Abstract
Introduction
Analysis of continuous physiological signals measured by pulse oximetry during sleep may provide a novel method to assess cardiovascular (CV) risk. The sleep period appears to be a particularly useful window for assessment.
Methods
Subjects (n=520, 346 males, age 55.0±13.4 yrs, BMI 29.9±6.1 kg/m2) were referred to five sleep centers in Germany and Sweden. CV risk factors were assessed and subjects were classified by the ESC/ESH risk matrix into five separate risk classes. The autonomic state indicator (ASI) algorithm extracted patterns of the peripheral pulse wave and SpO2 signal by amplitude and time/frequency analysis from the overnight digital photoplethysmographic recording and computed a CV risk score (range 0-1, ≥0.5 equals to high risk). Nine derived parameters (irregular pulse, RCDC, pulse rate variability, pulse wave variability, pulse propagation time, oxygen desaturations, duration of periodic symmetric desaturations and baseline SpO2) were used to determine the final score.
Results
In the validation group (n=390), the developed algorithm detected high CV risk (ESC/ESH scores 4 and 5) patients with a sensitivity of 74.5% and specificity of 76.4%. The area under the ROC curve was 0.80. The ASI CV risk score was elevated in patients with an already established CV endpoint (MI and/or stroke, n=50) compared with all other patients (0.73±0.27 vs. 0.42±0.34, p<0.001).
Conclusions
The ASI technique appears to provide a possibility to detect increased CV risk from a recording of physiological signals during sleep. The technique – based on a modified pulse oximeter – may be useful in both sleep and cardiovascular medicine.
- © 2012 ERS