Institutional members access full text with Ovid®

The use of overnight pulse wave analysis for recognition of cardiovascular risk factors and risk: a multicentric evaluation

Sommermeyer, Dirka,b; Zou, Dinga; Eder, Derek N.a; Hedner, Jana; Ficker, Joachim H.c; Randerath, Winfriedd; Priegnitz, Christinad; Penzel, Thomase; Fietze, Ingoe; Sanner, Berndf; Grote, Ludgera

doi: 10.1097/HJH.0000000000000039

Objectives: Conventional methods for cardiovascular disease risk stratification are based on quantification of recognized risk factors or assessment of biomarkers during the wake period. We evaluated an algorithm on the basis of a photoplethysmographic pulse wave recording during sleep for cardiovascular risk assessment.

Methods: Five hundred and twenty individuals (346 men, age 55.0 ±13.4 years, BMI 29.9 ± 6.1 kg/m2) with suspected sleep apnoea were randomly recruited at five sleep centres. Individual cardiovascular risk scores were calculated in accordance with established cardiovascular risk matrixes (ESH/ESC, Framingham, SCORE, PROCAM scores). A digital photoplethysmographic pulse wave signal was continuously recorded during the night using an oximeter sensor. An algorithm based on eight separate hypoxic and pulse wave derived parameters was trained in 130 individuals and validated in 390 individuals for low/high cardiovascular risk classification.

Results: All derived parameters were associated with elevated ESH/ESC risk in univariate analysis and five in the multiple logistic regression model [discrimination index C = 0.8, Chi-square (7) = 69, P <0.0001]. The combined algorithm detected high-risk patients (validation set, ESH/ESC risk classes 4 and 5) with a sensitivity, specificity, positive predictive value and negative predictive value of 74.5, 76.4, 69.0 and 81.0%, respectively. Significant associations were also found for the Framingham, SCORE and PROCAM scores. The computed risk scores in individuals with/without (n = 34/356) a previous history of cardiovascular event (myocardial infarction, transitory ischemic attack or stroke) were 0.71 ± 0.27 and 0.42 ± 0.34 (P <0.001), respectively.

Conclusion: Parameters derived from modified pulse oximetry during sleep may provide information on cardiovascular function. Combined signal analysis may be used for recognition of individuals with established cardiovascular risk in a sleep laboratory cohort.

aDepartment of Pulmonary Medicine, Sahlgrenska University Hospital, Gothenburg, Sweden

bSchool of Engineering and Architecture, SRH University, Heidelberg

cDepartment of Respiratory Medicine, Allergology and Sleep Medicine, Klinikum Nuernberg, Nuernberg

dDepartment of Pulmonary Medicine, Bethanien Hospital, Solingen

eDepartment of Cardiology, University Hospital Charité, Berlin

fDepartment of Pulmonary Medicine, Agaplesion Bethesda Krankenhaus Wuppertal, Wuppertal, Germany

Correspondence to Ludger Grote, MD, PhD, Center for Sleep and Wake Disorders, University of Gothenburg, Box 421, SE 405 30 Gothenburg, Sweden. Tel: +46 31 3421000; fax: +46 31 825207; e-mail:

Abbreviations: ASI, autonomic state indicator; ESH/ESC, European Society of Hypertension/European Society of Cardiology; OR, odds ratio; PASD, periodic and symmetric desaturations; PPT, pulse propagation time; PR-I, pulse rate acceleration index; PROCAM, Prospective Cardiovascular Muenster; PR_SpO2-I, difference between pulse rate acceleration index and hypoxia index; PWA-I, pulse wave attenuation index; RRPO, respiratory-related pulse oscillation; SCORE, Systematic Coronary Risk Evaluation; SpO2-I, hypoxia index; SpO2, oxygen saturation; TimeBelow90, time under 90% oxygen saturation

Received 20 June, 2013

Accepted 2 October, 2013

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Website (

© 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins