Ugur Bilge

The accuracy of autotitrating CPAP-determined residual apnea–hypopnea index

PurposeAutotitrating continuous positive airway pressure (auto-CPAP) devices not only titrate CPAP pressures but also measure residual respiratory events. The aim of the present study was to determine the accuracy of auto-CPAP-derived residual apnea–hypopnea index (AHI).MethodsWe studied 137 consecutive patients (72.3% men) with obstructive sleep apnea from January 2008 to December 2010 who underwent in-laboratory overnight polysomnography (PSG) using auto-CPAP. We excluded patients with comorbidities like congestive heart disease, chronic obstructive pulmonary disease, or hypoventilation syndromes and patients with central sleep apnea. Residual AHI obtained from the auto-CPAP device by smart card (CPAP-AHI) was compared simultaneously with AHI from an overnight PSG on auto-CPAP (PSG-AHI) using Bland–Altman analysis and Wilcoxon signed-rank test.ResultsThe mean AHI on the diagnostic study was 45.08 ± 1.8. During the titration, auto-CPAP markedly suppressed the respiratory events (PSG-AHI, 3.40 ± 0.20). On the other hand, CPAP-AHI was 3.35 ± 0.17. Bland–Altman analysis showed good agreement between auto-CPAP-AHI and PSG-AHI (AHI mean difference of 0.05, and the limits of agreement for the AHI were from +4.9 to −4.8). Two methods have also been compared with paired samples t test and no statistically significant difference was found (p > 0.05).ConclusionAuto-CPAP can identify residual respiratory events equivalent to the use of PSG in a selected population.

Elastic Light Single-Scattering Spectroscopy for the Detection of Cervical Precancerous Ex vivo

Potential application of elastic light single-scattering spectroscopy (ELSSS) for differentiating high-grade squamous intraepithelial lesions (HSIL) from non-HSIL tissues was investigated. An ELSSS system was used to acquire spectra from cervix tissues. A single-fiber optical probe with a diameter of 100 μm was used for both delivery and detection of white light to and from the cervix tissue. Spectroscopic measurements were acquired from 95 ex vivo biopsy samples of 60 pap smear positive patients and normal cervix tissue from 10 patients after hysterectomy were used as a negative control group. Spectroscopic results of 95 cervix biopsy were compared to the histopathology of the biopsy samples. Sensitivity and specificity of the ELSSS system in the differentiation of HSIL and non-HSIL tissues are 87.5% and 45.6%, respectively, for the pap smear and colposcopy positive biopsy samples. The ELSSS system has the potential for use in real-time diagnosis of HSIL tissues as an adjunct to Papanicolaou test (pap smear) and colposcopy.

Community-acquired pneumonia in patients with chronic obstructive pulmonary disease requiring admission to the intensive care unit: risk factors for mortality.

PURPOSE The aims of this study are to identify factors predicting mortality in patients with chronic obstructive pulmonary disease (COPD) and community-acquired pneumonia (CAP) requiring intensive care unit (ICU) admission and to examine whether noninvasive ventilation treatment reduces mortality. MATERIALS AND METHODS An analysis was performed on data from patients with CAP hospitalized in the ICUs of 19 different hospitals in Turkey between October 2008 and January 2011. Predictors of mortality were assessed by both univariate and multivariate statistical analyses. RESULTS Two hundred eleven patients with COPD and CAP were included. The overall ICU mortality was 23.9%. Noninvasive ventilation treatment (odds ratio [OR], 0.12; 95% confidence interval [CI], 0.03-0.49; P = .003), hypertension (OR, 0.13; 95% CI, 0.02-0.93; P = .042), bilateral infiltration (OR, 13.92; 95% CI, 2.94-65.84; P = .001), systemic corticosteroid treatment (OR, 0.19; 95% CI, 0.35-0.96; P = .045), length of ICU stay (OR, 0.65; 95% CI, 0.47-0.89; P = .007), and duration of invasive mechanical ventilation (OR, 1.11; 95% CI, 1.01-1.22; P = .032) were independent factors related to mortality. CONCLUSION Noninvasive ventilation, hypertension, systemic corticosteroid treatment, and shorter ICU stay are associated with reduced mortality, whereas bilateral infiltration and longer duration of invasive mechanical ventilation are associated with increased risk of mortality in patients with COPD and CAP requiring ICU admission.

Artificial neural network, genetic algorithm, and logistic regression applications for predicting renal colic in emergency settings

BackgroundLogistic regression is the most common statistical model for processing multivariate data in the medical literature. Artificial intelligence models like an artificial neural network (ANN) and genetic algorithm (GA) may also be useful to interpret medical data.AimsThe purpose of this study was to perform artificial intelligence models on a medical data sheet and compare to logistic regression.MethodsANN, GA, and logistic regression analysis were carried out on a data sheet of a previously published article regarding patients presenting to an emergency department with flank pain suspicious for renal colic.ResultsThe study population was composed of 227 patients: 176 patients had a diagnosis of urinary stone, while 51 ultimately had no calculus. The GA found two decision rules in predicting urinary stones. Rule 1 consisted of being male, pain not spreading to back, and no fever. In rule 2, pelvicaliceal dilatation on bedside ultrasonography replaced no fever. ANN, GA rule 1, GA rule 2, and logistic regression had a sensitivity of 94.9, 67.6, 56.8, and 95.5%, a specificity of 78.4, 76.47, 86.3, and 47.1%, a positive likelihood ratio of 4.4, 2.9, 4.1, and 1.8, and a negative likelihood ratio of 0.06, 0.42, 0.5, and 0.09, respectively. The area under the curve was found to be 0.867, 0.720, 0.715, and 0.713 for all applications, respectively.ConclusionData mining techniques such as ANN and GA can be used for predicting renal colic in emergency settings and to constitute clinical decision rules. They may be an alternative to conventional multivariate analysis applications used in biostatistics.

Mortality indicators in community-acquired pneumonia requiring intensive care in Turkey

BACKGROUND Severe community-acquired pneumonia (SCAP) is a fatal disease. This study was conducted to describe an outcome analysis of the intensive care units (ICUs) of Turkey. METHODS This study evaluated SCAP cases hospitalized in the ICUs of 19 different hospitals between October 2008 and January 2011. The cases of 413 patients admitted to the ICUs were retrospectively analyzed. RESULTS Overall 413 patients were included in the study and 129 (31.2%) died. It was found that bilateral pulmonary involvement (odds ratio (OR) 2.5, 95% confidence interval (CI) 1.1-5.7) and CAP PIRO score (OR 2, 95% CI 1.3-2.9) were independent risk factors for a higher in-ICU mortality, while arterial hypertension (OR 0.3, 95% CI 0.1-0.9) and the application of non-invasive ventilation (OR 0.2, 95% CI 0.1-0.5) decreased mortality. No culture of any kind was obtained for 90 (22%) patients during the entire course of the hospitalization. Blood, bronchoalveolar lavage, and non-bronchoscopic lavage cultures yielded enteric Gram-negatives (n=12), followed by Staphylococcus aureus (n=10), pneumococci (n=6), and Pseudomonas aeruginosa (n=6). For 22% of the patients, none of the culture methods were applied. CONCLUSIONS SCAP requiring ICU admission is associated with considerable mortality for ICU patients. Increased awareness appears essential for the microbiological diagnosis of this disease.

A web-based intensive care clinical decision support system: From design to evaluation

The aim of this study is to develop and evaluate a web-based clinical decision support system (CDSS) containing clinical guidelines and protocols that will support intensive care unit (ICU) providers in making decisions more effectively and quickly. First, a survey was carried out with 38 physicians in order to determine their preferences, needs and concerns regarding decision support tools. After the survey, guidelines were prepared by a group of specialists in ICU, and a medical informatician converted the guidelines into algorithm forms. Ten CDSS were developed using the algorithms, and placed onto the Intensive Care Decision Support Website (ICDSW). In order to evaluation of the website, 15 physicians were asked to answer 10 questions in 10 different scenarios first using a paper-based approach, then with ICDSW. When the answers were analyzed, it was found that the answers given by using ICDSW were significantly better than the paper-based approach (p <  0.001). However, there was no significant difference in terms of the time needed to answer the questions (p  =  0.138). The usability score of the website was 85.6  ±  8.89. The study demonstrated the successful implementation of an ICDSW within intensive care units.

Application of data mining techniques for detecting asymptomatic carotid artery stenosis

Asymptomatic carotid stenosis, one of the etiological factors for stroke, has several risk factors such as hypertension, cardiac morbidity, smoking, diabetes, and physical inactivity. Understanding and determining factors that predispose to asymptomatic carotid stenosis will help in the design of acute stroke trials and in prevention programs. The goal of this study is to explore rules and relationships that might be used to detect possible asymptomatic carotid stenosis by using data mining techniques. For this purpose, Genetic Algorithms (GAs), Logistic Regression (LR), and Chi-square tests have been applied to the patient dataset. Results of these tests have also been compared.

Detection of precancerous cervical conditions using elastic light single-scattering spectroscopy

We have investigated the potential application of elastic light single-scattering spectroscopy (ELSSS) as an adjunctive tool for screening of cervical precancerous lesions non-invasively and in real time. Ex-vivo measurements were performed on 95 cervix biopsy tissue of 60 patients. Normal cervix tissue from 10 patients after hysterectomy was used as a control group. Correlation between ELSSS spectra and histopathology results were investigated. It was found that the spectral slope was positive for all the spectra taken on normal cervix tissue samples from the control group. We assumed that if there is only one spectrum with a negative spectral slope among the all spectra taken on a biopsy specimen, the biopsy specimen is pathologically abnormal. This shows that pap smear and ELSSS results are in good agreement. Most biopsy tissue samples had both positive and negative spectral slopes. Therefore, we calculated the percentage of negative spectral slopes and hypothesized that this was correlated to dysplastic percentage of the epithelial tissue of the biopsy material. The ROC curve was calculated using the dysplastic percentage and high squamous intraepithelial lesion (HSIL) and low squamous intraepitherlial lesions (LSIL) biopsy specimens were differentiated from non HSIL and LSIL with a sensitivity and specificity of 70.4% and 66.7% respectively, with p < 0.05.

Antibiotic treatment outcomes in community-acquired pneumonia

Background/aim The optimal empiric antibiotic regimen for patients with community-acquired pneumonia (CAP) remains unclear. This study aimed to evaluate the clinical cure rate, mortality, and length of stay among patients hospitalized with community- acquired pneumonia in nonintensive care unit (ICU) wards and treated with a β-lactam, β-lactam and macrolide combination, or a fluoroquinolone. Materials and methods This prospective cohort study was performed using standardized web-based database sheets from January 2009 to September 2013 in nine tertiary care hospitals in Turkey. Results Six hundred and twenty-one consecutive patients were enrolled. A pathogen was identified in 78 (12.6%) patients. The most frequently isolated bacteria were S. pneumoniae (21.8%) and P. aeruginosa (19.2%). The clinical cure rate and length of stay were not different among patients treated with β-lactam, β-lactam and macrolide combination, and fluoroquinolone. Forty-seven patients (9.2%) died during the hospitalization period. There was no difference in survival among the three treatment groups. Conclusion In patients admitted to non-ICU hospital wards for CAP, there was no difference in clinical outcomes between β-lactam, β-lactam and macrolide combination, and fluoroquinolone regimens.

RLIP76 Gene Variants are not Associated with Drug Response in Turkish Epilepsy Patients.

RLIP76 Gene Variants are not Associated with Drug Response in Turkish Epilepsy Patients Approximately 30% of epileptic patients remain untreated, in spite of trials with maximum tolerable doses of more than one drug. The RalA binding protein 1 (RALBP1/RLIP76), a multifunctional, anti-apoptotic, multidrug transporter protein, has been proposed as being responsible for the drug resistance mechanism in epilepsy. We have investigated polymorphic differences in the coding regions and exon-intron boundaries of the RLIP76 gene, between 146 refractory and 155 non refractory epileptic patients in Turkey, using denaturing high performance liquid chromatography (HPLC) and sequencing analysis techniques. We have detected the following sequence variants: c.160-4G>A, c.187C>G, c.1562-38G>A, c.1670+107G>A, c.1670+93G>A, c.1670+96G>A, c.1670+100C>T, c.1670+130C>T, c.1670+131G>C, c.1670+140 G>C, and found no statistically significant correlation between allele frequencies and drug response status. We conclude that sequence variants of this gene are not involved in drug resistance in epilepsy.