Ezgi Yılmaz

Dr. Ezgi Yılmaz


Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
27 October 2025 in International Journal of Antimicrobial Agents
Verified
Journal Article
Open access
AAAbdullah Tarik Aslan
Abdullah Tarik Aslan
ETElif Seren Tanriverdi
Elif Seren Tanriverdi
SKSibel Yıldız Kaya
Sibel Yıldız Kaya
GDGozde Dalgan
Gozde Dalgan
NSNeşe Saltoğlu
Neşe Saltoğlu
Murat AkovaMurat Akova
Murat Akova
Due to underrepresentation of carbapenemase-producing Enterobacterales infections in randomized controlled trials with ceftazidime-avibactam (CZA) and high cost of CZA therapy, other appropriate antimicrobial therapies (OAAT) are still being used for OXA-48- or KPC-producing Enterobacterales infections in Türkiye. We conducted a multicentre retrospective 1:1 matched cohort study of patients who received ≥48 h of CZA or OAAT for documented OXA-48- or KPC-producing Enterobacterales infections. Patients were matched based on (1) the number of days (± 1 d) from the infection onset to the initiation of therapy, (2) INCREMENT-CPE score (± 1), (3) source of infection, (4) year of infectious episode, and (5) type of causative microorganism. From 5 Turkish university hospitals, 180 patients were enrolled. Baseline characteristics were all similar between treatment groups. At the time of treatment initiation, 63.9% of patients were in the intensive care unit, 35.6% had septic shock and 41.1% required mechanical ventilation support. Thirty-day mortality occurred in 35.6% (32/90) of patients treated with CZA and in 56.7% (51/90) of those receiving OAAT regimens (P = 0.004). Twenty-one-day clinical response was seen in 50% (45/90) and 26.7% (24/90) of patients receiving CZA and OAAT, respectively (P = 0.002). In multivariable logistic regression analyses, CZA treatment was associated with less likelihood of mortality (aOR = 0.37; 95% CI: 0.19-0.71; P = 0.003) and higher likelihood of 21-d clinical response (aOR = 3.32; 95% CI: 1.68-6.53; P < 0·001). Treatment with CZA is associated with more favourable clinical outcomes in treatment of OXA-48- or KPC-producing Enterobacterales infections. A randomized controlled trial is needed to confirm these results.

4 December 2024 in Journal of Clinical Medicine
Verified
Journal Article
Open access
PDF file
BCBahadır Ceylan
Bahadır Ceylan
OOOktay Olmuşçelik
Oktay Olmuşçelik
BKBanu Karaalioğlu
Banu Karaalioğlu
SCSule Ceylan
Sule Ceylan
MSMeyha Sahin
Meyha Sahin
AMAli Mert
Ali Mert
Background/Objectives: Studies attempting to predict the development of severe respiratory failure in patients with a COVID-19 infection using machine learning algorithms have yielded different results due to differences in variable selection. We aimed to predict the development of severe respiratory failure, defined as the need for high-flow oxygen support, continuous positive airway pressure, or mechanical ventilation, in patients with COVID-19, using machine learning algorithms to identify the most important variables in achieving this prediction. Methods: This retrospective, cross-sectional study included COVID-19 patients with mild respiratory failure (mostly receiving oxygen through a mask or nasal cannula). We used XGBoost, support vector machines, multi-layer perceptron, k-nearest neighbor, random forests, decision trees, logistic regression, and naïve Bayes methods to accurately predict severe respiratory failure in these patients. Results: A total of 320 patients (62.1% male; average age, 54.67 ± 15.82 years) were included in this study. During the follow-ups of these cases, 114 patients (35.6%) required high-level oxygen support, 67 (20.9%) required intensive care unit admission, and 43 (13.4%) died. The machine learning algorithms with the highest accuracy values were XGBoost, support vector machines, k-nearest neighbor, logistic regression, and multi-layer perceptron (0.7395, 0.7395, 0.7291, 0.7187, and 0.75, respectively). The method that obtained the highest ROC-AUC value was logistic regression (ROC-AUC = 0.7274). The best predictors of severe respiratory failure were a low lymphocyte count, a high computed tomography score in the right and left upper lung zones, an elevated neutrophil count, a small decrease in CRP levels on the third day of admission, a high Charlson comorbidity index score, and a high serum procalcitonin level. Conclusions: The development of severe respiratory failure in patients with COVID-19 could be successfully predicted using machine learning methods, especially logistic regression, and the best predictors of severe respiratory failure were the lymphocyte count and the degree of upper lung zone involvement.

30 April 2021 in Infectious Diseases and Clinical Microbiology
Verified
Journal Article
Closed access
SYserap yavuz
serap yavuz
SDSibel Dogan–Kaya
Sibel Dogan–Kaya
DDDenef Deniz
Denef Deniz
ETElif Tükenmez-Tigen
Elif Tükenmez-Tigen
SÖSerpil Öztürk
Serpil Öztürk
HEHaluk Eraksoy
Haluk Eraksoy
Objective: A simple and widely available test is needed to monitor the response to the treatment and predict adverse outcomes in patients with infective endocarditis (IE). We aimed to detect the role of C-reactive protein (CRP) level in the prognosis of IE. Materials and Methods: The patients diagnosed as probable or definite IE between 2015 and 2016 from five medical centers were included in the study. Clinical and laboratory features of the patients recorded on previously prepared forms retrospectively. Comparison of CRP levels between survived and dead patients were made by repeated-measures variance analysis. Receiver operating characteristic (ROC) curves at baseline, the first, second, third and fourth week of antimicrobial treatment were constructed to explain the ability of CRP level to predict mortality. Multivariate analysis was performed to predict the risk factors for mortality. Results: In total, 111 cases of IE were included. Serum CRP levels were higher in patients with mortality at every measurement. Comparing baseline CRP, lower than a double or triple decrease of serum CRP level at the third or fourth week of treatment or any increase under treatment were related to mortality. According to the ROC curve analysis, the highest overall accuracy in death prediction was at the fourth week CRP level (c=0.74); the CRP level of ≥50 mg/L was 86% sensitive and 78% specific to predict mortality. The presence of chronic renal failure (OR 14.386) and the fourth week CRP level (OR 1.016) were independent risk factors in terms of mortality. Conclusion: Being an easily performed and widely available test, CRP is a useful tool to follow up the response to the treatment and predict outcome among patients with IE. In the fourth week of treatment, a CRP level of ≥50 mg/L should be regarded as an alarming sign of poor outcome and should lead to investigation and appropriate management of complications. Keywords: Infective endocarditis, CRP, outcome

28 March 2024 in Klimik Dergisi/Klimik Journal
Verified
Journal Article
Open access
EYEzgi Yılmaz
Ezgi Yılmaz
SBSeniha Başaran
Seniha Başaran
SYSERAP SUZUK YILDIZ
SERAP SUZUK YILDIZ
SYserap yavuz
serap yavuz
ACAtahan Cagatay
Atahan Cagatay
HEHaluk Eraksoy
Haluk Eraksoy
Objective: Identifying the presence and type of carbapenemases is essential to determine the treatment choices for carbapenem-resistant Gram-negative bacilli (CRB). Genotypic characterization of CRB needs technical support and experienced staff and is not an option for most laboratories due to its high cost. For this reason, especially in countries with limited resources, cheap, reliable phenotypic methods are an alternative, do not require experience, and can be easily applied in daily practice. The goal of the study was to evaluate the performance of phenotypic methods for carbapenemase production in CRB and to form a simple algorithm to differentiate carbapenemase types such as blaOXA-48 or blaNDM, which are common in our country. Methods: The study included 16 consecutive carbapenem-resistant, Gram-negative bacteria. Simplified carbapenem inactivation methods (sCIM) and modified Hodge test (MHT) were performed. Genes responsible for carbapenemase production (blaOXA-48, blaNDM, and blaKPC) were detected by real-time polymerase chain reaction. Temocillin resistance and ceftazidime-avibactam disc diffusion test were also applied to define carbapenemase types. Results: The carbapenemase gene was detected in 12 of the 16 strains; sCIM positivity was found in 11, and MHT was positive in 10. Sensitivity for sCIM and MHT were 91.9% and 83.3%, respectively. All Enterobacterales strains were positive for sCIM, and blaOXA-48 was the most common carbapenemase. sCIM false negativity was detected for only one strain. High-level temocillin resistance (MIC >128 μg/mL) was present in all strains with blaOXA-48; it wasn’t detected in the strain carrying isolated blaNDM. Conclusion: sCIM positivity was present for all Enterobacterales, which were shown to carry the carbapenemase gene by RT-PCR. Our findings support the usage of sCIM in daily practice to screen for carbapenemase production in CRB.

26 June 2026 in International Journal of Infectious Diseases
Verified
Journal Article
Open access
EYEzgi Yılmaz
Ezgi Yılmaz
NKNuran Karabulut
Nuran Karabulut
AİAyşe İstanbullu-Tosun
Ayşe İstanbullu-Tosun
SASema Alacam
Sema Alacam
MSMeyha Sahin
Meyha Sahin
AMAli Mert
Ali Mert
False-positive HIV screen tests have clinical and social consequences when time constrains necessitate interventions before confirmatory results are available, such as in late phases of pregnancy. We conducted this study to test the performance of HIV screening in pregnancy and explore ways to address the challenges caused by false-positives. Cases with a reactive fourth-generation HIV screening assay were identified and the cut-off indices (COI), confirmatory HIV test results, and pregnancy status at the time of testing were acquired through digital archives of two tertiary centers. These data were used to calculate PPV with the standard and alternative COI thresholds. Among 904 reactive screens, 58.6% were false-positive. Pregnant women (n=112) had a markedly lower PPV than the rest (12.5% vs 45.5%, p<0.0001). COI distribution clusters were distinctly apart between false-and true-positive cases (median 1.9 vs 608.0; p<0.0001). Two COI-thresholds, 1.0 (standard) and 3.0 (explorative) were equally sensitive; however, the PPV were 41.4% and 70.1%, respectively. COI threshold of 15.9 had the best numeric balance with 99.2% sensitivity and 95.4% PPV. Reactive fourth-generation HIV screening among pregnant woman were mostly false-positive in a low-prevalence community. This study found alternative COI-thresholds to improve PPV without significantly jeopardizing sensitivity.

26 December 2024 in Flora the Journal of Infectious Diseases and Clinical Microbiology
Verified
Journal Article
Closed access
SASelda Aydın
Selda Aydın
MSMeyha Sahin
Meyha Sahin
EYEzgi Yılmaz
Ezgi Yılmaz
AMAli Mert
Ali Mert