Diabetes affects 1 in 11 people globally and requires monitoring of blood glucose levels as a part of its treatment. Initial blood glucose monitoring devices involved finger prick testing at regular intervals daily, a self-monitoring strategy reliant on patient compliance. Advances in technology have facilitated continuous glucose monitors (CGMs), (semi-) implantable devices that can monitor glucose levels in vivo and transmit the data to mobile applications. Majority of commercial CGMs are electrochemical in nature and meet the requirements for a commercial sensing implant. There are three main factors that limit CGMs for in vivo use – namely oxygen dependence, low molecular weight (LMW) materials and foreign body response (FBR) – which negatively affect the lifetime and accuracy of CGMs. This thesis aims to detail strategies to combat these issues, building on previous work performed in the field of enzymatic electrochemical glucose sensors. Chapter 2 details the use of design of experiments (DoE) to optimise enzyme electrode components – Osmium complex-based redox polymer, commercial glucose oxidising enzyme (glucose oxidase, GOx) and crosslinker (polyethylene glycol diglycidyl ether, PEGDGE). Previous work established high current and stability of a similar system which also incorporated acid treated multiwalled carbon nanotubes (MWCNTs) as a nanosupport. The MWCNTs enable high currents and surface coverages, but the quantities required were quite high, which can be detrimental for in vivo applications. In this chapter, the grafting of enzyme to nanosupport was carried out to allow minimisation of MWCNT amounts while retaining high currents and operational stability. DoE facilitated the determination of electrode component amounts for optimal current density and stability. The optimised enzyme electrodes show a current density of 3.18 ± 0.30 mA cm−2 , representing a 146% increase in current density in 50 mM phosphate-buffered saline at 37 °C containing 5 mM glucose when compared to similar systems where enzyme and nanosupport are not grafted to each other. Using the predictive DoE model, component amounts were then modified to minimise the quantity of the enzyme-MWCNT nanoconjugate, resulting in a biosensor which showed similar electrochemical behaviour and current density to the optimised system while using 93% less of the nanoconjugate Commercial GOx shows excellent behaviour for glucose oxidation but uses oxygen in its half reaction to regenerate. This is problematic as in vivo oxygen levels can fluctuate resulting in errors in measurement. Additionally, enzyme regeneration by oxygen reduction gives hydrogen peroxide as a product. Peroxide can cause enzyme instability as it oxidises the methionine residues of the enzyme, decreasing its activity. To combat this use, GOx was replaced with engineered cellobiose dehydrogenase (CDH) in Chapter 3. CDH is a dehydrogenase and thus does not use oxygen in its half reactions and has been modified to selectively choose glucose as its substrate. The enzyme electrodes comprising osmium complex-based redox polymer, CDH and PEGDGE were optimised with DoE, while a direct electron transfer (DET) based system was also optimised through conventional methods. The resulting sensors had sensitivities in the same order of magnitude as those in literature. Most importantly, sensor signals showed no difference in the presence and absence of oxygen. The sensors derived from CDH were shown to be specific to glucose over other clinically relevant in vivo sugars and selective, i.e., capable of glucose sensing in the presence of interfering species present in complex media. While no individual species is classified as an interferent in complex media they seemed to exhibit a cooperative effect resulting in a minimisation of current (43%). This is usually overcome with the use of polymer coatings. However, polymer coatings themselves lead to reduced sensor signals on their application, due to the formation of a diffusion barrier. Chapter 4 focuses on the design of polymer coatings to enable protection against biofouling while retaining current density. This was done by designing polymer coatings with a compatible epoxy crosslinking moiety on the polymer backbone that could crosslink with the redox polymer used in the sensing layer. The polymers selected in this study were zwitterionic in nature because of their inherent ability to minimise biofouling. Protein adsorption and cell adhesion studies, using fibrinogen and fibroblasts respectively, allowed a screening to select the most effective of the synthesised polymers for biofouling resistance. This poly(2- methacryloyloxyethyl phosphorylcholine-co-glycidyl methacrylate (MPC)-type polymer showed similar biofouling resistance compared to commercial polymer Lipidure with ~50% reduction in fibrinogen adsorption and ~80% reduction in fibroblast adhesion. When used as coatings for glucose biosensors fabricated in Chapter 3, MPC showed ability to resist protein adsorption while retaining current density similar to a non-coated system with 1.5-fold increase in sensitivity. MPC polymers showed ability to impart biofouling resistance while maintaining current signals. Nevertheless, their ability to resist LMW materials was not proved. In Chapter 5, a series of different protective strategies were explored to determine which approach would be the best to protect from biological and LMW interferences. Enzymatic scavenging using enzymes that target LMW species was investigated but showed inefficient scavenging, likely due to low enzymatic activity. Polymer multilayer approach with successive anionic and MPC polymer layers was utilised and showed the best potential. MPC as the outerlayer showed biofouling resistance whereas an anionic interlayer ([Poly(1-vinylimidazole-co-4-styrene sulfonic acid sodium salt hydrate], P(VI1 -SSNa1 )) inhibited anionic LMW species such as uric acid and ascorbic acid which cause interference. Moreover, due to the compatible crosslinking sites, the layers intermix at the boundary between them, minimising diffusional barrier and allowing current signals similar to a non-coated system. This multilayer protection system extends linear range and enables higher current and stability than a non-coated system in 50 mM phosphate-buffered saline and artificial plasma. Chapter 6 summarises the results from Chapter 2-5 and highlights the significant conclusions that can be drawn from these results. Future directions that can improve on the strategies in this thesis are also discussed.