National and Kapodistrian University of Athens
Ατομικό γενετικό υπόβαθρο για αυξημένη προδιάθεση σε μόλυνση ή/και λοίμωξη από ιούς και εξατομικευμένη θεραπευτική προσέγγιση
Abstract
dc:descriptionIntroduction: The COVID-19 pandemic presents a unique opportunity to understand the impact of individual genetic backgrounds on susceptibility to infections. Predominant risk factors such as age and comorbidities do not fully account for the considerable variability observed in infection susceptibility, disease severity, and patient outcomes. Prediction tools derived from Machine Learning (ML) offer a valuable means of elucidating the differences in outcomes among COVID-19 patients. Aim: This study aims to investigate the influence of genetic polymorphisms, along with individual demographic and clinical characteristics, on disease outcomes in COVID-19 patients. Additionally, it examines the significance of machine learning in disease prediction. By leveraging individual genetic profiles, clinicians can devise tailored treatment regimens for each patient. Materials and Methods: The study included 323 unvaccinated patients with confirmed SARS-CoV-2 infection. DNA extraction was conducted using specialized equipment, and genotyping was performed with specifically designed primers and probes. Data collected from questionnaires and medical records were input into several machine learning classification algorithms. The algorithm with the highest predictive accuracy was selected to develop a web-based disease prediction and outcome tool. Results: Among the patients, 30.3% exhibited mild symptoms, 32.8% were hospitalized in standard care, and 36.8% required intubation and admission to the ICU. The overall mortality rate was 19.5%, exclusively among intubated ICU patients. Notably, male gender and older age were predominant factors associated with hospitalization and adverse outcomes. Arterial hypertension emerged as the comorbidity with the most statistically significant difference between moderate and ICU patients (OR=6.46, p<0.001). Other significant comorbidities predictive of ICU admission included obesity (OR=3.25, p=0.001), diabetes mellitus (OR=4.06, p=0.001), chronic obstructive pulmonary disease (COPD) (OR=11.94, p=0.001), and malignancy (OR=4.08, p=0.001). Dyslipidemia and atrial fibrillation also exhibited a minor aggravating role (p=0.009 and p=0.003, respectively). Conversely, thyroid disease, bronchial asthma, neurological disease, and coronary artery disease did not demonstrate a statistically significant effect on outcomes. In terms of laboratory findings, the mean ferritin level upon admission was three times higher in patients who developed severe disease compared to those who did not require ICU care. A statistically significant likelihood was observed for ICU patients to present with ferritin levels exceeding 800 ng/mL (normal range: males: 12 to 300 ng/mL, females: 12 to 150 ng/mL). Minor statistically significant differences were also noted for D-dimer and glucose levels at the time of hospital admission. The genetic polymorphisms analyzed were found to have either protective or aggravating effects, particularly concerning the TMPRSS2, ACE and HFE genes. With regard to the TMPRSS2 gene, the polymorphisms rs469390, rs2070788, rs2298661 and rs1239970 demonstrate statistically significant differences when comparisons are made with regard to the outcome of the patient in the hospital and the probability of intubation and admission to an ICU, particularly when comorbidities such as arterial hypertension and dyslipidemia are present. In cases where dyslipidemia is present, the rs2070788 homozygous A/A + G/G genotype has been observed to increase the likelihood of ICU admission (OR = 0.26, p = 0.0223). Homozygous genotypes of the rs1800562 polymorphism (HFE gene) appear to have a similar deleterious effect. In particular, homozygous G/G and the G allele of rs1800562 in the gene comparison: patients hospitalised in clinic - patients without hospitalisation advocate for hospitalisation (OR=10.739, p=0.026). Conversely, heterozygous G/A protects the patient, leading to mild disease not requiring hospitalisation (OR=0.093, p=0.026). The predictive algorithm demonstrated strong capability in identifying patients who did not require hospitalization and those at risk of mortality among intubated patients; however, improvements are needed for predicting intermediate stages (hospitalization with survival). A total of 111 independent characteristics were evaluated, with age, sex, hypertension, obesity, and cancer comorbidity showing significant associations with severe COVID-19. Conclusions: Genetic polymorphisms implicated in SARS-CoV-2 infection play a crucial role in determining patient severity and outcomes, even in the presence of other risk factors. The individualized treatment of patients, coupled with the utilization of risk prediction algorithms, is integral to effective therapeutic strategies. The prognostic tool developed through this study may facilitate personalized treatment approaches and could encourage vulnerable populations to pursue vaccination, despite its non-mandatory status.
Degree
thesis:*- Grantor dc:publisher
- National and Kapodistrian University of Athens
- Year dc:date
- 2025
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Christodoulou, Avgi
- Χριστοδούλου, Αυγή
Subjects
dc:subject × 29- SARS-CoV-2
- Πολυμορφισμοί
- Γενετικό υπόβαθρο
- Μηχανική μάθηση
- Φύλο
- Υπέρταση
- Παχυσαρκία
- TMPRSS2
- rs469390
- rs2070788
- rs2298661
- rs1239970
- ACE
- rs1799752
- HFE
- rs1800562
- rs1799945
- Polymorphisms
- Genetic background
- Machine learning
- Gender
- Hypertension
- Obesity
- Βασική Ιατρική
- Ιατρική και Επιστήμες Υγείας
- Φαρμακολογία και Φαρμακευτική
- Basic Medicine
- Medical and Health Sciences
- Pharmacology and Pharmacy
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/58792