarin management requires regular monitoring, patient adherence, and well-established healthcare systems [10, 12]. In Ethiopia and other sub-Saharan African countries, anticoagulation control remains inconsistent due to variations in healthcare infrastructure, monitoring practices, and patient-related factors [4]. Several Ethiopian hospital-based studies have reported poor anticoagulation control among adult patients on warfarin therapy [13–16]. However, these studies show substantial variability in findings, with reported TTR values ranging approximately from 29% to 47.24%, indicating inconsistent results across different settings. Moreover, existing studies are limited by small sample sizes, single-center designs, and methodological heterogeneity, including differences in study populations and outcome measurements. Importantly, there is currently no national-level pooled estimate of anticoagulation control among adult patients receiving warfarin therapy in Ethiopia. Therefore, a systematic review and meta-analysis is essential to synthesize existing evidence, address inconsistencies, and generate a more precise and generalizable pooled estimate of anticoagulation control. This study aims to estimate the pooled estimate of adequate anticoagulation control among adult patients receiving warfarin therapy in Ethiopia. Methodology Study protocol and reporting The review protocol was registered in the PROSPERO database (registration number: CRD420251271413) and can be accessed at: https://www.crd.york.ac.uk/PROSPERO/view/CRD420251271413. To stay objective and prevent bias, we mapped out our entire plan and methods before we even started the review. Information sources and search strategies A comprehensive search of the literature was conducted from database inception to December 10, 2025, using PubMed/MEDLINE, Scopus, Embase, Web of Science, African Journals Online, and Google Scholar. With a geographic focus on Ethiopia, the search strategy combined free-text terms and Medical Subject Headings (MeSH) related to warfarin and anticoagulation management, such as warfarin, anticoagulation control, international normalized ratio, INR, time in therapeutic range, and TTR. Boolean operators “AND” and “OR” were applied to appropriately combine the search terms. The search string was structured as follows: (“warfarin” OR “vitamin K antagonist” OR “anticoagulant”) AND (“time in therapeutic range” OR “TTR” OR “INR” OR “anticoagulation control” OR “anticoagulation quality”) AND (Ethiopia OR Ethiopian). The searches were conducted independently by two investigators (FST and BGG), who also eliminated duplicates and checked titles and abstracts for eligibility. Potentially eligible records’ full texts were obtained and independently evaluated in relation to the inclusion criteria; differences were settled through conversation or by consulting a third reviewer (BSS). Study selection and data extraction The selection procedure (number of records identified, screened, excluded, and included) was documented using a PRISMA flow diagram. First author, year of publication, study design, study setting, sample size, and proportion reaching TTR ≥ 65% (or study criterion) were retrieved from each included study. The data was extracted using a standardized data abstraction format created by Microsoft Excel. The study design, population characteristics, sample size, first author, publication year, counsel and study setting, and study findings (% of time in therapeutic range and warfarin-related side effects) are all included in this tool. Two reviewers separately extracted the data using a piloted extraction form (FST and BSS); disagreements were settled by consensus. Quality assessment / risk of bias An adaptation of the Joanna Briggs Institute (JBI) Critical Appraisal Checklist for Analytical Cross-Sectional studies [17]. This tool evaluates various domains, including the clarity of the study population, the validity and reliability of the measurement of exposure and outcomes, and the appropriateness of the statistical analysis. To ensure a systematic interpretation of the results, each study was assigned a quality score based on the percentage of “Yes” responses to the JBI checklist items. Studies were categorized as high quality (score ≥ 75%), moderate quality (score 50–74%), and low quality (score < 50%). Any discrepancies in scoring between the two reviewers were resolved through discussion or by consulting a third reviewer. In the final synthesis, only studies categorized as moderate or high quality were included in the meta-analysis to minimize the risk of bias. Furthermore, the quality scores were utilized in a sensitivity analysis; we performed a “leave-one-out” analysis and a subgroup analysis based on study quality (high vs. moderate) to determine if the pooled estimate of anticoagulation control was significantly influenced by the methodological rigor of the individual studies. The details of quality assessment result was separately attached as Supplementary file 1. Outcome definition and data handling The proportion (percentage) of appropriate anticoagulation control determined by TTR was the primary outcome of the meta-analysis. We intended to use the Freeman–Tukey double arcsine transformation to examine proportions and back-transform pooled estimates for reporting because included studies report proportions that might be near to 0 or 1 and variance stabilizing transformations enhance meta-analytic performance for proportions. Between-study variance (τ²) was estimated with the DerSimonian–Laird random-effects model to account for anticipated heterogeneity. Heterogeneity was quantified using Cochran’s Q test and the I² statistic; I² values of 25%, 50% and 75% were interpreted as low, moderate and high heterogeneity, respectively. Data synthesis and statistical analysis Random-effects models (DerSimonian–Laird) were used for a meta-analysis of proportions [18]. 95% confidence intervals (CIs) for pooled estimate were computed and displayed as percentages. The proportions of TTR ≥ 65% (main analysis) and Mean TTR (%) were ascertained using independent meta-analyses. In this meta-analysis, subgroup and sensitivity analyses were performed to ensure the validity and depth of our findings. Given the anticipated clinical and methodological diversity among studies conducted in different regions of Ethiopia, subgroup analysis was conducted to explore potential sources of statistical heterogeneity (I2). By stratifying studies based on variables such as geographical region, study year, and sample size, we aimed to identify specific factors that might influence the prevalence of anticoagulation control and to provide more nuanced data for local healthcare settings. Moreover, we conducted sensitivity analysis to assess the robustness of our pooled estimate. A “leave-one-out” sensitivity analysis was performed by iteratively removing one study at a time and recalculating the pooled estimate. This was justified to ensure that our findings were stable and not overly influenced by a single outlier or a large-scale study. Furthermore, a quality-based sensitivity analysis was conducted to confirm that the inclusion of studies with different risk-of-bias scores did not significantly alter the primary conclusion of the review. Egger’s test and funnel plots were used to evaluate publication bias. Stata version 17 was used for all analyses. Eligibility criteria Observational studies conducted in Ethiopia among adult warfarin users reporting TTR were included. Additionally, only English-language research was considered; while studies lacking relevant outcomes or involving non-adult populations were excluded. Furthermore, studies that solely discussed other anticoagulation outcomes—such as patient knowledge, adherence, contentment, quality of life, economic outcomes, adverse drug events other than bleeding, and interactions with warfarin—were excluded. Qualitative research, review papers, unpublished works (thesis), case reports, case series, case-control studies, insufficient material in letters to the editor, author perspective, abstract proceedings, and expert comments were also excluded from the review. Results Identification of literature and findings The systematic search across electronic databases, including PubMed, Embase, Scopus, Web of Science, Google Scholar, and African Journals Online, yielded a total of 498 records. After the removal of 335 duplicate entries, 163 unique records remained for primary screening. During the initial screening of titles and abstracts, 96 records were excluded; of these, 72 did not meet the basic inclusion criteria and 24 were excluded as they were not conducted within the Ethiopian context. This left 67 articles for a comprehensive full-text eligibility assessment. Following the full-text review, 61 articles were excluded because they did not report the primary outcome of interest (Time in Therapeutic Range). Ultimately, 6 studies met all predefined eligibility criteria and were included in this systematic review and meta-analysis. The detailed transition from initial identification to final study inclusion is illustrated in the PRISMA flow diagram (Fig. 1). Fig. 1PRISMA flow diagram for study selection for systematic review [19] PRISMA flow diagram for study selection for systematic review [19] Characteristics and quality assessment of the studies A total of six studies comprising 2,022 participants were included in this systematic review and meta-analysis. The findings of the review reported the highest rate of poor anticoagulation control with warfarin medication (47.24%) and the lowest rate (29%). All six included studies employed retrospective cross-sectional designs and were conducted in hospital settings across different regions of Ethiopia, including Addis Ababa, Oromia, and Amhara. The study populations consisted of adult patients receiving warfarin therapy, with sample sizes ranging from 202 to 402 participants. The reported adequate anticoagulation control (TTR ≥ 65%) varied across studies, ranging from 29% to 47.24%. All of the included studies have low risk methodological quality, according to the modified JBI critical assessment procedure (Table 1). Table 1Characteristics and quality assessment of included studies for systematic reviewStudy IDRegionStudy designStudy settingStudy populationSample sizePrevalence of TTR ≥ 65%Quality scoreYimer et al.,2021 [14]Addis AbabaRetrospective StudyHospitalAdult Patients30042.038Addisu et al.,2025 [20]AmharaA Multicenter Cross-SectionalHospitalsOlder patients384295Teka et al.,2025 [13]OromiaRetrospective studyHospitalsAdult Patients40230.48Getachew et al.,2023 [15]Addis AbabaRetrospective StudyPrivate Cardiac CentersAdult Patients37447.245Masresha et al.,2021 [21]AmharaRetrospective Cross-SectionalHospitalAdult Patients20229.25Fenta et al.,2017 [16]Addis AbabaRetrospective Cross-SectionalHospitalAdult Patients360295 Table 1Characteristics and quality assessment of included studies for systematic reviewStudy IDRegionStudy designStudy settingStudy populationSample sizePrevalence of TTR ≥ 65%Quality scoreYimer et al.,2021 [14]Addis AbabaRetrospective StudyHospitalAdult Patients30042.038Addisu et al.,2025 [20]AmharaA Multicenter Cross-SectionalHospitalsOlder patients384295Teka et al.,2025 [13]OromiaRetrospective studyHospitalsAdult Patients40230.48Getachew et al.,2023 [15]Addis AbabaRetrospective StudyPrivate Cardiac CentersAdult Patients37447.245Masresha et al.,2021 [21]AmharaRetrospective Cross-SectionalHospitalAdult Patients20229.25Fenta et al.,2017 [16]Addis AbabaRetrospective Cross-SectionalHospitalAdult Patients360295 Characteristics and quality assessment of included studies for systematic review Pooled estimate of anticoagulation control with warfarin therapy among adult patients in Ethiopia This systematic review found a mean TTR of 34.5 ± 8.1% (29 to 47.24%). TTR of 34.30% (95% CI: 28.16–40.44) was the pooled estimate of anticoagulation control with warfarin therapy across 6 studies. The meta-analysis of this study revealed heterogeneity (I2 = 79.09%, p-value = 0.00), necessitating subgroup analysis (Fig. 2). Fig. 2Forest plot of the pooled estimate of anticoagulation control with warfarin therapy among adult patients in Ethiopia, 2026 Forest plot of the pooled estimate of anticoagulation control with warfarin therapy among adult patients in Ethiopia, 2026 Subgroup analysis of pooled estimate of anticoagulation control with warfarin therapy among adult patients in Ethiopia A subgroup analysis of the region variable was carried out. Subgroup analysis by region showed that the Addis Ababa region had the greatest pooled estimate of anticoagulation control with warfarin therapy, at TTR of 22.25% (95% CI: 12.32–32.17, I2 = 88.02%, p = 0.00), followed by the other regions (Amhara and Oromia) at TTR of 22.19% (95% CI: 8.85–35.52, I2 =93.74%, p = 0.00) (Fig. 3). Fig. 3Subgroup analysis of pooled estimate of anticoagulation control with warfarin therapy based on the region among adult patients in Ethiopia, 2026 Subgroup analysis of pooled estimate of anticoagulation control with warfarin therapy based on the region among adult patients in Ethiopia, 2026 Publication bias The potential for publication bias was evaluated using both qualitative and quantitative methods. Visual inspection of the funnel plot (Fig. 4) revealed some degree of asymmetry around the pooled estimate, particularly with an apparent imbalance among smaller studies. While such asymmetry can sometimes suggest publication bias, it not indicates; it is frequently observed in meta-analyses characterized by significant between-study heterogeneity or “small-study effects,” where results from smaller studies vary more widely due to limited precision. To provide a more objective and rigorous assessment, Egger’s linear regression test was performed. The statistical results demonstrated no significant small-study effects (β = −0.00, SE = 193.462, z = − 0.00, p = 1.000). Given that the p-value is significantly above the 0.05 threshold, we concluded that there is a low likelihood of substantial publication bias. The discrepancy between the visual funnel plot inspection and the Egger’s test results likely reflects the inherent methodological diversity and heterogeneity of the included studies in Ethiopia rather than a systematic exclusion of unpublished findings. Fig. 4Funnel plot of the pooled estimate of good anticoagulation control with warfarin therapy among adult patients in Ethiopia, 2026 Funnel plot of the pooled estimate of good anticoagulation control with warfarin therapy among adult patients in Ethiopia, 2026 Sensitivity analysis A leave-one-out sensitivity analysis was conducted to assess the robustness of the pooled estimate of anticoagulation control with warfarin therapy. Sequential exclusion of each individual study showed that the overall pooled estimate remained stable across all iterations, with no single study exerting a disproportionate influence on the summary effect size. Both the direction and magnitude of the pooled estimate were largely unchanged after removal of any study, and the resulting confidence intervals substantially overlapped with those of the primary analysis. Moreover, excluding individual studies did not produce meaningful changes in between-study heterogeneity, indicating that the observed heterogeneity was not driven by any specific study. Collectively, these findings support the robustness and reliability of the meta-analytic results (Fig. 5). Fig. 5Leave-one-out sensitivity analysis of the pooled estimate of anticoagulation control with warfarin therapy among adult patients in Ethiopia, 2026 Leave-one-out sensitivity analysis of the pooled estimate of anticoagulation control with warfarin therapy among adult patients in Ethiopia, 2026 Discussion This systematic review and meta-analysis synthesized evidence from multiple studies to provide the first national-level pooled estimate of anticoagulation control among adult patients on warfarin therapy in Ethiopia. Our findings reveal that the overall prevalence of adequate anticoagulation control, as measured by the TTR, remains alarmingly low across the country. This suboptimal level of control suggests that a significant majority of patients are at an elevated risk for both thromboembolic complications and major bleeding events. These results underscore a critical gap in current cardiovascular care and highlight the urgent need for standardized anticoagulation management protocols and improved monitoring infrastructure within the Ethiopian healthcare system. In this systematic review and meta-analysis, the pooled estimate of adequate anticoagulation control among Ethiopian adults was a TTR of 34.30% (95% CI: 28.16–40.44). This finding indicates that the vast majority of patients do not achieve recommended therapeutic targets. This TTR value is significantly lower than the thresholds of TTR ≥ 65% [22] and TTR > 70% recommended by the European Society of Cardiology guidelines [23]. Our results align with a recent meta-analysis of sub-Saharan Africa, which also found that a substantial proportion of patients fail to reach guideline-recommended TTR levels [24]. When compared with individual Ethiopian studies, our pooled estimate is consistent with the low anticoagulation quality reported in various regions. For instance, a retrospective study at a teaching hospital in Addis Ababa reported a TTR of only 29% [16]. Similarly, studies at the University of Gondar [21] and other teaching hospitals in the Amhara Region, Ethiopia [20], reported that only 29.2% and 29% of patients achieved target INR levels, respectively. Collectively, these findings confirm that anticoagulation control is generally poor in routine Ethiopian clinical practice. In contrast, anticoagulation control in high-income countries is substantially higher. Clinical trials and registries in those settings typically report mean TTR values between 55% and 70% [25–27]. This represents an absolute difference of approximately 20–35% points compared to our pooled estimate of 34.30%. This gap likely reflects more organized anticoagulation care systems. High-income settings benefit from dedicated clinics, frequent INR monitoring via point-of-care testing, standardized dosing algorithms, and electronic decision-support tools. They also maintain superior patient education and follow-up systems. Findings from other middle-income settings also surpass our results. For example, a multicenter study in Poland reported a TTR of 55% [28]. The variations between these findings and our own likely stem from differences in healthcare infrastructure and anticoagulation management practices between countries. The clinical and public health implications of these findings are significant. Data consistently link lower TTR with higher rates of stroke, thromboembolism, and major bleeding [29–34]. Therefore, the low pooled estimate observed here implies a substantial and preventable burden of adverse outcomes for Ethiopian patients. To address these discrepancies, several interventions that have improved TTR in other settings should be considered. These include establishing pharmacist-led anticoagulation clinics, scaling up point-of-care INR testing to shorten monitoring intervals, and implementing evidence-based dosing protocols. Furthermore, patient education programs focusing on adherence, diet, and drug interactions are essential. Finally, the selective use of DOACs could be considered for eligible patients with non-valvular conditions [35] DOACs eliminate the need for routine INR monitoring and offer more predictable pharmacokinetics. However, constraints such as high cost, limited access, and a lack of local safety evidence especially in patients with rheumatic or valvular disease must be addressed. Given the high prevalence of valvular heart disease in Ethiopia, many patients will continue to require vitamin K antagonists. Consequently, improving local warfarin management remains an urgent priority. Limitations of the study While this systematic review and meta-analysis provide the first pooled estimate of anticoagulation control in Ethiopia, several limitations must be acknowledged. First, the relatively small number of included studies (n = 6) may limit the statistical power of the meta-analysis and the robustness of the pooled estimate. With a limited study pool, the stability of the final result is more susceptible to the influence of individual study characteristics. Second, there was significant statistical heterogeneity observed across the included studies. This variability likely stems from differences in study populations, geographical settings (urban vs. rural), and the level of healthcare facilities (teaching vs. general hospitals). Such heterogeneity implies that the pooled TTR should be interpreted with caution, as clinical practices and patient adherence may vary across the diverse Ethiopian healthcare landscape. Third, the methods used to define and calculate TTR were not uniform across all primary studies. Differences in calculation approaches such as the Rosendaal method versus the percentage of tests in range can impact the comparability of data and may introduce measurement bias into the pooled results. Fourth, although we assessed publication bias using funnel plots and Egger’s test, these methods have reduced sensitivity when fewer than ten studies are included. Consequently, we cannot entirely rule out the possibility that smaller, unpublished studies with different outcomes may exist. Finally, our search was restricted to articles published in the English language, which may have excluded relevant local evidence published in other languages, potentially introducing a degree of language bias into our synthesis. Conclusion This systematic review and meta-analysis provides the first comprehensive pooled estimate of anticoagulation control among adult patients receiving warfarin therapy in Ethiopia, demonstrating that only about one-third of patients [34.30%) achieve adequate therapeutic control. This finding extends beyond individual primary studies by offering a nationally aggregated estimate and confirming that suboptimal anticoagulation control is a consistent and widespread challenge across different healthcare settings in the country. Furthermore, the study quantifies a substantial gap between anticoagulation control in Ethiopia and internationally recommended targets, highlighting an important public health concern. These findings underscore the urgent need for health system–level interventions, including strengthening anticoagulation monitoring capacity, establishing structured anticoagulation services, and improving patient follow-up and education. Future research should focus on well-designed prospective multicenter studies to better understand determinants of anticoagulation control, as well as interventional studies evaluating the effectiveness of strategies such as pharmacist-led anticoagulation clinics, point-of-care INR testing, and standardized dosing protocols. In addition, further studies are warranted to assess the feasibility, safety, and cost-effectiveness of direct oral anticoagulants in the Ethiopian context. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1