*4.1. Confirmation of Positive BC Results by NGS*

In only 3 of 25 cases were relevant pathogens detected by both BC and NGS. BC confirmed NGS results in two cases, yet was unable to detect one clinically relevant bacterium in one patient: *S. epidermidis* was isolated in two sets of BCs drawn from a CVC in one patient, whereas NGS detected three different bacteria. These BC results were considered catheter-related BSI according to clinical documentation and, consequently, the CVC was removed, and ongoing antimicrobial treatment continued.

A putative lack of pathogens in NGS analysis might be explained by a technical analysis algorithm. As described in a previous study, isolates might have been identified by NGS but not reported due to a low read count and stringent threshold settings during analysis [13]. Thus, low concentrations of pathogens prevent further analysis and might lead to disagreement between the two methods [24]. Further research will be necessary to address this question and possibly improve analysis algorithms.

#### *4.2. Defining Antimicrobial Therapy Using NGS*

This study examined the potential impact of NGS results on the choice of antimicrobial treatment. In three cases, four antimicrobial therapies were initiated following positive NGS results. In addition, NGS found pathogens in twelve cases, in which BC remained negative. In three of those cases, empirical therapy was deemed appropriate, and no adjustment seemed necessary. In some cases, the identification of additional pathogens in the bloodstream might have even wider implications, as demonstrated in one case. While conventional methods only found *S. aureus* in respiratory samples and BC remained negative, NGS detected *S. aureus* DNA in the bloodstream. This could have warranted a prolonged duration of antimicrobial therapy and further diagnostic procedures, such as echocardiography to assess for endocarditis. Although this study was not designed to demonstrate a significant benefit of NGS regarding therapy improvement, it can be hypothesized that additional NGS diagnostics may lead to the optimization of antimicrobial therapy in certain critically ill COVID-19 patients. However, given the high cost of NGS compared to standard diagnostics and the still unclear overall limited therapeutic benefit, the indication for the use of NGS should be carefully considered.

#### *4.3. Contamination*

An unusually high rate of contamination by CNS in BC diagnostics was observed in this study: CNS were detected in 16% of all specimens and in 57% of positive specimens, which is significantly higher than the usual false-positivity rate reported in our annual pathogen and resistance statistics. In two extensive reviews, performed before SARS-CoV-2 emerged, the overall contamination rates of BC were notably lower and ranged from only 0.6 to 12.5% [25,26]. During the pandemic, a general increase in contamination rate in specimens from COVID-19 individuals was observed, presumably caused by a high workload, newly trained staff, wearing full personal protective equipment, and time pressures [27–30]. Since this study was conducted during the peak of the second and third waves of the pandemic, these aspects may also have been major contributors to this study. In contrast, no CNS were identified in any specimen by NGS. This circumstance may reflect methodological differences. NGS analysis only targets cell-free DNA released by degradation processes or immune system interaction. In cases of contamination, bacterial cells remain mostly intact, avoiding DNA release, which consecutively leads to negative NGS results. This clearly differs from the BC methodology, in which vital bacteria are cultivated followed by positive signaling.

#### *4.4. Value of NGS in the Diagnosis of Fungal Infections*

Critically ill COVID-19 patients are at increased risk of developing secondary fungal infections such as COVID-19-associated pulmonary aspergillosis and candidemia [31–36]. In this study, one isolate of *C. parapsilosis* was found only by NGS. However, the result was not considered clinically relevant. According to current guidelines, the diagnosis of candidemia could have warranted additional interventions, such as CVC removal and ophthalmological examination [37]. However, recent studies reported inconsistent results regarding the benefit of NGS in the detection of systemic fungal infections and larger prospective studies will be needed to assess this question [13,17,18,24].

#### *4.5. Value of NGS in the Diagnosis of Viral Infections*

Reactivation of latent viruses is common in patients with sepsis and may be even more frequent in patients with severe COVID-19 [38–43]. However, tests for viral infections in clinical routines are lacking. In our study, the implementation of additional NGS analysis led to the identification of 8 viruses in 25 patients, which would have been missed by standard diagnostics. However, in the absence of a clinically apparent viral disease in most cases, the clinical relevance and therapeutic implications of these results remain unclear.

In two patients, treatment with aciclovir was initiated following the identification of HSV-1 by NGS. However, data regarding the prognostic implications of HSV reactivation in patients with COVID-19 is inconclusive, while the only randomized controlled trial on aciclovir treatment in non-COVID-19 patients found no benefit on morbidity or mortality [31,39,40,44,45]. Whether viral reactivation of HSV or CMV in critically ill patients reflects true viral disease (and, therefore, represents possible treatment strategies), or is merely indicative of an immunocompromised state remains controversial and requires further investigation.
