統計方法

Numerical variables were reported as means ± SDs. Numerical parameters with nonnormal distribution (serum leptin, adiponectin, IgE, allergen-specific IgE and ECP levels) were reported as medians and interquartile ranges and were log-transformed before analysis. Bivariate analyses were evaluated by using χ² and Mann-Whitney U or Student t tests. Correlations between log leptin or log adiponectin and various other parameters were assessed by using the Pearson correlation coefficient. To test which variables independently contribute to the variation of leptin and adiponectin levels, multivariate linear regression analysis was performed. Adjusted R² that corrected for the number of variables and the sample size was used to estimate the variance explained in each model. To investigate the relationships between sex (1, male; 0, female), age, BMI, leptin, adiponectin and the development of allergic rhinitis, a multivariate logistic regression analysis was done. Statistical analysis was performed with the SPSS software, version 7.5 for Windows (SPSS, Chicago, Ill). P < .05 was considered significant.

第三章 研究結果

第一節 描述性統計分析

3-1-1 過敏性鼻炎與正常小朋友兩組基本資料比較

Anthropometric parameters of the AR patients and the controls are shown in Table I. The mean values for age, weight, height, and BMI did not show any difference between the groups of allergic rhinitis and healthy children. There was no significant sex difference with respect to these parameters between and within the groups, either.

3-1-2 過敏性鼻炎與正常小朋友血清中瘦素及脂締素差異性

A significant elevation in serum leptin was observed in children with AR when compared with healthy children (p = 0.041) despite a similar BMI in both groups. There was a tendency toward higher leptin levels in AR boys compared with healthy boys with a borderline significance (P = 0.085) despite a similar BMI in both groups. In contrast, serum adiponectin levels in children with AR were significantly lower (p = 0.005) than those in healthy children in spite of a similar BMI in both groups. This significant difference was also apparent between the AR boys and healthy boys with similar BMI values (P = 0.009), whereas a similar difference was not observed between AR girls and healthy girls. No significant sex difference was detected in serum leptin and adiponectin levels among AR children, as well as among healthy children (Table I).

19

第三章 研究結果

第二節 推論性統計分析

3-2-1 影響血清中瘦素及脂締素濃度相關因子的探討

Log leptin levels of both AR and healthy children showed positive

correlations with BMI, but log adiponectin levels correlated negatively with BMI in both groups (Table II). A weak positive correlation existed between log leptin levels and age only in AR patients, whereas log adiponectin levels were significantly inversely related to age in healthy children (r = -0.53; P < 0.001) (Table II). In a multivariable regression model including BMI, sex, age, allergic rhinitis and log adiponectin with log leptin as the dependent variable, we

identified BMI (p < 0.0001), and allergic rhinitis (P = 0.040) as significant independent determinants of leptin levels, explaining 35 % of the variance (adjusted R²= 0.35; p < 0.0001; Table III). In a multivariable regression model including BMI, sex, age, allergic rhinitis and log leptin with log adiponetin as the dependent variable, BMI (P = 0.020), age (P = 0.031), and allergic rhinitis (P = 0.011) were significant independent determinants of adiponetin levels, explaining 20 % of the variance (adjusted R²=0.20; p = 0.0001; Table IV).

3-2-2 發生過敏性鼻炎與瘦素及脂締素濃度的相關性探討

When sex, age, BMI, log leptin and log adiponetin were all included in the logistic regression model, using allergic rhinitis as the dependent variable, only log leptin and log adiponetin were independent risk factors for the development

of allergic rhinitis, with their odds ratios being 27.06 (95% confidence interval (CI), 1.01-1209) and 13.14 (95% CI, 1.84-115.1), respectively (Table V).

3-2-3 血清中瘦素及脂締素濃度與過敏性發炎指標的關連性

A significantly negative but weak correlation was observed between log adiponectin and log ECP levels among AR children (r = -0.29; P = 0.036) but a significant correlation was not found between log leptin and log ECP. Both leptin and adiponectin levels did not correlate with total IgE or mite-specific IgE levels among AR children. There was no statistical correlation between leptin and adiponectin levels in both groups (Table II)

There was no statistically significant difference in the leptin and adiponetin levels between AR children with higher mite-specific IgE levels (valence 4-6) and those with lower mite-specific IgE levels (valence 1-3) (data not shown).

21

The aims of this study were to evaluate serum leptin and adiponectin levels in a group of children with AR before the initiation of therapy and to examine the relationship between leptin and adiponectin and allergic inflammatory markers in AR children.

Our results show that a significant elevation in serum leptin and a significant reduction in serum adiponectin were observed in children with AR when

compared with healthy children in spite of no difference in BMI levels, although these differences appeared to be far more significant in boys than girls. The lack of statistically significant differences in leptin and adiponectin levels between AR girls and healthy girls could be a result either of a small sample size to detect a difference or a true modifying effect of sex on leptin and adiponectin.

Higher levels of leptin with a borderline significance and significantly lower levels of adiponectin in AR boys but not in girls may suggest that both of the adipokines might be involved in the pathophysiologic mechanisms of childhood AR, which is also seen more frequently in boys than in girls (32, 33).

4-1-2 影響血清中瘦素及脂締素濃度相關因子的探討

Serum leptin levels show positive correlation and adiponectin levels

correlated inversely with BMI in healthy prepubertal children, as has also been shown in our study in both the AR children and the healthy children (14, 34).

Serum leptin levels are relatively low in prepubertal ages, showing a gradual rise in both sexes before the onset of puberty, followed by a significant increase during puberty in girls and a decrease in boys (34, 35). Our results showed that a weak correlation existed between leptin levels and age only in AR patients and in univariate linear regression, age was positively associated with leptin (P=

0.0006) which explained 10 % of the variation in leptin concentrations (data not shown). However, the effect of age on leptin levels was eliminated in a multiple regression model that included BMI, age, sex and AR. By contrast, as previous reports (37, 38), serum adiponectin levels were significantly inversely related to age during childhood and age was a significant independent determinant of adiponectin in a multiple regression model in our study. Leptin and adiponectin levels were found in similar concentrations in both sexes over the prepubertal years in some studies as has also been shown in children in our study (34, 38), whereas a significant sex difference in leptin levels was noted in other studies, with girls having higher leptin levels than boys despite the absence of difference in body weight (35, 36). Interestingly, in the multiple regression analysis, we found that sex correlated with plasma leptin with a borderline significance (p = 0.052) after the adjustment. The effect size of sex on leptin levels needs further study.

23

4-1-3 過敏性鼻炎與瘦素及脂締素濃度以及過敏性發炎指標的相關性以 及免疫學相關機轉的探討

Although the relationship among AR and adipokines cannot be adequately addressed in this study because of the small sample size, the results of

multivariant analysis suggested that leptin and adiponectin may have some effects on the regulation of allergic inflammation in the allergic rhinitis.

Whereas the primary physiologic function of leptin and adiponectin is in the regulation of metabolism, they also regulate immune function and are believed to be pro-inflammatory and anti-inflammatory, respectively. Both of them have been reported to be involved in a variety of inflammatory-autoimmune

conditions, such as inflammatory bowel disease (38, 39), rheumatoid arthritis (40), multiple sclerosis (41), Kawasaki disease (42) and endometriosis (43).

However, the pathogenic role played by adipokines in such disorders is far from understood. Few clinical studies have investigated the relationship between adipokines and atopic diseases, such as asthma and atopic dermatitis, with all of them focusing on leptin. In the study by Guler et al (27), serum leptin levels were found to be higher in children with atopic bronchial asthma when

compared with nonatopic asthmatic subjects and healthy children with similar BMI. They found serum leptin to be an independent predictive factor for asthma in children, especially in boys. Gurkan et al, reported that, in a small cohort (n=23) of children with mild to moderate asthma, treatment with inhaled budesonide resulted in a significant reduction in serum leptin concentrations

after 4 weeks (44). In a smaller group (n= 50), reported by Kimata, patients with IgE-associated atopic eczema/dermatitis syndrome (AEDS) had significantly elevated serum leptin levels related to steatohepatitis and hyperlipidemia (45).

To our knowledge, neither the association between adipokines and AR nor the role of adiponectin in atopic disorders has previously been studied clinically.

Furthermore, although altered systemic adipokine levels in atopic disorders are noted, the exact mechanism involved in the alteration in adipokines expression and the pathogenic role played by adipokines in T_H2-mediated inflammation has not been fully clarified. In two previous studies by Shore et al (25, 26),

sensitization and repeated airway challenge with OVA in BALB/cJ mice increased serum leptin but caused a reduction in serum adiponectin and a corresponding decrease in adipose tissue adiponectin mRNA expression.

Because adipocytes are the primary source of leptin and adiponectin, these data suggested that allergic responses in the airways are capable of altering adipocyte function. It has been well documented that proinflammatory cytokines including IL-1β, TNF-α and IL-6 are capable of inducing the release of leptin from

adipocytes but inhibiting adiponectin expression in adipocytes (46-48).

Furthermore they have been shown to contribute to allergic airway responses in mice (49-51) and serum levels of these cytokines are increased in asthma and allergic rhinitis (52). It is possible that proinflammatory cytokines might be involved in allergen-induced alterations in adipokines expression. However, there may be a further complexity in the regulation of adipokines expression in

25

allergic inflammation. As previous reports, the production of proinflammatory mediators is also markedly influenced by adipokines levels (7, 8, 15, 18). The interaction between proinflammatory cytokines, adipokines and other

inflammatory cells and mediators may be complex and remains to be further elucidated. The findings in the studies by Shore et al also indicate that

exogenous leptin augments OVA- induced airway hyperresponsiveness (AHR) to methacholine (MCh). The increase in OVA-induced airway responsiveness occurred in the absence of any effects of leptin on inflammatory cell influx or T_H2 cytokine expression but was associated with an increase in OVA-induced IgE production in the leptin treated mice. In contrast, exogenous adiponectin can attenuate OVA-induced changes in inflammatory cell influx and T_H2 cytokine expression in the lungs, as well as OVA-induced AHR. These investigators speculated that the ability of leptin to augment OVA-induced AHR may be related to the direct effects of leptin on B lymphocytes that contributed to the production of IgE and may be the result of increased activation of mast cells.

Guler et al also reported a weak but significant correlation between serum leptin and serum IgE among the children with asthma (r = 0.231, p = 0.019) (27).

However, the finding in the present study, indicating no significant correlation between serum leptin and serum IgE among children with AR, is in contrast to the findings by Shore et al and Guler et al. Indeed, discrepant results about the association between leptin and IgE existed. Kimata ever reported leptin can inhibit in vitro IgE and IgG4 production (45) and in a previous study of mice by

Hetland et al (53), leptin does not influence the IgE response to OVA in mice.

Further studies regarding the effects of leptin on IgE production are necessary.

The relationship between adiponectin and IgE is less studied. In the report by Shore et al, OVA challenge did not increase serum IgE levels in

adiponectin-treated mice, although IgE levels were greater in adiponectin- versus Tris-buffered saline (TBS)-treated mice, regardless of the PBS or OVA aerosol challenge. In the present study, there was no significant correlation between adiponectin and total IgE levels. Furthermore, we found both leptin and adiponectin levels did not correlate with mite allergen-specific IgE levels and there was no statistically significant difference in the leptin and adiponectin levels among AR children on the basis of allergen-specific IgE levels. In a

recent study by Conus el al (54), human eosinophils were found to express leptin surface receptors under in vitro and in vivo conditions, and leptin directly

activates eosinophils and delays spontaneous apoptosis of these cells. The authors speculated that leptin may contribute to eosinophil accumulation at inflammatory sites, and perhaps it might then also stimulate the release of proinflammatory mediators from these cells. In the present study, we did not find a significant correlation between serum leptin and serum ECP level, an activated eosinophil-derived inflammatory mediator and is significantly higher in AR patients than in controls (55). Interestingly, there was a significantly inverse but weak correlation between serum adiponectin and serum ECP level among the children with AR (r = -0.29; P = 0.036). According to the report by

27

Shore et al, this may be explained partly by an adiponectin-induced inhibition of T_H2 cytokine expression and eosinophil influx in the inflammatory sites. Further work is required to understand whether adiponectin has direct negative effects on eosinophil function and activation.

第四章 討論