INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: asthma; children; endogenous serum cortisol; FEV₁

Nocturnal cough, wheeze, and chest tightness interfere at least once a week with sleep in almost 50% of children with asthma (1). These symptoms result from overnight airflow limitation, as a consequence of an exaggeration of the normal circadian rhythm in airway caliber. The pathogenesis of nocturnal airway obstruction is still unclear and several mechanisms have been proposed. Possible causes include increased parasympathetic tone (2, 3), and increased airway inflammation associated with enhanced bronchoconstrictor mediator release (3). Decreased serum cortisol levels have also been suggested as a factor in the pathogenesis of nocturnal airway obstruction in adult asthma (10). Cortisol, a glucocorticoid produced by adrenal glands, shows a circadian rhythm with its peak at about 0800 h and its nadir at about midnight. These low cortisol levels at night precede the nocturnal fall in FEV₁ and may result in less suppression of airway inflammation with subsequent increased airflow limitation (13).

Increased airway responsiveness, which is predominantly caused by airway inflammation, is another phenomenon underlying nocturnal airflow limitation (5, 14). Airway hyperresponsiveness can be measured with various stimuli. Histamine and methacholine challenges cause airway obstruction mainly by their direct action on receptors of airway smooth muscles, whereas a stimulus such as adenosine 5'-monophosphate (AMP) is thought to act primarily on inflammatory cells. The latter stimulus thereby initiates processes that indirectly lead to smooth muscle contraction (14). A relationship between serum cortisol levels and increased airway responsiveness can be expected if either overall circadian cortisol levels are lower or if low cortisol levels, specifically at night, result in less suppression of airway inflammation.

The aim of this study was to determine whether endogenous serum cortisol levels are lower and/or the variation in endogenous 24-h serum cortisol levels is greater in children with asthma compared with normal control subjects. In addition, we assessed the relationship between serum cortisol levels and airway obstruction and hyperresponsiveness in children with asthma. Therefore, we measured serum cortisol and FEV₁ at six time points over a 24-h period, blood eosinophils at 0400 and 1600, and airway responsiveness to methacholine and AMP at 0400 and 1600.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects with Asthma

We studied 28 children with stable asthma, aged 7 to 16 yr, recruited at the pediatric outpatient clinic of the University Hospital of Groningen (Groningen, The Netherlands). All children had a history of episodic wheezing on exposure to allergens or nonallergic stimuli. In addition to the clinical diagnosis of asthma, the other inclusion criteria were as follows: allergy for at least one aeroallergen determined by a positive specific immunoglobulin E (IgE) response to house dust mite, cat, dog, grasses, and trees, and an increased level of total serum IgE (Kabi Pharmacia, Woerden, The Netherlands). All children had an increased airway responsiveness to histamine (PC₂₀ histamine [provocative concentration of histamine causing a 20% fall in FEV₁]  8 mg/ml) and used inhaled corticosteroids as maintenance medication before the study. Patients entered the study after they had stopped inhaled corticosteroid for 2 wk, and long-acting ₂ agonists for 5 d before the onset of the study. Short-acting ₂ agonists were withheld for 8 h before the study. No oral corticosteroids were used for at least 2 mo before the study.

Subjects without Asthma

Eighteen age-matched healthy control subjects were studied. Selection criteria were as follows: (1) no history of asthma symptoms, either for the child or for the family in the first line, (2) normal total eosinophil count (0.0-0.4 × 10⁹/L), (3) no demonstrated allergy for aeroallergens, (4) a normal level of total serum IgE (< 50 KU/L for 7- 10 yr, < 100 KU/L for > 10 yr), and (5) no increased airway responsiveness to histamine (PC₂₀ histamine > 16 mg/ml).

Informed consent was obtained from all children and their parents. The study was approved by the Medical Ethical Committee of the University Hospital of Groningen.

Study Design

We used FEV₁ variation (highest minus lowest FEV₁ value divided by the mean FEV₁ value) for classification of asthmatic children with and without nocturnal worsening of asthma. The group of asthmatic children with a variation in FEV₁  15% was defined to have nocturnal worsening of asthma. The mean daily dose of inhaled corticosteroids they used before the study was 420 µg for the children with and 380 µg for those without nocturnal worsening of their asthma.

Children arrived at the hospital at 1930 and remained until 1700 the next day. An intravenous line was placed to take blood samples. The intravenous line was kept open with a continuous infusion of 0.9% saline with a portable CADD+ pump (Graseby Medical BV, Rosmalen, The Netherlands). Children were awakened 10 min before the measurements at 0000 and 0400. Bedtime was between 2100 and 2200. The children with asthma attended the hospital a second day for the AMP inhalation challenge at 1600 and 0400, 24 h after the last measurement of the first day.

Measurements

Serum cortisol was measured by a chemiluminescence immunoassay method (University Hospital Groningen) every 4 h over a 24-h period, starting at 2000. At 0400 and 1600 blood eosinophils were counted in a Coulter counter (Beckman Coulter, Fullerton, CA). FEV₁ values were obtained every 4 h over a 24-h period by a pneumotachograph (Masterscreen I.O.S.; Jaeger, Wurtzburg, Germany). The FEV₁ was measured until three reproducible recordings were obtained, with the best of three being used for analysis. Reference values for the FEV₁ are those of Zapletal and coworkers (15). FEV₁ values before the methacholine challenge at 0400 and 1600 were used for the asthma group. The inhalation provocation tests were performed with a nebulizer (DeVilbiss, Somerset, PA) attached to a French-Rosenthal dosimeter. Solutions of methacholine (Chemie Brunschwig, Basel, Switzerland) and AMP (Sigma, St. Louis, MO) were administered at room temperature as aerosols. After inhalation of 0.9% sodium chloride solution, doubling concentrations of methacholine bromide (0.15-39.3 mg/ml) or AMP (0.04-160 mg/ml) were inhaled. The challenge was stopped when the FEV₁ had fallen by > 20% of the prechallenge level or when the highest concentration had been administered. PD₂₀ values were calculated by linear interpolation between the last two data points.

Statistical Analysis

Data were analyzed with the Statistical Package for Social Sciences (SPSS, Chicago, IL) for DOS version 5.0 and Windows version 9.0. Normal distribution was checked visually by probability plots of the residuals and tested formally by the Kolmogorov-Smirnov test. Differences between groups were tested by parametric or nonparametric test as appropriate. The Student t test was used to test differences in FEV₁ %pred (FEV₁ as a percentage of the predicted value) between two groups. One-way analysis of variance (ANOVA) was used to test differences in mean FEV₁ %pred between three groups. Nonparametric tests (Mann-Whitney U test or Kruskall-Wallis test) were used to test differences between groups in terms of levels of cortisol and PD₂₀ methacholine and PD₂₀ AMP at 0400 and 1600. The effect of mean cortisol level over 24 h on FEV₁ at 2000, 0000, 0400, 0800, 1200, and 1600 on eosinophil count at 0400 and 1600, and on methacholine and AMP responsiveness at 0400 and 1600 was estimated by means of multiple linear regression. All regression analyses were performed for all dependent variables separately, with additional adjustment for age and sex. The effect of the 24-h variation in cortisol [cortisol_24h = (highest  lowest cortisol value)/mean cortisol value] on FEV₁, eosinophil count, and methacholine and AMP responsiveness at the various time points was estimated in the same way. A p value < 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Twenty-eight children with asthma (16 boys), mean age 13.1 ± 1.8 yr, were included. Ten children (6 boys), mean age 13.1 ± 1.4 yr, had nocturnal asthma defined as an FEV₁ variation  15% (NA⁺, mean FEV₁ variation 24.9 [15.9-45.4]%), and 18 had an FEV₁ variation < 15% (NA, mean FEV₁ variation 8.9 [(5.1-14.9]%). Eighteen children (7 boys) were included in the healthy control group, mean age 13.1 ± 1.9 yr. There was a significant difference between the three groups in mean FEV₁ %pred (SD) over a 24-h period, the NA⁺ group having the lowest values (86.7 [10.1]%pred), the NA group having intermediate values (92.2 [11.2]%pred), and the control subjects having the highest values (101.6 [6.8]%pred) (p < 0.001). The same pattern was seen at 0400 and 1600 (Table 1).

Cortisol Measurements

Mean serum cortisol levels were lower in the asthma group than in the control group, and this was significant at midnight (p < 0.05) (Figure 1). The latter was true both for the groups with and without nocturnal worsening of asthma compared with the control group. At 0800 and 1200, the NA⁺ group had lower cortisol levels than the control subjects as well (p < 0.05). Although the NA⁺ group had lower mean cortisol values than the NA group, differences did not reach significance (Figure 2). No significant differences between the groups were found regarding the mean serum cortisol level over 24 h, or regarding the mean 24-h variation in serum cortisol (cortisol_24h) and cortisol_16000400/mean cortisol value, respectively.

View larger version (17K): [in this window] [in a new window]   Figure 1.   Mean (SEM) level of cortisol (nM) at six different time points in the asthma and control groups. *Asthma versus control subjects, p < 0.05.  Control subjects; asthmatics.

View larger version (16K): [in this window] [in a new window]   Figure 2.   Mean (SEM) level of cortisol (nM) at six different time points in the nocturnal asthma and nonnocturnal asthma groups. NA; NA+.

Numbers of Eosinophils

Median numbers of eosinophils were significantly different between the three groups both at 0400 (0.79 vs. 0.61 vs. 0.29 × 10⁹/L in asthmatic children with and without nocturnal asthma and control subjects, respectively; p < 0.05) and at 1600 (0.68 vs. 0.39 vs. 0.21 × 10⁹/L, respectively; p < 0.05) (Table 1).

Hyperresponsiveness

Median PD₂₀ methacholine was lower in the NA⁺ than in the NA group, which was significant at 0400 (1.0 and 3.2 mg/ml, respectively; p < 0.05; Table 1), but not at 1600 (1.5 and 3.9 mg/ml, p = 0.07).

Median PD₂₀ AMP was significantly lower in the NA⁺ group than in the NA group at 0400 (13.5 and 65.3 mg/ml, respectively; p < 0.05; Table 1), but not at 1600 (36.0 and 89.8 mg/ml). Individual PD₂₀ methacholine and PD₂₀ AMP values at 1600 and 0400 for both groups are shown in Figures 3 and 4, respectively.

View larger version (17K): [in this window] [in a new window]   Figure 3.   Bronchial hyperresponsiveness to methacholine. Shown are individual PD20 methacholine values at 1600 and 0400 of the nocturnal asthma and nonnocturnal asthma groups.

View larger version (18K): [in this window] [in a new window]   Figure 4.   Bronchial hyperresponsiveness to AMP. Shown are individual PD20 AMP values at 1600 and 0400 of the nocturnal asthma and nonnocturnal asthma groups.

Mean 24-H Serum Cortisol Level

A higher mean serum cortisol level over 24 h was significantly associated with a higher mean FEV₁ %pred value at 0400, 0800, and 2000 (p < 0.05) in the asthma group as a whole (Table 2). The mean 24-h level of cortisol also tended to be associated with the FEV₁ %pred at 1200 and 16.00 (p < 0.10).

Neither hyperresponsiveness to AMP and methacholine, nor the number of eosinophils, was significantly associated with mean 24-h serum cortisol levels. We found no significant associations between the mean 24-h level of cortisol and the FEV₁ %pred in the control group.

Variation in Cortisol over 24 H

There was no significant association between the 24-h cortisol variation and FEV₁ %pred values in the whole asthma group (Table 3). In the asthmatic children without nocturnal worsening of asthma a higher 24-h cortisol variation was associated with a lower FEV₁ %pred, and this was also the case in healthy control subjects (Table 3). These associations were significant at all time points except for 0800 (p = 0.15) and 1200 (p = 0.08) in the NA group (Table 3).

In the asthma group, a higher 24-h variation in serum cortisol tended to be associated with higher numbers of eosinophils at 0400 and 1600 (p < 0.1). In the NA group a higher 24-h variation in serum cortisol was significantly associated with higher numbers of eosinophils at both 0400 and 1600.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study of 28 children with asthma and 18 healthy control subjects was designed to identify whether the mean 24-h level of endogenous serum cortisol is lower and/or the 24-h variation is larger in children with asthma compared with normal control subjects. We found that the children with asthma had lower serum cortisol levels at all time points measured compared with the control group, and this difference was statistically significant at midnight. Specifically, children with nocturnal asthma had lower cortisol levels than the healthy control group, and not only at midnight, but also at 0800 and 1200. A lower mean 24-h cortisol level in the asthma group was associated with a significantly lower FEV₁ %pred at 0400, 0800, and 2000, whereas this association did not exist in healthy control subjects. A higher 24-h cortisol variation was associated with a lower FEV₁ %pred at all time points in both the healthy children and the asthmatic children without nocturnal airway obstruction.

Our results suggest a relationship between serum cortisol level on the one hand, and the overall level of FEV₁ and nocturnal fall in FEV₁ on the other hand; that is, the higher the serum cortisol the better the lung function. A relationship between serum cortisol levels and lung function has previously been found in adults. Kraft and coworkers (12) compared serum cortisol values in three similar adult groups. They found that the cortisol levels in the nocturnal asthma group were significantly higher than in the other groups. This difference was completely explained by a significantly higher value at 2000 in the nocturnal asthma group. In contrast to adults, we found a significant negative association between the mean 24-h level of cortisol and the level of FEV₁ %pred at almost all time points. This is compatible with one other epidemiological study showing that adults with asthma have lower serum cortisol levels than healthy adults. Moreover, in accordance with our observation, they found an association between low cortisol levels and nocturnal asthma (13). Together these data support the notion that higher endogenous cortisol levels are associated with better lung function values, and that they are of importance especially with respect to better nocturnal FEV₁.

A few published studies have addressed the circadian variation in serum cortisol. Szefler and coworkers (16) measured plasma cortisol in 7 nocturnal and 10 non-nocturnal adults with asthma and in 10 healthy volunteers at 4 A.M. and 4 P.M. They found no significant differences between the groups. Haen and coworkers (17) studied plasma cortisol at 4-h intervals for 24 h in 10, untreated male patients with asthma and 8 healthy men. Contrary to our results, they did find a significantly higher 24-h mean serum cortisol level and a lower drop at night in their patients with asthma. The discrepancies in results may stem from the fact that we have studied children, whereas Haen and coworkers have investigated adults. A clear relationship between age and cortisol levels in adults has been reported (18) and aging was associated with higher evening cortisol levels, which became apparent after midlife. The morning maximum values remained stable across all age ranges (18).

An important observation in our study is that the significantly lower cortisol levels over 24 h were found particularly in our asthmatic children with a nocturnal fall in their FEV₁. It is not clear from this study whether this is a causal relation. To test this hypothesis it would be necessary to perform a study with cortisol replacement. However, studies investigating the effects of inhaled and oral corticosteroids on nocturnal asthma have clearly shown a beneficial effect, which points in the same direction and lends support to the hypothesis that a lower cortisol level may induce a higher susceptibility to nocturnal asthma. This effect of a lower mean 24-h cortisol level on the presence of nocturnal asthma may well occur via an effect on airway wall inflammation. Herrscher and coworkers (19) investigated the effect of endogenous cortisol on the IgE-dependent cutaneous response. They found that a lower cortisol level was associated with increased allergic inflammation by affecting the expression of cellular events at the late-phase sites. Earlier findings of our group have shown that airway wall inflammation is more extensive in those subjects with asthma who experience symptoms at night than in those who do not have nocturnal airway obstruction, and yet the difference in the extent of inflammation between 0400 and 1600 is comparable in both groups (3, 20). We interpret the above-described results as suggesting that lower cortisol levels at both day and night amplify the circadian rhythm of yet another factor that inhibits airway obstruction at night. One such factor may well be the physiologic nadir of -adrenergic receptor function at night, hence the excellent response of nocturnal symptoms and airway obstruction to long-acting -agonists (21).

We hypothesized that lower endogenous cortisol levels, or greater 24-h swings in endogenous cortisol, are associated with a lower degree of protection against stimuli inducing airway wall inflammation. Indirect evidence of increased inflammation in nocturnal asthma is provided by the observation that bronchial responsiveness is more severe in asthmatic subjects with nocturnal airway obstruction than in those without (14, 22, 23). In this study we measured responsiveness to methacholine and AMP at 0400 and 1600. We used both stimuli to find out whether direct and/or indirect bronchial responsiveness played a role in the nocturnal worsening of FEV₁ values in children with asthma, and whether the level of endogenous cortisol influenced these two types of airway hyperresponsiveness. We found indeed that hyperresponsiveness to both methacholine and AMP was worse in the NA⁺ group, significantly so at 0400. Furthermore, in six of eight children with nocturnal worsening of their asthma the PD₂₀ AMP decreased during the night, but the PD₂₀ methacholine did not decrease. However, we found no association between endogenous mean 24-h cortisol levels or the 24-h variation in cortisol levels and the degree of bronchial responsiveness. Thus, we could not suggest a relationship between endogenous cortisol and either muscarinic receptor function on airway smooth muscle cells, or with inflammatory cell processes such as mast cell activation, which is under the influence of AMP.

We realize that interpretation of our results could be hampered by the fact that cortisol is a stress hormone that can be influenced by several factors. The literature provides several studies showing effects of the use of inhaled corticosteroids, exercise, sleep, and psychological factors on levels of cortisol (24). These factors accounted only to a small extent for the variance in serum cortisol measures. One of the most important findings in the study by Dahl and coworkers (24) was that sleep continuity resulted in lower nocturnal cortisol levels. This underscores the finding that the observed low cortisol levels at night in our study are not the result of sleep discontinuity.

When reviewing the relationship between nocturnal adrenal function and the use of inhaled corticosteroids, Kallenbach and coworkers (27) concluded that the majority of the studies performed in children do not show hypothalamic-pituitary- adrenal cortical axis suppression, when dosages of inhaled corticosteroids ranging from 300 to 800 µg/d are being used. Others confirmed this (27) and indicate that dosages of inhaled corticosteroids up to and including 800 µg/d do not suppress the hypothalamic-pituitary-adrenal cortical axis. Because the children in our study used inhaled corticosteroids ranging from 100 to 500 µg/d before they stopped their medication (2 wk before they entered the study), it seems unlikely that the serum cortisol values in our study were influenced.

In conclusion, we found that children with asthma had lower serum cortisol levels compared with healthy control subjects at all six time points of the day that we measured cortisol. A lower mean 24-h cortisol level was associated with lower FEV₁ values in the asthmatic population, and significantly so at 0400, 0800, and 2000. A higher circadian variation in cortisol was associated with lower FEV₁ levels in healthy volunteers and asthmatic subjects without nocturnal asthma. Furthermore, we found more severe hyperresponsiveness to methacholine and AMP and higher numbers of peripheral blood eosinophils in the asthmatic children with nocturnal airway obstruction. We found no relationship between those parameters and the level and/or variation of serum cortisol. Our data suggest that cortisol concentrations indeed contribute to both overall lower FEV₁ values in asthma, especially at night, contributing to the presence of nocturnal airway obstruction in asthma.