Clinical development of a new patient-dedicated arterial blood analyser

Frequent measurement of blood gases is a familiar routine in the operating theatre, intensive care unit and other critical care settings. Over the last two years we have been involved in the clinical development of Proxima, a device that will challenge the way that healthcare workers carry out arterial blood analysis and could consign walking a syringe of blood to the blood gas machine to history.

The Proxima system is based on a miniaturised disposable blood gas and electrolyte analyser (microanalyser) that is integrated into the patient’s arterial line and can be used for multiple measurements (Figure 1). When a blood measurement is required, blood is drawn up the arterial line into the device. This will give a result within about 60 seconds, permitting rapid feedback of changes in the cardiopulmonary system or metabolism. The blood can then be returned to the patient within a closed system. Because measurement will be available at the bedside or in theatre, there should be a reduction in the time and cost associated with taking a sample when compared with a desktop blood gas machine or laboratory. This should be particularly useful at times when a critical care area is busy or dealing with multiple unstable patients at one time.

As well as providing rapid readings, this approach should reduce both infection risk to the patient and blood loss. The volume of blood that is lost during blood gas analysis is a particular problem in neonates or paediatric patients.  Proxima should be of benefit to staff by reducing the infection risk to them and the time taken to draw and process a sample; they will no longer need to leave the bedside to obtain a result.

The technology behind the Proxima is dependant upon two silicon chips. One chip is devoted to processing of the outputs from the second chip – the sensor array chip. Both chips sit within a custom‑designed flow cell through which blood samples are taken, as well as used for calibration and flush solutions (Figure 2).

The sensor chip uses miniaturised versions of the electrochemical sensors found within a conventional blood gas analyser. Each sensor is of the order of 0.25–0.4mm in diameter and are fabricated using a custom designed machine that places a microscopic drop of liquid containing a few nanolitres of material into a well on the chip.  

The sensor chip can measure pH, pCO2, pO2, potassium, sodium, ionised calcium, glucose, lactate and haematocrit. From these, standard calculated variables, such as bicarbonate and base excess, can be reported. There is flexibility over the panel of sensors that is used and it will be possible to swap or add sensors to suit different patient groups or clinical applications.  The company plans to start with a simple critical care panel and add further sensors over time.

The sensor and electronics chips are packaged into a microanalyser device that is small (30 mm x 30 mm x 12 mm) and light, weighing just 20g. As with a pressure transducer, the microanalyser is attached by a cable to a bedside monitor, used for recording and displaying results.

Through the development process, the Proxima system has undergone a series of clinical studies as steps from an in-vitro to a full ex-vivo measurement device. Initially these focused on investigating and validating aspects of patient safety, moving on to demonstrating laboratory analyser performance in a patient-attached device. All trials were undertaken at the West Suffolk Hospital, Bury St Edmunds, UK. 

In November 2010 we completed the first clinical study of the Proxima system attached to human subjects. The primary endpoints were to demonstrate safety and to investigate how the device performed when attached to a patient and exposed to un-heparinised venous blood. 

The study used four healthy volunteers and took place at the Clinical Trials Unit at West Suffolk Hospital. The device was connected to a venous access device for a maximum of two hours. Up to seven blood samples were drawn into the device every 20 minutes. Each successive sample had a longer dwell time, in which blood remained within the flow cell. This was followed by flushing with a Plasma-lyte flush. Signals from pH, potassium, pCO2, pO2, glucose, HCT and temperature sensors were monitored.   

No concerns over safety were raised and no deterioration or sensor failure was noted during the trial. On examination, the Proxima microanalyser did not contain any significant fouling from blood within the sample chamber.  The monitored readings from the device were not calibrated as blood gas readings, but the signals obtained demonstrated no unexpected or spurious signals from patient movement or grounding issues.

In this first human exposure, the device proved to have no 

safety concerns and to be suitable for further investigations in a patient‑attached mode.

The second clinical study was also in healthy volunteers using venous blood. Having established in the first study that drawing un-heparinised blood into the Proxima led to no undesirable effects, we set out to investigate the effect of extended exposure, by drawing up to 20 blood samples into the device, representative of the number of samples that could be tested using the final product.

Thirteen healthy participants were recruited to the second study. The Proxima devices were fitted to a venous access for up to four hours, during which time up to 20 blood samples were drawn into the device and flushed to waste. A rudimentary calibration system was implemented to start to investigate the effects of un-heparinised blood exposure on analytical performance by comparison to another blood gas analyser, the i-STAT®.   

We were able to further demonstrate safety and usability within a clinical environment. The level of agreement between the Proxima and the i-STAT® across all parameters was satisfactory. This trial provided us with confidence to proceed to a study using routine surgical and critical care patients. 

The third clinical study was an open, non-randomised, study to provide a comparison of data obtained from the Proxima 2 device with those obtained from a blood gas analyser in patients who require the insertion of an arterial line. This was carried out at two centres in West Suffolk and Queen Elizabeth Hospital, Birmingham. The study was conducted to obtain the method comparison data required for regulatory approval for the device from the FDA and European regulatory bodies. The participants were seventeen patients undergoing major operations, including upper and lower gastrointestinal tract, neuro- and liver surgery as well as patients in the intensive therapy ward. The Proxima functioned as an in vitro device with both calibrants and blood being drawn to waste. The study was completed in November 2011 and the Proxima 2 system was CE marked as a diagnostic product in December 2011.                                                                                                                                           

This trial provided quantitative data over a wider physiological range than could be achieved with healthy volunteers. It also extended the time envelope of device exposure to blood and drugs within the circulation and provided a realistic picture of how the device performs in a clinical setting.

Following PROX003, Sphere Medical plans to carry out further development, with the goal of obtaining regulatory approval for the Proxima as a medical device in which blood samples and calibration solutions will be re-infused into the patient’s circulation. The current arterial line flush bag will be replaced by a dual purpose flush and calibration fluid in a gas impermeable bag. This will integrate measurement and calibration within a simplified version of the routine workflow steps of drawing an arterial blood sample.

The clinical trials to date have focused on demonstrating the safety and performance of the Proxima system.  Looking forward, there is significant potential for Proxima to impact the delivery of critical care in West Suffolk in a number of areas.

Fast results mean the ability to quickly make decisions and also check that those decisions are having the desired effect. Although this is not required for much of the time, this could have a real impact on efficiently stabilising patients or weaning from mechanical ventilation. 

The volume of blood loss that is associated with blood gas analysis is a particular problem in some patient groups, for example in neonates or paediatric patients.  

The incidence of bloodstream infections is fortunately very low but the effects can be devastating and expensive. By avoiding opening the arterial line for each measurement, the risk of introducing infection by this route is reduced. Furthermore, the system reduces the amount of blood handling by staff with its attendant risks.

At West Suffolk we have a blood gas analyser sited in the ITU which is also used for samples from theatre. The Proxima system should be of benefit to staff by reducing the time taken to take and process a sample. They will no longer need to leave the bed side to obtain a result or run samples in from theatre. 

 

This micro-analyser could be of great usefulness in challenging environments such as hospital transfers and retrieval and transfer of trauma cases. It could also function as a back-up to conventional analysers.

The Proxima system is a patient‑dedicated arterial blood analyser for use in the intensive care unit and operating room. It comprises a disposable multi-parameter microanalyser integrated into an existing arterial or venous line on the patient. All blood is returned into the patient after the measurements and the introduction of the Proxima into clinical practice could signal the end for transporting blood-filled syringes around the hospital.