Tamara van Donge

Chapter 3 52 Opportunities: how to close these knowledge gaps In the current digital landscape, computerized protocols and electronic medical records can help diagnose and provide warnings based on abnormal vital signs, which can alert the physician. The implementation of a uniform series of diagnostic criteria and a protocol driven therapy can improve and unify making the correct diagnosis. 48 In today’s clinical practice the diagnosis of neonatal sepsis is supported by results from e.g. nonspecific markers (C-reactive protein) and the identification of potential causative organisms by molecular methods (polymerase chain reaction). 3 Non-specific screening measures such as cytokines (IL-6, IL-8, TNF-α, IFN-ɣ) and cell surface antigens are also considered as biomarkers in sepsis. 3 A recent study showed that presepsin (P-SEP) can be an accurate marker for identifying neonatal sepsis. 49,50 Serum P-SEP decreases during the course of antibiotic treatment suggesting its possible role in monitoring therapeutic response during follow-up periods. 51 Although newer rapid diagnostics can offer possibilities they are not yet validated and ready for clinical use. Furthermore, in recent decades novel techniques have provided information about the process of target site distribution. Microdialysis provides direct measurement of concentrations of unbound antibiotics at the site of action when the site of infection is not the bloodstream. 32,52 Measurement of the free (unbound) drug concentration in the interstitial fluid is better correlated with the antimicrobial efficacy, compared to concentration measurements in plasma. The principle is based on a concentration difference between two fluid compartments across a semi-permeable membrane. Microdialysis offers a useful sampling tool which can quantify the unbound antibiotic at infection sites. 27 In contrast to the MIC breakpoint, which reflects the sensitivity of a pathogen at only one time point, bacterial kill-curves can offer more detailed information about the killing activity as a function over time and might even be used to identify the presence of resistant subpopulations. 53,54 Bacterial kill-curves are very labor intensive and until the method is automated and widely implementable, this approach might not be practical. 54 In addition, dose adjustment of antibiotics is crucial, especially since the administration of an inefficacious (too low) dose is likely to be problematic for patient outcome and antibiotic resistance. Although the physiological mechanisms of augmented renal clearance are not yet (fully) understood, it has not only been observed in critically ill adults, but also in pediatric patients. 44,55 Consequently, there is a real need for reliable measurements of renal function in the pediatric patient population. Overall, the search for dosing regimens and drug combinations for effective antibacterial activities can benefit tremendously from modeling and simulation strategies when there is on-going communication and exchange between research groups and clinicians. 56