Pediatric Pulmonology
The Bronchodilator Response as a Predictor of Inhaled Corticosteroid Responsiveness in Asthmatic Children With Normal Baseline Spirometry Stanley P. Galant,
MD,
1
* Tricia Morphew, MS,2 Olga Guijon,
MD,
1
and Linh Pham,
1 MD
Summary. Rationale: Although inhaled corticosteroids (ICS) are considered first line controller therapy in children with persistent asthma, heterogeneity of the ICS response can be an important clinical problem. The purpose of this study is to determine the value of the bronchodilator response (BDR) in identifying the ICS responder and establish the optimal BDR cut-point that could be particularly useful in the clinic setting when baseline spirometry is normal. Methods: Mexican American asthmatic children, 5–18 years, with normal baseline spirometry who required low dose (step 2), or medium dose (step 3) ICS therapy were evaluated by skin prick test for atopy, and preand post-bronchodilator spirometry. ICS responders were defined by a 7.5% improvement in the FEV1 following 4–6 weeks of therapy. The optimal cut-point was determined by Receiver Operator Characteristic (ROC) curves as the best balance between sensitivity and specificity. Results: There were 34.8% of the 132 study patients who were ICS responders. ROC curves showed the BDR 10% to be an optimal cut-point with sensitivity 46%, specificity 76%, positive predictive value (PPV) 50%, and negative predictive value (NPV) 72%. Atopic females with a BDR 10% had a PPV of 73%. Conclusions: The composite phenotype of female gender, atopic, and the BDR of 10% identified 73% as ICS responders compared to 50% in our overall population with a BDR of 10% alone, with minimal false positives. We suggest that the BDR in conjunction with gender and atopic status be considered as potentially useful predictors of the ICS responder, particularly when baseline spirometry is normal. Pediatr Pulmonol. ß 2014 Wiley Periodicals, Inc. Key words: biomarkers; phenotypes; hispanic; atopy; gender. Funding source: none reported
INTRODUCTION
Inhaled corticosteroids (ICS) are recommended as first line long-term controller therapy in patients of all ages with persistent asthma.1 However, a number of studies have suggested heterogeneity in ICS responders with a considerable number of children and adults who do not respond well in terms of either spirometric parameters,2–5 or clinical outcomes.6,7 Thus, there is a real need to establish phenotypic characteristics that would help identify the ICS responder. Potential phenotypes have included clinical parameters,6,8 baseline spirometry,2,9 assessment of airway hyperreactivity,2,9 and biomarkers which reflect eosinophilic inflammation.10 The bronchodilator response (BDR) which reflects many of these same biomarkers, particularly exhaled nitric oxide (eNO),10,11 sputum,12 and bronchial eosinophilia,13 has also been assessed in a limited number of studies, particularly in children. Sharma et al.14 reported that those with a persistent BDR of 10% or 12% showed poor clinical and spirometric outcomes, particularly in a subset who did not receive ICS over a 4-year evaluation period, while Tantisera et al.9 showed that the baseline BDR was an independent predictor of improveß 2014 Wiley Periodicals, Inc.
ment in forced expiratory volume in 1 sec (FEV1) 4 years later, most significantly in those receiving ICS therapy compared to placebo. In terms of short term studies, evaluating ICS responsiveness over a 6- to 8-week period, Szefler et al.2 reported that the BDR could identify children who responded better to ICS than a leukotriene receptor antagonist (LTRA), but not as a predictor of a 7.5% improvement in FEV1, while Gould et al.15 evaluating ICS
1 Breathmobile, CHOC Children’s Hospital, Orange County, Orange, California. 2 Biostatistician, Southern California Chapter of Asthma and Allergy Foundation of America, Los Angeles, California.
Conflict of interest: None.
Correspondence to: Stanley P. Galant, MD, 1201 West La Veta, Orange, CA 92868. E-mail: [email protected] Received 20 March 2013; Accepted 16 October 2013. DOI 10.1002/ppul.22957 Published online in Wiley Online Library (wileyonlinelibrary.com).
2
Galant et al.
responsiveness in African-American adolescents and adults found the BDR to be the only factor consistently associated with improvement in FEV1 and asthma control. A limiting factor in these studies however, has been entry criteria requiring evidence of airway hyperreactivity demonstrated by either a BDR 12% after maximal bronchodilitation stimulation, or response to bronchoconstrictor stimuli. 2,9,15 These requirements define, a priori, a select subset, not necessarily the spectrum of asthmatic children most likely to be encountered in the office setting. In addition, although an association has been suggested between the BDR and ICS responsiveness, an optimal BDR cut-point which could be useful for the clinician, particularly, in those with normal baseline spirometry, has not been previously established. We have previously shown in Mexican American children, who are the predominant ethnic population in our inner-city asthma program, that the BDR was useful in identifying asthmatic children,16 and in a cross sectional analysis that there was a significant relationship between the BDR phenotype 10% and poor asthma control in a controller naive population with normal baseline spirometry.17 Most asthmatic children who are asymptomatic at the time of the visit have normal baseline spirometry, regardless of severity classification.18 Despite these observations none of the previous studies characterizing phenotypes which might predict ICS responsiveness in terms of improvement of FEV1 have evaluated asthmatics with normal baseline spirometry.2–4,9 The purpose of our study was to determine the value of the BDR phenotype as a potentially useful clinical tool in identifying the ICS responder in asthmatic children with normal baseline spirometry. In addition, we wanted to establish an optimal BDR cut-point which could be utilized by the provider in the clinical setting.
METHODS Subjects
Children participating in a school-based, low-income, asthma mobile van program, the BreathmobileTM, were referred for care during the 6-year period 2004–2011 by school nurses, community public health clinics, response to flyers, and a validated asthma questionnaire.19 Criteria for the diagnosis of asthma by the physician included a history of recurrent coughing, wheezing, shortness of breath with rest or exercise, symptomatic improvement after bronchodilator, and exclusion of other diagnoses.1 Inclusion criteria included children of Mexican American descent, normal baseline spirometry (FEV1 and FEV1/ FVC ratio 80% predicted), clinical evidence of mild to moderate persistent or poorly controlled asthma that required low dose (Step 2) or medium dose (Step 3) ICS Pediatric Pulmonology
therapy, ages 5–18 years, ability to perform reproducible spirometry, avoidance of bronchodilators 4–6 hr prior to spirometry, and ICS and oral corticosteroid-free for 4–6 weeks prior to entry. ICS therapy was restricted to beclomethasone HFA or fluticasone HFA at baseline. The doses and schedules for these preparations were those recommended by the NAEPP guidelines for Step 2 or Step 3 therapy.1 This protocol was reviewed by the Children’s Hospital of Orange County Institutional Review Board as an expedited review and determined that a formal informed consent was not required because the retrospective data used by the study was de-identified before the analysis, and the BDR was considered standard for assessing pulmonary function. However, all patients signed an institutional consent that permits standard evaluation after appropriate explanation. Spirometry
Pulmonary function testing was attempted in children 5 years of age or older in a standing position as previously described.16,17 Spirometric results were included in the analysis only if the child completed at least three baseline FVC maneuvers that met the American Thoracic Society criteria in a maximum of six attempts and was able to successfully complete post-bronchodilator spirometry.20 The best spirometric measures of at least three attempts were recorded for analysis, including forced vital capacity (FVC), FEV1, FEV1/FVC ratio, and the forced expiratory flow 25–75% predicted (FEF25–75 % predicted). Postbronchodilator spirometry was evaluated 10 min after administering two puffs (180 mg) from an albuterol meter dose inhaler with a spacer. Complete and acceptable spirometric measures were compared with the Knudson/ Intermountain Thoracic Society normal predictive values and adjusted for ethnic values on the basis of a parent report of ethnicity or race.21 The BDR was calculated as follows: ((FEV1 L postbronchodilatorFEV1 L prebronchodilator)/prebronchodilator FEV1) 100.16,17 Data Entry
All patient data were entered in a standardized format into Asma-Trax, a computer-based data entry program, developed and trademarked by Loran Clement, formerly of Southern California Breathmobile Program.19 The data included age, gender, body mass index (BMI), ethnicity, allergen specific skin test, baseline Step 2 or Step 3 ICS therapy, duration of asthma, adherence to medication, and baseline pulmonary function, including the BDR. Atopic Status
Screening skin prick tests to common indoor and outdoor inhalant allergens were performed by a nurse and
Bronchodilator Response Predicts Inhaled Corticosteroid Responsiveness TM
assessed by the Breathmobile physician. A positive skin test was defined as a wheal 3 mm larger than the negative saline solution control. Defining ICS Responsiveness
ICS responsiveness was defined as a 7.5% improvement in FEV1 after a 4–6-week ICS treatment period.2 We chose spirometry (i.e., FEV1) as the outcome variable because this measurement is reproducible, and objective, while the clinical history may be unreliable.22,23 Furthermore, the FEV1 in childhood is related to continued asthma symptoms in adulthood,24 predicts subsequent exacerbations in children,25 and is used as an impairment criterion for asthma severity and control classification by NAEPP guidelines.1 In addition, improvement of the FEV1 has previously been utilized as the primary outcome variable in assessing the heterogeneity of ICS responsiveness of several ICS preparations,3 in studies comparing the effectiveness of ICS with a LTRA,2 and the FEV1 response defines systemic steroid sensitive and resistant asthma.26,27 Finally, Martin et al.4 showed in adults that clinical success of long term ICS treatment could be based on ICS responsiveness determined by improvement in the FEV1 which was best predicted by the baseline BDR in a 6-week trial. Adherence
Adherence with ICS therapy was determined by the self reported history. The physician defined adherence based on this history as taking the ICS 75% of the time the therapy was prescribed.
3
Statistical Analysis
The distribution of patient characteristics in our study population was described in terms of valid percent of patients with defined trait for categorical variables and as average (mean, SD) observed in population for continuous variables (Table 1). Chi-square test statistic (categorical factors) and ANOVA (continuous factors) were utilized to assess significance of distributional differences in characteristics between patients responsive and nonresponsive to ICS therapy (responsiveness defined as 7.5% improvement in FEV1 L from baseline to visit#2). Distributions of bronchodilator response at baseline and change in FEV1 (L) from baseline to visit#2 were described graphically using histograms. Univariable and multivariable logistic regression analyses were used to establish for each factor the unadjusted and adjusted odds ratio (OR) reflecting the likelihood patients with defined characteristic compared to the reference category who will respond to ICS therapy (7.5% change). The multivariate analysis applied stepwise selection criteria to determine a final model of phenotypic characteristics that contribute to prediction of patient responsiveness to ICS therapy. The study was not adequately powered to examine potential influence (interactive effect) of non-significant covariates excluded from the final model on the relationship between significant predictors and response to therapy (outcome). Receiver Operator Characteristic (ROC) curves described visually the validity of BDR across thresholds to classify patients responsive to ICS therapy. Utility of BDR further considered contributory factors determined
TABLE 1— Characteristics of Study Population 1. ~ in FEV1L from baseline to visit#2 Characteristics (valid column %) Age (years), mean SD Female sex BMI percentile, mean SD Atopic Duration of asthma
The Bronchodilator Response as a Predictor of Inhaled Corticosteroid Responsiveness in Asthmatic Children With Normal Baseline Spirometry Stanley P. Galant,
MD,
1
* Tricia Morphew, MS,2 Olga Guijon,
MD,
1
and Linh Pham,
1 MD
Summary. Rationale: Although inhaled corticosteroids (ICS) are considered first line controller therapy in children with persistent asthma, heterogeneity of the ICS response can be an important clinical problem. The purpose of this study is to determine the value of the bronchodilator response (BDR) in identifying the ICS responder and establish the optimal BDR cut-point that could be particularly useful in the clinic setting when baseline spirometry is normal. Methods: Mexican American asthmatic children, 5–18 years, with normal baseline spirometry who required low dose (step 2), or medium dose (step 3) ICS therapy were evaluated by skin prick test for atopy, and preand post-bronchodilator spirometry. ICS responders were defined by a 7.5% improvement in the FEV1 following 4–6 weeks of therapy. The optimal cut-point was determined by Receiver Operator Characteristic (ROC) curves as the best balance between sensitivity and specificity. Results: There were 34.8% of the 132 study patients who were ICS responders. ROC curves showed the BDR 10% to be an optimal cut-point with sensitivity 46%, specificity 76%, positive predictive value (PPV) 50%, and negative predictive value (NPV) 72%. Atopic females with a BDR 10% had a PPV of 73%. Conclusions: The composite phenotype of female gender, atopic, and the BDR of 10% identified 73% as ICS responders compared to 50% in our overall population with a BDR of 10% alone, with minimal false positives. We suggest that the BDR in conjunction with gender and atopic status be considered as potentially useful predictors of the ICS responder, particularly when baseline spirometry is normal. Pediatr Pulmonol. ß 2014 Wiley Periodicals, Inc. Key words: biomarkers; phenotypes; hispanic; atopy; gender. Funding source: none reported
INTRODUCTION
Inhaled corticosteroids (ICS) are recommended as first line long-term controller therapy in patients of all ages with persistent asthma.1 However, a number of studies have suggested heterogeneity in ICS responders with a considerable number of children and adults who do not respond well in terms of either spirometric parameters,2–5 or clinical outcomes.6,7 Thus, there is a real need to establish phenotypic characteristics that would help identify the ICS responder. Potential phenotypes have included clinical parameters,6,8 baseline spirometry,2,9 assessment of airway hyperreactivity,2,9 and biomarkers which reflect eosinophilic inflammation.10 The bronchodilator response (BDR) which reflects many of these same biomarkers, particularly exhaled nitric oxide (eNO),10,11 sputum,12 and bronchial eosinophilia,13 has also been assessed in a limited number of studies, particularly in children. Sharma et al.14 reported that those with a persistent BDR of 10% or 12% showed poor clinical and spirometric outcomes, particularly in a subset who did not receive ICS over a 4-year evaluation period, while Tantisera et al.9 showed that the baseline BDR was an independent predictor of improveß 2014 Wiley Periodicals, Inc.
ment in forced expiratory volume in 1 sec (FEV1) 4 years later, most significantly in those receiving ICS therapy compared to placebo. In terms of short term studies, evaluating ICS responsiveness over a 6- to 8-week period, Szefler et al.2 reported that the BDR could identify children who responded better to ICS than a leukotriene receptor antagonist (LTRA), but not as a predictor of a 7.5% improvement in FEV1, while Gould et al.15 evaluating ICS
1 Breathmobile, CHOC Children’s Hospital, Orange County, Orange, California. 2 Biostatistician, Southern California Chapter of Asthma and Allergy Foundation of America, Los Angeles, California.
Conflict of interest: None.
Correspondence to: Stanley P. Galant, MD, 1201 West La Veta, Orange, CA 92868. E-mail: [email protected] Received 20 March 2013; Accepted 16 October 2013. DOI 10.1002/ppul.22957 Published online in Wiley Online Library (wileyonlinelibrary.com).
2
Galant et al.
responsiveness in African-American adolescents and adults found the BDR to be the only factor consistently associated with improvement in FEV1 and asthma control. A limiting factor in these studies however, has been entry criteria requiring evidence of airway hyperreactivity demonstrated by either a BDR 12% after maximal bronchodilitation stimulation, or response to bronchoconstrictor stimuli. 2,9,15 These requirements define, a priori, a select subset, not necessarily the spectrum of asthmatic children most likely to be encountered in the office setting. In addition, although an association has been suggested between the BDR and ICS responsiveness, an optimal BDR cut-point which could be useful for the clinician, particularly, in those with normal baseline spirometry, has not been previously established. We have previously shown in Mexican American children, who are the predominant ethnic population in our inner-city asthma program, that the BDR was useful in identifying asthmatic children,16 and in a cross sectional analysis that there was a significant relationship between the BDR phenotype 10% and poor asthma control in a controller naive population with normal baseline spirometry.17 Most asthmatic children who are asymptomatic at the time of the visit have normal baseline spirometry, regardless of severity classification.18 Despite these observations none of the previous studies characterizing phenotypes which might predict ICS responsiveness in terms of improvement of FEV1 have evaluated asthmatics with normal baseline spirometry.2–4,9 The purpose of our study was to determine the value of the BDR phenotype as a potentially useful clinical tool in identifying the ICS responder in asthmatic children with normal baseline spirometry. In addition, we wanted to establish an optimal BDR cut-point which could be utilized by the provider in the clinical setting.
METHODS Subjects
Children participating in a school-based, low-income, asthma mobile van program, the BreathmobileTM, were referred for care during the 6-year period 2004–2011 by school nurses, community public health clinics, response to flyers, and a validated asthma questionnaire.19 Criteria for the diagnosis of asthma by the physician included a history of recurrent coughing, wheezing, shortness of breath with rest or exercise, symptomatic improvement after bronchodilator, and exclusion of other diagnoses.1 Inclusion criteria included children of Mexican American descent, normal baseline spirometry (FEV1 and FEV1/ FVC ratio 80% predicted), clinical evidence of mild to moderate persistent or poorly controlled asthma that required low dose (Step 2) or medium dose (Step 3) ICS Pediatric Pulmonology
therapy, ages 5–18 years, ability to perform reproducible spirometry, avoidance of bronchodilators 4–6 hr prior to spirometry, and ICS and oral corticosteroid-free for 4–6 weeks prior to entry. ICS therapy was restricted to beclomethasone HFA or fluticasone HFA at baseline. The doses and schedules for these preparations were those recommended by the NAEPP guidelines for Step 2 or Step 3 therapy.1 This protocol was reviewed by the Children’s Hospital of Orange County Institutional Review Board as an expedited review and determined that a formal informed consent was not required because the retrospective data used by the study was de-identified before the analysis, and the BDR was considered standard for assessing pulmonary function. However, all patients signed an institutional consent that permits standard evaluation after appropriate explanation. Spirometry
Pulmonary function testing was attempted in children 5 years of age or older in a standing position as previously described.16,17 Spirometric results were included in the analysis only if the child completed at least three baseline FVC maneuvers that met the American Thoracic Society criteria in a maximum of six attempts and was able to successfully complete post-bronchodilator spirometry.20 The best spirometric measures of at least three attempts were recorded for analysis, including forced vital capacity (FVC), FEV1, FEV1/FVC ratio, and the forced expiratory flow 25–75% predicted (FEF25–75 % predicted). Postbronchodilator spirometry was evaluated 10 min after administering two puffs (180 mg) from an albuterol meter dose inhaler with a spacer. Complete and acceptable spirometric measures were compared with the Knudson/ Intermountain Thoracic Society normal predictive values and adjusted for ethnic values on the basis of a parent report of ethnicity or race.21 The BDR was calculated as follows: ((FEV1 L postbronchodilatorFEV1 L prebronchodilator)/prebronchodilator FEV1) 100.16,17 Data Entry
All patient data were entered in a standardized format into Asma-Trax, a computer-based data entry program, developed and trademarked by Loran Clement, formerly of Southern California Breathmobile Program.19 The data included age, gender, body mass index (BMI), ethnicity, allergen specific skin test, baseline Step 2 or Step 3 ICS therapy, duration of asthma, adherence to medication, and baseline pulmonary function, including the BDR. Atopic Status
Screening skin prick tests to common indoor and outdoor inhalant allergens were performed by a nurse and
Bronchodilator Response Predicts Inhaled Corticosteroid Responsiveness TM
assessed by the Breathmobile physician. A positive skin test was defined as a wheal 3 mm larger than the negative saline solution control. Defining ICS Responsiveness
ICS responsiveness was defined as a 7.5% improvement in FEV1 after a 4–6-week ICS treatment period.2 We chose spirometry (i.e., FEV1) as the outcome variable because this measurement is reproducible, and objective, while the clinical history may be unreliable.22,23 Furthermore, the FEV1 in childhood is related to continued asthma symptoms in adulthood,24 predicts subsequent exacerbations in children,25 and is used as an impairment criterion for asthma severity and control classification by NAEPP guidelines.1 In addition, improvement of the FEV1 has previously been utilized as the primary outcome variable in assessing the heterogeneity of ICS responsiveness of several ICS preparations,3 in studies comparing the effectiveness of ICS with a LTRA,2 and the FEV1 response defines systemic steroid sensitive and resistant asthma.26,27 Finally, Martin et al.4 showed in adults that clinical success of long term ICS treatment could be based on ICS responsiveness determined by improvement in the FEV1 which was best predicted by the baseline BDR in a 6-week trial. Adherence
Adherence with ICS therapy was determined by the self reported history. The physician defined adherence based on this history as taking the ICS 75% of the time the therapy was prescribed.
3
Statistical Analysis
The distribution of patient characteristics in our study population was described in terms of valid percent of patients with defined trait for categorical variables and as average (mean, SD) observed in population for continuous variables (Table 1). Chi-square test statistic (categorical factors) and ANOVA (continuous factors) were utilized to assess significance of distributional differences in characteristics between patients responsive and nonresponsive to ICS therapy (responsiveness defined as 7.5% improvement in FEV1 L from baseline to visit#2). Distributions of bronchodilator response at baseline and change in FEV1 (L) from baseline to visit#2 were described graphically using histograms. Univariable and multivariable logistic regression analyses were used to establish for each factor the unadjusted and adjusted odds ratio (OR) reflecting the likelihood patients with defined characteristic compared to the reference category who will respond to ICS therapy (7.5% change). The multivariate analysis applied stepwise selection criteria to determine a final model of phenotypic characteristics that contribute to prediction of patient responsiveness to ICS therapy. The study was not adequately powered to examine potential influence (interactive effect) of non-significant covariates excluded from the final model on the relationship between significant predictors and response to therapy (outcome). Receiver Operator Characteristic (ROC) curves described visually the validity of BDR across thresholds to classify patients responsive to ICS therapy. Utility of BDR further considered contributory factors determined
TABLE 1— Characteristics of Study Population 1. ~ in FEV1L from baseline to visit#2 Characteristics (valid column %) Age (years), mean SD Female sex BMI percentile, mean SD Atopic Duration of asthma
