Abstract
Numerous studies have demonstrated a diurnal rhythm in indices of pulmonary function in both healthy subjects and subjects with asthma, with minima occurring during the night. To determine whether such diurnal changes are caused by an endogenous circadian rhythm or by diurnal alterations in behavior or the environment, we measured indices of pulmonary function throughout a “constant routine” protocol designed to unmask underlying circadian rhythms. After two acclimation days in the laboratory, 10 healthy adults maintained relaxed wakefulness in a semirecumbent posture in a constant environment with low light (10 lux) for 41 h. Measurements of FEV1, FEVC, PEF, blood cortisol, and core body temperature (CBT) were performed every 2 h. Results of cosinor analysis of group data aligned to CBT circadian minimum revealed significant circadian variations in FEV1 and FEV1/FEVC, cortisol, and CBT, and lack of significant circadian variations in FEVC and PEF. The ranges (peak to trough) of mean circadian changes in spirometric variables were 2.0–3.2% of the mesor. The circadian minima of all variables occurred within the usual sleep period (although subjects remained awake). Because of differences in phase relationships between CBT and pulmonary function among subjects, the circadian rhythms within subjects were generally larger than the group average circadian changes, being significant for FEV1/FEVC in 5 of 10 subjects and for PEF in 6 of 10 subjects. Sleep deprivation (24 h) failed to cause a significant change in any pulmonary function variable (when controlled for circadian phase). Thus, endogenous circadian rhythms contribute to diurnal changes in pulmonary function in healthy subjects.
Numerous studies have demonstrated a diurnal rhythm in indices of pulmonary function. The clearest examples of this phenomenon are the existence of patients with asthma who experience a nocturnal worsening of pulmonary function, asthmatic symptoms, and sleep quality, and the observation of a diurnal variation in the death rate among patients with asthma with the peak occurring during the early morning hours (e.g., [1–11]). However, such studies were not able to determine whether these diurnal variations are caused by an endogenous circadian rhythm (related to the output of the circadian pacemaker), diurnal alterations in behavior (e.g., activity, meals, posture, or sleep) or diurnal alterations in the environment (e.g., temperature, light, or allergen exposure).
To try and separate effects of sleep from effects of circadian rhythms, some studies have used total sleep deprivation, waking subjects from sleep, or changing the sleep schedule from night to day (3, 5, 12-19). While the results of these studies are quite variable, overall the evidence is suggestive of independent influences on pulmonary function from the circadian pacemaker and from diurnal behaviors. However, clear assessment of the separate contributions of circadian and sleep-dependent influences on pulmonary function has been hampered by the following general limitations of previous studies: (1) there has been lack of control of environment or behavior during protocols; (2) measurements occurred at few circadian phases (e.g., 4:00 p.m. and 4:00 a.m.), with the likelihood of peaks and troughs being missed; (3) markers of the endogenous circadian rhythm (such as core body temperature) were not measured; and (4) the underlying circadian rhythm may have changed throughout the protocol because of uncontrolled light exposure and behaviors. In addition, sleep deprivation studies have their limitations because altered pulmonary function could be caused by underlying circadian rhythms, by sleep deprivation itself (e.g., the buildup of a sleep-promoting factor affecting bronchiolar smooth muscle tone), or by the altered behaviors that occur as a consequence of imposing nighttime wakefulness (e.g., increased activity, stress, social interaction, food intake, ambient light and noise, or altered posture). Furthermore, there can be an interaction of these last two effects because altered behaviors and environment can alter the phase of the circadian rhythms. For example, the increased light exposure during the sleep deprivation period will likely shift the phase of the circadian rhythm because subjects are exposed to light during their usual sleep hours when the resetting of the endogenous circadian pacemaker is most sensitive to the effects of even low-level room light (20, 21).
To address some of the limitations of previous studies, the current study has made pulmonary function measurements in healthy humans throughout an established “constant routine” protocol (e.g., [22, 23]) that is specifically designed to “unmask” underlying circadian rhythms by making measurements under constant behavioral and environmental conditions during prolonged wakefulness. We sought to determine the magnitude of any endogenous circadian rhythm of indices of pulmonary function. A secondary aim was to determine the “pure” effect of sleep deprivation on indices of pulmonary function—while controlling for circadian phase. We ensured that our subjects maintained relatively constant behavior (relaxed wakefulness), a constant posture (semirecumbent), ate regular small meals (uniform snacks every 2 h), and stayed in an environment with constant room temperature and very low light (10 lux). Additional improvements of the current protocol relative to previous studies include an extended screening period with a stable sleep/wake schedule, two full days and nights in the laboratory for acclimation, frequent measurements (every 2 h for 41 h), concurrent measurement of robust markers of the endogenous circadian pacemaker (core body temperature and plasma cortisol), and electroencephalographic (EEG) confirmation of wakefulness throughout the protocol. In this initial experiment, we chose to study healthy adult subjects to provide a baseline for any future studies in patient groups.
The study was approved by the internal review board at Brigham and Women's Hospital (Boston, MA). Subjects were informed of the procedures and possible risks, and each subject provided written consent to participate. The study was conducted with 10 healthy, adult male subjects: their age was 23.7 ± 3.9 yr (mean ± SD), their height was 176.3 ± 6.4 cm, their weight was 73.9 ± 11.3 kg (body mass index [BMI] = 23.8 ± 3.3 kg m−2), and they had normal pulmonary function: vital capacity, 5.5 ± 0.9 L (106 ± 12% pred), forced expiratory volume in 1 s (FEV1), 4.5 ± 0.8 L (105 ± 11% pred), and peak expiratory flow (PEF), 10.2 ± 1.9 L s−1 (110 ± 12% pred). The subjects did not regularly take medications, and they were free from all medications (including over-the-counter medications) for at least 1 wk before initiating the study. Individuals with evidence of significant psychopathology were excluded.
To ensure a stable circadian baseline before entry into the laboratory, subjects were asked to establish a regular sleep–wake schedule with their habitual bedtimes and wake times varying by no more than 1 h each day for at least 2 wk before the laboratory study. This schedule was verified by time-logged telephone calls of bedtimes and wake times for 2 wk as well as activity monitoring for the 3–7 d before entering the laboratory (wrist-worn Actigraph; Ambulatory Monitoring, Ardsley, NY). One week before the study, subjects abstained from caffeinated drinks and smoking.
Subjects were then studied in an individual laboratory suite isolated from sunlight and external time cues, including clocks, radios, televisions, visitors, and mail. Subjects maintained contact with staff members who were trained to avoid communicating the time of day. Laboratory temperature was maintained at approximately 23° C. The laboratory protocol included two initial baseline days (16 h wakefulness with light level at 150 lux followed by an 8-h sleep opportunity with light level < 1 lux) that enabled the subjects to acclimate to the environment, to become familiarized with the equipment and respiratory test (see below), and to assist in achieving a relative steady state. Each day, subjects had three meals and one snack composed of approximately 25% fat, 50% carbohydrate, and 25% protein. They also received 3.5 L of fluids per 24-h period. The overall calories were calculated by the Harris–Benedict formula with an activity factor of 1.4 (24). The sleep periods also served to ensure the absence of any primary sleep disorders by performance of standard overnight polysomnography with surface electrodes for two electroencephalograms, electrooculograms, submental electromyogram, respiratory inductance plethysmography, airflow, and arterial oxygen saturation. All signals were recorded on an electroencephalograph (Nicolet Biomedical, Madison, WI).
Subjects woke up at their usual time on the morning of the third day in the laboratory and remained in bed and awake in a semirecumbent position for 41 h in an established “constant routine” protocol (22, 23). This constant routine protocol was performed under the same time-isolation conditions as the “laboratory baseline monitoring.” During the constant routine, the light level was 10 lux, such that it had little or no entraining effect on the circadian pacemaker (20). Small identical snacks and 292 ml of fluids were given every 2 h with an overall composition that was similar to the baseline nutrition. Experimenters were present in the laboratory throughout the constant routine to ensure that the subjects remained awake, and this was verified from continuous recordings of two electroencephalograms, two electrooculograms and a submental electromyogram (Nicolet Biomedical). All necessary adjustments were performed during the first hour of wakefulness (e.g., change to semirecumbent posture, reapplication of electrodes).
Measurement of standard markers of the circadian pacemaker. Two robust markers of the endogenous circadian pacemaker were measured throughout the 4-d stay in the laboratory: core body temperature (CBT) was recorded continuously with a rectal temperature sensor (Yellow Springs Instrument Company, Yellow Springs, OH) and plasma cortisol concentration was determined from 1.75 ml of blood sampled from an indwelling venous catheter every 30 min. Cortisol analysis was performed with a paramagnetic, chemoluminescent immunoassay (Beckman Coulter, Miami, FL).
Respiratory measurements. Starting 1 h after awakening, the subjects performed forced expiratory maneuvers at 2-h intervals. The measurement was performed while subjects breathed through a mouthpiece with a nose-clip in place. Expired volumes and flows were measured with calibrated pulmonary function equipment, using a turbine for volume measurement (OxyconAlpha system; Erich Jaeger GmbH, Würzburg, Germany). The following variables were measured for each forced expiratory maneuver: FEV1, forced expiratory vital capacity (FEVC), FEV1/FEVC, and PEF. For each variable in each individual, the highest value obtained in the three best trials at each measurement period was used in subsequent analyses.
To further ensure that circadian measurements were made under basal conditions, the first 5 h of constant routine data was excluded from all analyses to eliminate any residual effects of sleep on pulmonary function variables (22, 23). Spirometric variables were corrected to btps units. For group averages, data for each subject were expressed in terms of a percentage deviation from the 24-h mean (centered around the CBT minimum).
Analysis of circadian rhythmicity. The circadian phase and period of the CBT rhythm were estimated for each subject by a two-harmonic regression analysis of the temperature data, using a nonlinear least-squares method (25). This established technique constrains the circadian period to be within the normal physiological range of 24.0 to 24.3 h (26). Allowing this small degree of flexibility of circadian period results in a more accurate estimate of circadian phase. Thereafter, for comparison among variables sampled at different frequencies, 2-h averages of all variables were calculated, centered around the minimum of CBT for every subject. The circadian phases and amplitudes of all variables were then estimated by cosinor analysis, using the following equation (27): y = M + A · cos (ω · t + φ) + s · t (where y = selected variable; M = mesor; A = amplitude of circadian rhythm [peak minus mean]; ω = 2π/period of circadian rhythm [with the circadian period derived for each subject by initial nonlinear analysis of CBT data as described above]; φ = phase of circadian rhythm [relative to CBT] = −ω · t; t = time [h]; s = underlying linear trend [h−1]). The linear trend term (s · t) represents any effects of prolonged wakefulness. Similar analyses were performed on the group data with each individual's data aligned with respect to their minimum CBT, and the subject's data averaged at each circadian phase. These group data were subjected to cosinor analysis, using the average period of subjects' rhythms derived from the initial nonlinear analyses of CBT data as described above (i.e., 24.08 h). The phase shift between the circadian rhythms of CBT and the other variables was estimated as the time lag between minima of the fitted rhythms. Also, since there were substantial differences in phase between the circadian rhythms of CBT and the indices of pulmonary function among subjects, an additional (post hoc) group analysis was performed with each individual's pulmonary function data aligned with respect to the minimum of their pulmonary function variable derived from the individual cosinor analyses (i.e., rather than aligned with respect to their minimum CBT).
The existence of significant circadian rhythms was determined by applying the cosinor analysis on the detrended data, i.e., y − s · t. For this analysis, there were 16 data points for group data (i.e., 15 degrees of freedom) and 18 data points for individual subjects (i.e., 17 degrees of freedom). Significant circadian oscillations were acknowledged to have occurred if p < 0.05 for F3/13 (group analysis) or F2/15 (individuals' analyses).
Analysis of the effects of sleep deprivation, while controlling for circadian phase. To assess the “pure” effect of 24-h sleep deprivation from the 41-h constant routine data while controlling for circadian phase, data at specific circadian phases in the first circadian cycle were compared with data at the same circadian phase in the subsequent circadian cycle. For group mean analysis, each subject's data were aligned with respect to their wake-up time (rather than their CBT minimum). The first 5 h of constant routine data was discarded from analysis to ensure a steady state (as previously noted), and two-way repeated measures analyses of variance were performed comparing six pairs of measurements with each pair separated by 24 h sleep deprivation (repeated measures = six phases of the circadian cycle × 2 d) (SAS statistical software; SAS Institute, Cary, NC). Significant effects of sleep deprivation were acknowledged to have occurred if p < 0.05.
Assessment of overnight change in pulmonary function in the absence of sleep. Numerous studies report overnight decrements in indices of pulmonary function (3, 16). These changes could be attributable to the occurrence of sleep or to factors related to the passage of time, including an underlying circadian rhythm. To determine whether this overnight decrement in pulmonary function also occurs without sleep, data were compared across the 8-h period during the constant routine that represented the beginning and end of the usual sleep period (paired t tests).
No subject had any sleep-related breathing disorder detected during the baseline nights. All subjects remained awake throughout the constant routine protocol, according to standard criteria (28).
Results of the group data aligned according to the circadian minimum of CBT are shown in Table 1 and Figure 1. The cosinor analyses of detrended group mean data explained between 11 and 72% of the total variation in the spirometric data, and 95% of the variation in the established markers of circadian rhythms, i.e., CBT and cortisol, respectively. These analyses revealed that there was a significant circadian variation in group mean FEV1 (p = 0.020) and FEV1/FEVC (p = 0.008), and a lack of significant circadian variation in group mean FEVC (p = 0.073) and PEF (p = 0.483). The ranges (peak to trough = twice the amplitude) of mean circadian changes of FEV1 and FEV1/FEVC were 3.5 and 2.6% of the mesor, respectively. The ranges of mean circadian changes of FEVC and PEF were 2.2 and 2.2% of the mesor, respectively. The circadian minima of these spirometric variables occurred in the group mean data within the usual sleep period (even though subjects remained awake), but usually well before the circadian minimum of CBT (Table 1 and Figure 1).
| Variable | Mesor | Amplitude | Trend | Phase Difference versus CBT (h) | ||||
|---|---|---|---|---|---|---|---|---|
| FEV1 | 4.36 L | 0.07 L | 0.003 L h−1 | −4.2 | ||||
| FEVC | 5.58 L | 0.07 L | −0.001 L h−1 | −6.8 | ||||
| FEV1/FEVC | 79.4% | 1.1% | 0.048% h−1 | −2.9 | ||||
| PEF | 10.1 L s−1 | 0.1 L s−1 | −0.017 L s−1 h−1 | −0.2 | ||||
| Cortisol | 9.4 μg dl−1 | 5.5 μg dl−1 | 0.062 μg dl−1 h−1 | −7.9 | ||||
| CBT | 37.2° C | 0.3° C | −0.001° C h−1 | 0 |
Fig. 1. Circadian rhythms of pulmonary function detected by a constant routine protocol. Shown are the group mean levels (± SEM) of forced expiratory volume in 1 s (FEV1), forced expiratory vital capacity (FEVC), FEV1/FEVC, peak expiratory flow (PEF), plasma cortisol concentration, and core body temperature (CBT). Individual data (n = 10) were aligned with respect to the reference circadian rhythm (minimum of core body temperature, CBTmin) and averaged. The SEMs of all data are shown, but are small and indistinguishable from the mean at most time points for cortisol and CBT. Wider variation occurred in the other variables. The ordinate is expressed as deviation from the 24-h mean (left) and in absolute units (right). The 24-h data, collected during the constant routine and centered around the CBTmin, are “double plotted” for ease of visualizing circadian rhythms (i.e., the 24-h data on the left are reproduced on the right). The abscissa is expressed in degrees (with CBTmin assigned a phase of zero degrees) and in relative clock hour. The gray bars on the abscissa represent the time of the subjects' usual sleep episode (although sleep did not occur during this constant routine). Small differences occur in comparison with Table 1 because the values in Table 1 were derived from the cosinor analysis, whereas the actual mean data are plotted here.
[More] [Minimize]Differences among subjects in the circadian phase of their pulmonary function indices meant that circadian rhythm amplitudes within individual subjects were generally larger than the group average. From the individual analyses, the range among subjects of the maximal circadian changes across the circadian cycle was 1.9 to 18.2% of the mesor for FEV1 (median, 3.0%); 0.2 to 8.9% for FEVC (median, 2.6%); 0.5 to 16.6% for FEV1/FEVC (median, 4.95%); and 0.3 to 35.7% for PEF (median, 8.8%). Thus, many individuals showed prominent circadian rhythms in FEVC and in PEF even though the group average waveform failed to reach significance. Indeed, when individuals were considered, the greatest circadian variability emerged in PEF. The results of the individual analyses are shown in Figure 2. Figure 2 also indicates the range of individual variation across the 41-h constant routine protocol for each variable in order to illustrate the ratio between the amplitude of each individual's circadian rhythms relative to the inherent variability in that measurement. It can be seen that, on average, the circadian variation (as assessed by the cosinor model) accounts for less than half of the overall variability in the data. With the data of each individual aligned with respect to the minimum of their pulmonary function variable derived from the individual cosinor analyses (i.e., rather than aligned with respect to their CBT minimum), significant circadian variations were detected for both FEV1/FVC and PEF. This can be seen on the right-hand side of Figure 2. Figure 3 shows examples of individual subject PEF rhythms, demonstrating the marked differences in phase between the circadian rhythms of CBT and PEF.
Fig. 2. Amplitudes (peak to trough) of individual subject circadian rhythms derived from cosinor analysis (solid bar); range of raw data across the constant routine (open thin bar), and circadian amplitude and inherent variability (range) of the group mean data aligned according to the phase of each individual's cosinor analysis for each variable (i.e., rather than aligned with CBTmin). Shown are forced expiratory volume in 1 s (FEV1), forced expiratory vital capacity (FEVC), FEV1/FEVC, and peak expiratory flow (PEF). The ordinates are expressed as deviations from the 24-h means. Significant circadian rhythms are indicated: *p < 0.05; **p < 0.01; ***p < 0.001.
[More] [Minimize]
Fig. 3. For three example subjects, peak expiratory flow data (PEF) and cosinor model output aligned to core body temperature minimum (CBTmin, dashed vertical line) are shown. R2 = proportion of variance in data explained by each cosinor model fit. The abscissa shows the time of day (clock hour) for individual subjects. The bars on the abscissa represent the time of the usual sleep episode for each subject (although sleep did not occur during this constant routine). Note substantial differences among subjects in the phase relationships between PEF and CBTmin. Phase differences among subjects also existed for PEF in relation to habitual waketime, and for PEF in relation to actual clock time.
[More] [Minimize]Individual correlation analyses were performed to determine possible causal links among variables across the constant routine. The correlation matrix is presented in Table 2. These correlations reveal that in at least 8 of 10 subjects there were no significant correlations (at lag time zero) between either cortisol or CBT and the three primary indices of pulmonary function (FEV1, FEVC, and PEF). Also, there was no significant correlation between FEVC and PEF in 7 of 10 subjects, and no significant correlation between FEV1 and PEF in 6 of 10 subjects.
| Variable | FEV1 | FEVC | FEV1/FEVC | PEF | Cortisol | CBT | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| FEV1 | — | 7 | 6 | 4 | 0 | 1 | ||||||
| FEVC | 0.5255 | — | 3 | 3 | 1 | 2 | ||||||
| FEV1/FEVC | 0.5820 | −0.3305 | — | 3 | 2 | 4 | ||||||
| PEF | 0.4085 | 0.2980 | 0.2840 | — | 2 | 2 | ||||||
| Cortisol | 0.3035 | 0.1070 | 0.1160 | −0.0885 | — | 3 | ||||||
| CBT | 0.0450 | −0.0505 | 0.2580 | 0.1145 | −0.3420 | — |
The mean data before and after 24-h sleep deprivation are shown in Figure 4 and the results of the statistical analyses are presented in Table 3. In this strictly controlled protocol, 24-h sleep deprivation failed to cause a significant change in any variable for the group. In addition, comparisons of individual changes revealed that at least 7 of the 10 subjects had no significant change in any of the spirometric variables after 24-h sleep deprivation, consistent with the results of the group analyses.
Fig. 4. Effect of 24-h sleep deprivation on forced expiratory volume in 1 s (FEV1), forced expiratory vital capacity (FEVC), FEV1/FEVC, and peak expiratory flow (PEF). Individual data are shown before and after 24 h of sleep deprivation (mean of six measurements before and six measurements after 24 h of sleep deprivation are connected by lines). The group mean values are shown as short horizontal lines. Repeated measures ANOVAs that controlled for circadian phase failed to find significant effects of 24-h sleep deprivation on any of these variables.
[More] [Minimize]| Hours 6–11 after Awakening | Hours 36–41 after Awakening | Mean 24-h Difference | % Change | p Value (ANOVA) | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| FEV1, L | 4.39 ± 0.75 | 4.44 ± 0.73 | 0.05 ± 0.16 | 1.2 ± 4.0 | 0.3772 | |||||
| FEVC, L | 5.52 ± 0.89 | 5.53 ± 0.81 | 0.01 ± 0.21 | 0.5 ± 3.8 | 0.9212 | |||||
| FEV1/FEVC, % | 80.4 ± 6.6 | 81.2 ± 7.0 | 0.75 ± 3.21 | 1.0 ± 4.1 | 0.5141 | |||||
| PEF, L s−1 | 10.0 ± 1.6 | 9.4 ± 1.6 | −0.57 ± 1.05 | −5.3 ± 9.2 | 0.1204 | |||||
| Cortisol, μg dl−1 | 7.45 ± 2.08 | 8.71 ± 1.64 | 1.25 ± 1.35 | 19.8 ± 17.1 | 0.0168 | |||||
| CBT, °C | 37.26 ± 0.17 | 37.22 ± 0.16 | −0.05 ± 0.10 | 0.1866 |
There was no significant change in any pulmonary function variable in these healthy awake subjects across the 8-h period that represented the beginning and end of the usual sleep period. The group mean changes (means = SEM) over this 8-h usual sleep period were 2.2 ± 2.8% for FEV1; 0.8 ± 1.8% for FEVC; 3.0 ± 4.2% for FEV1/FEVC; and −0.2 ± 7.6% for PEF.
The current study is the first to use an established technique to determine the degree to which the endogenous circadian pacemaker contributes to the previously reported diurnal changes in pulmonary function (e.g., [3, 5, 12–19]). Our results on group mean data when aligned with CBT minimum revealed small but significant circadian variations in FEV1 and FEV1/ FEVC, and a lack of significant circadian variation in FEVC and PEF, in healthy adults. The circadian minima of spirometric variables occurred within the usual sleep period (even though subjects remained awake). The individual subjects had larger amplitude circadian rhythms in these variables as well as large individual differences in phase between the circadian rhythms of CBT and the indices of pulmonary function. These results are qualitatively similar to but quantitatively less than the findings of most of the diurnal studies mentioned above and suggest that an endogenous circadian rhythm partly contributes to the diurnal rhythms in pulmonary function in healthy subjects.
Although established circadian protocols were not used previously, some studies have tried to separate the effects of circadian rhythms from the effects of sleep on pulmonary function by using total sleep deprivation (3, 5, 16, 19), waking subjects from sleep (3, 13, 17, 18), changing the sleep schedule from night to day over the short term (3, 15), and shift work (12, 14). For example, Hetzel and Clark (3) found that some patients with asthma had similar overnight reductions in PEF when sleep was either permitted or prevented, and yet there was another subgroup of subjects in whom PEF fell only after they had been allowed to sleep. Ballard and coworkers (19) measured lower airway resistance in a similar protocol and found that resistance fell when sleep was prevented, but the reduction in resistance was greater when sleep was permitted in patients with asthma (but not in healthy subjects). Despite the differences in lower airway resistance, the fall in FEV1 overnight was similar both with and without sleep in the patients with asthma. These studies provide some evidence of the existence of independent influences on pulmonary function from the circadian pacemaker and from diurnal behaviors including sleep.
In assessing the amplitude of the endogenous circadian rhythm of pulmonary function in the laboratory (current study) versus the amplitude of the diurnal rhythm of pulmonary function in the home (previous studies) it is important to recognize that various authors have reported amplitudes of circadian rhythms in different ways. We have used the traditional approach in circadian biology of aligning subjects according to the phase of a robust circadian marker (CBT) and averaging (Figure 1), and by reporting the circadian amplitude in the traditional way as the difference between the mean and the peak (Table 1). As noted above, this approach will certainly give lower values for circadian amplitude when compared with other studies that (1) perform cosinor analysis on individual data before averaging ([4, 29]; and Figure 2 of current study); (2) report circadian amplitude as the peak-to-trough value from cosinor analysis rather than the mean-to-peak value (4); or (3) express the circadian amplitude as the absolute range in the raw data (e.g., [30]; and Figure 2 of current study).
When considering the endogenous circadian rhythms alone, group average waveforms in the current study indicate the existence of relatively small-amplitude rhythms in pulmonary function in healthy subjects (peak to trough, 2.2 to 3.5%). However, because of variability among subjects in the phases of the pulmonary function rhythms in relation to the circadian phase of CBT, individuals had larger rhythms than the group average, particularly for FEV1/FEVC (median, 4.9%) and PEF (median, 8.8%). Thus, aligning the data of each subject according to the circadian phase of CBT before averaging the pulmonary function data resulted in a smaller group rhythm in pulmonary function than when aligning the data for each subject according to the circadian phase of pulmonary function before averaging. This discrepancy could reflect real physiological variation among subjects and/or inaccuracies in the estimation of the circadian phase of either pulmonary function or CBT (note, 1 standard deviation of circadian phase estimation from constant routine data = 0.6 h) (30).
Other studies that used cosinor analysis of individual data have yielded results qualitatively similar to those of the current study. For example, a wide distribution in phase was also observed in the study by Hetzel and Clark (4), with the acrophase (peak) in PEF ranging widely over the period between 2:00 p.m. and 10:00 p.m. From data collected at home from ambulant, healthy subjects, previous studies found the circadian range (i.e., peak to trough) in PEF to be 8.3 to 12.5% of the mesor (4, 29), with the highest values in the afternoon and the lowest values in the early morning hours. D'Alonzo and coworkers (31) demonstrated that the absolute diurnal variability in the FEV1 and PEF could be underestimated by 55–80% by making only two measurements separated by 12 h, and underestimated by 20–45% by making four measurements separated by 8 h, in comparison with measurements separated by 2 h for 24 h. Thus, although these other studies made sporadic measurements at times that are expected to coincide approximately with the maximal diurnal peak-to-trough variability, it is possible that had they made more frequent measurements, larger rhythms would have been encountered. With these considerations in mind, it is evident that the overall diurnal change in pulmonary function in ambulant healthy subjects in the home (previous studies) is on the whole somewhat larger than the endogenous circadian rhythm of pulmonary function detected in the current study with subjects in a constant environment with controlled behaviors.
A secondary aim of the present study was to determine the “pure” effect of sleep deprivation on indices of pulmonary function, while controlling for circadian phase. Cooper and Phillips (32) and Chen and Tang (33) found in healthy subjects that 24 h of sleep deprivation caused a significant decline in maximal voluntary ventilation and inspiratory muscle endurance, but less than a 2% change in FEV1 and FEVC. Phillips and coworkers (34) found that patients with chronic obstructive pulmonary disease (COPD) had a 5–6% significant fall in both FEV1 and FEVC following 24 h of sleep deprivation. It is plausible that these earlier findings are an artifact of the experimental protocols, which did not control behavior or environment during the imposed nighttime wakefulness, such as activity, stress, social interaction, food intake, ambient light and noise, or altered posture. For example, the increased light exposure during the sleep deprivation period will likely shift the phase of the circadian rhythm (with consequences on pulmonary function) because subjects are exposed to light during their usual sleep hours when the resetting of the endogenous circadian pacemaker is most sensitive to the effects of even low-level room light (20, 21). In addition, none of these studies measured markers of the underlying circadian rhythm to ensure that this did not change between measurements. In the current study, in which we controlled behavior and environment during the sleep deprivation period (including maintaining the light exposure at < 10 lux so that there would be no effect on the endogenous circadian pacemaker), we were able to assess the effect on pulmonary function of 24 h of sleep deprivation at numerous points across the circadian cycle. From our findings it is clear that sleep deprivation has little to no independent effect on pulmonary function in healthy subjects.
Previous studies have found an overnight decline in indices of pulmonary function in patients with asthma, with less of a decline if sleep was not permitted (e.g., [3, 16]). For instance, there was a 33% overnight decline in PEF when sleep occurred between measurements at 10:00 p.m. and 6:00 a.m. but only a 20% decline overnight when sleep did not occur (16). Nonetheless, the decline overnight without sleep occurred in all 11 patients and was still highly significant (16). Those data suggest that both sleep and circadian rhythms independently affect pulmonary function in patients with asthma. In contrast, in the current highly controlled study, we found no significant change in any pulmonary function variable in healthy subjects who remained awake across the 8-h period that represented the beginning and end of the usual sleep period. Indeed, some subjects even increased PEF over the usual sleep period, presumably because of the underlying circadian influence (e.g., Figure 3). Although we did not assess the independent effect of sleep on pulmonary function in our subjects, these observations suggest that the superposition of sleep and circadian influences on pulmonary function would not necessarily result in an overall decline in pulmonary function because in some healthy subjects these two influences could cancel out.
Circadian scientists have developed specific protocols to “unmask” underlying endogenous circadian rhythms from the influences of varied environments and behavior (e.g., [22, 23]). For instance, the constant routine protocol has been effectively used to determine the existence and magnitude of endogenous circadian rhythms and sleep deprivation effects on numerous physiological and neuropsychological variables including core body temperature, plasma melatonin, plasma cortisol, vigilance, and cognitive performance. In terms of quantifying the endogenous circadian component of pulmonary function, the improvements of the current study over previous studies include (1) control of environment and behavior during protocols, (2) increased frequency of measurements in order to avoid missing the peaks and troughs of the circadian rhythm (i.e., measurements every 2 h across the entire circadian cycle), (3) measurement of known markers of the output of the endogenous circadian pacemaker (core body temperature and cortisol), and (4) ensuring that the protocol does not induce changes in the underlying circadian rhythm by controlled light exposure and behaviors. Despite the numerous differences in protocol between the previous diurnal studies and the current circadian study, circadian rhythms in pulmonary function are qualitatively similar among studies. However, our study suggests that the diurnal effects noted in the other studies are larger than can be expected from influences by the endogenous circadian pacemaker alone, suggesting that posture, sleep, or other changes in behavior or the environment have affected the results of previous studies.
There are some limitations to our study. Chief among these is the relevance of the data to the clinical condition of asthma. Previous studies have found increasingly larger diurnal ranges in PEF in smokers and patients with chronic obstructive lung diseases (29), but the greatest diurnal rhythm is detected in patients with asthma (e.g., peak-to-trough variability = 50.9% of the mean; [4]). A subpopulation of patients with asthma experience a worsening of symptoms overnight (6-8). This has resulted in the classification of “nocturnal asthma,” defined as “a variable exacerbation of the underlying asthma condition associated with increases in symptoms, need for medication, airway responsiveness, and/or worsening of lung function” (35). Nonetheless, it has been suggested that nocturnal asthma is merely an indication of worsening asthma rather than a separate disease entity (9-11). Similarly, there may be a spectrum from healthy normal subjects to patients with asthma. We note that although all our healthy volunteer subjects had normal pulmonary function during baseline days, we did not perform methacholine or cold air challenges in our subjects, raising the remote possibility that a subclinical degree of asthma could be present in some of our “healthy subjects.” While patients with asthma exhibit larger than normal diurnal rhythms in bronchiolar smooth muscle tone, lung volume, bronchial reactivity, airway resistance, and airway inflammation (reviewed in [35]), the fact that diurnal changes in indices of pulmonary function have similar time courses in healthy and asthmatic subjects (e.g., [3]) does suggest that data from healthy subjects may have relevance to nocturnal asthma.
A second limitation of the current study is that it provides little information on the underlying mechanisms of the circadian changes in pulmonary function. Nonetheless, some information may be gleaned from phase relationships among variables. The group average rhythms (aligned to the minimum of CBT) of the FEV1 and FEV1/FEVC were phase advanced by ∼ 3–4 h with respect to CBT and phase delayed by ∼ 4–5 h with respect to blood cortisol rhythm (Table 1). In addition, individual correlation analyses (Table 2) reveal no systematic relationship between either cortisol or CBT and the three primary indices of pulmonary function, suggesting that circadian changes in pulmonary function were not caused by circadian rhythms in either CBT or cortisol, at least in these healthy subjects. However, to be certain that these variables are unrelated, it will be necessary to experimentally adjust them.
Troyanov and coworkers (36) found FEV1 and PEF to be equally predictive of airway caliber in a heterogeneous group of normal subjects and subjects with asthma. However, in our healthy subjects throughout the constant routine protocol, PEF has a larger circadian rhythm than FEV1 (Figure 2). Certainly, effort dependence of PEF but not FEV1 is a possible explanation. However, we doubt that this contributed to our results, as we found that 24-h sleep deprivation did not result in a decline in PEF. Thus, it is possible that FEV1 and PEF track together well only when pulmonary function changes are large (as occurs in asthma). If the circadian changes in PEF observed in the current study are caused by changes in airway caliber, this suggests that PEF may be a more sensitive indicator. However, this result in our healthy subjects is somewhat surprising because most evidence suggests that FEV1 is more sensitive than PEF in detecting airflow obstruction (reviewed in [37]). Perhaps the most important question concerning underlying mechanisms is why circadian rhythms cause dangerous changes in bronchial diameter in subjects with asthma, but little change in normal individuals. There are numerous physiological variables that could either directly or indirectly produce diurnal rhythms in the severity of asthma, including circulating cortisol, catecholamines, vagal activity, airway inflammation, or even abnormalities in circadian or sleep/wake cycles. Thus, to begin to distinguish among these mechanisms, it will be valuable in future to build on the current work in healthy subjects by studying patients with asthma in established circadian protocols.
The current study is the first to use an established technique to determine the degree to which the endogenous circadian pacemaker contributes to the previously reported diurnal changes in pulmonary function. We found that sleep deprivation did not have an independent effect on pulmonary function. Our data indicate that healthy adults have a small but significant circadian variation in FEV1 and FEV1/FEVC. The results for FEVC and PEF were more ambiguous because many individuals had prominent circadian rhythms in these variables, but differences in phase among subjects meant that the group mean data did not reach significance when individual data were aligned to a robust circadian marker, namely core body temperature. The circadian minima of spirometric variables occurred within the usual sleep period (even though subjects remained awake). These data are consistent with the findings of previous diurnal studies, suggesting that the endogenous circadian pacemaker is partly responsible for the diurnal rhythms of pulmonary function.
The authors thank the 10 volunteer subjects who cheerfully completed this arduous protocol, Charles A. Czeisler for advice concerning the constant routine protocol, Hance Oliver for data acquisition and analysis software and for running preliminary studies, David Rimmer for assistance with protocol integration and for help with data acquisition, Elita Harvey for help with data acquisition and analysis, the technical staff of the General Clinical Research Center for providing 24-h monitoring of our subjects, and Johnette Kao for screening the volunteer subjects.
Supported by Grant NIH HL62149 to Steven A. Shea and by NIH NCRR GCRC M01 RR02635 to the Brigham and Women's Hospital General Clinical Research Center. Christina M. Spengler was kindly supported by a Fellowship of the Swiss Foundation for Medical and Biological Research. Oxycon Alpha equipment was provided by Jaeger/MedPoint Technologies.
| 1. | Cochrane G. M., Clark T. J. H.A survey of asthma mortality in patients between ages 35 and 65 in the greater London hospitals in 1971. Thorax301975300315 |
| 2. | Hetzel M. R., Clark T. J. H., Branthwaite M. A.Asthma: analysis of sudden deaths and ventilatory arrests in hospital. Br. Med. J.11977808811 |
| 3. | Hetzel M. R., Clark T. J. H.Does sleep cause nocturnal asthma? Thorax341979749754 |
| 4. | Hetzel M. R., Clark T. J. H.Comparison of normal and asthmatic circadian rhythms in peak expiratory flow rate. Thorax351980732738 |
| 5. | Bagg L. R., Hughes D. T.Diurnal variation in peak expiratory flow in asthmatics. Eur. J. Respir. Dis.611980298302 |
| 6. | Peiffer C., Marsac J., Lockhart A.Chronobiological study of the relationship between dyspnoea and airway obstruction in symptomatic asthmatic subjects. Clin. Sci.771989237244 |
| 7. | Dethlefsen U., Repges R.Neues Therapieprinzip bei nächtlichem Asthma. Asthma Klin.8019854447 |
| 8. | Turner-Warwick M.Epidemiology of nocturnal asthma. Am. J. Med.85198868 |
| 9. | D'Alonzo G. E., Smolensky M. H.Chronophysiologic determinants of asthma. Ann. N.Y. Acad. Sci.6181991123139 |
| 10. | Postma D. S., Oosterhoff Y., van Aalderen W. M. C., Kauffman H. F., Wempe J. B., Koeter G. H.Inflammation in nocturnal asthma. Am. J. Respir. Crit. Care Med.1501994S83S86 |
| 11. | Oosterhoff Y., Kauffman H. F., Rutgers B., Zijlstra F. J., Koeter G. H., Postma D. S.Inflammatory cell number and mediators in bronchoalveolar lavage fluid and peripheral blood in subjects with asthma with increased nocturnal airways narrowing. J. Allergy Clin. Immunol.961995219229 |
| 12. | Guberan E., Williams M. K., Walford J., Smith M. M.Circadian variation of F.E.V. in shift workers. Br. J. Intern. Med.261969121125 |
| 13. | Reinberg A., Gervais P.Circadian rhythms in respiratory functions, with special reference to human chronophysiology and chronopharmacology. Bull. Physiopathol. Respir.81972663677 |
| 14. | Clark T. J., Hetzel M. R.Diurnal variation of asthma. Br. J. Dis. Chest7119778792 |
| 15. | Connolly C. K.The effect of bronchodilators on diurnal rhythms in airway obstruction. Br. J. Dis. Chest751981197203 |
| 16. | Catterall J. R., Rhind G. B., Stewart I. C., Whyte K. F., Shapiro C. M.Effect of sleep deprivation on overnight bronchoconstriction in nocturnal asthma. Thorax411986676680 |
| 17. | Smolensky M. H., Barnes P. J., Reinberg A., McGovern J. P.Chronobiology and asthma: I. Day–night differences in bronchial patency and dyspnea and circadian rhythm dependencies. J. Asthma231986321343 |
| 18. | Morrison J. F., Pearson S. B., Dean H. G.Parasympathetic nervous system in nocturnal asthma. Br. Med. J.296198814271429 |
| 19. | Ballard R. D., Saathoff M. C., Patel D. K., Kelly P. L., Martin R. J.Effect of sleep on nocturnal bronchoconstriction and ventilatory patterns in asthmatics. J. Appl. Physiol.671989243249 |
| 20. | Boivin D. B., Duffy J. F., Kronauer R. E., Czeisler C. A.Dose–response relationships for resetting of human circadian clock by light. Nature (London)3791996540542 |
| 21. | Czeisler C. A., Kronauer R. E., Allan J. S., Duffy J. F., Jewett M. E., Brown E. N., Ronda J. M.Bright light induction of strong (type 0) resetting of the human circadian pacemaker. Science244198913281333 |
| 22. | Mills J. N., Minors D. S., Waterhouse J. M.Adaptation to abrupt time shifts of the oscillator[s] controlling human circadian rhythms. J. Physiol. Lond.2851978455470 |
| 23. | Czeisler C. A., Allan J. S., Strogatz S. H., Ronda J. M., Sanchez R., Rios C. D., Freitag W. O., Richardson G. S., Kronauer R. E.Bright light resets the human circadian pacemaker independent of the timing of the sleep–wake cycle. Science2331986667671 |
| 24. | Harris, J. A., and F. G. Benedict. 1919. Standard Basal Metabolism Constants for Physiologists and Clinicians: a Biometric Study of Basal Metabolism in Man. J. B. Lippincott, Philadelphia. |
| 25. | Brown E. N., Czeisler C. A.The statistical analysis of circadian phase and amplitude in constant-routine core-temperature data. J. Biol. Rhythms71992177202 |
| 26. | Czeisler C. A., Duffy J. F., Shanahan T. L., Brown E. N., Mitchell J. F., Rimmer D. W., Ronda J. M., Silva E. J., Allan J. S., Emens J. S., Dijk D. J., Kronauer R. E.Stability, precision, and near-24-hour period of the human circadian pacemaker. Science284199921772181 |
| 27. | Nelson W., Tong Y. L., Lee J.-K., Halberg F.Methods for cosinor rhythmometry. Chronobiologia61979305323 |
| 28. | Rechtschaffen, A., and A. Kales. 1968. A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Subjects. Public Health Service, U.S. Government Printing Office, Washington, DC. |
| 29. | Casele R., Pasqualetti P.Cosinor analysis of circadian peak expiratory flow variability in normal subjects, passive smokers, patients with chronic obstructive pulmonary disease and patients with interstitial lung disease. Respiration641997251256 |
| 30. | Gershengorn, H., E. B. Klerman, and R. E. Kronauer. 1998. Circadian phase assessment from plasma melatonin. A comparison of measures. In Proceedings of the Sixth Meeting of the Society for Research of Biological Rhythms, Jacksonville, Florida, May 6–10. Society for Research of Biological Rhythms. 120. [Abstract 170] |
| 31. | D'Alonzo G. E., Steinijans V. W., Keller A.Measurements of morning and evening airflow grossly underestimate the circadian variability of FEV1 and peak expiratory flow rate in asthma. Am. J. Respir. Crit. Care Med.152199510971099 |
| 32. | Cooper K. R., Phillips B. A.Effect of short-term sleep loss on breathing. J. Appl. Physiol.531982855858 |
| 33. | Chen H. I., Tang Y. R.Sleep loss impairs inspiratory muscle endurance. Am. Rev. Respir. Dis.1401989907909 |
| 34. | Phillips B. A., Cooper K. R., Burke T. V.The effect of sleep loss on breathing in chronic obstructive pulmonary disease. Chest9119872932 |
| 35. | Martin R. J., Banks-Schlegel S.Chronobiology of asthma. Am. J. Respir. Crit. Care Med.158199810021007 |
| 36. | Troyanov S., Ghezzo H., Cartier A., Malo J. L.Comparison of circadian variations using FEV1 and peak expiratory flow rates among normal and asthmatic subjects. Thorax491995775780 |
| 37. | Jain P., Kavuru M. S., Emerman C. L., Ahmad M.Utility of peak expiratory flow monitoring. Chest1141998861875 |
Present address for Christina M. Spengler, Ph.D.: Exercise Physiology, Institute for Human Movement Sciences, Swiss Federal Institute of Technology and University of Zurich, Winterthurerstrasse 190, 8057 Zurich, Switzerland. E-mail: spengler@physiol.
