Having a biosensor inside the body is of immense value in the diagnosis, management and treatment of patients. Implanting a biomarker-specific biosensor can reduce the ‘lag-time’ that exists due to taking serum samples of the patients, and sending them to the lab and getting the results back. Cutting down on this ‘lag-time’ would allow for a quicker response, eventually improving patient outcomes. The continuous measurements of samples in-vivo will also allow for tracking of specific trends or patterns which could be used to identify pathologies before they arise [20]. The ability to monitor patients through non-invasive means, rather than requiring constant invasive procedures like blood draws, is highly beneficial for both clinicians and patients. It provides clinicians with vital information while allowing patients greater bodily autonomy and independence from repetitive invasive interventions. Moreover, implanting pressure sensors is also of great value as it can allow for continuous monitoring of changes in pressure and volume and offer quick feedback control when necessary [21].

Both types of biosensors (for detecting biochemical parameters and physiological parameters) may yield better patient outcomes if they were to be implanted and coupled with MCSDs. Currently, there is a lack of studies on the coupling of biomarker-specific biosensors with MCSDs. The existing pre-clinical research offers means to detect cardiac biomarkers in-vitro without coupling to MCSDs. This research can be applied in the future for in-vivo testing and coupling with MCSDs. As for pressure sensors, the existing research does offer pre-clinical testing of coupling the sensor with LVADs, in both in-vitro and in-vivo settings. Moreover, studies such as the INTELLECT 2-HF study [22] in the USA and MONITOR-HF [23] in the Netherlands have investigated the use of CardioMEMS monitoring system to measure pulmonary artery pressure in patients with LVADs patients, affirming the feasibility and safety of combining pressure sensors with LVADs. Both reported a significant reduction in heart failure hospitalisations and improvement in quality of life [22, 23]. However, large prospective clinical trials on pairing pressure sensors with LVADs are yet to be conducted.

Recent advancements in developing in-vitro and in-vivo models to test new heart failure devices like biosensors and measure parameters like intracardiac pressures have provided a foundation for future clinical research to build on. For example, Andrew Malone et al. have described a mock circulatory loop that could both mimic the cardiac cycle and features two independently controlled cardiac chambers to fully simulate the blood flow and pressures of the LA and the LV [24]. Meskin et al. have recently published their work on a novel mock circulatory loop that can mimic systolic and diastolic functions of both the LA and the LV [25]. In addition to the challenges of developing large in-vivo disease models [26], considerations around animal welfare in MedTech research has further motivated the development of these in-vitro platforms. Such robust mock circulatory test beds allow for prototyping and evaluation of MCSDs while implementing the 3Rs principles — replacing, reducing, and refining the use of animals in research.

Specific biomarkers can detect the ongoing issues with LVADs in-situ, for example, hemolysis that may lead to multiorgan failure. Timely detection of the rise in certain biomarkers may help in early intervention and prevention of any catastrophic event. Rodgers IL et al. retrospectively reviewed the levels of carboxyhemoglobin and methemoglobin in two patients with LVADs who developed significant haemolytic anemia and the levels of both the biomarkers were > 2%, indicative of the ongoing hemolysis [27].

Biomarker-specific biosensors function to detect specific target markers, such as interleukin-10 (IL-10) and tumour necrosis factor-α (TNFα). In general, they are fabricated by immobilising a specific bioreceptor onto a suitable substrate, which is placed on a transducer that recognises the interaction between the biomarker and the bioreceptor [28, 29]. The transducer is then able to convert this biochemical interaction into an electrical signal that may be processed and displayed, or trigger a feedback control mechanism, if necessary.

Two studies, both by Baraket et al. [30, 31], offered exemplary mechanisms to detect cytokines that appear due to post-LVAD implantation inflammation. Although LVADs are composed of biologically inert material, there still remains the risk of an inflammatory response at the site of the implantation of these foreign materials [30]. LVADs, particularly pulsatile flow LVADs, may lead to selective reduction in CD4 + T cells and increased apoptosis of CD4 + and CD8 + T cells. Such defects in adaptive immunity increase the risk of infection and sepsis [32]. This immune response hinders the functioning of an LVAD, reducing its efficacy, and eventually leading to worse patient outcomes.

In the first study [30], a fully integrated electrochemical biosensor was developed to detect IL-1b and IL-10 in-vitro, at minute concentrations. This biosensor was made by placing gold microelectrodes (MEs) onto a silicon substrate. The MEs were then functionalised with anti-IL-1b and anti-IL-10. This biosensor as a whole unit was then incubated in solutions with varying concentrations of IL-1b, IL-10, as well as IL-6 (just to determine the selectivity of the biosensor). The outcome of this study showed that the developed biosensor was highly sensitive for detecting the cytokines at minute concentrations ranging from 1 to 15 pg mL−1 [30], and selective with no interference with the other cytokines present. These findings offer an opportunity for the detection of other pro- and anti-inflammatory cytokines by using more MEs. The biosensor itself can conduct multiple measurements simultaneously, offering real-time information about many different parameters. The biosensor developed is biocompatible and could be implanted alongside LVADs. The measured cytokine levels are used to infer the inflammatory state of patients and adjust the management plan accordingly.

Another label-free biosensor developed by Baraket et al. [31] aimed to detect TNFα in patients with inflammation following LVAD implantation. The study was done in-vitro but offers mechanisms for future work to be done in-vivo. The biosensor consists of a gold surface deposited on a flexible polyimide substrate [31]. It was then functionalised by immobilising the corresponding antibodies to TNFα onto the gold electrodes. This whole biosensor unit was then placed in between the two parts of a Teflon cell to perform the electrochemical measurements. To determine the selectivity, the biosensor was exposed to different concentrations of TNFα, alongside IL-1 and IL-10. The novel biosensor developed was highly sensitive in the detection of TNFα at a range of 0.1 pg/mL to 0.5 ng/mL [31]. There was some interference as a result of non-specific binding, however, this was minimal as good selectivity was observed in the presence of other cytokines. The application of this biosensor to measure different inflammatory cytokines in plasma samples of patients after LVAD implantation is promising. Given that it is biocompatible, implanting this biosensor alongside the LVAD may yield positive patient outcomes over time.

Another category of biosensors includes those designed to detect physiological parameters, such as pressure and volume (See Fig. 2 for a schematic of how they work). Changes in the blood pressure or volume following an LVAD implantation can lead to dangerous consequences, such as arrhythmias, suction events, or even reduced efficacy of the LVAD [12, 33]. As such, it is important to accurately measure the physiological changes within the body in real-time to offer quick response, and hence, better patient outcomes. Four studies reviewed here offer mechanisms to detect the physiologic parameters in patients with LVAD implantation and are tested in a variety of in-vitro, in-vivo, and human settings. Two studies also explore the potential for autonomous feedback control.

Illustration to show left atrial pressure sensor and ramped speed test

Hubbert et al. [34] describe a Titan left atrial pressure (LAP) monitoring system with an implanted Micro-Electro-Mechanical Systems (MEMS) pressure sensor. The system communicates wirelessly with an external monitor to provide a real-time LAP monitoring. Precise measurement of LAP can guide therapy to optimise the balance between pump preload and pump speed, as illustrated in Fig. 2 [34]. For this reason, the aim of this study was to report the first human application of this wireless pressure sensor in patients who are undergoing LVAD support. The Titan wireless monitoring system consists of two parts: an implantable telemetric sensor, which contains the MEMS sensor and an external monitor [34]. Four patients underwent surgery for LVAD implantation, and the LA sensor was introduced on the border between the LA and the right upper pulmonary vein. To validate the accuracy of the Titan system, simultaneous LA pressure measurements were recorded using both the wireless Titan sensor and a standard fluid-filled catheter placed in the LA for reference. The two pressure readings were compared in the first 1–2 postoperative days showing excellent correlation with only 1–3 mmHg difference between the Titan system and the fluid-filled catheter. To test the function of the Titan LAP sensor, the recordings of the four patients were monitored in the ICU as well as on the ward. Three of those patients were sent home while remotely monitoring the LAP [34]. To determine the LAP at which the LVAD pump functions at optimal speed, echocardiographic ramped speed testing was performed in each patient after discharge from the ICU and at postoperative follow-up some weeks later [34]. This study was successful in accurately measuring LAP values in patients with LVAD implantation. The pressures recorded while the three patients were discharged at home were accurately measured and the signals were sent through the internet to the hospital where a nurse would analyse them daily. The battery-free operation of this sensor also allows for long-term function without the need for maintenance or battery replacement [34]. Through echocardiographic ramped speed testing, the optimal LVAD pump speed for each patient was achieved at a LAP of 8 mmHg [34]. This indicates that daily monitoring of LAP can provide a very useful tool for assessing the pump function. With no adverse effects reported, this study demonstrates promising results for future clinical implementation [34]. Future work should include developing feedback mechanisms to allow automatic adjustment of pump speed [34].

Fritz et al. [33] describe the development of a semi-conductor strain gage inlet pressure sensor that can detect a suction event in the inlet of a Tesla style LVAD. A Tesla style LVAD is different than other conventional pumps in that it is a shear flow pump that utilises a series of rapidly rotating discs to impart momentum in the fluid. The aim of the study was to create and design a sensor, which could monitor pressure changes throughout the cardiac cycle, and control the LVAD accordingly [33]. The sensor developed is made from titanium and the transducer is incorporated into the lumen of the titanium islet connector of the LVAD. Testing of this system included finite element analysis, sensitivity testing, drift testing, hysteresis testing, and step testing before testing the system in in-vitro and in-vivo settings [33]. In-vivo testing included a 5-h study in which the Tesla style LVAD with the inlet pressure sensor was conducted on a 105 kg calf. The inlet pressure signal from the calf was input into a custom control algorithm to automatically change the pump speed [33]. When a suction event (defined as a large drop in inlet pressure) was detected, the control algorithm instructed the pump to reduce the pump speed [33]. Drift testing revealed that over a four-week period, offset drift varied between − 180 mmHg and 140 mmHg. Unfortunately, drift of this degree precludes the use of this sensor to accurately measure physiological blood pressure [33]. Although the system could not accurately detect physiological blood pressure, it has displayed success in detecting suction events and reducing the pump speed based on the control algorithm input [33].

Zhou et al. [12] describe an optical biosensor, based on the Fabry-Pérot interferometer principle, which is fabricated onto an LVAD. The Fabry-Pérot interferometer utilises two optically flat and partially reflecting surfaces arranged in parallel to form a chamber. By using a flexible diaphragm for one of the reflective surfaces, the pressure difference can be determined by the fringes generated from the optical output changes that which occur when the distance between the mirrors change [12]. The aims of this study were to develop an LVAD pressure sensor that can detect pressure changes up to 100 mmHg, respond to the changes within 10 ms, achieve a sensitivity of within 2 mmHg, be suitable for long term implantation, and not interfere with the hemodynamic environment inside the LVAD. The biosensor developed is placed on the inlet of the LVAD and detects the LV pressure using MEMS technology in patients with heart failure [12]. Parylene-C is the material used in this sensor as it is suitable for long-term implantation and is recognised as being a good structural material. The Parylene-C diaphragm is integrated directly onto the inlet shell of the LVAD and pressure change is determined by the number of fringes. The biosensor displayed success in detecting various changes in pressure within the range of 100 mmHg with a detection resolution of 1 mmHg. With larger parylene-C diaphragm sizes, the biosensor becomes more sensitive to pressure changes [12]. The sensor successfully records the pressure changes within 2 ms. This study demonstrates the development of a biosensor, which can be coupled to LVADs, to detect changes in pressure; however, further development still needs to be made in order to allow for feedback control of the LVAD. LVAD pump speed algorithms are yet to be developed that would change the speed based on the pressure signals. Ideally, both the LVAD and the biosensor power supply should be integrated. However, this remains a challenge [12].

Qawasma and Daud [35] describe an LVAD that is designed to pump blood from the LV to the aorta using an integrated Arduino Microcontroller (AMC). The aim of this study was to design a system for heart failure patients so that the LVAD pump speed could be changed based on the present physiologic status of the patient detected by an integrated ECG sensor. The ECG sensor relays its signal to the AMC through a conditional circuit and the system is designed to analyse the R-wave. The R-wave provides information about two key factors: the heart’s ability to provide the body with sufficient amount of blood and the required flow rate that needs to be achieved by the system [35]. The system adjusts the pump speed based on the time between the two successive R-waves [35]. The microcontroller analyses the number of beats and then alters the pump speed accordingly. The pump is also fitted with inflow and outflow valves, which control the direction of the blood flow [35]. A safe-mode has also been integrated into the system; so if the ECG signal is lost, or, in case of system failure, the LVAD will function at 75 beats per minute (BPM), which is predetermined in the microcontroller [35]. This system is able to respond accurately to the conditions detected from the ECG feedback. The integrated ECG successfully relays information in order to control the speed of the LVAD motor. The system accurately detects tachycardia, bradycardia and normal HR, and adjusts the flow rate accordingly [35].