Ministry of Education, Science and Technological Development
National Repository of Dissertations in Serbia (NaRDuS)Not a member yet
15259 research outputs found
Sort by
Application of statistical approaches in the assessment of rate and extent of absorption and development of a population pharmacokinetic model for clopidogrel and its metabolite clopidogrel carboxylic acid from generic products
No full textKlopidogrel je inhibitor agregacije trombocita, prolek koji se posle p.o. primene brzo resorbuje i intenzivno metaboliše, pre svega presistemski u jetri, pri čemu se veći deo primenjene doze konvertuje do neaktivnog metabolita klopidogrel karboksilne kiseline, a manji deo, oko 15%, do aktivnog metabolita klopidogrel tiola, preko koga klopidogrel ostvaruje svoj efekat. Na tržištu je prisutno više generičkih lekova, za koje je potvrđena biološka ekvivalentnost sa istim referentnim lekom, dok između samih generičkih lekova nije direktno potvrđena. Budući da klopidogrel ima varijabilnu farmakokinetiku, kao i odgovor na terapiju ovim lekom, za kliničku praksu mogu biti značajne informacije o eventualnim razlikama između generičkih lekova u brzini i stepenu resorpcije, zbog kojih kod pacijenata potencijalno može da se poveća rizik od pojave neželjenih reakcija. Cilj prvog dela disertacije bio je poređenje brzine i stepena resorpcije između 19 generičkih lekova klopidogrela, metodama indirektnog i direktnog poređenja, pri čemu su korišćeni rezultati studija biološke ekvivalentnosti koji uključuju primarne i/ili sekundarne podatke za klopidogrel i klopidogrel karboksilnu kiselinu. Cilj drugog dela istraživanja bio je karakterizacija procesa resorpcije i dispozicije klopidogrela, kao i njegove konverzije do klopidogrel karboksilne kiseline, kroz razvoj i validaciju združenog populacionog farmakokinetičkog modela, uz razmatranje faktora varijabilnosti, korišćenjem koncentracija klopidogrela i klopidogrel karboksilne kiseline i drugih dostupnih relevantnih primarnih podataka iz dve studije biološke ekvivalentnosti. Metodom prilagođenog indirektnog poređenja obrađeni su podaci za klopidogrel i klopidogrel karboksilnu kiselinu iz 19 studija biološke ekvivalentnosti i utvrđene su kombinacije generičkih lekova klopidogrela koje se mogu smatrati međusobno biološki ekvivalentnim i potencijalno zamenjivim u kliničkoj praksi (76%), kao i one za koje se zamena ne preporučuje. Direktnim poređenjem farmakokinetičkih parametara iz dve studije biološke ekvivalentnosti korišćenjem Student t-testa, utvrđena je sličnost između generičkih lekova ispitivanih u ovim studijama u pogledu stepena resorpcije kopidogrela i njegove konverzije do klopidogrel karboksilne kiseline, ali ne i u pogledu brzine ovih procesa. Budući da je klinički značaj stepena resorpcije veći u odnosu na brzinu ovog procesa, može se smatrati da između ova dva generička leka nema statistički značajne razlike u biološkoj raspoloživosti. Metodom nelinearnog modelovanja kombinovaniah efekata razvijen je i validiran združeni semi-fiziološki populacioni farmakokinetički model kopidogrela i klopidogrel karboksilne kiseline, koji uključuje linearnu resorpciju sa 2 tranzitna prostora, hepatički prostor za opisivanje presistemskog metabolizma do metabolita, jednoprostorni model za klopidogrel, dvoprostorni model za klopidogrel karboksilnu kiselinu, linearnu eliminaciju ovog metbolita, alomerijsko skaliranje klirensa i volumena distribucije. Razvijeni model daje čvrstu osnovu za potencijalna buduća poboljšanja pri građenju složenijeg modela. Sprovedeno istraživanje koristi savremene metode analize sekundarnih i/ili primarnih podataka dostupnih za generičke lekove klopidogrela i pruža dodatne informacije o međusobnoj zamenjivosti lekova, koje mogu biti od koristi u kliničkoj praksi, te predstavlja iskorak u pravcu primene alternativnih metoda u svrhu poređenja biološke raspoloživosti lekova kod kojih nije pogodan standardni pristup predviđen aktuelnom regulativom.Clopidogrel, a platelet aggregation inhibitor, is a prodrug that undergoes rapid absorption and extensive hepatic first-pass metabolism after p.o. administration. The majority of the dose is metabolized into the inactive metabolite clopidogrel carboxylic acid, while only about 15% is converted into the active metabolite clopidogrel thiol, responsible for the clopidogrel therapeutic effect. Given its variable pharmacokinetics and therapeutic response, understanding differences in the rate and extent of absorption between generic products is crucial, as these differences could potentially impact patient safety or therapeutic outcomes. Although generic products are individually bioequivalent to the reference drug, bioequivalence between generics has not been directly assessed. The dissertation aimed to address this gap through two main objectives. In the first part of the dissertation the aim was to compare the rate and extent of absorption between 19 generic clopidogrel drugs, using indirect and direct comparison methods, analyzing primary and/or secondary data for clopidogrel and clopidogrel carboxylic acid originated from the bioequivalence studies. The aim of the second part of the research was to characterize the process of absorption and disposition of clopidogrel, as well as its conversion to clopidogrel carboxylic acid, through the development and validation of a joint population pharmacokinetic model, considering also variability factors. Clopidogrel and clopidogrel carboxylic acid concentrations as well as other available relevant primary data from two bioequivalence studies were used for model development. Data for clopidogrel and clopidogrel carboxylic acid from 19 bioequivalence studies were analyzed using adjusted indirect comparison method, which allowed the identification of combinations of generic clopidogrel drugs that can be considered mutually bioequivalent and potentially interchangeable in clinical practice (76%), as well as those for which substitution is not recommended. A direct comparison of pharmacokinetic parameters from two bioequivalence studies using Student t-test showed similarity between the generic drugs from these studies in the extent of clopidogrel absorption and its conversion to clopidogrel carboxylic acid, but not in terms of the rate of these processes. Since the extent of absorption is clinically more important compared to the rate of this process, these differences were deemed not statistically significant. A joint semi-physiological population pharmacokinetic model for clopidogrel and clopidogrel carboxylic acid was developed and validated using the nonlinear mixed effects modeling approach. Model included linear absorption with 2 transit compartments, hepatic compartment for describing clopidogrel presystemic metabolism to metabolites, one-compartment model for clopidogrel, two-compartment model for clopidogrel carboxyl acid, linear elimination of this metabolite, allomeric scaling of clearance and volume of distribution. The developed model provides a solid base for potential future improvements in building a more complex model. This research uses state-of-the-art methods for analyzing secondary and/or primary data available for generic clopidogrel drugs and provides additional information on interchangeability of these drugs, especially for cases where traditional regulatory approaches may not suffice. It provides additional data that can inform clinical decision-making and optimize the use of generic clopidogrel products in practice
Examination of interactions of ellagic acid's secondary metabolites (urolithins) with albumins and extracellular vesicles
No full textUrolitini (URO) su sekundarni metaboliti elaginske kiseline nastale u digestivnom traktu pod dejstvom crevne flore na hranu bogatu elagitaninima. Zbog svoje lipofilne strukture URO prolaze biološke barijere i na taj način dospevaju u cirkulaciju. U cirkulaciji se nalaze u vidu konjugata sa glukuronskom kiselinom ili se vezuju za albumine kako bi se transportovali do tkiva u kojima ispoljavaju antioksidativnu, antiinflamatornu i antikancersku aktivnost.
U doktorskoj disertaciji je ispitivano vezivanje urolitina, kao i njihovih glukuronida za serumske albumine kao proteinske nosače metodom fluorescentne spektroskopije. Mehanizam gašenja fluorescencije i konstante vezivanja (Kb) određeni su titracijom rastvora proteina fiksne koncentracije različitim koncentracijama liganda uz merenje intenziteta fluorescencije. Variranjem temperature izračunati su termodinamički parametri interakcija. Promene sekundarnih struktura proteina su praćene primenom infracrvene spektroskopije sa Furijeovom transformacijom i cirkularnim dihroizmom. Rezultati spektroskopskih analiza su upotpunjeni molekulskim dokingom kojim su predviđena mesta vezivanja URO za serumske albumine. Dobijeni rezultati omogućili su bolje razumevanje mehanizma vezivanja i transporta, odnosno farmakodinamiku ovih biološki aktivnih jedinjenja. U ovoj disertaciji su ispitivane i metode inkorporacije URO A, predstavnika klase urolitina, u ekstracelularne vezikule (EV) dobijene iz komercijalnih humanih ćelijskih linija. Ekstracelularne vezikule predstavljaju dobre neproteinske nanonosače zbog svoje veličine, strukture i mogućnosti prolaska različitih bioloških barijera. Značaj inkorporacije URO A u EV je ciljani transport, kao i zaštita ovih jedinjenja od oksidacije i konjugacije u cirkulaciji.Urolithins (UROs) are secondary metabolites of ellagic acid, formed in the digestive tract by the action of intestinal flora on ellagitannin-rich foods. Due to their lipophilic structure, UROs can cross biological barriers and enter circulation. In circulation, UROs exist as conjugates with glucuronic acid or sulfates or bind to albumins. Albumins facilitate their transport to tissues, where they exhibit antioxidant, anti-inflammatory, and anti-cancer activities.
The binding of urolithins and their glucuronides to serum albumins, as protein carriers, was investigated in this doctoral dissertation using fluorescence spectroscopy. The fluorescence quenching mechanisms and binding constants (Kb) were determined by titrating a fixed-concentration protein solution with varying ligand concentrations and measuring fluorescence intensity. By varying the temperature, the thermodynamic parameters of the interactions were calculated. Changes in the secondary structure of proteins were monitored using Fourier-transform infrared spectroscopy and circular dichroism. The results of spectroscopic analyses were complemented by molecular docking, which predicted the binding sites of UROs on serum albumins. These findings contribute to a better understanding of the binding and transport mechanisms, as well as the pharmacodynamics of these biologically active compounds. Furthermore, this dissertation explored methods for incorporating URO A, a representative of the urolithin class, into extracellular vesicles (EVs) derived from commercial human cell lines. EVs are promising non-protein nanocarriers due to their size, structure, and ability to traverse various biological barriers. The incorporation of URO A into EVs is significant for achieving targeted delivery and protecting these compounds from oxidation and conjugation during circulation
Development and application of HILIC MS/MS method for determination of acrylamide in selected food samples with human health risk assessment
No full textNaučni cilj ove doktorske disertacije bio je razvoj i primena HILIC-MS/MS metode za određivanje akrilamida u odabranim uzorcima hrane sa procenom rizika po zdravlje ljudi usled izloženosti akrilamidu, pri čemu su razmatrani nekarcinogeni, karcinogeni i mutageni rizici. Razvijena je i validirana brza, jednostavna, osetljiva i robustna metoda za određivanje akrilamida u različitim prehrambenim proizvodima primenom tečne hromatografije zasnovane na hidrofilnim interakcijama spregnute sa tandemnom masenom spektrometrijom. Za pouzdanu validaciju metode pripremljen je kontrolni uzorak hleba sa dodatkom enzima asparaginaze, čime je prevaziđen nedostatak matriksa bez akrilamida. Validirani postupak pokazao je dobre analitičke performanse sa niskim granicama detekcije i kvantifikacije uz zadovoljavajući rikaveri i preciznost metode. Metoda je primenjena za određivanje sadržaja akrilamida u uzorcima sa srpskog tržišta: različitim vrstama hleba, odabranim finim pekarskim proizvodima, čipsevima i dečijoj hrani na bazi cerealija. U dečijoj hrani na bazi cerealija sadržaj akrilamida je bio ispod granice detekcije. Utvrđeno je da većina ispitanih uzoraka hleba ima koncentracije akrilamida ispod regulatornih graničnih vrednosti. Najviše koncentracije akrilamida detektovane su u integralnom pšeničnom hlebu sa semenkama, a najniže u ražanom hlebu. Ispitivanje distribucije akrilamida u hlebu pokazalo je značajno veće koncentracije u kori nego u sredini hleba. Eksperimentalno je potvrđena direktna zavisnost sadržaja akrilamida od vremena tostiranja, pri čemu je utvrđeno da tostiranje duže od 5 minuta rezultira značajnim povećanjem mutagenog rizika. U čipsu su izmerene najviše koncentracije akrilamida među svim ispitanim proizvodima. Procena rizika po ljudsko zdravlje pokazala je da ne postoji nekarcinogeni rizik, ali je utvrđen potencijalni mutageni rizik za sve ispitane uzorke, sa posebno visokim vrednostima za čips i tostirani hleb. Sa aspekta karcinogenog rizika, samo određene vrste hleba, neki uzorci biskvita i grisina pokazali su zadovoljavajući nivo bezbednosti. Ovo istraživanje predstavlja prvu procenu rizika po zdravlje stanovnika Srbije usled izloženosti akrilamidu putem hrane i pruža značajnu naučnu osnovu za razvoj strategija za smanjenje izloženosti ovom kontaminantu.The scientific objective of this doctoral dissertation was to develop and apply a HILIC-MS/MS method for the determination of acrylamide in selected food samples, along with an assessment of human health risks associated with acrylamide exposure, considering non-carcinogenic, carcinogenic, and mutagenic risks. A rapid, simple, sensitive, and robust method was developed and validated for the quantification of acrylamide in various food products using hydrophilic interaction liquid chromatography coupled with tandem mass spectrometry. To enable reliable method validation, a control bread sample was prepared using the asparaginase enzyme, thereby overcoming the lack of an acrylamide-free matrix. The validated method demonstrated good analytical performance, with low limits of detection and quantification, and satisfactory recovery and precision. The method was applied to determine acrylamide content in samples from the Serbian market, including various types of bread, selected fine bakery products, potato chips, and cereal-based baby food. In cereal-based baby food, acrylamide was found to be below the detection limit. Most analyzed bread samples contained acrylamide concentrations below regulatory threshold values. The highest acrylamide levels were detected in wholegrain wheat bread with seeds, while the lowest were found in rye bread. Distribution analysis revealed significantly higher acrylamide concentrations in bread crust compared to crumb. A direct relationship between toasting time and acrylamide formation was experimentally confirmed, with toasting longer than 5 minutes resulting in a substantial increase in mutagenic risk. Potato chips had the highest acrylamide concentrations among all tested products. The human health risk assessment indicated no non-carcinogenic risk but identified a potential mutagenic risk for all analyzed samples, with particularly high values for chips and extensively toasted bread. From the perspective of carcinogenic risk, only certain types of bread, as well as some biscuit and snacks samples, demonstrated an acceptable level of safety. This study presents the first health risk assessment for the Serbian population related to dietary acrylamide exposure and provides a valuable scientific basis for the development of strategies aimed at reducing exposure to this contaminant
E-education model on social networks based on big data analytics
No full textПредмет истраживања докторске дисертације је развој модела е-образовања на друштвеним мрежама заснованог на big data аналитици. Централна питања која се испитују у доктoрској дисертацији тичу се могућности анализе активности е-учења на друштвеним мрежама и у систему за управљање учењем применом big data аналитике, као и и унапређења образовног процеса на основу резултата анализе.
У систему е-образовања, који интегрише друштвене мреже и систем за управљање учењем, генерише се велика количина података који могу да се прикупљају и анализирају у реалном времену. Примена техника big data аналитике и анализе друштвених мрежа над образовним подацима омогућава унапређење образовних садржаја, препознавање индивидуалних потреба студената и предикцију исхода учења.
Модел обухвата инфраструктуру, сервисе и активности учења у систему е-образовања на друштвеним мрежама и процес интеграције компонената. Моделом се дефинише тип и начин прикупљања података у систему, избор и интеграција big data инфраструктуре и аналитике.
Евалуација развијеног модела треба да покаже да ли интеграција друштвених мрежа у е-образовање и примена big data аналитике утиче на повећање ефективности, ефикасности и квалитета образовног процеса. Модел ће бити евалуиран на Академији техничко-уметничких струковних студија у Београду на Одсеку Висока школа електротехнике и рачунарства.The subject of the doctoral dissertation is development of е-education model on social networks based on big data analytics. The central problem examined in the doctoral dissertation is the possibility of analysing e-education on social networks and learning management system activities by applying big data analytics and enhancing education process by using the analysis results.
A large amount of data is generated in e-education system that integrate social networks and learning management system. These data can be collected and analysed in real time. The application of big data analytics techniques and social network analysis over educational data, enables educational content development, recognizing individual student' needs and learning outcomes prediction.
The model should involve learning infrastructure, services and activities in e-education system on social networks and process of component integration. Using the model, type and way of data gathering in the system, choice and integration of big data infrastructure and analytics, is defined.
Evaluation of the developed model should show wether integration of social networks in e-education and usage of big data analytics has effect on effectivity increase, efficiency and quality of educational process. The model will be evaluated in the Academy of Technical and Art Applied Studies Belgrade School of Electrical and Computer Engineering
Toxic effects of lead, cadmium, mercury and arsenic mixture in a subacute exposure model for rats: protective effect of probiotics
No full textSažetak
U okviru ove doktorske disertacije ispitivana je subakutna toksičnost smeše metal(oid)a olova (Pb),
kadmijuma (Cd), žive (Hg) i arsena (As) kod pacova, kao i utvrđivanje potencijalnog protektivnog
efekta probiotika na umanjenje štetnih efekata ispitivane smeše metal(oid)a. Studija je sprovedena
na mužjacima Wistar pacova podeljenim u 8 grupa po 5 jedinki od kojih je jedna grupa bila
kontrolna, a 7 grupa je tretirano oralnim putem tokom 28 dana rastućim dozama smeše (MIX:
Pb+Cd+Hg+As), od najnižeg nivoa doze - MIX 1 do najvišeg nivoa doze - MIX 5 (mg/kg t.m./dan)
(Pb: 0,003, 0,01, 0,1, 0,3, 1; Cd: 0,01, 0,03, 0,3, 0,9, 3; Hg: 0,0002, 0,0006, 0,006, 0,018, 0,06; As:
0,002, 0,006, 0,06, 0,18, 0,6), kao i monokomponentnim (5,16×10⁸ CFU/kg/dan; Saccharomyces
cerevisiae var. boulardii) i multikomponentnim probiotikom (8,78×10⁸ CFU/kg/dan;
Saccharomyces cerevisiae var. boulardii, Lactobacillus rhamnosus, Lactobacillus plantarum LP
6595, i Lactobacillus plantarum HEAL9) uz smešu najvišeg doznog nivoa MIX 5. Ispitani su nivoi
toksičnih metal(oid)a u krvi, hematološki, biohemijski, hormonski i parametri redoks statusa, kao i
nivoi bioelementa, dok su organi analizirani histopatološki, uz određivanje toksičnih metal(oid)a,
bioelemenata i redoks parametara, kao i aktivnosti acetilholinesteraze u mozgu. Modelovan je odnos
doza-odgovor i određene su Benchmark doze za toksične efekte pojedinačnih metal(oid)a u smeši.
Značajan porast nivoa Pb, Cd Hg i As u krvi uočen je samo u grupi kojoj je davana najviša doza,
dok je akumulacija toksičnih metala dobijena prvenstveno u jetri, bubrezima, mozgu i femuru.
Izlaganje najnižoj dozi ispitivane smeše metala dovodi do značanjih promena parametara anemije u
krvi, odnosno sniženja broja eritrocita, hemoglobina, hematokrita, i nivoa gvožđa, kao i značajnog
povećanja parametra oksidativnog stresa, O2
.-, PAB i AOPP. Tkiva ispoljavaju različitu osetljivost
na oksidativni stres izazvan rastućim dozama smeše toksičnih metala (Pb, Cd, Hg, As).
Najizraženije promene zabeležene su mozgu i plućima, dok je u jetri dobijeno smanjenje SOD i
GSH pri većini doznih nivoa, dok je MDA snižen u srednjem i dva najviša dozna nivoa u
bubrezima. Dodatno, više doze smeše Pb, Cd, Hg i As značajno su smanjile aktivnost enzima AChE
u mozgu. Histopatološka analiza ukazuje na progresivna oštećenja većine organa pri izloženosti
smeši metal(a), sa uočljivijim efektima pri većim doznim nivoima. Benchmark analiza utvrdila je
postojanje doza-odgovor pri čemu je najniža BMD vrednost za nivo srednjeg volumena eritrocita
(MCV), koji predstavlja kritični toksični efekat za svaki metal(oid) u ispitivanoj smeši i iznosi
3,15E-04 mg Pb/kg t.m./dan, 1,05E-03 mg Cd/kg t.m./dan, 2,10E-05 mg Hg/kg t.m./dan i 2,10E-04
mg As/kg t.m./dan. Probiotici, monokomponentni i mulitkomponentni, uticali su značajno na
smanjenje Cd, Hg i As u krvi, kao i umanjenje pojedinih metala u jetri, bubrezima, plućima i
femuru. Primena oba probiotika dovodi do normalizovanja izmenjenih vrednosti hematoloških
parametara i TAS parametra u krvi, dok primena multikomponentnog probiotika normalizuje
vrednosti TAS u jetri, IMA u plućima, TOS i O2.- u pankreasu i MDA u testisima i aktivnost
enzima acetilholinesteraze u mozgu koji su značajno promenjeni pod uticajem smeše ispitivanih
metal(oid)a. Dobijeni rezultati ukazuju da je mutikomponentni probiotik efikasniji u umanjenju
štetnih efekata ispitivane smeše olova, kadmijuma, žive i arsena u odnosu na monokomponentni
probiotik..In this doctoral dissertation, the subacute toxicity of the metal(oid) mixture of lead (Pb), cadmium
(Cd), mercury (Hg) and arsenic (As) in rats was examined, as well as the determination of the
potential protective effect of probiotics on reducing the harmful effects of the investigated
metal(oid) mixture. The study was conducted on male Wistar rats divided into 8 groups of 5
individuals, one group was a control group, and 7 groups were treated orally for 28 days with
increasing doses of the mixture (MIX: Pb+Cd+Hg+As), from the lowest dose level - MIX 1 to the
highest dose level - MIX 5 (mg/kg bw/day) (Pb: 0.003, 0.01, 0.1, 0.3, 1; Cd: 0.01, 0.9, 3; Hg:
0.0006, 0.018, 0.06; As: 0.002, 0.006, 0.18, 0.6). (5.16×10⁸ CFU/kg/day; Saccharomyces cerevisiae
var. boulardii) and a multicomponent probiotic (8.78×10⁸ CFU/kg/day; Saccharomyces cerevisiae
var. boulardii, Lactobacillus rhamnosus, Lactobacillus plantarum LP 6595, and Lactobacillus
plantarum HEAL9) with the mixture of the highest dose level MIX 5. The levels of toxic
metal(oids) in the blood, hematological, biochemical, hormonal and redox status parameters were
examined, as well as the levels of bioelements, while the organs were analyzed histopathologically,
with the determination of toxic metal(oids), bioelements and redox parameters, as well as
acetylcholinesterase activity in the brain. A dose-response relationship was modeled and
Benchmark doses for the toxic effects of individual metal(oids) in the mixture were determined. A
significant increase in the levels of Pb, Cd, Hg and As in the blood was observed only in the group
that received the highest dose, while the accumulation of toxic metals was obtained primarily in the
liver, kidneys, brain and femur. Exposure to the lowest dose of the examined metal mixture leads to
significant changes in the parameters of anemia in the blood, i.e. a decrease in the number of
erythrocytes, hemoglobin, hematocrit, and iron levels, as well as a significant increase in the
parameters of oxidative stress, O2.-, PAB and AOPP. Tissues exhibit different sensitivity to
oxidative stress caused by increasing doses of a mixture of toxic metals (Pb, Cd, Hg, As). The most
pronounced changes were recorded in the brain and lungs, while SOD and GSH were reduced in the
liver at most dose levels, while MDA was decreased in the middle and two highest dose levels in
the kidneys. Additionally, higher doses of a mixture of Pb, Cd, Hg and As significantly decreased
the activity of the AChE enzyme in the brain. Histopathological analysis indicates progressive
damage to most organs upon exposure to the metal(s) mixture, with more noticeable effects at
higher dose levels. Benchmark analysis established the existence of a dose-response, where the
lowest BMD value for the mean erythrocyte volume (MCV) level, which represents the critical
toxic effect for each metal(oid) in the tested mixture, is 3.15E-04 mg Pb/kg bw/day, 1.05E-03 mg
Cd/kg bw/day, 2.10E-05 mg Hg/kg bw/day and 2.10E-04 mg As/kg body weight/day. Probiotics,
monocomponent and multicomponent, had a significant effect on the reduction of Cd, Hg and As in
the blood, as well as the reduction of certain metals in the liver, kidneys, lungs and femur. The use
of both probiotics leads to the normalization of the changed values of hematological parameters and
the TAS parameter in the blood, while the use of the multicomponent probiotic normalizes the
values of TAS in the liver, IMA in the lungs, TOS and O2.- in the pancreas and MDA in the
testicles and the activity of the acetylcholinesterase enzyme in the brain, which were significantly
changed under the influence of the mixture of investigated metal(oids). The obtained results indicate
that the multi-component probiotic is more effective in reducing the harmful effects of the examined
mixture of lead, cadmium, mercury and arsenic compared to the mono-component probiotic
Population pharmacokinetics of linezolid and correlation with efficacy and safety parameters in patients with acute respiratory distress syndrome on veno-venous extracorporal membrane oxygenation
No full textVeno-venska ekstrakorporalna membranska oksigenacija (V-V ECMO) je sve više zastupljena u zbrinjavanju respiratorne insuficijencije kod kritično oboljelih. Ipak, uticaj ovog oblika terapijske podrške na farmakokinetiku (FK) i doziranje lijekova, posebno linezolida, još uvijek nije dovoljno istražena oblast. Cilj istraživanja je da se procijeni opravdanost primjene viših doza linezolida, u poređenju sa standardnim, kod kritično oboljelih sa akutnim respiratornim distres sindromom (ARDS) uzrokovanim COVID-19, koji su istovremeno bili na V-V ECMO podršci. U istraživanje su prospektivno uključeni bolesnici sa dijagnostikovanim COVID-19 i ARDS uz terapijski model vvECMO podrške sa istovremenom intravenskom primjenom linezolida 600 mg/8 h. Kriterijumi za isključivanje su bili: osobe mlađe od 18 godina, poznata alergija na linezolid, trudnoća, terapijska izmjena plazme u posljednjih 24 sata i nadomjesna bubrežna terapija. Za analizu FK linezolida od svakog bolesnika u stanju ravnoteže lijeka prikupljeno je po šest uzoraka krvi u definisanim vremenskim tačkama. Prikupljeni su svi demografski i klinički podaci neophodni za procjenu uticaja istih na FK linezolida. Nelinearnim modeliranjem kombinovanih efekata razvijen je i validiran populacioni FK model, korišten za Monte Karlo simulacije (5000 virtualnih bolesnika) za generisanje individualnih FK parametara i koncentracijskih profila nakon režima doziranja 600 mg/8 h i 600 mg/12 h. Za procjenu vjerovatnoće postizanja ciljnih vrijednosti farmakokinetičkog-farmakodinamičkog (FK-FD) indeksa (PTA) korišteni su sljedeći targeti: 85%T>MIC, fAUC24/MIC≥80 i fAUC24/MIC≥100. Izračunata je i kumulativna frakcija odgovora (CFR) za različite Gram-pozitivne bakterije. Praćeni su i bezbjednosni aspekti na osnovu promjene nivoa trombocita i razvoj trombocitopenije u odnosu na vrijeme započinjanja vvECMO i linezolida uz istovremenu primjenu. Ukupno su analizirane 53 koncentracije linezolida. Pokazana je visoka korelacija izmjeđu Cmin i AUC24. FK parametri linezolida nisu značajno odstupali u odnosu na ne-ECMO bolesnike i nije pronađen značajan uticaja kovarijata na FK parameter. Nakon doznog režima linezolida 600 mg/8 h predviđeno je jednako i veće postizanje FK-FD ciljne vrijednosti 85%T>MIC za MIC=2 mg/L kod 90%, dok je kod bolesnika koji su primali standardni dozni režim 85%T>MIC zabilježen kod dvije trećine bolesnika. fAUC24/MIC≥80 postignut je kod skoro tri puta većeg broja bolesnika primjenom linezolida 600 mg/8 h za istu vrijednost MIC-a. Prisustvo trombocitopenije sa značajnim smanjenjem broja tombocita zabilježeno je kod ukupno 81,8% bolesnika. Rezultati istraživanja ukazaju da dozni režim linezolida 600 mg/8 h u odnosu na standardni ima prednost kod istovremene primjene vvECMO terapijskog modela za postizanje FK-FD ciljnih vrijednosti linezolida i dobijanja adekvatnog terapijskog odgovora. Terapijsko praćenje koncentracije linezolida u serumu i broja trombocita je neophodno kako bi se zadovoljili bezbjednosni aspekti uz maksimalan terapijski efekat linezolida kod kritično oboljelih sa ARDS na V-V ECMO podršci.Veno-venous extracorporeal membrane oxygenation (V-V ECMO) is increasingly used in managing respiratory failure in critically ill patients. However, the impact of this form of therapeutic support on the pharmacokinetics (PK) and dosing of medications, especially linezolid, remains an insufficiently explored area. This study aims to evaluate the rationale for administering higher doses of linezolid, compared to standard doses, in critically ill patients with acute respiratory distress syndrome (ARDS) caused by COVID-19, who are concurrently supported by V-V ECMO. The study prospectively included patients with COVID-19, ARDS, on a therapeutic V-V ECMO support model with concurrent intravenous administration of linezolid 600 mg every 8 hours. Exclusion criteria were: individuals under 18 years, known allergy to linezolid, pregnancy, therapeutic plasma exchange in the last 24 hours, and renal replacement therapy. For the analysis of linezolid PK, six blood samples were collected from each patient at steady state at defined time points. All demographic and clinical data necessary for assessing their impact on linezolid PK were collected. A population PK model was developed and validated using nonlinear modeling of combined effects, which was then employed in Monte Carlo simulations (5,000 virtual patients) to generate individual PK parameters and concentration profiles after dosing regimens of 600 mg every 8 hours and 600 mg every 12 hours. The probability of achieving target pharmacokinetic-pharmacodynamic (PK-PD) index values (PTA) was assessed using the following targets: 85%T>MIC, fAUC24/MIC≥80 i fAUC24/MIC≥100. The cumulative fraction of response (CFR) was also calculated for various Gram-positive bacteria. Safety aspects were monitored based on changes in platelet levels and development of thrombocytopenia in relation to the initiation of V-V ECMO and linezolid with concurrent application. A total of 53 linezolid concentrations were analyzed. A high correlation was found between Cmin and AUC24. Linezolid PK parameters did not significantly deviate from those in non-ECMO patients, and no significant impact of covariates on PK parameters was found. After the linezolid dosing regimen of 600 mg every 8 hours, the probability of achieving PK-PD target value of 85%T>MIC for MIC =2 mg/L was predicted to be 90%, while two-thirds of the patients on the standard dosing regimen reached 85%T>MIC. fAUC24/MIC≥80 was achieved in nearly three times more patients with the 600 mg every 8 hours linezolid regimen for the same MIC value. Thrombocytopenia occurred in 81.8% of patients with a significant reduction in platelet count. This study indicates that the linezolid dosing regimen of 600 mg every 8 hours, compared to the standard, is advantageous when using the V-V ECMO therapeutic model to achieve linezolid PK-PD target values and adequate therapeutic response. Therapeutic monitoring of linezolid serum concentrations and platelet counts is necessary to satisfy safety aspects while maximizing therapeutic effects in critically ill patients with ARDS on V-V ECMO
Hardware-software support for data stream processing in cloud-fog computing concept
No full textVelika zastupljenost i popularnost IoT uređaja, na svetskom nivou, doveli su do
generisanja ogromne količine podataka, dok se sa druge strane pojavilo Cloud
računarstvo koje poseduje snažnu podrška za procesiranje tih podataka.
Prosleđivanjem podataka od IoT uređaja do Cloud okruženja formirana je jako pogodna
kombinacija, pri čemu se podrazumeva da je obrada podataka sasvim centralizovana.
Motivacija i glavni cilj u ovoj disertaciji jeste pružanje dodatnog uvida u
drugačiji koncept koji se oslanja na decentralizovanu obradu podataka. Kao pogodan
pristup korišćeno je Fog računarstvo kako bi se zadaci procesiranja podataka
preneli sa nivoa Cloud računarstva na nivo bliži izvorima podataka. Kao primer
obrade podataka korišćene su aplikacije koje se zasnivaju na modelima dubokog
učenja, gde je konkretno realizovan model koji pripada tipu konvolucijskih
neuronskih mreža. Navedene aplikacije su služile za klasifikaciju slika u oblasti
pametne poljoprivrede (Smart Agriculture) sa ciljem da se ukaže na mogućnosti
efikasnije detekcije stanja posmatranog objekta i primene odgovarajućih mera.
Predstavljeni su postupci za treniranje i proveru modela klasifikacije, kao i
njegovo konvertovanje u optimizovane modele koji mogu da se izvršavaju na uređajima
sa ograničnim hardverskim resursima. Glavna varijanta je podrazumevala pripremu
modela za izvršavanje na FPGA kolu formirajući tako hardverske akceleratore za
klasifikaciju slika.
Sa primnjenim modelom klasifikacije koji je prenet za izvršavanje u okviru
Fog računarstava dobijene su vrednosti sa prihvatljivim kašnjenjem. Što znači da su
ostvarene obrade podataka u realnom vremenu, dok su redukovane vrednosti
opterećenja servera u pogledu obima prenetih podataka i energetske potrošnje.
U slučaju potrebe za procenom sistema Fog računarstva većeg obima modelovana
je Cloud–Fog struktura korišćenjem iFogSim projekta za simulaciju. Pri tome su
uzeti u obzir prioriteti aplikacija i vršene su procene o prihvatljivom opterećenju
korisničkih aplikacija koje omogućava dobijanje rezultata u realnom vremenu.The large presence and popularity of IoT devices, at the global level, have led to the
generation of a huge amount of data, also Cloud computing has emerged with strong support
for processing that data. By forwarding data from IoT devices to the Cloud environment, a
very suitable combination has been formed, whereby it is assumed that data processing is
completely centralized.
The motivation and main goal of this dissertation is to provide additional insight into a
different concept that relies on decentralized data processing. Fog computing was used as a
suitable approach to transfer data processing tasks from the Cloud computing level to a level
closer to the data sources. As an example of data processing, applications based on deep
learning models were used, where specifically the type of convolutional neural networks was
used for model implementation. The above applications were used for image classification as
a segment of Smart Agriculture to indicate the possibilities of more efficient detection of the
state of the observed object and the application of appropriate measures.
Procedures for training and testing the classification model are presented, as well as its
conversion into optimized models that can be executed on devices with limited hardware
resources. The main variant involved preparing the model for execution on an FPGA circuit,
thus forming hardware accelerators for image classification.
With the pre-trained classification model, executed within Fog computing, acceptable
result delay has been obtained. This means that data processing was achieved in real time,
while the load on the server was reduced such as the volume of transferred data and energy
consumption.
In case of need for evaluation of a larger scale Fog computing system, a Cloud-Fog
structure was modeled using the iFogSim simulation project. Application priorities were taken
into account and estimates were made of the acceptable load of user applications that enable
results in real time
Application of Hybrid Methods Combining Recurrent Neural Networks and Modified Metaheuristics for Time-Series Forecasting
No full textTime-series forecasting is becoming very important in many different industries, including cybersecurity, healthcare, energy, finance and cloud computing. This thesis explores the use of advanced metaheuristic algorithms together with recurrent neural networks (RNNs) to produce hybrid models enhancing the accuracy, stability, and generalizing ability of forecasting. We address problems including nonlinear behavior, seasonality, and noise in time-series data by utilizing the benefits of both Recurrent Neural Networks (RNNs) and metaheuristics.
This work uses an enhanced Moth Flame Optimizer (MFO) to improve the optimization of logistic regression models, which boosts the accuracy of spam email classification. Furthermore, the efficiency of hybrid models in capturing intricate seasonal trends is demonstrated by the use of an Enhanced Sine Cosine Algorithm (ESCA) in conjunction with Long Short-Term Memory (LSTM) networks for wind energy ahead-of-time prediction. With a modified Reptile Search Algorithm, Convolutional Neural Networks (CNNs) and Extreme Gradient Boosting (XGBoost) show potential in hybrid techniques in enhancing the security of IoT networks in the field of cybersecurity.
Moreover, the practical benefits in the field of healthcare are shown by the identification of seizures in electroencephalogram (EEG) data using metaheuristic-optimized recurrent neural networks (RNNs). The study on improving the performance of LSTM networks utilizing the Enhanced Harris Hawks Optimization Algorithm for crude oil price prediction underlines the need to integrate different methodologies in financial modelling. Using modified metaheuristic-optimized LSTM networks for earthquake magnitude prediction emphasizes the several applications of these hybrid approaches in preventing disasters.
Additionally, this study addresses the challenges associated with cloud instance pricing, which are increased by the complex nature of budgeting due to fluctuating demand and changing costs. A modified optimizer is introduced and tested on cloud instance data using multi-headed recurrent architectures, specifically LSTM and Gated Recurrent Unit (GRU) networks. The GRU model, enhanced by this technique, attained remarkable forecasting IV precision, validated through statistical analysis and revealed by using artificial intelligence techniques to clarify feature importance.
The thesis significantly improves the current degree of knowledge in time-series data prediction, resulting in new possibilities for the application of hybrid methods across various areas. By addressing the limitations of previous approaches and using the powers of RNNs and metaheuristics, this work helps predictive modelling and its useful applications to advance