Sepsis is a severe and potentially life-threatening condition resulting from an uncontrolled immune response to infection. Rapid identification of pathogens is essential for enhancing clinical outcomes. While conventional blood culture diagnostics remain the gold standard in clinical practice, they are limited by lengthy turnaround times and decreased sensitivity in low-biomass infections. This study examined how the yield and purity of DNA extracted from human whole blood that has been spiked with bacteria or synthetic microbial DNA affect sequencing-based microbial detection using Oxford Nanopore MinION technology. Manual and semi-automated extraction workflows utilising the QIAamp DNA Mini Kit were assessed for bacterial cultures, and the Quick-DNA Miniprep kit was employed for samples spiked with blood. Subsequent taxonomic and antimicrobial resistance analyses were conducted using Kraken2, EPI2ME, and BV-BRC workflows. Semi-automated extraction has enhanced workflow efficiency and diminished operator-dependent variability compared to manual extraction methods. Gram-negative E. coli demonstrated higher DNA yields and more consistent purity than Gram-positive S. epidermidis, which can be attributed to differences in cell wall structure and lysis requirements. An updated statistical evaluation and outlier handling revealed variability across different sample types. Sequencing of blood spike-in samples facilitated taxonomic classification and the detection of antimicrobial resistance genes, including fosA, mph(K), blaZ, and several efflux-associated genes. However, samples with low bacterial biomass or a high background of host DNA exhibited reduced microbial read recovery and limited detection of AMR genes. Additionally, the presence of occasional low-level taxa, such as S. aureus, likely indicated contamination or misassignment of reads. The study emphasises that DNA extraction quality significantly affects the effectiveness of nanopore-based microbial detection. It highlights both the potential and challenges of rapid, culture-independent metagenomic diagnostics for bloodstream infections, stressing the need to optimise extraction methods to improve accuracy and reliability in clinical use.