Martin Jahn, Hans-Georg Ortlepp, Martin Schädel, Olaf Brodersen and Kurt J. G. Schmailzl
Non-invasive blood pressure monitoring in the ear
Cardiovascular diseases, including arterial hypertension, are among the most frequent causes of death worldwide (1). In Germany, despite the positive trend of recent years, only about half of the population suffering from hypertension has blood pressure values in the target range, with one in five having no knowledge of their disease (2). Driven by this, in the last two decades the opinion has increasingly prevailed that the diagnosis and treatment of hypertension should be based on ambulatory blood pressure measurements, especially 24h monitoring (3; 4; 5; 6). This is opposed by low patient acceptance, since wearing the upper arm cuff over a longer period of time is often perceived as disturbing.
A solution can be offered by miniaturized sensor systems integrated into accessories, clothing or medical aids, which are already used to monitor selected vital parameters. Wireless communication with central servers, mostly via mobile devices, enables the data to be evaluated promptly by a telemedicine centre or the attending physician himself. This enables preventive and therapeutic measures to be taken at an early stage as well as optimal monitoring of high-risk patients. In addition, the patient receives a better overview of his health condition and can influence his therapy more independently, as is already the case with diabetics.
Motivated by this development, a demonstrator for a comfortable, cuff-free sensor for blood pressure monitoring was developed within the frame of the digilog research project (7). It is attached to the outer auditory canal, more precisely the tragus, via an earmould. With photoplethysmography (PPG), a widespread, purely optical measuring method for the non-invasive recording of vital parameters is applied. In addition to the planned application, the heart rate, its variability and blood oxygen saturation can be monitored. Thereby, the central measuring position on the head allows critical changes in oxygen saturation to be detected more than 30 seconds earlier than is possible, for example, on the extremities.
PPG is based on the light absorption of blood in the near infrared and visible range, from which conclusions can be drawn about changes in blood volume in the vessels. This allows the blood pressure waves to be detected which spread in the vascular system as a result of cardiac activity. The speed at which the pressure waves propagate depends on the stiffness of the vessels, which in turn is a function of the blood pressure – the quantity actually sought. The pulse wave velocity is usually determined over the time between two measuring points on the body. For the in-the-ear system, however, only one sensor is required, since a patented method is used here that is based on the extraction of a transit time parameter from the measured PPG signals. These result from the superposition of the originally emitted pressure wave and the part of it reflected at the arterial branches. Since the latter has covered a greater distance in the body, the pulse wave velocity can in principle be determined from the difference in transit time between the two partial waves.
The transition of the pressure wave from the aorta, via the peripheral arteries, into the finer blood vessels, however, leads to a change in the measured signal form. Therefore, the original waveform of the central pulse pressure wave must first be reconstructed from the PPG data using suitable models. This requires high signal quality, which has been achieved by optimizing the optical components and electronics of the ear sensor system. In addition, the exact measuring position should be individually adapted to the ear of the test person. Once the reconstruction has been completed, the obtained pulse curves can be broken down into their components. The relative change in blood pressure can be determined from the resulting differences in running time. A one-point calibration using an established reference method is required in order to obtain the course of the absolute blood pressure. To check the system, this should be repeated at longer intervals.
In self-tests with provoked blood pressure changes, it has already been shown that the blood pressure curve determined by an ear sensor is in good agreement with the reference data of a cuff-based measurement. However, a quantitative evaluation of the system requires further clinical studies. A first step in this direction was taken in cooperation with the cardiology department of the Ruppiner Kliniken. During a cardiac catheter examination, the in-the-ear sensor was tested parallel to the invasive blood pressure measurement. After performing a single referencing of the ear sensor via the catheter measurement, at 226 mmHg systolic, blood pressure changes between 150 and 250 mmHg could be continuously tracked. The deviation from other point wise catheter measurements was 10 mmHg (RMSE) or 15 mmHg (maximum). Details can be found in the digital appendix.
The measuring position at the ear has clear advantages over conventional blood pressure monitors. The wearing comfort is considerably higher because the extremities are free of measuring equipment. In addition, the measurement of the cuff-free sensor is hardly noticed after only a few minutes, especially when custom-made earmoulds are used. The central position in the external auditory canal ensures reliable blood circulation and thus stable signals, whereas the blood supply to the extremities depends heavily on the ambient temperature. In addition, there are comparatively few vascular branches between the aortic arch – the point of superposition of both partial waves – and the ear, so that the signal reconstruction and the resulting differences in running time can be determined much more precisely than, for example, on the finger. Furthermore, the results indicate that the ear measurement is more robust against changes in body position and movement artifacts. The only external sources of interference are speaking and chewing movements, the influence of which can be reduced by an adapted signal evaluation. The size of the overall system is mainly limited by the data logger currently in use. For testing purposes, a miniaturized version was realized, which is attached behind the ear and has an interface for wireless transmission. In this case, however, continuous operation is limited to a maximum of one hour, in contrast to approximately eight hours with the previous solution. To ensure long-term monitoring, new concepts for energy optimization and intermittent recording must be developed in future work.
Altogether, the in-the-ear sensor represents a promising alternative for long-term blood pressure monitoring, which simultaneously enables the collection of further important vital parameters. As a supplement to other mobile sensors, ECG or blood glucose monitors, an even more comprehensive picture of the patient’s state of health can be acquired, which will further increase the benefit of eHealth systems as a diagnostic and therapeutic tool.