Important influence of respiration on human R-R interval ...€¦ · Important influence of respiration on hu- man R-R interval power spectra is largely ignored. J. Appl. Phys- iol

Important influence of respiration on human R-R interval power spectra is largely ignored

TROY E. BROWN, LARRY A. BEIGHTOL, JUNKEN KOH, AND DWAIN L. ECKBERG Departments of Medicine and Physiology, Hunter Holmes McGuire Department of Veterans Affairs Medical Center and Medical College of Virginia, Richmond, Virginia 23249

BROWN,TROY E., LARRY A. BEIGHTOL,JUNKENKOH,AND DWAIN L. ECKBERG. Important influence of respiration on human R-R interval power spectra is largely ignored. J. Appl. Phys- iol. 75(5): 2310-2317,1993.-Frequency-domain analyses of R- R intervals are used widely to estimate levels of autonomic neural traffic to the human heart. Because respiration modulates autonomic activity, we determined for nine healthy subjects the influence of breathing frequency and tidal volume on R-R interval power spectra (fast-Fourier transform method). We also surveyed published literature to determine current practices in this burgeoning field of scientific inquiry. Supine subjects breathed at rates of 6, 7.5, 10, 15, 17.1, 20, and 24 breaths/min and with nominal tidal volumes of 1,000 and 1,500 ml. R-R interval power at respiratory and low (0.06-0.14 Hz) frequencies declined significantly as breathing frequency increased. R-R interval power at respiratory frequencies was significantly greater at a tidal volume of 1,500 than 1,000 ml. Nei- ther breathing frequency nor tidal volume influenced average R-R intervals significantly. Our review of studies reporting human R-R interval power spectra showed that 51% of the studies controlled respiratory rate, 11% controlled tidal volume, and 11% controlled both respiratory rate and tidal volume. The major implications of our analyses are that breathing parameters strongly influence low-frequency as well as respiratory frequency R-R interval power spectra and that this influence is largely ignored in published research.

heart rate; respiration rate; tidal volume; respiratory sinus arrhythmia

FLUCTUATIONSOFR-RINTERVALS orheartratesareused widely as indexes of the level of autonomic traffic to the heart. This usage has its origin in a study published by Katona and Jih (18), who showed that in anesthetized dogs with constant breathing rates respiration-related peak-valley R-R interval fluctuations (respiratory sinus arrhythmia) are related linearly to absolute vagal firing rates. On the basis of this empirical observation, R-R interval fluctuations have been promoted as faithful noninvasive quantitative estimates of human vagal-car- disc nerve traffic. In recent years, the time-domain method used by Katona and Jih has been supplanted by frequency-domain methods. Some workers consider power spectral density or autoregressive approaches to

2310

be superior to time-domain analyses of R-R intervals because the former approaches exclude random fluctuations and may provide quantitative information on the level of sympathetic- as well as vagal-cardiac neural outflow (27).

Several studies have sh own that respirato ry rate and tidal volume exert major influ .ences on R-R interval or heart rate fluctuations (8, 16, 21, 37). This genre of research is important because if fluctuations of R-R intervals are to be taken as valid noninvasive indexes of autonomic neural traffic, then they should reflect such traffic faithfully and should not be influenced importantly by respiratory-autonomic interactions unrelated to net neural outflow. Our study had two goals. First, we determined the effects of different voluntarily controlled respiratory rates and tidal volumes on R-R in terval power spectra. Second, we surveyed published literature to learn how workers in this field have dealt with the potential influence of respiration on these measurements. Our results show that respiration exerts a strong influence on low-frequency as well as respiratory frequency R-R interval power spectra and that this influence is largely ignored in published research.

METHODS

Subjects. Nine healthy subjects (8 men and 1 woman; age 23-32 yr) were studied in recumbency. Six of the subjects had been studied previously (8), and three were studied prospectively. The prospective study was identical to the earlier study in all important details (see below). This research was approved by the human investi- gation committees of the Hunter Holmes McGuire De- partment of Veterans Affairs Medical Center and the Medical College of Virginia. All volunteers gave their in- formed written consent to participate in this research.

Measurements. We recorded the electrocardiogram, integrated tidal volume [Fleisch pneumotachograph for the retrospective study and transit-time ultrasonic breath analyzer (GHG Medizin-Elektronik, Zurich, Switzerland) for the prospective study], and end-tidal CO, concentration (infrared analyzer) on FM tape.

Respiratory control. Subjects breathed with an auditory

by 10.220.33.1 on March 30, 2017

http://jap.physiology.org/D

ownloaded from

http://jap.physiology.org/

RESPIRATORY MODULATION OF AUTONOMIC OUTFLOW 2311

signal at rates of 6, 7.5, 10, 15, 17.1, 20, and 24 breaths/ min. They controlled their tidal volumes at nominal levels of 1,000 and 1,500 ml according to signals displayed on a calibrated oscilloscope. Subjects wore a face mask connected to the pneumotachograph or transit-time ultrasonic breath analyzer and a three-way respiratory valve (Ewald Koegel, San Antonio, TX, for the retrospective study and Hans Rudolph, Kansas City, MO, for the prospective study).

Experimental protocol. Initial measurements were obtained at the lowest ventilation rate, 6 breaths/min with a tidal volume of 1,000 ml. The end-tidal CO, concentration measured at this level of ventilation was maintained at all other breathing frequencies and tidal volumes by addition of CO, to the breathing line. The remaining respiratory intervals at the two tidal volumes were used in random sequence. Data for each respiratory rate and tidal volume were recorded during one continuous 128-s period. Breathing was not controlled between data collection periods.

Data analyses. Electrocardiograms and tidal volumes were digitized (CODAS, Dataq Instruments, Akron, OH) with a Compaq 386/25 microcomputer. Power spectral analyses were derived from R-R interval time series with a custom program developed for use with a software package (DADiSP, DSP Development, Cambridge, MA). The technique used for estimation of R-R interval power was based on the Welch algorithm of averaging periodograms (42), which was implemented according to the method of Rabiner et al. (31). A 128-s time series of beat- to-beat R-R intervals was fitted to a cubic spline function, interpolated at 8 Hz to obtain equidistant time intervals, and divided into seven equal overlapping seg- ments. Each segment was in turn detrended, Hanning filtered, and fast-Fourier transformed to its frequency representation. The modified periodograms were aver- aged to produce the spectrum estimate. The method we used yielded a frequency resolution of 0.0078 Hz and a flat response over the frequencies of interest.

We measured total absolute R-R interval power at low and respiratory frequencies over fixed frequency ranges. We determined the respiratory frequency power bandwidth as follows: 1) power spectral analyses were performed on each respiratory signal, 2) the respiratory bandwidths were determined (95% of peak respiratory power), and 3) the average bandwidth at each respiratory frequency was calculated (all were 0.08 Hz). The 0.08-Hz bandwidth derived from analysis of respiration was used to integrate R-R intervals. (Thus, R-R interval spectral power was integrated at frequencies to.04 Hz and centered on each respiratory frequency.) A fixed bandwidth, centered at 0.1 Hz (0.06-0.14 Hz), was also used for low- frequency spectral area calculations.’ We calculated gain-frequency responses to construct Bode plots as

’ In this article, we use the term respiratory frequency to signify R-R interval power at respiratory rates and low frequency to signify power between 0.06 and 0.14 Hz. The 0.06- to 0.14-Hz range was chosen to encompass R-R interval frequencies that are considered to reflect sympathetic outflow (24). Resting humans also have R-R interval spectral power at lower frequencies (~0.06 Hz), which are thought to reflect thermoregulatory oscillations (20). We did not evaluate the lowest portion of each subject’s power spectra.

20 log( output /input). Inpu t was the tidal volume in liters, and output was the total absol .ute power of the respiratory frequency or low-frequency portion of the R- R interval power spectrum.

We evaluated data for normality with the Kolmo- gorov-Smirnov test (22). We made statistical compari- sons on normally distributed data with repeated-mea- sures analysis of variance to test for significant individual differences among mean integrated R-R interval power at different respiratory rates and tidal volumes. Then we used Scheffe’s test to identify significantly different means. We considered differences significant when P 5 0.05. Two-tailed tests were used.

Literature reuiew. We searched the National Library of Medicine Medline data base for all references between 1966 and the first quarter of 1991 to “power spectra” or “power spectrum” and “heart rate” or “R-R interval” in humans. We examined the articles identified by this search, references cited in those articles, and other references known to us to determine how often respiratory rate a .nd tidal volume were mea sured and controlled.2 Then we searched the Science Citation Index (Institute for Scientific Information, Philadelphia, PA) to determine how many times each of these references was cited from 1974 to 1991. We report results from the most influential articles, those cited 210 times, individually.

RESULTS

Experimental data. Figure 1 depicts the digitized electrocardiogram, tidal volume, and R-R interval signals from one subject during breathing at respiratory rate of 6,15, or 24 breaths/min and at a nominal tidal volume of 1,000 ml. The rhythmic R-R interval changes during controlled breathing in this subject were similar to those observed in other subjects. In this subject, R-R interval variability decreased as breathing frequency increased; at the slowest breathing rate, maximum R-R intervals were longer and minimum R-R intervals were shorter than at the two fastest breathing rates.

Subjects maintained average respiratory rates within 0.3% of targeted levels and tidal volumes at -7% below targeted levels. End-tidal CO, concentrations were similar (P > 0.05) at all breathing frequencies and tidal volumes and ranged between 4.7 and 6%. In all subjects, end-tidal CO, was maintained within 0.25% of targeted levels. Figure 2 shows average R-R intervals at all breathing rates and both tidal volumes. There were no statistically significant differences among mean R-R intervals (which ranged between 0.905 and 1.039 s) at different respiratory rates and tidal volumes. All data were normally distributed.

Figure 3 shows average R-R interval power spectral waveforms at all respiratory rates and at a nominal tidal

by 10.220.33.1 on March 30, 2017


ownloaded from


2312 RESPIRATORY MODULATION OF AUTONOMIC OUTFLOW

1.2 n

cll

0 L

1.0

2

l t 0.8 Qf I

f~ 0.6

0 20 30 0 4 8 12

time (s) 0 2.5 5 7.5

FIG. 1. Experimental records from 1 subject. Respiratory rates shown are 6 (A), 15 (B), and 24 breaths/min (C). R-R interval variability was greatest at slowest breathing rate.

1.50

ul

z 1.25

L a,

c '- 1.00 CY

I fx

c 0.75

F/i

E

0.50

volume of 1,000 ml. At each breathing interval, R-R in- 17%; range, 4-33s; P < 0.05). There was a highly signifi- terval power spectra were concentrated at the breathing cant linear correlation between R-R interval power at frequency. Spectral power declined precipitously when respiratory frequencies and R-R interval standard devia- the respiratory rate was increased from 10 to 15 tions (r = 0.87 at 1,000 ml and 0.93 at 1,500 ml). breaths/min. Power at low frequencies (Fig. 4C) also declined signifi-

Figure 4 depicts average integrated respiratory frequency and low-frequency R-R interval power and gain. Power at the respiratory frequency (Fig. 4A) declined significantly as respiratory rate increased (P < 0.001); a major reduction occurred between respiratory rates of 7.5 and 15 breaths/min at both tidal volumes (P < 0.05). Respiratory frequency power was significantly greater at tidal volumes of 1,500 than 1,000 ml (by an average of

.b 0 1000 ml 0 1500 ml

5 10 15 20 25 //- 31

respiratory rate (breaths/min) 2

FIG. 3. Average R-R interval power spectra at respiratory rates of 6, FIG. 2. Mean (+-SE) of R-R interval at each respiratory rate and 7.5, 10, 15, 17.1,>0, and 24 breaths/m& and nominal tidal volume of

tidal volume. There were no statistically significant differences among 1,000 ml. Respiratory frequency power declines as respiratory rate in- mean R-R intervals. creases.

by 10.220.33.1 on March 30, 2017


ownloaded from



5 5 15 25

-25 n

respiratory rate (breaths/min)

cantly as respiratory rate increased (P < 0.001). In- creases of low-frequency power at rates of 510 breaths/ min resulted from overlap of respiratory frequency and low-frequency bands (Fig. 3). Low-frequency power at the two tidal volumes was similar (P > 0.05).

The gain-frequency response (Bode plot) of the respiratory frequency component of R-R interval power spectra documented significantly decreasing system gain with increasing respiratory rate (P < 0.001; Fig. 4B). A nonlinear segmented model fit (SAS Institute, Cary, NC) analysis suggested that the gain-respiratory rate response relationship was best described by a plateau region between 6 and 10 breaths/min (0.10-0.17 Hz) for both tidal volumes. The plateau region was succeeded by roll-offs of 10.6 dB/decade for the l,OOO-ml tidal volume and 14.7 dB/decade for the 1,500-ml tidal volume. Gain of respiratory frequency R-R interval power spectra at the lowest breathing frequencies (6-10 breaths/min) was significantly greater than that at all other breathing frequencies (15-24 breaths/min) for both tidal volumes (P < 0.05). Respiratory frequency system gain was significantly greater at tidal volumes of 1,500 than 1,000 ml (by an average of 7%; range, 4-13%; P < 0.001).

The gain-frequency response of the low-frequency component of the R-R interval power spectra also documented significantly decreasing system gain with increasing respiratory rate (P < 0.001; Fig. 40). Gain of low-frequency R-R interval power spectra at the two lowest breathing frequencies (6 and 7.5 breaths/min) was significantly greater than that at the highest breathing frequencies (lo-24 breathsimin) for both tidal volumes (P < 0.001). Gain at 10 breaths/min also was significantly greater than that at higher breathing frequencies (15-24 breaths/min; P < 0.05). Low-frequency system

FIG. 4. Effects of respiratory rate on R-R interval power spectra at nominal tidal volumes of 1,000 and 1,500 ml. Data are means + SE. A: respiratory frequency power significantly declined as respiratory rate increased (P < 0.001). Power was greater at tidal volumes of 1,500 than 1,000 ml (P < 0.05). B: respiratory frequency gain also significantly declined with increasing respiratory rate (P < 0.001). * There was significant reduction in gain from lower (6- 10 breathsimin) to higher breathing rates (15 24 breaths/min; P < 0.05). Gain was significantly higher (P < 0.001) at 1,500- than l,OOO- ml tidal volume. C: low-frequency power significantly declined as respiratory rate increased (P < 0.001). Power at 2 tidal volumes was similar. D: low-frequency gain also significantly decreased as respiratory rate increased (P < 0.001). ** Gain at 2 lower breathing frequencies (6 and 7.5 breathsimin) was greater than at all other breathing frequencies (lo-24 breaths/min; P < 0.001). * Additionally, lo- breaths/min respiratory rate gain was significantly greater than higher respiratory rates (15-24 breathsjmin; P < 0.05). Values for 2 tidal volumes were similar.

gains at the two tidal volumes were nearly identical (P > 0.05).

Literature review. Our medical literature search identified 147 articles that reported use of R-R interval or heart rate power spectral analysis in humans.3 Table 1 summarizes these articles. Respiratory rate and tidal volume were controlled in only 11% of studies in which control was possible.

Our Science Citation Index search indicated that 23 of the 147 articles were cited 210 times. Table 2 summarizes these articles. There were 133 citations of the article by Pagani et al. (27) and 233 citations of the remaining articles. Four articles (6, 19, 29, 33) dealt with methods and modeling and so were excluded. Thirteen of the remaining 19 articles (68.4%) included measurements of respiratory rate. Three articles (4, 25, 34) involved ambulatory subjects and thus were excluded. Four of the remaining 16 studies (25%) included measurements of tidal volume. Control of respiratory rate and tidal volume was impossible for studies that involved ambulatory subjects and infants and for those that involved psycho- logical treatments. Of the remaining eight studies in which respiratory rate and tidal volume could have been controlled, respiratory rate was controlled in six (75%) and tidal volume was controlled in one (12.5%).

DISCUSSION

We studied the influence of respiratory rate and tidal volume on human R-R interval power spectra and sur-

3 A complete list of the articles uncovered by our search can be obtained from the National Auxiliary Publications Service, c/o Micro- fiche Publications, PO Box 3513, Grand Central Station, NY 10017 (NAPS No. 05059).

by 10.220.33.1 on March 30, 2017


ownloaded from



TABLE 1. Review of publications reporting R-R interval power spectra at fewer respiratory rates (5 vs. 7 in our power spectra in humans

Articles reviewed Possibility of respiratory rate measurement

Measured respiratory rate

study), at more (6 vs. 2 in our study) and different (4OO- 1,400 vs. 1,000 and 1,500 ml in our study) tidal volumes,

n Percentages and at different experimentally maintained average end-

147 tidal CO, concentrations (3.5-4.6 vs. 4.7-6% in our 119 study).

61 51 Other studies also document a dependence of R-R in- Possibility of tidal volume measurement

Measured tidal volume Possibility of respiratory rate control

Controlled respiratory rate Possibility of tidal volume control

90 20 65 24 65

22

37

terval (or heart rate) fluctuations on respiratory parameters. The study most similar to ours is that of Hirsch and Bishop (16), who characterized heart rate fluctuations in the time domain as respiration-related peak-valley R-R

Controlled tidal volume 7 11 Controlled respiratory rate and tidal volume

intervals (the longest minus the shortest R-R interval 7 11 during each breath). Hirsch and Bishop derived Bode

n, no. of publications; Percentages, percentage of published studies plots very similar to ours with a plateau region, a corner in which possibility of measurement or control was realized. frequency at 0.11 Hz, and a linear roll-off of 21 dB/de-

cade. They also showed that peak-valley R-R interval veyed the medical literature to learn how scientists who changes and the steepness of the roll-off portion are di- employ power spectral analyses in their research have rectly proportional to tidal volume. Our more limited dealt with respiration. We reached three major conclu- data (Fig. 4B) are concordant with the last observation. sions. First, respiratory rate and tidal volume strongly Hirsch and Bishop did not indicate if the relationship influence low-frequency as well as respiratory frequency between tidal volume and respiratory R-R interval R-R interval power. Second, because the breathing pa- changes was statistically significant. However, their Fig. rameters we studied did not alter average R-R intervals, 4 [and data published by Eckoldt and Schubert (lo)] sug- variations of respiration distribute autonomic outflow gests that tidal volume may be an important determinant within the respiratory cycle but do not alter net levels of of such changes: in one subject, sixfold increases of tidal autonomic traffic. Thus, vagal activity during the various volume led to approximately threefold increases of slow- phases of the respiratory cycle varied considerably, whereas the overall tonic level remained constant. Third, most of the scientists who performed frequency-domain R-R interval analyses neither controlled nor accounted for the strong influence of respiration on their results.

Respiratory influences on R-R interval power spectra at respiratory frequencies. Breathing rates of - 10 breaths/ min or less yielded maximum R-R interval power. Breathing rates of >lO breaths/min yielded reduced levels of R-R interval power in an inverse relationship to breathing rate (Figs. 3 and 4, A and B). However, over the range we studied, breathing rate did not significantly alter average R-R intervals (Fig. 2). Our study effectively doubles the published literature on R-R interval power spectra in subjects who breathed at multiple experimentally controlled breathing rates and tidal aware of only 0 ne similar article (37).

volumes; we are

Selman et al. (37), in an excellent but rarely cited study, obtained results that differed from ours qualita- tively and quantitatively. First, they found an inverse relationship between tidal volume and system gain. We found a direct relationship (Fig. 4B). Second, they obtained a Bode plot described by a distinct “M shape”

breath peak-valley R-R interval intercepts. Saul et al. (35), who used frequency-domain methods

similar to ours but a broad-band respiratory frequency input, also documented a critical dependence of R-R interval power spectra on breathing-rate. Although the

TABLE 2. Influential R-R interval power spectral research on humans

Respiratory Parameters

Authors No. of

Citations

Measured Controlled

Rate Volume Rate Volume

Pagani et al. (27) 133 Saul et al. (34) 33 Gordon et al. (13) 30 Myers et al. (25) 30 Porges et al. (29) 25 Rohmert et al. (32) 25 Hyndman and Gregory (17) 24 Baselli et al. (2) 23 De Boer et al. (7) 23 Eckberg et al. (9) 22 Walter and Porges (41) 22 Gordon et al. (12) 18

+ + + + x x + x x - x x x x x - - X - - X + + +

- + + + + + X + - X

X X

such that system gains were much higher at breathing Rompelman et ai. (33) 17 x x x x

rates of 10 and 16 than at 12 breaths/min. In our study Lishneretal= (23) 16 + - + -

(Figs. 3 and 4), there was only a hint of an M shape. ~~~n~~~~P EgreY(ll) Third, they did not define a convincing roll-off of system Kit-ey et al’ tlg)

ti + - x x - - - -

gain at higher breathing frequencies (probably because Nugent and*Finley (26) 12 x x x x 12 + - x x

they studied only 5 breathing frequencies). We found a Vybiral et al. (40) 12 + - + -

steady decline of R-R interval power at breathing rates Bigger et al. (4) 11 - - x x

above -10 breaths/min. Fourth, they did not report low- ~~~~~~~“,$~) 11 x x x x

frequency power separately. The methods used by Sel- Shannon et g, (38) 10 - - x x

. 10 + - + - man et al. differed from ours in several respects. They studied subjects whose ages spanned a wider range (26- Reference no. is given in parentheses. +, Parameter measured or

52 vs. 23-32 yr in our study). Their subjects were seated controlled; -, parameter not measured or controlled; X, research de- sign prevented measurement or control of narameter or article was

and ours were supine. They measured R-R interval devoted to methods or modeling.

by 10.220.33.1 on March 30, 2017


ownloaded from



shapes of the transfer function of Saul et al. and our Bode plot are similar, there may be quantitative differences. We found (Fig. 4B) that the plateau height and the roll-off slope are influenced significantly by tidal volume. In the study of Saul et al., tidal volumes and respiratory rates covaried; therefore, the separate influence of tidal volume on the height of the plateau and the slope of the roll-off could not be determined.

Respiratory influences on R-R interval power spectra at low frequencies. A major rationale for study of low-frequency R-R interval fluctuations is that they reflect levels of sympathetic traffic to the heart. Although sympathetic nerve traffic also fluctuates at respiratory frequencies (9, 30), R-R interval responses to such rapid changes of sympathetic activity are small because of the sluggishness of adrenergically mediated sinus node responses. Our results (Fig. 4, C and D) indicate that slow breathing provokes major increases of low-frequency (0.06-0.14 Hz) R-R interval power. In literature involv- ing R-R interval (or heart rate) power spectra, the terms respiratory frequency and high frequency are used inter- changeably. Surprisingly, this literature does not antici- pate that during periods of slow breathing respiratory frequency and low frequency overlap. Obviously, as breathing rate slows, the possibility that fluctuations of sympathetic nerve traffic contribute to respiratory frequency R-R interval changes increases (however, probably not as a step function). Under such circumstances, the view that changes of sympathetic outflow do not contribute to respiratory frequency power becomes increas- ingly untenable. It is highly unlikely that the greatly aug- mented low-frequency R-R interval power we recorded at slow breathing rates (Figs. 3 and 4C) reflects increased sympathetic neural outflow. Seals et al. (36) found that, although rapid breathing redistributes directly measured human (muscle, not cardiac) sympathetic nerve activity within the various phases of the respiratory cycle, it does not alter the net tonic level of sympathetic outflow within the overall cycle.

R-R interval power spectra as noninvasive indexes of vagal-cardiac nerve traffic. Kollai and Mizsei (21) argued persuasively that respiratory rate and tidal volume should be factored into estimates of human vagal-cardiac nerve traffic. They compared peak-valley R-R interval changes during uncontrolled breathing and R-R interval shortening provoked by large-dose intravenous atropine (the current “gold standard” for quantification of resting vagal-cardiac nerve traffic in humans). They found a weak but significant (r = 0.61) linear correlation between peak-valley R-R intervals and atropine-induced R-R interval shortening. However, they also found that when respiratory interval and tidal volume were factoredin as coregressors this correlation was much stronger (r = 0.93).

Our study raises new questions about the validity of R-R interval power spectra as quantitative indexes of vagal-cardiac neural traffic within individual subjects. Our data suggest that even with explicit knowledge (and control) of respiration, respiratory sinus arrhythmia may not reflect tonic vagal-cardiac neural traffic. Our subjects varied the magnitude of their respiratory frequency R-R interval power -IO-fold simply by changing their breathing rates (Fig. 3). Despite huge changes of R-R

interval power, however, there were only trivial changes of mean R-R intervals (Fig. 2). This observation is related to one made by Grossman et al. (14), who found that in ,&blocked subjects respiration-induced changes of peak-valley R-R intervals were not associated with changes of mean R-R intervals. We suggest that within subjects different breathing frequencies and depths distribute vagal firing within the respiratory cycle but do not alter the net level of vagal outflow.

Literature review. We identified and reviewed 147 articles that reported measurements of human R-R interval (or heart rate) power spectra to determine the prevailing practices of workers in this field. Only about one-half of the studies involved measurement of breathing rate and only about one-third involved control of breathing rate. We also determined the number of times each of the 147 articles had been cited to identify the most influential studies in this field. Authors of the most heavily cited studies (Table 2) were more likely to recognize the importance of measurement and control of respiratory rate than those of less cited studies. However, very few (only 11%) of the studies involved control of tidal volume.

Pagani et al. (27), authors of the most heavily cited article in this field, discount the importance of control of respiration for frequency-domain analyses of R-R intervals. They argue that voluntary control of breathing increases power at respiratory frequencies (and net vagal outflow) and decreases power at low frequencies. This conclusion was based on a comparison of R-R interval power (as delimited by an autoregressive algorithm) at respiratory frequencies during spontaneous breathing (at an average rate of 14.4 breaths/min) and during controlled breathing (at an average rate of 19.8 breaths/ min). The approach of Pagani et al. may have been flawed by their comparison of R-R interval power spectra at different breathing rates. However, two other groups compared respiration-related R-R interval fluctuations during controlled and spontaneous breathing at the same breathing rates and concluded that voluntary control of breathing does not alter R-R interval fluctuations. Hirsch and Bishop (16) found that the plateau frequency intercept in 17 subjects was similar during controlled and spontaneous breathing. Eckberg et al. (9) found that respiration-related R-R interval fluctuations (and muscle sympathetic nerve traffic) in eight subjects were nearly superimposable during controlled and spontaneous breathing.

Potential limitations. First, use of fast-Fourier transform methods for power spectral analysis requires that signals be stationary. The mathematical concept of stationarity may not be applicable to the highly variable signals that can be obtained from conscious human volunteers, and we did not test our data for stationarity. However, our subjects maintained exquisite control of breathing frequency during data collection periods. Moreover, our results (Fig. 4) are very similar to those of Hirsch and Bishop (16), who used time-domain methods. Second, in our analysis, actual tidal volumes used by our subjects (-930 and 1,400 ml) were larger than the average tidal volumes studied by others. However, Bendixen et al. (3) showed that almost one-half of the healthy women they studied had resting tidal volumes as high as the smaller tidal volume we used.

by 10.220.33.1 on March 30, 2017


ownloaded from



Implications of this study. Frequency-domain analyses of R-R intervals represent sophisticated attempts to estimate, noninvasively, levels of efferent autonomic neural traffic to the human heart. Our analysis shows that respiration influences both respiratory frequency and low-frequency R-R interval power spectra. We pro- pose that, since the influence of respiration is so strong and pervasive, interpretation of R-R interval power spectra must be grounded on understanding of the respiratory patterns that shaped them.

It may be inappropriate to use R-R interval power spectra obtained during brief recording sessions to gauge levels of efferent autonomic neural traffic if breathing rate and tidal volume are neither measured nor controlled. During brief recording sessions, breathing patterns of healthy subjects are highly variable (5,39), and a substantial percentage of breaths may occur at frequencies <9 and >20 breaths/min (15). Although the majority of healthy supine young subjects have respiratory power in the roll-off region, where small changes of breathing rate are likely to provoke large changes of respiratory frequency R-R interval power, some of the subjects have substantial respiratory power in the plateau region (40 breaths/min; J. A. Taylor, T. E. Brown, and D. L. Eck- berg, unpublished data), where breathing rate is likely to sharply augment low-frequency R-R interval power.

In conclusion, we studied the effects of respiratory rate and tidal volume on R-R interval power spectra. Our results indicate that respiratory parameters strongly influence both low-frequency and respiratory frequency R-R interval power spectra. The major implication of these results is that respiration must be controlled if R-R interval power spectra are to be interpretable. Our survey of published articles in this field shows that the great majority do not meet this minimum condition.

NOTE ADDED IN PROOF

We call attention to the article by Novak et al. (%a) that was published while our study was in press. Novak et al. plotted Wigner distributions of R-R intervals, tidal volumes, and sys- tolic and diastolic pressures during continuously varying periods of slow to rapid breathing. Although the subjects in this study did not control their tidal volumes, their R-R interval Wigner distributions bear a striking similarity to the average data we depict in Fig. 3.

We thank Paul Grossman, Abhijit Patwardhan, and J. Andrew Tay- lor for their critical reviews of this manuscript.

This research was supported by National Heart, Lung, and Blood Institute Grant HL-‘22296, a grant from the Department of Veterans Affairs, and National Aeronautics and Space Administration Grants NAS9-16046 and NAG9-412.

Present addresses: J. Koh, Dept. of Anesthesia, Kure National Hos- pital, Kure, Hiroshima 737, Japan; T. E. Brown, Krug Life Sciences, 1290 Hercules, Suite 120, Houston, TX 77058.

Address for reprint requests: D. L. Eckberg, Cardiovascular Physiol- ogy, Hunter Holmes McGuire Dept. of VA Medical Center, 1201 Broad Rock Blvd., Richmond, VA 23249.

Received 25 January 1993; accepted in final form 25 May 1993.

REFERENCES

1. BARTOLI, F., G. BASELLI, AND S. CERUTTI. AR identification and spectral estimate applied to the R-R interval measurements. Int. J. Bio-Med. Comput. 16: 201-215. 1985.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

BASELLI, G., S. CERUTTI, S. CIVARDI, F. LOMBARDI, A. MALLIANI, M. MERRI, M. PAGANI, AND G. RIZZO. Heart rate variability signal processing: a quantitative approach as an aid to diagnosis in cardiovascular pathologies. Int. J. Bio-Med. Comput. 20: 51-70, 1987. BENDIXEN, H. H., G. M. SMITH, AND J. MEAD. Pattern of ventilation in young adults. J. Appl. Physiol. 19: 195-198, 1964. BIGGER, J. T., JR., M. T. LA ROVERE, R. C. STEINMAN, J. L. FLEISS, J. N. ROTTMAN, L. M. ROLNITZKY, AND P. J. SCHWARTZ. Compari- son of baroreflex sensitivity and heart period variability after myo- cardial infarction. J. Am. Coil. Cardiol. 14: 1511-1518, 1989. DAVIS, J. N., AND D. STAGG. Interrelationships of the volume and time components of individual breaths in resting man. J. Physiol. Lond. 245: 481-498, 1975. DE BOER, R. W., J. M. KAREMAKER, AND J. STRACKEE. Descrip- tion of heart-rate variability data in accordance with a physiological model for the genesis of heartbeats. Psychophysiology 22: l47- 155, 1985. DE BOER, R. W., J. M. KAREMAKER, AND J. STRACKEE. Relation- ships between short-term blood-pressure fluctuations and heart- rate variability in resting subjects. I. A spectral analysis approach. Med. Biol. Eng. Comput. 23: 352-358, 1985. ECKBERG, D. L. Human sinus arrhythmia as an index of vagal cardiac outflow. J. Appl. Physiol. 54: 961-966, 1983. ECKBERG, D. L., C. NERHED, AND B. G. WALLIN. Respiratory modulation of muscle sympathetic and vagal cardiac outflow in man. J. Physiol. Lond. 365: 181-196, 1985. ECKOLDT, K., AND E. SCHUBERT. Zum Einfluss der Atemtiefe auf die Sinusarrhythmie des Herzens. Acta Biol. Med. Ger. 34: 767-771, 1975. GIDDENS, D. P., AND R. I. KITNEY. Neonatal heart rate variability and its relation to respiration. J. Theor. Biol. 113: 759-780, 1985. GORDON, D., D. P. SOUTHALL, D. H. KELLY, A. WILSON, S. AK- SELROD, J. RICHARDS, B. KENET, R. KENET, R. J. COHEN, AND D. C. SHANNON. Analysis of heart rate and respiratory patterns in sudden infant death syndrome victims and control infants. Pediatr. Res. 20: 680-684, 1986. GORDON, D., R. J. COHEN, D. KELLY, S. AKSELROD, AND D. C. SHANNON. Sudden infant death syndrome: abnormalities in short term fluctuations in heart rate and respiratory activity. Pediatr. Res. 18: 921-926, 1984. GROSSMAN, P., J. KAREMAKER, AND W. WIELING. Prediction of tonic parasympathetic cardiac control using respiratory sinus arrhythmia: the need for respiratory control. Psychophysiology 28: 201-216, 1991. GROSSMAN, P., AND K. WIENTJES. Respiratory sinus arrhythmia and parasympathetic cardiac control: some basic issues concerning quantification, applications and implications. In: Cardiorespiratory and Cardiosomatic Psychophysiology, edited by P. Grossman, K. H. L. Janssen, and D. Vaitl. New York: Plenum, 1986, p. 117- 138. HIRSCH, J. A., AND B. BISHOP. Respiratory sinus arrhythmia in humans: how breathing pattern modulates heart rate. Am. J. Phys- ioZ. 241 (Heart Circ. Physiol. 10): H620-H629, 1981. HYNDMAN, B. W., AND J. R. GREGORY. Spectral analysis of sinus arrhythmia during mental loading. Ergonomics 18: 255-270, 1975. KATONA, P. G., AND F. JIH. Respiratory sinus arrhythmia: noninvasive measure of parasympathetic cardiac control. J. Appl. Phys- iol. 39: 801-805, 1975. KITNEY, R., D. LINKENS, A. SELMAN, AND A. MCDONALD. The interaction between heart rate and respiration. Part II. Nonlinear analysis based on computer modelling. Automedica Lond. 4: 141- 153,1982. KITNEY, R. I. Proceedings: entrainment of the human RR interval by thermal stimuli. J. Physiol. Lond. 252: 37P-38P, 1975. KOLLAI, M., AND G. MIZSEI. Respiratory sinus arrhythmia is a limited measure of cardiac parasympathetic control in man. J. Physiol. Lond. 424: 329-342,199O. LILLIEFORS, H. W. On the Kolmogorov-Smirnov test for normality with mean and variance unknown. J. Am. Stat. Assoc. 62: 399-402, 1967. LISHNER, M., S. AKSELROD, V. MOR AVI, 0. Oz, M. DIVON, AND M. RAVID. Spectral analysis of heart rate fluctuations. A non-invasive, sensitive method for the early diagnosis of autonomic neuropathy in diabetes mellitus. J. Auton. New. Syst. 19: 119-125, 1987. MALLIANI, A., M. PAGANI, F. LOMBARDI, AND S. CERUTTI. Cardio-

by 10.220.33.1 on March 30, 2017


ownloaded from



vascular neural regulation explored in the frequency domain. Cir- 32. culation 84: 482-492, 1991.

25. MYERS, G. A., G. J. MARTIN, N. M. MAGID, P. S. BARNETT, J. W. SCHAAD, J. S. WEISS, M. LESCH, AND D. H. SINGER. Power spectral 33. analysis of heart rate variability in sudden cardiac death: comparison to other methods. IEEE Trans. Biomed. Eng. 33: 1149-1156, 1986.

ROHMERT, W., W. LAURIG, U. PHILIPP, AND H. LUCZAK. Heart rate variability and work-load measurement. Ergonomics 16: 33-44, 1973.

%a.NOvAK, V., P. NOVAK, J. DE CHAMPLAIN, A. R. LE BLANC, R. MAR- 34.

ROMPELMAN, O., A. J. R. M. COENEN, AND R. I. KITNEY. Measure- ment of heart-rate variability. Part I. Comparative study of heart- rate variability analysis methods. Med. Biol. Eng. Comput. 15: 233- 239, 1977. SAUL, J. P., Y. ARAI, R. D. BERGER, L. S. LILLY, W. S. COLUCCI, AND R. J. COHEN. Assessment of autonomic regulation in chronic congestive heart failure by heart rate spectral analysis. Am. J. Car- dial. 61: 1292-1299, 1988. SAUL, J. P., R. D. BERGER, M. H. CHEN, AND R. J. COHEN. Transfer function analysis of autonomic regulation. II. Respiratory sinus arrhythmia. Am. J. Physiol. 256 (Heart Circ. Physiol. 25): H153-H161, 1989. SEALS, D. R., N. 0. SUWARNO, AND J. A. DEMPSEY. Influence of lung volume on sympathetic nerve discharge in normal humans. Circ. Res. 67: 130-141, 1990. SELMAN, A., A. MCDONALD, R. KITNEY, AND D. LINKENS. The interaction between heart rate and respiration. Part I. Experimen- tal studies in man. Automedica Land. 4: 131-139, 1982. SHANNON, D. C., D. W. CARLEY, AND H. BENSON. Aging of modulation of heart rate. Am. J. Physiol. 253 (Heart Circ. Physiol. 22): H874-H877,1987. TOBIN, M. J., M. J. MADOR, S. M. GUENTHER, R. F. LODATO, AND M. A. SACKNER. Variability of resting respiratory drive and timing in healthy subjects. J. Appl. Physiol. 65: 309-317, 1988. VYBIRAL, T., R. J. BRYG, M. E. MADDENS, AND W. E. BODEN. Ef- fect of passive tilt on sympathetic and parasympathetic components of heart rate variability in normal subjects. Am. J. Cardiol. 63: 1117-1120, 1989. WALTER, G. F., AND S. W. PORGES. Heart rate and respiratory responses as a function of task difficulty: the use of discriminant analysis in the selection of psychologically sensitive physiological responses. Psychophysiology 13: 563-571, 1976. WELCH, P. D. The use of fast Fourier transform for the estimation of power spectra: a method based on time averaging over short, modified periodograms. IEEE Trans. Audio Electroaccoust. 15: 70- 73, 1967.

26.

27.

28.

29.

30.

31.

TIN, AND R. NADEAU. Influence of respiration on heart rate and blood pressure fluctuations. J. Appl. Physiol. 74: 617-626, 1993. NUGENT, S. T., AND J. P. FINLEY. Spectral analysis of periodic and normal breathing in infants. IEEE Trans. Biomed. Eng. 30: 672- 35. 675, 1983. PAGANI, M., F. LOMBARDI, S. GUZZETTI, 0. RIMOLDI, R. FURLAN, P. PIZZINELLI, G. SANDRONE, G. MALFATTO, S. DELL’ORTO, E. PICCALUGA, M. TURIEL, G. BASELLI, S. CERUTTI, AND A. MAL- 36. LIANI. Power spectral analysis of heart rate and arterial pressure variabilities as a marker of sympatho-vagal interaction in man and conscious dog. Circ. Res. 59: 178-193, 1986. 37.

PAGANI, M., F. LOMBARDI, S. GUZZETTI, G. SANDRONE, 0. RI- MOLDI, G. MALFATTO, S. CERUTTI, AND A. MALLIANI. Power spectral density of heart rate variability as an index of sympatho-vagal 38* interaction in normal and hypertensive subjects. J. Hypertens. 2, Suppl. 3: 383-385, 1984. PORGES, S. W., R. E. BOHRER, M. N. CHEUNG, F. DRASGOW, P. M. 3g* MCCABE, AND G. KEREN. New time-series statistic for detecting rhythmic co-occurrence in the frequency domain: the weighted co- herence and its application to psychophysiological research. Psy-

4. ’

chol. Bull. 88: 580-587, 1980. PORTER, T. R., D. L. ECKBERG, J. M. FRITSCH, R. F. REA, L. A. BEIGHTOL, J. F. SCHMEDTJE, JR., AND P. K. MOHANTY. Autonomic pathophysiology in heart failure patients. Sympathetic-cholinergic

41 ’

interrelations. J. Clin. Invest. 85: 1362-1371, 1990. RABINER, L. R., R. W. SCHAFER, AND D. DLUGOS. Periodogram method for power spectrum estimation. In: Programs for Digital 42. Signal Processing, edited by Digital Signal Processing Committee, IEEE Acoustics, Speech, and Signal Processing Sot. New York: IEEE, 1979, p. 2.1-l-2.1-10. by 10.220.33.1 on M

arch 30, 2017http://jap.physiology.org/

Dow

nloaded from


Documents

Important influence of respiration on human R-R interval ...€¦ · Important influence of respiration on hu- man R-R interval power spectra is largely ignored. J. Appl. Phys- iol