Biosensors can effectively be used to monitor multiple metabolites such as glucose, lactate, ATP and drugs in the human body. Continuous monitoring of these metabolites is essential for patients with chronic or critical conditions. Moreover, this can be used to tune the dosage of a drug for each individual patient, in order to achieve personalized therapy. Implantable medical devices (IMDs) based on biosensors are emerging as a valid alternative for blood tests in laboratories. They can provide continuous monitoring while reduce the test costs. The potentiostat plays a fundamental role in modern biosensors. A potentiostat is an electronic device that controls the electrochemical cell, using three electrodes, and runs the electrochemical measurement. In particular the IMDs require a low-power, fully-integrated, and autonomous potentiostats to control and readout the biosensors. This thesis describes two integrated circuits (ICs) to control and readout multi-target biosensors: LOPHIC and ARIC. They enable chronoamperometry and cyclic voltammetrymeasurements and consume sub-mW power. The design, implementation, characterisation, and validation with biosensors are presented for each IC. To support the calibration of the biosensors with environmental parameters, ARIC includes circuitry to measure the pHand temperature of the analyte through an Iridiumoxide pH sensor and an off-chip resistor-temperature detector (RTD). In particular, novel circuits to convert resistor value into digital are designed for RTD readout. ARIC is integrated into two IMDs aimed for health-care monitoring and personalized therapy. The control and readout of the embedded sensor arrays have been successfully achieved, thanks to ARIC, and validated for glucose and paracetamol measurements while it is remotely powered through an inductive link. To ensure the security and privacy of IMDs, a lightweight cryptographic system (LCS) is presented. This is the first ASIC implementation of a cryptosystem for IMDs, and is integrated into ARIC. The resulting system provides a unique and fundamental capability by immediately encrypting and signing the sensor data upon its creation within the body. Nano-structures such as Carbon nanotubes have been widely used to improve the sensitivity of the biosensors. However, in most of the cases, they introduce more noise into the measurements and produce a large background current. In this thesis the noise of the sensors incorporating CNTs is studied for the first time. The effect of CNTs as well as sensor geometry on the signal to noise ratio of the sensors is investigated experimentally. To remove the background current of the sensors, a differential readout scheme has been proposed. In particular, a novel differential readout IC is designed and implemented that measures inputcurrents within a wide dynamic range and produces a digital output that corresponds to the -informative- redox current of the biosensor.