Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Oct 20.
Published in final edited form as: Physiol Meas. 2020 Jun 30;41(6):065004. doi: 10.1088/1361-6579/ab8e12

Characterization of Lung to Finger Circulation Time in Sleep Study Assessment: The Multi-Ethnic Study of Atherosclerosis

Younghoon Kwon 1, Sara Mariani 2, Sneha R Gadi 1, David R Jacobs Jr 3, Naresh M Punjabi 4, Michelle L Reid, Ali Azarbarzin 2, Andrew D Wellman 2, Susan Redline 2
PMCID: PMC9583690  NIHMSID: NIHMS1751488  PMID: 32344384

Abstract

Introduction

Lung to finger circulation time (LFCT) measured from sleep studies may reflect underlying cardiac dysfunction. We aimed to examine the distribution of LFCT in community-dwelling men and women in order to better understand the factors determining LFCT between and within subjects.

Methods

We included participants of the Multi-Ethnic Study of Atherosclerosis (MESA) Sleep with polysomnography-based evidence of sleep apnea (defined by apnea hypopnea index>15/hr). In a randomly selected subset of the analytical dataset, we tested an automated LFCT measurement method against the visual method. Using the automated method we then scored LFCTs from all eligible respiratory events for all included participants. A multiple regression model was constructed to determine factors independently associated with average LFCT across subjects. We also explored factors that are associated with LFCT within subjects using linear mixed-effect models.

Results

In a subset of the cohort (N=39) there was a high correlation in average LFCT obtained by automated and visual methods (r = 0.96). In the analysis of 596 participants, men [19.6 (2.8)] (vs. women [17.9 (2.7) sec], p<0.0001) and older age (>69 (vs. ≤ 69) had longer average LFCT (19.4 [2.8] vs. 18.5 [2.9] sec, p<0.0001). These associations persisted in multivariable analysis. No association was found with body habitus. Within subject analysis revealed trivial associations between apnea/hypopnea duration, apnea (vs. hypopnea), nadir O2 saturation and sleep stages (NREM vs. REM) and individual LFCT.

Conclusion

Automated LFCT measurement was highly correlated with visual-based LFCT measurement. In this group of community dwelling adults, male sex and older age were associated with higher average LFCT.

Keywords: circulation, polysomnography, sleep study, sleep apnea

Introduction

Sleep disordered breathing (SDB) is an important cardiovascular risk factor. However, most commonly used clinical indices, apnea hypopnea index (AHI), despite its value in quantification of respiratory event, do not differentiate well which individuals with SDB are most likely to have poor cardiovascular outcomes. Commonly measured, but currently underutilized, parameters from sleep studies may better characterize the cardiovascular risks associated with SDB.

“Circulation time” (Ct) can be derived from polysomnography (PSG) studies, and has been shown to correlate with cardiac output and thus may be a marker for underlying cardiovascular dysfunction or alternatively, represent a physiological index reflective of cardiovascular susceptibility.(Kwon et al., 2015) The Ct was first introduced as a clinical test to assess heart failure (HF) nearly a century ago (Blumgart, 1931, Blumgart and Weiss, 1927, Blumgart and Weiss, 1928). While rarely being used in the current era, its core concept that the Ct is mainly influenced by cardiac output and central blood volume is still utilized in modern medicine particularly with advanced imaging techniques.(Shors et al., 2003, Muller et al., 2000, Puskas and Schuierer, 1996, Bae, 2010, John et al., 2016) Similar to intravenous dyes that were used as indicators for Ct measurement in the past, O2 can be used as an indicator allowing for Ct measurement in the presence of dynamic O2 change.(Kasravi et al., 1998, Wexler et al., 1946) Based on this concept, in the presence of apneic events accompanying desaturation, Oxygen (O2) Ct can be derived from measuring the time interval between the start of recovery breath and the nadir point of desaturation (O2 travel time). A delay in lung-to-chemoreceptor O2 Ct is both a hallmark and an important pathophysiologic basis of central sleep apnea/Cheyne Stokes Respiration that is commonly observed in HF.(Sands and Owens, 2015, Oldenburg et al., 2007) A number of studies have suggested prolonged sleep study-derived Ct is an index of cardiac dysfunction in patients with HF and that lung to periphery Ct is inversely related to cardiac output.(Hall et al., 1996, Kasravi et al., 1998, Kwon et al., 2014, Kwon et al., 2015) In examining this in patients with HF, lung to ear Ct based on central apneas was used considering the proximity of ear to carotid body.(Hall et al., 1996, Ryan and Bradley, 2005) However, in clinical practice, O2 saturation is typically measured in finger tips for convenience. Furthermore, the distribution and characteristics of the sleep study-derived Ct based on much more common OSA events have not been previously examined in general population without HF.

Moreover, its use and clinical application has been limited by lack of our understanding of how this measure varies across populations and how it may vary within individuals across the sleep period. The main purpose of this study was to further our understanding of PSG-derived Ct by describing variation in community dwelling men and women without known cardiac impairment and with moderate or more severe SDB.

Methods

Study population

The Multi-Ethnic Study of Atherosclerosis (MESA) is a multi-site cohort study of community-dwelling men and women aged 45–84 years without known cardiovascular disease (history of coronary heart disease, HF, or stroke) at enrollment in 2000–2002.(Bild et al., 2002) A subset of participants participated in the MESA Sleep ancillary study (October 2010 through February 2013) and underwent PSG shortly following MESA exam 5 (April 2010 through February 2012).

An overnight in-home PSG (Compumedics Ltd., Abbostville, Australia) was conducted and scored as detailed before (Chen et al., 2015). All recordings were centrally scored by certified scorers masked to all other data. The AHI was defined as the sum of all apneas plus hypopneas with a ≥ 4% desaturation. For each relevant airflow signal), quality was prospectively evaluated by the MESA Sleep Reading Center staff (data available from the National Sleep Research Resource; sleepdata.org).(Dean et al., 2016) We discarded recordings with nasal air flow signal quality rated lower than 3, meaning that the signal had to be of good quality for 50% or more of the sleep time.

Since the O2 Ct measurement method requires presence of a respiratory (apnea or hypopnea) with subsequent desaturation, we only included participants with sufficient quantity of respiratory events as indicated by AHI ≥15/hr. We also limited the analytical sample to individuals without reduced cardiac systolic function, as indicated by left ventricular ejection fraction <50% by cardiac magnetic resonance imaging MESA exam 5. The research protocols were approved by the Institutional Review Boards at each participating institution and all participants gave written informed consent.

Measurement of Lung to Finger Circulation Time

Ct is defined as the time it takes for a given indicator to reach the place of detection, and its measurement is based on the “indication-dilution” method. Lung to finger Ct (LFCT) can be measured from data routinely collected in sleep studies by using peripheral O2 saturation as an indicator, the lung as an “injection” site, and the fingertip as a detection site. Thus, by measuring time interval between initiation of breathing (O2 entry) after apneic episode and the nadir point of O2 saturation (SpO2) oxygen transport time (i.e., Ct) can be estimated (Figure 1). Using conveniently sampled randomly selected PSGs from participants included in this study (N=39), we first developed and tested an algorithm that, starting from visually annotated event start and endpoints, automatically refines the detection of the end of apnea/hypopnea using the nasal pressure signal and detects the following nadir SpO2 using the finger pulse oximeter signal, thereby calculating LFCT. Since events that do not show abrupt delineation between apnea/hypopnea and hyperpnea, thereby making it difficult to determine the starting point of LFCT, for both automated and visual measurement, only respiratory events with distinct demarcation of termination of apnea/hypopnea (or initiation of hyperpnea) followed by clear and monotonic desaturation ≥4% and re-saturation were chosen. LFCT measurement was restricted to obstructive respiratory events (i.e., events with reduced airflow but respiratory effort on thoracic and/or abdominal inductance bands). Average LFCT was derived from all eligible LFCT measures per participant.

Figure 1.

Figure 1.

Open in a new tab

Illustration of lung to finger circulation time (LFCT) measurement.

For validation of the automatic algorithm, we automatically scored all eligible LFCT measures per recording in the conveniently sampled subset of 39 participants (range: 8~275 events per participant). Then, following automated scoring, LFCT was reviewed and edited by one investigator. We eliminated non-physiological outlier values that were caused by measurement error (LFCT>50 sec). We computed intra-individual correlations between measures scored with the two approaches; then, we averaged LFCT over those events in a given study and compared automatically estimated average LFCT values (in sec) with visually measured average LFCT.

Method validation

We compared LFCT measures for respiratory events that were scored both by the automated algorithm and the visual scorer. Events were matched using a distance criterion of no more than 10 sec. A single scorer visually measured LFCT over each sleep study that had annotations from automated LFCT measurement. Visually measured LFCT was used as a gold standard. We computed Pearson correlation coefficients between LFCT values computed with the two methods for each subject and used paired t test to test the hypothesis that average LFCT between the two methods were not different.

Statistical analysis

Using the tested algorithm, we derived automated LFCT among MESA participants who had an AHI≥15/hr. OSA was further classified into moderate vs. severe based on obstructive AHI (15~30 vs. >30/hr) to assess the effect of OSA severity on LFCT. Those with missing LFCT or covariates were excluded.

A multiple regression model was constructed to determine factors independently associated with average LFCT across individuals. This model included age (dichotomous by median age), sex, height, weight and either obstructive AHI (OAHI) or square root transformed OAHI considering the non-normal distribution of OAHI. Another model additionally including arousal index as a measure of sleep quality (neuronoal noise) was tested.(Dvir et al., 2018) We also explored factors that are associated with LFCT across the sleep period within individuals. Candidate exposure variables included, in turn: duration of preceding apnea/hypopnea event (respiratory event length in seconds), nadir SpO2, sleep position (supine vs. non-supine) and sleep stages (REM [rapid eye movement] vs. NREM [non-rapid eye movement sleep]) all linked to the respective LFCT. To effectively account for individual variability, we used linear mixed-effect models by designating the intercept and a candidate exposure variable of interest as random effects, using individual as a grouping variable. These models were adjusted for age, race, sex, height, weight and OAHI. Values were expressed as mean (SD) unless specified otherwise. P< 0.05 was considered statistically significant. All statistical analyses were performed using MATLAB (the Mathworks, Natick MA).

Results

Method validation: LFCT by visual vs. automated measurement

In 39 sample records, on average there were 40.1 LFCT matching measures per record identified both visually and automatically by the algorithm (Figure 2). Correlations within individuals ranged between 0.34 and 1, with an average value over the 39 recordings of r=0.84. Correlations of average LFCT (one LFCT value per individual) obtained by the two methods was r = 0.96 (p< 0.0001) (Figure 3). There was no significant difference in the mean of average LFCT between the two methods when using matching events (Visual: 19.4 [3.7] s; automated: 19.3 [3.7] p=0.6).

Figure 2.

Figure 2.

Open in a new tab

Histogram: Distribution of average lung to finger circulation time (LFCT in sec) by automatic vs. visual methods in validation cohort (N=39).

Figure 3.

Figure 3.

Open in a new tab

Correlation of average LFCT between automatic and visual methods in validation sample (N=39).

LFCT distribution in entire cohort

Baseline characteristics of all participants (N = 596) and by average LFCT are depicted in Table 1. The cohort is an ethnically diverse older middle age men and women (mean age 69 [9] years, male 60.4%). LFCT was normally distributed with mean and median average LFCT of 19.4 (3.1) and 19.4 sec respectively (Figure 4). Older age (≥69 years [i.e., middle old and very old group] vs. < 69 [i.e., young old),(Forman et al., 1992) male and higher height were associated with longer average LFCT (Figure 5). With regard to sleep parameter, those in the longer average LFCT had higher arousal index (Low LFCT vs. high LFCT: 28.6[13.8] vs. 31.4[13.9] sec, p=0.015). Average SD of LFCT per subject was 4.8 (1.2) sec [Apnea vs. Hypopnea; 3.7 [2.0] vs. 5.1 [0.9] sec].

Table 1.

Baseline characteristics of the study participants (N=596)

All (N=596) Low LFCT (N=298) High LFCT (N=298) P value
Age 69.0 (8.9) 68.0 (8.6) 70.0 (9.1) 0.004
Sex (male) 360 (60.4%) 145 (48.7%) 215 (72.2%) <.0001
BMI (kg/m2) 30.6 (5.9) 31.2 (6.1) 30.1 (5.5) 0.014
Height (cm) 166.6 (10.1) 165.0 (10.2) 168.3 (9.7) <.0001
Weight 188.0 (40.7) 187.5 (41.0) 188.4 (40.6) 0.95
Hypertension medication 344 (57.7%) 161 (54.0%) 183 (61.4%) 0.07
Race, N (%) 0.89
 White 198 (33.2 %) 95 (31.9%) 103 (34.6%)
 Chinese American 82 (13.8 %) 38 (12.8%) 44 (14.8%)
 African American 154 (25.8 %) 91 (30.5%) 63 (21.1%)
 Hispanic 162 (27.2%) 74 (24.8%) 88 (29.5%)
Obstructive AHI 31.7 (16.7) 31.0 (16.7) 32.5 (16.8) 0.18
Mean SpO2 saturation 93.4 (2.0) 93.2 (2.1) 93.5 (1.8) 0.08
Number of arousals 169.7 (82.0) 163.2 (82.3) 176.2 (81.3) 0.06
Arousal index (/hour) 30.0 (13.9) 28.6 (13.8) 31.4 (13.9) 0.015
Total sleep time (hour) 345.8 (84.8) 349.4 (87.2) 342.2 (4.8) 0.30
Open in a new tab

Low vs. High LFCT is based on the median value of 19.4 sec. Comparisons between two groups were made using Rank-Sum test. AHI, apnea hypopnea index; SpO2, oxygen saturation

Figure 4.

Figure 4.

Open in a new tab

Histogram: Distribution of average lung to finger circulation time (LFCT in sec) in entire cohort (N=596).

Figure 5.

Figure 5.

Figure 5.

Open in a new tab

Box-plot of average LFCT A) by sex, B) by sex and age group in entire cohort (N=596)

Factors associated with average LFCT across subjects

On average there were 105.2 [range: 7~645] LFCTs measured per subject. Males had higher LFCT than females (19.6 [2.8] vs. 17.9 [2.7] sec, p<0.0001) (Figure 3At). Individuals older than age>69 years (vs. ≤ 69) had longer average LFCT (19.4 [2.8] vs. 18.5 [2.9] sec, p<0.0001) (Figure 3B). Overall, weak correlations with LFCT were found with age (r=0.18, p<0.0001), height (r=0.18, < 0.0001) and OAHI (r=0.10, p=0.012). No statistically significant correlations were observed with either weight or mean SpO2. Multivariable analysis showed that age, sex and OAHI remained significantly associated with LFCT. Height was not significant in multivariable models (Table 2). Results remained similar when square root transformed OAHI was included in the model. A model that additionally included arousal index yielded similar results (data not shown). Analyses by sleep stages, REM vs. NREM, did not significantly change the results.

Table 2.

Between Individual Associations between demographic, anthropometric and AHI values and average lung to finger circulation time LFC

Beta estimate (sec) SE P value
Age>69 yrs (vs. ≤69) 0.98 0.23 <0.0001
Sex (Male) (vs. female) 1.32 0.32 <0.0001
Height (per 1cm) 0.02 0.02 0.29
Weight (per 1lbs) −0.002 0.004 0.63
Obstructive AHI (square root- transformed per 1/hr) 0.19 0.09 0.024
Open in a new tab

Multivariable model adjusting for age (dichotomous by median age), sex, height, weight and square root transformed obstructive apnea hypopnea index (OAHI)

LFCT distribution within subjects

Multivariate mixed model analysis results for within individual analyses are shown in Table 3. LFCT was positively associated with apneas (vs hypopneas) and events in REM sleep. Longer preceding apneic event duration and higher nadir SpO2.were associated with shorter LFCT. Sleep position was not associated with LFCT but sleep stage was associated with LFCT with LFCT in REM (vs. NREM) sleep exhibiting slightly longer LFCT.

Table 3.

Within- individual factors associated with lung to finger circulation time (LFCT) measures

Beta estimate (in sec) SE P value
Event duration (per 1 sec increase) −0.0621 0.0038 < 0.001
Nadir SpO2
(per 1% higher)		 −0.1025 0.0103 < 0.001
Type of event = apnea (vs. hypopnea) 2.6162 0.0904 <0.001
Sleep position
(supine vs. non-supine)		 0.1143 0.1157 0.32
Sleep stage
(REM vs. NREM)		 1.119 0.1083 < 0.001
Open in a new tab

Results from multivariable mixed model. For each candidate exposure, the model evaluated its relationship with lung to finger circulation time (LFCT) (the outcome), grouping by subject and with random effects for slope and intercept. Each model was adjusted for age (continuous), race, sex, height, weight and apnea hypopnea index. SpO2, oxygen saturation; REM, rapid eye movement sleep; NREM, Non-REM sleep.

Discussion

We aimed to characterize the distribution of sleep study-derived average LFCT in community dwelling adults and to determine factors associated with LFCT both between and within individuals. To our best of knowledge, this study represents the first in-depth study to describe sleep study-derived LFCT in people with OSA in general population.

To effectively achieve the aim of the study, we first developed an automated LFCT measurement tool and validated it against the gold standard visual LFCT measurement in subsample of the cohort before we applied to entire cohort. We found that overall the automated LFCT scoring method showed a high correlation of average LFCT with that by visually measured method in a sample of our study cohort. Moreover, there was an excellent agreement of average LFCT by two methods. Average LFCT was normally distributed in this cohort with the mean average LFCT value of 19.4 sec. Across individuals, average LFCT was longer in men, with older age and more severe OSA. Within individuals, LFCT varied with sleep state (longer in REM), with respiratory event subtype (longer with obstructive apneas vs hypopneas), degree of O2 desaturation (longer with lower nadir), and respiratory event duration (longer with shorter event duration). This finding suggests that LFCT is influenced by various factors independent of cardiac function. As such, simple comparison of a random LFCT between patients is not logical.

Two prior clinic-based studies showed that sleep study-derived Ct is prolonged in patients with OSA and HF compared with OSA only.(Kwon et al., 2014, Ryan and Bradley, 2005) A study by Kwon et al demonstrated that the LFCT can be reliably measured from PSG and provided a glimpse into the distribution of the LFCT in patients without HF.(Kwon et al., 2014) In that study, average LFCT was 17.3 (4.5) sec, which is slightly lower than that of 19.4 (3.1) sec from the current study. It is important to note that the average age represented in that study was much younger (50.1 vs. 69 years old) and that only an average of 10 consecutive LFCTs from N2 sleep was used to derive average LFCT.(Kwon et al., 2014) Thus, the difference of average LFCT between the two studies may be due to the difference in cohort characteristics as well as the methodology of the average LFCT derivation. However, our results are consistent with that study in that male sex and older age (middle old and very old) were associated with longer LFCT.(Forman et al., 1992) Longer Ct in men has been described in prior investigations and may be related to larger blood volume of men.(Willems et al., 1971) Longer Ct in older age has also been described.(Brandfonbrener et al., 1955, Willems et al., 1971) One possible explanation of longer Ct with older age is due to decrease in cardiac output with aging.(Brandfonbrener et al., 1955) Higher OAHI, a measure of OSA, was associated with longer LFCT. This finding is not easily explainable and cannot be inferred in our study but may be attributed to subtle difference in cardiac output across the OSA severity at the time of PSG (i.e., more severe OSA, lower cardiac output). Although subjects with taller height had longer LFCT, height was no longer a significant factor in multivariable analysis suggesting that the association was likely confounded by sex. However, it is worth noting such a small difference between the age groups and sex may not be clinically meaningful. Within individuals, there was significant variability in LFCT as evidenced by an average LFCT SD of 5 sec in the context of overall average LFCT of 19.4 sec (coefficient of variation: 0.26). We found that in a given study, respiratory event length, nadir SpO2, apnea (vs. hypopnea) and sleep stage were all associated with LFCT. For example, each 1 sec increment in respiratory event duration was associated with 0.06 sec shorter LFCT and each 1% lower nadir saturation following respiratory event was associated with 0.1 sec longer LFCT. As with the inter-individual analysis, despite statistical significance, the strength of the association observed for within individual analysis seems trivial considering typical distributions of sleep apnea duration and nadir desaturation and as such, the practical implication of the association is uncertain.(Leppänen et al., 2016) Apnea events conferred longer average LFCT compared with hypopneas. LFCT measurement of hypopnea is more subject to imprecise demarcation of LFCT starting point given less prominent transition between the end of respiratory event and hyperpnea (i.e., starting point of LFCT measurement). Thus, shorter LFCT with hypopnea may have resulted from possible systematic delayed detection of the starting point by our algorithm. It is noteworthy that visual-based LFCT measurement would be also subject to the similar challenge for hypopnea events. On the other hand, longer LFCT with apnea may be related to more profound hemodynamic impact leading to longer Ct as compared with hypopnea events. Because of these reasons, it may be important to distinguish LFCT based on the event type. Slightly longer LFCT in REM sleep suggests that stroke volume may be more suppressed because of higher systemic vascular resistance resulting from more robust sympathetic activity. One prior study comparing hemodynamic changes between REM vs. NREM sleep reported inconsistent variable changes in stroke volume and cardiac output across sleep stages.(Khatri and Freis, 1967) Although not measured in our study (and would be nearly impossible to measure in real time) variable cardiac output and pulmonary/peripheral circulation from event to event would be some of the important underlying factors responsible for the variability.(Krinsky et al., 1998, Tolle et al., 1983) On the other hand, it is important to recognize that this difference of LFCT between REM and NREM sleep may be related to measurement challenge in REM sleep due to erratic breathing pattern, which makes it difficult to determine the starting point of LFCT measurement.

While the clinical utility of sleep study-derived LFCT with regards to underlying cardiac impairment cannot be supported by this study alone and requires more rigorous evaluation, our findings provide useful reference data for future investigation. Specifically, the distribution of average LFCT shown in our study have a potential to enable more meaningful LFCT comparisons across patients in clinical practice. Using the LFCT distribution data of this study, a clinician may determine whether an average LFCT derived from sleep study of their patient is meaningfully prolonged or not, and if deemed abnormal, it would warrant further consideration for evaluating underlying cardiac impairment or implementing more aggressive therapy of OSA as part of cardiovascular risk preventive measure. More importantly, since LFCT may represent a novel index of poor cardiovascular outcomes, detailed description about the distribution and the associated factors of LFCT can be useful in our continuing endeavor to search for a PSG metrics that may serve as an alternative or additional tool to assess severity of OSA.

The results of this study also highlight the feasibility of automated LFCT measurement method. Although random samples of LFCT can be readily assessed in patients with OSA by visual inspection from sleep study, such an approach can be limited given the variability of LFCT within a given study as shown in our study. As such, the demonstration of automated LFCT measurement introduced in this study offer a preview of convenient and systematic LFCT measurement. In particular, artificial intelligence-based algorithm may improve the accuracy of the measurement and offers the potential to be further tailored to suit more user-specific needs.

A number of limitations are worth acknowledging. LFCT as measured by our method can be influenced by the inherent instrumental delay of the pulse oximetry. However, since identical devices were used for all participants, this effect would be similar across studies and therefore would be unlikely to influence the differences across the subjects. While those with low left ventricle ejection fraction were excluded from this study, there could have been subjects with preserved ejection fraction but with low cardiac output state at the time of sleep study or individuals with subclinical or unrecognized cardiac impairment. Future studies should examine LFCT distribution in other cohorts to test the consistency of our findings and more importantly, investigate the role of sleep study-derived LFCT as a physiologic biomarker for underlying cardiovascular abnormality.

In conclusion, we introduced an automated LFCT measurement to measure average LFCT and characterized the distribution of average LFCT as well as LFCT in a given subject in community dwelling adults free of cardiovascular disease.

Support:

American Academy of Sleep Medicine Foundation (162-FP-17), NIH R21HL140432, N01-HC-95159, N01-HC-95160, N01-HC-95161, N01-HC-95162, N01-HC-95163, N01-HC-95164, N01-HC-95165, N01-HC-95166, N01-HC-95167, N01-HC-95168 and N01-HC-95169 from the NHBLI, by grants UL1-TR-000040, UL1-RR-025005 from NCRR, R01HL098433 (MESA Sleep). SR was also partly support by NHLBI R35 HL135818

Footnotes

Conflicts of Interest: None

References

  1. BAE KT 2010. Intravenous contrast medium administration and scan timing at CT: considerations and approaches. Radiology, 256, 32–61. [DOI] [PubMed] [Google Scholar]
  2. BILD DE, BLUEMKE DA, BURKE GL, DETRANO R, DIEZ ROUX AV, FOLSOM AR, GREENLAND P, JACOB DR JR., KRONMAL R, LIU K, NELSON JC, O’LEARY D, SAAD MF, SHEA S, SZKLO M & TRACY RP 2002. Multi-ethnic study of atherosclerosis: objectives and design. Am J Epidemiol, 156, 871–81. [DOI] [PubMed] [Google Scholar]
  3. BLUMGART H 1931. The velocity of blood flow in health and disease. The velocity of blood flow in man and its relation to other measurements of the circulation. Medicine 10. [Google Scholar]
  4. BLUMGART HL & WEISS S 1927. STUDIES ON THE VELOCITY OF BLOOD FLOW: VII. The Pulmonary Circulation Time in Normal Resting Individuals. J Clin Invest, 4, 399–425. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. BLUMGART HL & WEISS S 1928. CLINICAL STUDIES ON THE VELOCITY OF BLOOD FLOW: IX. The Pulmonary Circulation Time, the Velocity of Venous Blood Flow to the Heart, and Related Aspects of the Circulation in Patients with Cardiovascular Disease. J Clin Invest, 5, 343–77. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. BRANDFONBRENER M, LANDOWNE M & SHOCK NW 1955. Changes in cardiac output with age. Circulation, 12, 557–66. [DOI] [PubMed] [Google Scholar]
  7. CHEN X, WANG R, ZEE P, LUTSEY PL, JAVAHERI S, ALCANTARA C, JACKSON CL, WILLIAMS MA & REDLINE S 2015. Racial/Ethnic Differences in Sleep Disturbances: The Multi-Ethnic Study of Atherosclerosis (MESA). Sleep, 38, 877–88. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. DEAN DA 2ND, GOLDBERGER AL, MUELLER R, KIM M, RUESCHMAN M, MOBLEY D, SAHOO SS, JAYAPANDIAN CP, CUI, MORRICAL MG, SUROVEC S, ZHANG GQ & REDLINE 2016. Scaling Up Scientific Discovery in Sleep Medicine: The National Sleep Research Resource. Sleep, 39, 1151–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. DVIR H, ELBAZ I, HAVLIN S, APPELBAUM L, IVANOV PC & BARTSCH RP 2018. Neuronal noise as an origin of sleep arousals and its role in sudden infant death syndrome. Sci Adv, 4, eaar6277. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. FORMAN DE, BERMAN AD, MCCABE CH, BAIM DS & WEI JY 1992. PTCA in the elderly: the “young-old” versus the “old-old”. J Am Geriatr Soc, 40, 19–22. [DOI] [PubMed] [Google Scholar]
  11. HALL MJ, XIE A, RUTHERFORD R, ANDO S, FLORAS JS & BRADLEY TD 1996. Cycle length of periodic breathing in patients with and without heart failure. Am J Respir Crit Care Med, 154, 376–81. [DOI] [PubMed] [Google Scholar]
  12. JOHN G, PLATON A, POLETTI P-A, PERRIER A & BENDJELID K 2016. Contrast Circulation Time to Assess Right Ventricular Dysfunction in Pulmonary Embolism: A Retrospective Pilot Study. PloS one, 11, e0159674– e0159674. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. KASRAVI B, BOEHMER JP & LEUENBERGER UA 1998. A noninvasive method for estimating cardiac output using lung to finger circulation time of oxygen. Am J Cardiol, 82, 915–7. [DOI] [PubMed] [Google Scholar]
  14. KHATRI IM & FREIS ED 1967. Hemodynamic changes during sleep. J Appl Physiol, 22, 867–73. [DOI] [PubMed] [Google Scholar]
  15. KRINSKY GA, KAMINER E, LEE VS, ROFSKY NM & WEINREB JC 1998. The effects of apnea on timing examinations for optimization of gadolinium-enhanced MRA of the thoracic aorta and arch vessels. J Comput Assist Tomogr, 22, 677–81. [DOI] [PubMed] [Google Scholar]
  16. KWON Y, KHAN T, PRITZKER M & IBER C 2014. Circulation time measurement from sleep studies in patients with obstructive sleep apnea. J Clin Sleep Med, 10, 759–65, 765A. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. KWON Y, VAN’T HOF J, ROY SS, BACHE RJ & DAS G 2015. A Novel Method for Assessing Cardiac Output With the Use of Oxygen Circulation Time. J Card Fail. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. LEPPÄNEN T, TÖYRÄS J, MURAJA-MURRO A, KUPARI S, TIIHONEN P, MERVAALA E & KULKAS A 2016. Length of Individual Apnea Events Is Increased by Supine Position and Modulated by Severity of Obstructive Sleep Apnea. Sleep disorders, 2016, 9645347–9645347. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. MULLER HM, TRIPOLT MB, REHAK PH, GROELL R, RIENMULLER R & TSCHELIESSNIGG KH 2000. Noninvasive measurement of pulmonary vascular resistances by assessment of cardiac output and pulmonary transit time. Invest Radiol, 35, 727–31. [DOI] [PubMed] [Google Scholar]
  20. OLDENBURG O, LAMP B, FABER L, TESCHLER H, HORSTKOTTE D & TOPFER V 2007. Sleep-disordered breathing in patients with symptomatic heart failure: a contemporary study of prevalence in and characteristics of 700 patients. Eur J Heart Fail, 9, 251–7. [DOI] [PubMed] [Google Scholar]
  21. PUSKAS Z & SCHUIERER G 1996. [Determination of blood circulation time for optimizing contrast medium administration in CT angiography]. Radiologe, 36, 750–7. [DOI] [PubMed] [Google Scholar]
  22. RYAN CM & BRADLEY TD 2005. Periodicity of obstructive sleep apnea in patients with and without heart failure. Chest, 127, 536–42. [DOI] [PubMed] [Google Scholar]
  23. SANDS SA & OWENS RL 2015. Congestive heart failure and central sleep apnea. Critical care clinics, 31, 473–495. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. SHORS SM, COTTS WG, PAVLOVIC-SURJANCEV B, FRANCOIS CJ, GHEORGHIADE M & FINN JP 2003. Heart failure: evaluation of cardiopulmonary transit times with time-resolved MR angiography. Radiology, 229, 743–8. [DOI] [PubMed] [Google Scholar]
  25. TOLLE FA, JUDY WV, YU PL & MARKAND ON 1983. Reduced stroke volume related to pleural pressure in obstructive sleep apnea. J Appl Physiol Respir Environ Exerc Physiol, 55, 1718–24. [DOI] [PubMed] [Google Scholar]
  26. WEXLER J, WHITTENBERGER JL & HIMMELFARB S 1946. An Objective Method for Determining Circulation Time from Pulmonary to Systemic Capillaries by the Use of the Oximeter. J Clin Invest, 25, 447–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. WILLEMS JL, ROELANDT JR, VAN DE VEL HR & JOOSSENS JV 1971. The circulation time in the aged. American Journal of Cardiology, 27, 155–161. [DOI] [PubMed] [Google Scholar]

RESOURCES