EARSATS-19: In-ear Measurement of Blood Oxygen Saturation in COVID-19 Follow up

Non-invasive ambulatory estimation of blood oxygen saturation (SpO2) has emerged as an
important clinical requirement to detect hypoxia (low blood oxygen) in the delayed
post-infective phase of COVID- 19, where dangerous hypoxia may occur in the absence of
subjective breathlessness. This now commonly termed 'silent' hypoxia presents an immediate
clinical driver.

The COVID-19 pandemic places substantial pressure on clinical resources with ambulatory
patients who were often discharged home to be cared for on virtual wards with daily telephone
consults. Now the initial surge has settled, we are beginning to understand the challenge of
following up patients who have had COVID-19 at volumes that mean doing this in a timely
fashion is not possible (BTS Guidance 6 weeks).

There is mounting concern that the huge number of COVID-19 survivors will be at increased
risk of chronic lung diseases such as pulmonary fibrosis, where early capture of patterns of
low SpO2 could unearth desperately needed insights for preventing disease progression. The
inconveniences, impracticalities and inaccuracies of conventional measurement of SpO2 (the
percentage unit of blood oxygen) using a conventional finger probe for discrete readings are
well established. A user-friendly, remote continuous SpO2 monitoring tool would therefore
allow the rapid identification of any deterioration post discharge and also track longer-term
recovery, allowing the accurate triage of patients to limited follow up resources.

Hearables and In-Ear Photoplethysmography

Reflective photoplethysmography (PPG), which uses a changes in the ratio between red and
infrared light to estimate SpO2, can be measured from any site with skin that has
vasculature, but for SpO2 measurement has conventionally been from fingers or sometimes the
earlobe. The recent interest in the development of Hearables9 has promoted the ear canal as a
preferred site for the measurement of vital signs in digital health technology. Indeed, the
ear canal, which acts as a shield from external electrical noise (resembling a nature-built
Faraday cage), represents a unique opportunity for physiological measurements, due to its
general fixed position relative to the heart (ear-electrocardiogram) and proximity to the
brain (ear-electroencephalogram). Relative to limbs, the head moves less in daily life and
this is one property that has helped establish the feasibility of wearables to detect Ear-ECG
and Ear-EEG.

Unlike the finger and ear lobe PPG signal, in-ear PPG has shown that the ear canal, owing to
being a narrow cavity with consistent blood supply, offers a signal which is stable and
resistant to changes in blood volume occurring during hypothermia . This attribute has
established the ear canal as a preferred site for accurate core body temperature measurement.
In-ear PPG has also been shown to be far more sensitive than earlobe and finger PPG to
amplitude variations that arise from respiration, thus allowing for a better measurement of
respiratory rate. Furthermore, a significant delay has been evidenced between earlobe pulse
oximetry and pulse oximetry on the hand or the foot for detection of hypoxemia (low levels of
blood oxygen)..

Previous work led by Professor Mandic's group in the Faculty of Engineering showed that in a
study of 14 healthy volunteers in-ear PPG compared with the current standard of care - the
finger clip -- the in-ear sensor shows non-inferiority and potential superiority with regard
to detection of a mean blood oxygen delay (the time it took from detecting minimal blood
oxygen change in the ear to detecting minimal blood oxygen in the finger) reduction of 12.4
seconds, thus being faster at identifying acute desaturation.

RATIONALE FOR CURRENT STUDY

This project sets out to establish the feasibility of SpO2 measurement from the ear canal as
a convenient site for monitoring in the post COVID-19 follow up patient population, and
perform a comprehensive comparison with the right index finger - the conventional clinical
measurement site. The initial phase of the study will start with a total of n = 60 patients;
30 patients, who will be attending hospital as part of their routine clinical post COVID-19
follow up, which already includes SpO2 monitoring and a six minute walk test (6MWT). This
study will also recruit a further 30 patients with chronic lung disease also attending for
these routine outpatient investigations.

The continuous monitoring and rapid detection of desaturation can provide earlier warning of
deterioration or detailed tracking of an improving trajectory. Current alternative monitoring
strategies are reactive and all revolve around the classical finger or earlobe clip, which
are less sensitive and present a number of challenges for continual monitoring. Living daily
life with a finger clip, or even an earlobe clip, is problematic with much higher signal
noise from movement artefacts with the finger clip compared to the ear bud placed within a
stable head position.

From a patient perspective, the ear bud form factor that will be used in this study is
familiar to them, being similar to an in-ear headphone, and will be much less intrusive to
daily activities than a finger clip. Therefore the potential for continuous monitoring during
the active day and the night time is much more user-friendly, and vital for the younger and
potentially more active COVID-19 patients. Continuous monitoring would allow the development
of protocolised management pathways to provide increasing levels of supported self-care to
patients, freeing up vital resources during future pandemics.

The data generated in this pilot feasibility study would allow scaling to prepare for larger
research projects in any subsequent surge discharge and follow up. The next stage would move
to outpatient monitoring, with ambulatory patients going home with the device and the
incorporation of further COVID patient populations such as inpatients being monitored for
deterioration, where these devices could provide accurate continuous monitoring of SpO2,
detecting any deterioration before they develop breathlessness and their condition becomes
life threatening. This project therefore directly complements the NHS' strategic priority for
early supported discharge monitoring for COVID-19 positive patients and follow up for
long-term complications.

The COVID-19 follow up clinic cohort (n = 30) is ideally placed to evaluate this technology;
from a quantitative point of view, a significant number of these patients will likely
experience desaturation during the 6MWT, serving as an opportunity to test the device's
ability to detect this. Qualitatively, since many of these patients will have either had
hospital admissions or home monitoring -- and therefore lived experience of SpO2 monitoring
-- they are uniquely qualified to offer valuable input on user experience. Similarly, the
chronic lung disease cohort (n = 30), attending for routine OP baseline SpO2 measurement and
6MWT are also well-suited to this study.

In the long term this technology will also prove an invaluable device for the home monitoring
of respiratory patients. Beyond the current pandemic there exists a general quest for more
personalised health data, such that pulse oximetry measurement of capillary oxygen saturation
(SpO2) will likely expand into both the clinical and consumer market of wearable health
technology in the near future.