Main

It has become widely accepted that symptoms and severity of bronchial asthma are related to features of inflammation. Because inhaled glucocorticosteroids have marked antiinflammatory effects in the bronchial mucosa these drugs are increasingly being recommended for treatment of moderate or mild asthma. The risk of systemic adverse effects of inhaled glucocorticosteroids have conventionally been assessed by measures of hypothalamic-pituitary-adrenal function or by assessment of longitudinal growth(1–3). We recently evaluated various auxologic and biochemical measures of growth, bone and collagen turnover in the assessment of systemic activity of exogeneous glucocorticosteroids in children with mild asthma(4–8). Short-term growth suppressive effects of systemic and inhaled glucocorticosteroids could be accurately and sensitively detected by knemometry, a noninvasive technique for measurement of lower leg length(4–6). Dose-related suppressive effects of inhaled glucocorticosteroids on knemometric growth rates were observed(5). Serum osteocalcin, a noncollageneous protein released to the circulation during bone formation, was suppressed during treatment with systemic but not during treatment with inhaled glucocorticosteroids(7). Serum concentrations of the growth hormone-dependent IGF-I and IGFBP-3 were not affected by systemic or inhaled glucocorticosteroids(8, 9). Dose-related glucocorticosteroid-induced suppressive effects have been found on serum markers of type I collagen formation and degradation, PICP and ICTP, and on the serum marker of type III collagen formation, the PIIINP(8–14). Furthermore, short-term knemometry has proven more sensitive than measurement of urine free cortisol excretion in detecting the systemic activity of inhaled glucocorticosteroids(15). We have no knowledge, however, about the sensitivity of any of the biochemical serum markers compared with urine cortisol excretion or knemometry. The aim of the present study was to evaluate whether there is any correlation between overnight urine free cortisol excretion and short-term knemometric growth rates, respectively, and serum concentrations of IGF-I, IGFBP-3, osteocalcin, PICP, ICTP, and PIIINP in asthmatic children treated with inhaled fluticasone propionate, 200 μg, and beclomethasone dipropionate, 400 and 800 μg/d, and to compare these measures during the three treatments.

METHODS

The protocol was part of a clinical study comparing short-term growth in asthmatic children during treatment with fluticasone propionate and beclomethasone dipropionate(16). Seventeen children who completed the growth study and who had blood or urine samples taken during all active periods of the trial were included in the present analysis. Patient characteristics are given in Table 1. The children were outpatients in a secondary referral center. One girl was stage 2 according to Tanner's rating of puberty(17), the others were preadolescents. All had mild asthma requiring treatment only as needed with inhaled β-2 stimulants. None had received treatment with glucocorticosteroids either by mouth or inhaled for 2 mo before the study, and no other drug was taken during the study period.

Table 1 Characteristics of the study group
Full size table

The study design was a double blind, cross-over trial with three active treatment periods and two wash-out periods. After a run in period of 4 d the children were randomized to treatment with fluticasone propionate, 200μg/d, or beclomethasone dipropionate, 400 or 800 μg/d. Treatment order was allocated by a computerized randomization scheme prepared in balanced blocks. All periods were 15 d long. The drug was taken in the morning and in the evening as one blister from a dry powder Diskhaler (GlaxoWellcome). The children were instructed to brush their teeth after the inhalation and to spit out the rinsing water. To assess and encourage optimum compliance the number of consumed blisters was counted at every visit.

Knemometry was performed twice weekly and pulmonary function was monitored throughout the study as previously described(16). The technical error of the knemometer was 0.09 mm.

Urine analyses. Urine was collected during the night of the last Saturday (d 12 ± 2 d) of each of the three glucocorticosteroid treatment periods. The children were instructed to empty the bladder before going to bed and to collect urine during the night including the morning urine before breakfast at approximately 0800 h. Aliqouts for determination of creatinine and cortisol were collected from the overnight urine samples after weighing and thorough mixing of the whole volume. The aliquots were frozen until the end of the trial when free cortisol concentrations were measured by RIA (Farmos Diagnostica, Turku, Finland)(18). The intra- and interassay coefficients of variation were 1.4 and 7.1%, respectively. The cortisol concentrations were standardized in micrograms/g of creatinine(19). There are no available normal rages for overnight urine free cortisol excretion in children.

Serum analyses. Blood samples for analysis of IGF-I, IGFBP-3, osteocalcin, and the markers of collagen turnover were taken at d 15 of each active treatment period. The sampels were obtained at approximately the same time (i.e. within 1 h) in the afternoon (between 1500 and 1800 h). The samples were centrifuged at 3000 rpm for 10 min within 1 h after they were taken. Serum was frozen and stored at -20°C and batch assayed at the completion of the study.

IGF-I was analyzed by a specific RIA as described previously(20), with some modifications(21). Serum was extracted by acid/ethanol and cryoprecipitated before analysis to remove interfering binding proteins. A monoiodinated isomer of the truncated IGF-I (Tyr-31-des(1-3)IGF-I) was used as the radioligand. Inter- and intraassay variations were 8.7 and 3.9%, respectively (at a bound/free(B/B0) ratio of 0.40).

IGFBP-3 was analyzed by a specific RIA originally described by Blum et al.(22). Materials for the assay were obtained from Mediagnost GmbH, Tübingen, Germany. Inter- and intraassay variations were 10.0 and 7.2%, respectively (at a B/B0 ratio of 0.40). Normal serum ranges for IGF-I and IGFBP-3 have been provided(21, 23).

Osteocalcin was determined by a specific RIA(24). Inter- and intraassay variations were 8.9 and 4.2%, respectively. There are no normal serum ranges available for the assay in this age group.

PICP, ICTP, and PIIINP were determined by RIAs based on human antigen(Orion Diagnostica, Finland)(25–27). The inter- and intraassay variations were 7.1 and 6.8% (at a B/B0 ratio of 0.30) for PICP determinations, 5.9 and 5.2% (at a B/B0 ratio of 0.30) for ICTP, and 5.5 and 5.4%, respectively, for PIIINP (at a B/B0 ratio of 0.42). Normal serum ranges for PICP and PIIINP based on these assays are available(28), and preliminary ICTP standards have been published(9)

Statistical analysis. On the basis of all measurements in the three treatment periods, lower leg growth rates were calculated for each period and expressed as millimeters/wk. Knemometric growth rates, creatinine standardized cortisol data, and serum markers during treatment periods were compared using analysis of variance techniques adjusting for subject, period, and treatment effects.

No carryover effects were found, but statistically significant patient and period effects on all outcome measures were detected. A statistically significant interaction between treatment and period was observed on the lower leg growth rate, and IGF-I, IGFBP-3, and ICTP serum levels. For parameters with statistically significant overall treatment effect, contrast analysis between the treatments were performed using analysis of variance. All data but urine free cortisol/creatinine × 106, osteocalcin, ICTP, and PIIINP concentrations conformed to the normal distribution consumption. These four parameters were logarithmic transformed before statistical analysis. Treatment contrast analysis for logarithmic transformed parameters results in a ratio between the adjusted geometric means of the two treatments, whereas treatment contrast analysis for parameters not transformed before analysis results in the difference between the adjusted means of the two treatments. Furthermore, data from the three treatment periods were pooled before correlation analyses were performed. However, similar results to those presented were found when the treatments and periods were analyzed separately. The statistical analyses were carried out using the SAS package (version 6.10). Correlation analyses were performed by Pearson's test(29).

RESULTS

All children participated in the knemometry measurements and the collection of urine; however, two girls refused to have blood samples taken. No significant variations were seen in pulmonary function or symptom severity between the various treatments(16).

Figure 1 shows individual values of lower leg growth rate, urine free cortisol/creatinine × 106, and the markers of the IGF axis and collagen turnover during the three treatments. After adjustment for patient and period effects and for the treatment and period interaction, no significant overall effects of the inhaled glucocorticosteroids were detected on IGF-I (F = 0.81, p = 0.46), IGFBP-3 (F = 0.33, p = 0.72), or osteocalcin(F = 0.64, p = 0.54) concentrations. However, overall statistically significant suppressive effects were observed on lower leg growth rates (F = 12.43, p = 0.0002), urine free cortisol/creatinine × 106 (F = 8.36, p = 0.002), and on PICP (F = 5.32, p = 0.01), ICTP(F = 11.71, p = 0.0003), and PIIINP (F = 12.17,p = 0.0002) concentrations. Table 2 presents the results of the contrast analyses of the five parameters in which overall statistically significant effects were found.

Figure 1
figure 1

Individual lower leg growth rates and urine free cortisol/creatinine × 106 in 17 children and serum markers of the IGF axis and collagen turnover in 15 children during double blind treatment with fluticasone propionate, 200 μg, and beclomethasone dipropionate, 400 and 800 μg/d.

Full size image
Table 2 Contrast analyses of urine free cortisol(UFC)/creatinine × 106 and lower leg growth rates in 17 children and of PICP and ICTP and PIIINP in 15 children during double blind treatment with fluticasone propionate (FP), 200 μg, and beclomethasone dipropionate (BDP), 400 and 800 μg/d
Full size table

Figure 2 compares lower leg growth rates with the serum parameters. Test results are given in the legend. Statistically significant correlations between knemometric growth rates and IGF-I, IGFBP-3, PICP, ICTP, and PIIINP were found. No statistically significant correlations between urine free cortisol/creatinine × 106 and knemometric growth rates(Pearson's correlation coefficient = 0.18, p = 0.20) or between urine free cortisol/creatinine × 106 and any of the serum markers were detected (Pearson's correlation coefficient, P): IGF-I(0.05;0.76), IGFBP-3 (0.13; 0.41), osteocalcin (0.21; 0.16), PICP (0.08; 0.61), ICTP (0.14; 0.35), and PIIINP (-0.02; 0.90).

Figure 2
figure 2

Relation between lower leg growth rate and serum markers of the IGF axis and collagen turnover in 15 children during double blind treatment with fluticasone propionate, 200 μg, and beclomethasone dipropionate, 400 and 800 μg/d. The growth rates correlated statistically significantly (Pearson's correlation coefficient; P) with IGF-I(0.41; 0.006), IGFBP-3 (0.35; 0.02), PICP (0.44; 0.003), ICTP (0.35; 0.001), and PIIINP (0.46; 0.002) but not with osteocalcin (0.25; 0.10).

Full size image

DISCUSSION

Adverse effects of fluticasone propionate, 200 μg, on the evaluated parameters cannot be ruled out from the present data because run-in or placebo comparisons were not included in the trial. Efficacy studies have indicated that fluticasone propionate, 200 μg/d, is clinically equivalent to beclomethasone dipropionate, 400 μg/d, in controlling symptoms of asthma(30, 31). Therefore, we decided to compare the systemic effects of these two doses which would normally be considered standard pediatric doses capable of controlling asthma in most children. Furthermore, we added a period of treatment with beclomethasone dipropionate, 800 μg/d, because we wanted to include a dose at which suppressive effects on lower leg growth rate and PICP, ICTP, and PIIINP would be expected(5, 9, 10) to ensure that the study design was sensitive enough to detect statistically significant effects.

In an earlier study of similar design, 24-h urine free cortisol excretion was not reduced by inhaled budesonide, 200, 400, or 800 μg, from a metered dose inhaler with a spacer(15). In the present study overnight urine free cortisol concentrations were suppressed during treatment with both 400 and 800 μg of dry powder beclomethasone dipropionate. There are no data for comparison of 24 h and overnight urine cortisol excretion as measures of systemic glucocorticosteroid activity. The differences between the two drugs might be explained by a greater systemic activity of dry powder beclomethasone dipropionate than of budesonide from a metered dose inhaler with a spacer(32, 33). The finding of similar suppressive effects of the two doses of beclomethasone suggests that the top or less steep part of the dose-response curve with respect to suppression of cortisol excretion had already been reached at 400 μg/d as was the case with lower leg growth. In contrast, studies using integrated plasma or urinary concentrations of cortisol or cortisol metabolites for assessment of glucocorticosteroid-induced effects on the hypothalamic-pituitary-adrenal feed-back system have found dose-related suppression in the dose range from 200 to 900 μg of beclomethasone(1, 2). These observations suggest that integrated measures may be more sensitive than measurement of overnight urine free cortisol excretion which profits from being a simpler and noninvasive method. However, although no correlation was found, urine free cortisol levels seemed to be as sensitive as knemometry and serum markers of type I and III collagen turnover for assessing the systemic activity of inhaled beclomethasone dipropionate.

In analogy with previous findings, the inhaled glucocorticosteroid treatment had no effects on IGF-I, IGFBP-3, or on osteocalcin levels(7, 9). So, these parameters seem less sensitive for assessing glucocorticosteroid suppressive effects on growth(8, 9). Osteocalcin has been reported to be suppressed in adults(34) and in children(14) treated with high doses of inhaled glucocorticosteroids. In our previous studies we have not been able to detect any effects on osteocalcin levels of budesonide doses in the range from 200 to 800 μg/d from a metered dose inhaler with a spacer(7, 10). Because a high interassay difference in osteocalcin measurements has been reported(35), this could be due to an insensitive assay. However, statistically significant suppression of osteocalcin was observed in children treated with 2.5 and 5 mg of prednisolone(7).

Type I collagen is the most abundant collagen type in connective tissues and accounts for more than 90% of the organic matrix of bone and more than 70% of the dermis(36). Variations in concentrations of PICP and ICTP are considered to reflect fluctuations in synthesis and degradation of type I collagen throughout the body but primarily in bone(37). Type III collagen is present in dense and loose connective tissues throughout the body and constitutes 10-15% of skin collagen, whereas little is found in bone(27). Serum levels of PIIINP are assumed to reflect the synthesis of type III collagen in tissues throughout the body(27). Serum levels of PICP, ICTP, and PIIINP are sensitive measures of exogeneous glucocorticosteroid effects on collagen turnover in children(8, 9, 11). However, inflammatory conditions may influence these markers. Inflammation may be associated with increased PIIINP serum levels(38). PICP concentrations may be decreased in patients with rheumatoid arthritis(39). The influence of inflammation on the serum markers of collagen turnover has not been assessed in children with asthma. In the present study intrapulmonary inflammatory activity was monitored indirectly by pulmonary function measurements, recordings of symptom severity, and consumption of β-2 agonists(16). Because no variations in these parameters were observed during the study, it is unlikely that variations in intrapulmonary inflammation could explain the variations seen in the markers of collagen turnover. The observed reductions in serum levels of the collagen markers, therefore, should be taken to indicate suppression of type I and III collagen turnover.

ICTP and PIIINP levels were significantly reduced during the 400 μg of beclomethasone dipropionate period, whereas PICP was not. Only ICTP levels were able to discriminate between all three treatments. This is in agreement with other observations indicating that the sensitivity of ICTP for detection of glucocorticosteroid suppressive effects on collagen turnover is high(9, 40). PICP seems less sensitive than ICTP and PIIINP. This difference may be due to differences in metabolism. ICTP is extracted from the circulation by the kidneys and PICP and PIIINP by the liver sinusoidal endothelial cells(41, 42). PICP is taken up via the mannose receptor and PIIINP by the scavenger receptor. Intravenous glucocorticosteroids do not affect the hepatic clearance of PIIINP(43); however, we have no knowledge as to whether exogeneous glucocorticosteroids may affect the metabolism of PICP.

The observed correlations between short-term lower leg growth rates and IGF-I, IGFBP-3, and the collagen markers correspond with previous reports of correlation between lower leg growth and growth in height and of correlation between these serum measures and statural growth in normal children(21, 23, 28, 44). Although some of the observed correlation coefficients were low, when taken together with the results of the contrast analyses, they suggest that reduced lower leg growth during treatment with inhaled glucocorticosteroids is associated with suppression of type I and type III collagen turnover. This has recently been supported by observations of correlation between inhaled glucocorticosteroid-suppressed lower leg growth rates and suppressed urine excretion of pyridinoline and deoxypyridinoline (our unpublished data).

In summary, compared with fluticasone propionate, 200 μg, beclomethasone dipropionate, 400 and 800 μg/d, suppressed urine free cortisol excretion, lower leg growth rates, and serum PICP, ICTP, and PIIINP. Only serum ICTP levels discriminated between all three treatments. Urine free cortisol excretion did not correlate with lower leg growth rates or any of the serum measures. Reduced lower leg growth rates in children treated with inhaled glucocorticosteroids seem to be associated with suppressive effects on type I and type III collagen turnover.