1.3. Applications of Biosensor
1.3.1. Health Care
Measurement of Metabolites The initial impetus for advancing sensor technology came from health care area, where it is now generally recognized that measurements of blood gases, ions and metabolites are often essential and allow a better estimation of the metabolic state of a patient. In intensive care units for example, patients frequently show rapid variations in biochemical levels that require an urgent remedial action. Also, in less severe patient handling, more successful treatment can be achieved by obtaining instant assays. At present, the list of the most commonly required instant analyses is not extensive. In practice, these assays are performed by analytical laboratories, where discrete samples are analyzed, frequently using the more traditional analytical techniques.
Market Potential. There is an increasing demand for inexpensive and reliable sensors to allow not only routine monitoring in the central or satellite laboratory, but also analysis with greater patient contact, such as in the hospital ward, emergency rooms, and operating rooms. Ultimately, patients themselves should be able to use biosensors in the monitoring and control of some treatable condition, such as diabetes. It is probably true to say that the major biosensor market may be found where an immediate assay is required. If the cost of laboratory maintenance are counted with the direct analytical costs, then low-cost biosensor devices can be desirable in the whole spectrum of analytical applications from hospital to home.
Diabetes. The 'classic' and most widely explored example of closed-loop drugcontrol is probably to be found in the development of an artificial pancreas. Diabetic patients have a relative or absolute lack of insulin, a polypeptide hormone produced by the beta-cells of the pancreas, which is essential to the metabolism of a number of carbon sources. This deficiency causes various metabolic abnormalities, including higher than normal blood glucose levels. For such patients, insulin must be supplied externally. This has usually been achieved by subcutaneous injection, but fine control is difficult and hyperglycaemia cannot be totally avoided, or even hypoglycaemia is sometimes induced, causing impaired consciousness and the serious long-term complications to tissue associated with this intermittent low glucose condition.
Insulin Therapy. Better methods for the treatment of insulin-dependent diabetes havebeen sought and infusion systems for continuous insulin delivery have been developed. However, regardless of the method of insulin therapy, its induction must be made in response to information on the current blood glucose levels in the patient. Three schemes are possible (Fig. 1.6), the first two dependent on discrete manual glucose measurement and the third a 'closed-loop' system, where insulin delivery is controlled by the output of a glucose sensor which is integrated with the insulin infuser. In the former case, glucose has been estimated on 'finger-prick' blood samples with a colorimetric test strip or more recently with an amperometric 'pen'-size biosensor device by the patient themselves. Obviously these diagnostic kits must be easily portable, very simple to use and require the minimum of expert interpretation. However, even with the ability to monitor current glucose levels, intensive conventional insulin therapy requires multiple daily injections and is unable to anticipate future states between each application, where diet and exercise may require modification of the insulin dose. For example, it was shown that administration of glucose by subcutaneous injection, 60 min before a meal provides the best glucose/insulin management.
Artificial Pancreas. The introduction of a closed-loop system,
where integrated glucose measurements provide feedback control on a pre-programmed
insulin administration based on habitual requirement, would therefore relieve
the patient of frequent assay requirements and perhaps more desirably frequent
injections. Ultimately, the closed-loop system becomes an artificial pancreas,
where the glycaemic control is achieved through an implantable glucose
sensor. Obviously, the requirements for this sensor are very different
to those for the discrete measurement kits. As summarized in Table 1.4,
the prolonged life-time and biocompatibility represent the major requirements.
1.3.2. Industrial Process Control
Bioreactor Control. Real-time monitoring of carbon sources, dissolved gases,. in fermentation processes (Fig. 1.7a) could lead to optimization of the procedure giving increased yields at decreased materials cost. While real-time monitoring with feedback control involving automated systems does exist, currently only a few common variables are measured on-line (e.g. pH, temperature, CO2, O2)) which are often only indirectly related with the process under control.
Fig. 1.6 Schemes for insulin therapy.
Table 1.4. Seven requirements for an implantable glucose sensor.
Three different methods of controlling a bioreactor are:
On-Line Control. Method 3 is most desirable, which allows the process to follow an ideal pre-programmed fermentation profile to give maximum output. However, many problems exist with on-line measurements including in situ sterilization, sensor life-time, sensor fouling, etc. Some of the problems can be overcome if the sensor is situated so that the sample is run to waste, but this causes a volume loss, which can be particularly critical with small volume fermentations.
Off-Line Control. Although Method 3 may be the ultimate aim, considerable advantage can be gained in moving from Method 1 to Method 2 giving a rapid analysis and thus enabling finer control of the fermentation. The demands of the sensor are perhaps not as stringent in Method 2 as in Method 3.
Benefits of Control. The benefits which are achievable with process-control technology are considerable:
1.3.3. Military Applications
Dip Stick Test. The requirement for rapid analysis can also be anticipated in military applications. The US army, for example, have looked at dipstick tests
Fig. 1.7. Comparison of sensing modes: (a) bioreactor;
(b) clinical applications; (c) military or environmental monitoring.
Table 1.4 Summary of potential applications for biosensors
based on monoclonal antibodies. While these dipsticks are stable and highly specific (to Q-fever, nerve agents, yellow rain fungus, soman, etc.) they are frequently two-step analyses taking up to 20 min to run. Such a time lapse is not always suited to battlefield diagnostics; the resulting consequences are suggested in Fig. 1.7(c).
A particularly promising approach to this unknown hazard detection seems
to be via acetylcholine receptor systems. It has been calculated that with
this biorecognition system, a matrix of 13-20 proteins are required to give
95% certainty of all toxin detection.
1.3.4. Environmental Monitoring
Air and Water Monitoring. Another assay situation which may involve a considerable degree of the unknown is that of environmental monitoring. The primary measurement media here will be water or air, but the variety of target analytes is vast. At sites of potential pollution, such as in factory effluent, it would be desirable to install on-line real-time monitoring and alarm, targeted at specific analytes, but in many cases random or discrete monitoring of both target species or general hazardous compounds would be sufficient. The possible analytes include biological oxygen demand (BOD) which provides a good indication of pollution, atmospheric acidity, and river water pH, detergent, herbicides, and fertilizers (organophosphates, nitrates, etc.). The survey of market potential has identified the increasing significance of this area and this is now substantiated by a strong interest from industry. The potential applications of biosensors are summarized in Table 1.4.
Tuning to Application. The potential for biosensor technology is enormous and is likely to revolutionize analysis and control of biological systems. It is possible therefore to identify very different analytical requirements and biosensor developments must be viewed under this constraint. It is often tempting to expect a single sensor targeted at a particular analyte, to be equally applicable to on-line closed-loop operation in a fermenter and pin-prick blood samples. In practice, however, the parallel development of several types of sensor, frequently employing very different measurement parameters is a more realistic.