Assessing differences in inhaled corticosteroid response by self-reported race-ethnicity and genetic ancestry among individuals with asthma

Abstract

Background

Inhaled corticosteroids (ICS) are the preferred treatment for achieving asthma control. However, little is known regarding the factors contributing to treatment response and whether treatment response differs by population group.

Objective

To assess behavioral, socio-demographic, and genetic factors related to ICS response among African American and European American individuals with asthma.

Methods

Study participants were part of the Study of Asthma Phenotypes and Pharmacogenomic Interactions by Race-ethnicity (SAPPHIRE). The analytic sample included individuals with asthma aged 12–56 years, >12% bronchodilator reversibility, and a percent of predicted forced expiratory volume at one second (FEV₁) between 40–90%. Participants received 6 weeks of inhaled beclomethasone dipropionate. The primary measure of ICS response was a change in Asthma Control Test (ACT) score; the secondary measure was the change in pre-bronchodilator FEV₁. Adherence was measured with electronic monitors. Genetic ancestry was estimated for African American participants using genome-wide genotype data.

Results

There were 339 study participants; 242 self-identified as African American and 97 as European American. Baseline ACT score, percent of predicted FEV₁, degree of bronchodilator response, and ICS adherence were significantly associated with ICS response. A baseline ACT score ≤19 was useful in identifying those who would respond as evidenced by the significant dose-response relationship with ICS adherence. Neither self-reported race-ethnicity among all participants nor proportion of African ancestry among African American participants was associated with ICS responsiveness.

Conclusions

Our findings suggest that baseline lung function measures and self-reported asthma control predict ICS response, whereas self-reported race-ethnicity and genetic ancestry do not.

INTRODUCTION

Repeated studies have demonstrated that long-term adherence to inhaled corticosteroid (ICS) medication among individuals with asthma is poor.^1–3 The consequence of this non-adherence is an increased risk for severe asthma exacerbations, including the need for hospitalization.^{4, 5} Unaccounted for differences in medication adherence can also confound the relationship between treatment and response,⁶ thereby obscuring the true effect of treatment. For example, previously described differences in ICS adherence between African American and European American individuals could result in inaccurate inferences regarding treatment effect.^{7, 8}

We have previously shown that after accounting for medication adherence, ICS response did not vary among African American individuals by genetic ancestry (i.e., the proportion of African ancestry vs. European ancestry),⁹ These findings implied that as a whole, a genetic contribution to treatment-related differences between African American and European American individuals with asthma is likely to be small.

In the current analysis, we extended our earlier findings by expanding our study population of African American individuals and by including a group of European American individuals with asthma. This expanded study population allowed us to directly assess whether ICS treatment response differed by either self-reported race-ethnicity or genetic ancestry. All study individuals were part of the Study for Asthma Phenotypes and Pharmacogenomic Interactions by Race-ethnicity (SAPPHIRE), had similar enrollment criteria, and received the same 6-week ICS treatment course.

METHODS

Study Population

The study population was comprised of participants in the Study of Asthma Phenotypes and Pharmacogenomic Interactions by Race-ethnicity (SAPPHIRE), which was approved by the Institutional Review Board at Henry Ford Health System. Potentially eligible patients were first identified through the medical record based on their age (12–56 years), prior diagnosis of asthma, and lack of excluding diagnoses (i.e., chronic obstructive pulmonary disease and/or congestive heart failure). Patients who met these criteria and who lived in southeast Michigan were invited for a clinical evaluation. Study participants (or their guardian in the case of minors) provided written consent prior to the evaluation and study enrollment. The evaluation consisted of a staff-administered survey, vital sign and anthropomorphic measurements, and lung function testing. Race-ethnicity was based on patient self-report at the initial study visit.

Lung function testing was performed using a Fleisch type pneumotachometer in accordance with 2005 American Thoracic Society/European Respiratory Society guidelines for spirometry.¹⁰ Bronchodilator response was defined as a >12% improvement in the forced expiratory volume at 1 second (FEV₁) following albuterol administration. Albuterol sulfate hydrofluoroalkane (360 micrograms [mcg] – 4 puffs) was administered by metered-dose inhaler using a spacer device (AeroChamber plus Flow-Vu, Monaghan Medical Corp., Plattsburgh, NY) and lung function was measured 15 minutes later. Patients with ≤12% improvement in FEV₁ following the first albuterol dose received a second dose (360 mcg for those ≥18 years and 180 mcg for those <18 years) and lung function was re-measured after another 15-minute wait.

In order to receive 6-weeks of observed ICS treatment, study participants had to meet the following additional criteria: a measured FEV₁ that was between 40–90% of the predicted value (based on age, sex, height, and race-ethnicity),¹¹ >12% maximum bronchodilator reversibility, no smoking in the preceding year and <10 pack-years total smoking, no inhaled or systemic corticosteroids used in the preceding 4-week period, and not pregnant and not planning to become pregnant during the 6 weeks of ICS treatment. Patients in the SAPPHIRE cohort, who both met these criteria and agreed to treatment, received a 6-week course of beclomethasone dipropionate hydrofluoroalkane (320 mcg per day administered as 2 puffs twice-a-day). Patients self-administered the ICS medication via MDI and the use of the previously described spacer device. At the end of the 6-week treatment period, patients returned to complete another staff-administered questionnaire and undergo lung function testing.

Assessment of exposure and outcomes

Patient medication adherence was assessed using a DOSER-CT device (Meditrack, Easton, MA). The DOSER-CT attached to the MDI, and it counted each time that the inhaler was actuated. Adherence was calculated as the total number of recorded actuations divided by the product of the number of days between visits and 4 (the prescribed number of ICS puffs per day).

To assess changes in the level of asthma control, we measured the difference in the Asthma Control Test (ACT – QualityMetric, Lincoln, RI) responses before and after the 6-week course of ICS therapy. We also assessed changes in pre-bronchodilator FEV₁ between these time points.

Genome-wide genotype data were collected for the African American SAPPHIRE participants in the treatment trial using commercial arrays (Affymetrix Inc., Santa Clara, CA). We have previously used these data to calculate the proportion of West African ancestry (heretofore, called African ancestry) in these participants.¹² Briefly, the software program, LAMP, was used to estimate ancestry at each locus.^{13, 14} We then estimated each individual’s global genetic ancestry (i.e., proportion of African ancestry) as the proportion of African alleles among genotyped autosomal single nucleotide polymorphism locations.

Statistical analysis

The primary outcome of the study was ICS response as measured by the change in ACT score between the initial visit and the 6-week follow-up visit after ICS treatment (i.e., the composite ACT_(6-weeks)-ACT_(initial)). The secondary outcome was the percent change in pre-bronchodilator FEV₁ between these time points (i.e., FEV_1(6-weeks)-FEV_1(initial)/FEV_1(initial)). Linear regression was used to assess the relationship between both outcome variables and the following dependent variables: patient age, sex (coded male=0, female=1), self-reported race-ethnicity (coded European American=0, African American=1), body mass index (BMI), baseline percent of predicted FEV₁, baseline ACT score, and ICS adherence. Both age and BMI were modeled for each unit increase (i.e., year and kilogram per meter squared, respectively), but were aggregated into 10-unit increments for presentation in the tables. Based on the original validation of the ACT by Nathan et al.,¹⁵ we also dichotomized the composite ACT score at a cut point of 19 (i.e., patients with scores ≤19 were considered to have “not controlled” or “uncontrolled” asthma and individuals with scores ≥20 were considered “controlled”). Similarly, we dichotomized lung function at 70% of predicted FEV1 based on the midpoint for persistent moderate severity asthma in the current U.S. asthma guidelines.¹⁶ Therefore, we stratified our models by both baseline ACT score (i.e., ≤19 and ≥20) and baseline percent of predicted FEV₁ (i.e., <70% and 70–90% [90% was the upper limit in the treatment group]) in order to assess the relationship between exposure variables and outcomes within strata that are considered to separate clinically meaningful differences in asthma control and severity. Separate adjusted models limited to African American participants were used to assess the relationship between African ancestry and ICS response.

As a post hoc analysis, we assessed factors associated with achieving self-reported asthma control (ACT score ≥20) after 6 weeks of ICS treatment among all individuals whose asthma was not controlled at baseline (ACT score ≤19). Logistic regression was used to assess the likelihood of achieving control as a function of the following variables: patient age, sex (coded male=0, female=1), self-reported race-ethnicity (coded European American=0, African American=1), body mass index (BMI), baseline percent of predicted FEV₁, and ICS adherence.

Analyses were performed with SAS statistical computing software (SAS Institute Inc., Cary, NC).¹⁷ A P-value <0.05 was considered to be statistically significant.

RESULTS

Three hundred thirty-nine participants in the SAPPHIRE cohort met the criteria and completed 6-weeks of observed ICS treatment; 242 enrollees identified themselves as African American, and 97 identified as European American. The characteristics of those individuals before and after stratification by race-ethnicity are shown in Table 1. When compared with European American participants, African American study individuals were significantly younger (mean of 32.5 years vs. 36.8 years), had a higher BMI (mean of 32.8 vs. 29.4), and reported less well controlled asthma (mean ACT score of 18.1 vs. 20.0). African American individuals also had lower ICS adherence (mean 0.76 vs. 0.84), implying that on average African Americans took 76% of their prescribed study dose as compared with 84% in European American participants. The average estimated proportion of African ancestry in the African American participants was 79.9% (± 9.9%, standard deviation [SD]), and the distribution is shown in Figure E1 of the online supplemental materials.

Table 1.

Characteristics of SAPPHIRE study participants (n=339)^*

Variable	Overall (n = 339)	African American (n = 242)	European American (n = 97)	P-Value†
Age in years – mean ± SD	33.7 ± 13.8	32.5 ± 13.1	36.8 ± 14.9	<0.01
Female – no. (%)	198 (58.4)	148 (61.2)	50 (51.6)	0.11
Self-reported race-ethnicity – no. (%)				--
African American	242 (71.4)	242 (100.0)	--	--
European American	97 (28.6)	--	97 (100.0)	--
Proportion of African Ancestry	--	79.9 ± 9.9	--	--
Body mass index in kg/m2 – mean ± SD†	31.9 ± 9.5	32.8 ± 9.6	29.4 ± 8.8	<0.01
Asthma Control Test Score at enrollment – mean ± SD‡	18.6 ± 5.1	18.1 ± 5.2	20.0 ± 4.5	<0.01
Percent of predicted FEV1 at enrollment – mean ± SD	72.9 ± 12.7	73.5 ± 12.4	71.2 ± 13.2	0.13
Bronchodilator response at enrollment – mean ± SD§	20.5 ±15.3	20.4 ± 13.1	20.6 ± 19.8	0.91
Change in ACT score – mean ± SD||	3.1 ± 4.8	3.3 ± 4.9	2.4 ± 4.6	0.11
Percent change in pre-bronchodilator FEV1 – mean ± SD§	11.3 ± 18.1	10.6 ± 15.0	12.9 ± 24.2	0.28
ICS adherence – mean ± SD, median, and interquartile range¶	.79 ± .21, .84, .66 – .94	.76 ± .22, .80, .61 – .93	.84 ± .17, .89, .79 – .96	<0.01

SAPPHIRE denotes the Study of Asthma Phenotypes and Pharmacogenomic Interactions by Race-ethnicity; SD, standard deviation; kg/m², weight in kilograms divided by squared height in meters; FEV₁, the forced expiratory volume at 1 second; ACT, Asthma Control Test, and ICS, inhaled corticosteroid.

^*Race-ethnicity was determined by participant self-report

^†P-value for the comparison of African American and European American individuals.

^‡ACT scores ≤19 are considered poor asthma control, whereas those ≥20 are considered good control.

^§Bronchodilator response is measured as the percent change in FEV1 following administration of inhaled albuterol.

^||Measured as the change between values measured at the time of enrollment and after 6 weeks of inhaled corticosteroid treatment.

^¶The mean, median, and interquartile range represent the proportion of the prescribed amount of ICS taken over 6 weeks of treatment.

Over the 6-week course of ICS treatment, average improvement in ACT score and FEV₁ was similar among African American and European American participants. The improvement in ACT score was 3.3 and 2.4 points for African American and European American individuals, respectively (P= 0.11). Similarly, pre-bronchodilator FEV₁ improved by 10.6% and 12.9%, respectively (P=0.28).

These above findings were supported in Table 2, which examined the factors associated with change in ACT score and FEV₁ over the course of treatment. Self-identified race-ethnicity was not associated with change in ACT score or FEV₁ even after accounting for other variables including ICS adherence. After adjusting for all of the variables shown, baseline ACT score (parameter estimate [β]=−0.71, P<0.01), baseline percent of predicted FEV₁ (β=0.09, P<0.01), degree of bronchodilator response (β=0.05, P<0.01), and ICS adherence (β=2.38, P<0.01) were significantly associated with the change in ACT score. In contrast, only degree of bronchodilator response was significantly associated with the change in FEV₁ with treatment in the multivariable model (β=0.72, P<0.01).

Table 2.

Predictors of inhaled corticosteroid response measured as the change in the Asthma Control Test score and the change in lung function over 6 weeks of treatment.^*

	Measure of ICS response
	Change in ACT score*				Percent change in FEV1†
	Univariable parameter estimate (β)	P-value	Multivariable parameter estimate (β)‡	P-value	Univariable parameter estimate (β)	P-value	Multivariable parameter estimate (β)‡	P-value
Age in years§	−0.10 ± 0.19	0.62	−0.19 ± 0.14	0.17	−0.07 ± 0.72	0.92	−0.68 ± 0.57	0.24
Sex||	0.43 ± 0.53	0.42	0.18 ± 0.36	0.62	1.09 ± 1.99	0.59	1.21± 1.52	0.43
Race-ethnicity¶	0.93 ± 0.58	0.11	−0.57 ± 0.40	0.16	-2.36 ± 2.17	0.28	−2.52 ± 1.69	0.14
Body mass index§	0.35 ± 0.28	0.20	−0.01 ± 0.19	0.96	0.40 ± 1.04	0.70	0.45 ± 0.81	0.58
ACT score at baseline	−0.68 ± 0.04	<0.01	−0.71 ± 0.04	<0.01	−1.02 ± 0.19	<0.01	−0.24 ± 0.16	0.12
Percent of predicted FEV1 at baseline	−0.03 ± 0.02	0.21	0.09 ± 0.02	<0.01	−0.65 ± 0.07	<0.01	−0.13 ± 0.07	0.07
Bronchodilator response at baseline	0.09 ± 0.02	<0.01	0.05 ± 0.01	<0.01	0.80 ± 0.05	<0.01	0.71 ± 0.06	<0.01
ICS adherence	2.23 ± 1.26	0.08	2.38 ± 0.84	<0.01	3.78 ± 4.75	0.43	5.68 ± 3.55	0.11

ACT denotes Asthma Control Test, FEV₁, the forced expiratory volume at one second; and ICS, inhaled corticosteroid.

^*ICS response was measured as the numeric change in the composite ACT score from enrollment to follow-up after 6 weeks of ICS treatment.

^†ICS response was measured as the percent change in FEV₁ from enrollment to follow-up after 6 weeks of ICS treatment (i.e., FEV_1(6-weeks)-FEV_1(initial)/FEV_1(initial)).

^‡Adjusted for all other variables listed.

^§These variables were modeled for a single unit increase (i.e., year and kilogram per meter squared for age and BMI, respectively), but are aggregated here to show the effect of a 10-unit increase.

^||Referent is male (i.e., coded male=0, female=1)

^¶Referent is European American (i.e., coded European American=0, African American=1)

Because the effects of ICS adherence may differ with underlying asthma control or lung function, we stratified our analyses according to baseline ACT score (i.e., uncontrolled asthma [ACT score ≤19] and controlled asthma [ACT score ≥20]) and initial percent of predicted FEV₁ (i.e., <70% and ≥70%) (Table 3). ICS adherence was found to be a significant predictor of ACT improvement among individuals with uncontrolled asthma at baseline (β=3.95, P=0.04). Although ICS adherence had a consistent and positive association with FEV₁ improvement in both lung function strata (percent for predicted FEV₁ <70% and ≥70%), this association did not reach statistical significance in either group (β=5.35, P=0.49 and β=4.19, P=0.22, respectively). Both percent of predicted FEV₁ at baseline (β=0.11, P<0.01) and degree of bronchodilator response (β=0.09, P<0.01) were significantly associated with a change in ACT score among those whose asthma was initially uncontrolled (ACT ≤19). Bronchodilator response was associated with FEV₁ improvement among those in both lung function strata (P<0.01 for those with an initial percent of predicted FEV₁<70% and ≥70%). In none of these stratified models was self-reported race-ethnicity significantly associated with the change in ACT score or change in FEV₁.

Table 3.

Predictors of inhaled corticosteroid response stratified by baseline ACT score and baseline FEV₁.^*

	Measure of ICS response
	Change in ACT score*				Percent change in FEV1†
	Initial ACT score ≤19		Initial ACT score ≥20		Initial percent of predicted FEV1 <70%		Initial percent of predicted FEV1 ≥70%
	Parameter estimate (β)‡	P-value	Parameter estimate (β)‡	P-value	Parameter estimate (β)‡	P-value	Parameter estimate (β)‡	P-value
Age in years§	−0.35 ± 0.33	0.29	0.06 ± 0.14	0.70	−0.32 ± 1.35	0.81	−0.61 ± 0.52	0.24
Sex||	0.28 ± 0.83	0.73	0.33 ± 0.40	0.40	1.64 ± 3.38	0.63	1.01 ± 1.41	0.48
Race-ethnicity¶	0.48 ± 1.00	0.63	−0.17 ± 0.41	0.69	−5.70 ± 3.76	0.13	−0.84 ± 1.57	0.59
Body mass index§	−0.07 ± 0.43	0.88	0.13 ± 0.22	0.55	−0.26 ± 1.58	0.87	0.96 ± 0.84	0.25
ACT score at baseline	--	--	--	--	−0.63 ± 0.32	0.06	0.002 ± 0.16	0.99
Percent of predicted FEV1 at baseline	0.11 ± 0.04	<0.01	0.03 ± 0.02	0.16	--	--	--	--
Bronchodilator response at baseline	0.09 ± 0.03	<0.01	0.02 ± 0.03	0.56	0.74 ± 0.08	<0.01	0.63 ± 0.09	<0.01
ICS adherence	3.95 ± 1.94	0.04	1.41 ± 0.95	0.14	5.35 ± 7.69	0.49	4.19 ± 3.38	0.22

ACT denotes Asthma Control Test, FEV₁, the forced expiratory volume at one second; and ICS, inhaled corticosteroid.

^*ICS response was measured as the numeric change in the composite ACT score from enrollment to follow-up after 6 weeks of ICS treatment. The analysis was stratified by what was clinically considered to be uncontrolled asthma (ACT score≤19) and controlled asthma (ACT score≥20) at baseline.

^†ICS response was measured as the percent change in FEV₁ from enrollment to follow-up after 6 weeks of ICS treatment (i.e., FEV_1(6-weeks)-FEV_1(initial)/FEV_1(initial)). The analysis was stratified by percent of predicted FEV₁ measured at baseline (i.e., a cut-point of 70% of predicted FEV₁, which is the midpoint for persistent moderate severity asthma in the current U.S. asthma guidelines).¹⁶

^‡Adjusted for all other variables listed.

^§These variables were modeled for a single unit increase (i.e., year and kilogram per meter squared for age and BMI, respectively), but are aggregated here to show the effect of a 10-unit increase.

^||Referent is male (i.e., coded male=0, female=1)

^¶Referent is European American (i.e., coded European American=0, African American=1)

We assessed whether African ancestry was a predictor of ICS responsiveness (Table 4). Proportion of African ancestry was unrelated to ICS response as measured by both the change in ACT score and the change in FEV₁. African ancestry was similarly not related to ICS response even after stratifying by baseline level of asthma control and lung function.

Table 4.

Assessment of genetic ancestry as a predictor of inhaled corticosteroid response among African American individuals with asthma.

	Measure of ICS response
	Change in ACT score*						Percent change in FEV1†
	All individuals		Initial ACT score ≤19		Initial ACT score ≥20		All individuals		Initial percent of predicted FEV1 <70%		Initial percent of predicted FEV1 ≥70%
	Parameter estimate (β)‡	P- value	Parameter estimate (β)‡	P- value	Parameter estimate (β)‡	P- value	Parameter estimate (β)‡	P- value	Parameter estimate (β)‡	P- value	Parameter estimate (β)‡	P- value
Age in years§	−0.44 ± 0.19	0.02	−0.85 ± 0.38	0.03	0.04 ± 0.23	0.86	−0.27 ± 0.72	0.71	2.08 ± 1.77	0.25	−0.67 ± 0.71	0.35
Sex||	0.54 ± 0.49	0.27	0.57 ± 0.95	0.55	0.88 ± 0.59	0.14	1.73 ± 1.86	0.35	3.07 ± 4.22	0.47	1.04 ± 1.89	0.58
Proportion of African Ancestry¶	1.63 ± 2.30	0.48	1.37 ± 4.69	0.77	0.67 ± 2.71	0.81	2.96 ± 8.75	0.74	4.80 ± 22.52	0.83	2.45 ± 8.51	0.77
Body mass index§	0.05 ± 0.26	0.84	0.13 ± 0.49	0.79	−0.01 ± 0.31	0.98	−0.37 ± 0.97	0.70	−1.38 ± 1.99	0.49	−0.08 ± 1.06	0.94
ACT score at baseline	−0.69 ± 0.05	<0.01	--	--	--	--	−0.17 ± 0.18	0.35	−0.71 ± 0.41	0.09	0.15 ± 0.20	0.44
Percent of predicted FEV1 at baseline	0.09 ± 0.03	<0.01	0.06 ± 0.04	0.19	0.06 ± 0.04	0.14	−0.10 ± 0.10	0.29	--	--	--	--
Bronchodilator response at baseline	0.07 ± 0.02	<0.01	0.07 ± 0.04	0.07	0.07 ± 0.04	0.05	0.52 ± 0.09	<0.01	0.52 ± 0.13	<0.01	0.56 ± 0.12	<0.01
ICS adherence	3.92 ± 1.12	<0.01	7.84 ± 2.22	<0.01	2.45 ± 1.37	0.08	7.18 ± 4.28	0.10	11.53 ± 9.08	0.21	4.56 ± 4.64	0.33

FEV₁ denotes forced expiratory volume at one second; ACT, Asthma Control Test; and ICS, inhaled corticosteroid.

^*ICS response was measured as the numeric change in the composite ACT score from enrollment to follow-up after 6 weeks of ICS treatment. The ACT is stratified by what is clinically considered to be uncontrolled asthma (ACT score≤19) and controlled asthma (ACT score≥20).

^†ICS response was measured as the percent change in FEV₁ from enrollment to follow-up after 6 weeks of ICS treatment (i.e., FEV_1(6-weeks)-FEV_1(initial)/FEV_1(initial)). The analysis was stratified by percent of predicted FEV₁ measured at baseline (i.e., a cut-point of 70% of predicted FEV₁, which is the midpoint for persistent moderate severity asthma in the current U.S. asthma guidelines).¹⁶

^‡Adjusted for all other variables listed.

^§These variables were modeled for a single unit increase (i.e., year and kilogram per meter squared for age and BMI, respectively), but are aggregated here to show the effect of a 10-unit increase.

^||Referent is male (i.e., coded male=0, female=1)

^¶Modeled as the effect of each percent increase in the proportion of overall African ancestry, as measured using genome-wide genotype data.

As a post hoc analysis, we assessed factors associated with the likelihood of achieving self-reported asthma control (ACT score ≥20) after 6 weeks of ICS treatment among all individuals whose asthma was not controlled at baseline (initial ACT score ≤19). As shown in Table E1 of the online supplemental material, age, initial percent of predicted FEV₁, and degree of bronchodilator response were associated with the likelihood of reporting controlled asthma at the 6-week treatment follow-up.

DISCUSSION

Few studies have described the relationship between ICS treatment and the change in longitudinal measures of lung function and asthma control among African American individuals when compared to European American individuals. Here we demonstrate that neither self-identified race-ethnicity nor genetic ancestry were associated with ICS treatment response as defined by either a change in ACT score or FEV₁. This study builds upon our earlier analysis in African American individuals alone showing no relationship between African ancestry and change in FEV₁.⁹

The average improvement in FEV₁ was 11.3% in our overall study population. This is similar to our previous study where we observed a 11.6% improvement in FEV₁ after 6 weeks of ICS therapy.⁹ This magnitude of FEV₁ improvement is similar to that described by some,¹⁸ but higher than seen by others.^{19, 20} Part of these differences may be due to the relatively high level of medication adherence in the current study. We observed an average adherence of 79%, which is higher than that usually seen in unselected populations of ICS treated asthma patients.^{3, 4} This high level of adherence may be due to participants knowing that adherence was being monitored, as has been observed in other studies where patients were conscious of adherence monitoring.^21–23

While we found that neither self-identified race-ethnicity nor ancestry were associated with ICS response, our earlier work showed African ancestry to be associated with asthma exacerbations,²⁴ nocturnal asthma,¹² and lung function.²⁵ Perhaps this indicates that while the genetic determinates of ICS controller response do not differ among African American and European American individuals, determinates of intrinsic disease severity do. The implication here would be that genetic ancestry has little independent contribution to drug response beyond that already captured through baseline measures of asthma severity, asthma control, and lung function. Another potentially important implication of our findings is that differences in corticosteroid response are unlikely to account for the between group differences in asthma control and complication rates observed for African American and European American individuals on a population level.^{26, 27}

It is important to note that the lack of association for overall African ancestry should not be interpreted as an absence of population-specific pharmacogenomics variants. Genetic variants that influence drug response may occur exclusively or at different frequencies between population groups.^{28, 29} However, in a recent review of the pharmacogenomics of ICS response,³⁰ none of the existing genome-wide association studies included substantial numbers of African American individuals. Therefore, it is not known whether the risk variants that have been identified to date are generalizable to multiple population groups. Our results do not address the generalizability of existing pharmacogenomic associations. Rather, our study implies that the sum effect of genetic variants influencing ICS response is likely to be similar among individuals of African and European descent.

Our study is not without other limitations. First, the individuals in our study were all members of a single large health system in southeast Michigan; therefore, the findings from our study may not be generalizable to other patient populations within the US. However, the proportion of West African ancestry estimated for the individuals in our study is similar to that reported for other African American groups throughout the U.S.^{31, 32} Second, since we did not have admixture estimates in our European American participants, we could not assess the effect of ancestry within this group. However, there is no a priori reason to suspect that the effect of genetic ancestry would have differed between groups, and the lower degree of continental ancestral variation in European Americans would have required a much greater number of individuals to perform the same assessment.^{33, 34} Third, while this study did include both African American and European American individuals, the latter comprised a much smaller number of patients. Consequently, additional replication is needed to bolster our findings.

In this burgeoning era of personalized genomics, there is an increased effort to target therapies to individuals most likely to respond to treatment. Differences in medication response and treatment-related side effects by race-ethnicity have now been noted for a number of medications.^35–37 Fortunately, African American individuals, who as a group suffer disproportionately from asthma complications,^{38, 39} did not demonstrate response differences to ICS medication, the cornerstone treatment for persistent asthma. African American and European American individuals appeared to equally enjoy the benefits of ICS treatment for improving asthma control and lung function. In identifying individuals most likely to benefit from treatment, our study suggested that arguably more mundane factors, such as medication adherence, level of bronchodilator responsiveness, baseline lung function, and patient-reported asthma control, were consistently predictive of ICS response. Therefore, while pharmacogenomics may eventually pave the way for more targeted asthma treatment, fundamental characteristics of disease severity/control and management remain primary concerns for selecting and optimizing treatment.¹⁶

Clinical Implications.

• Neither self-reported race-ethnicity nor African ancestry appear to be major drivers of ICS treatment response, strongly suggesting that this cornerstone therapy is equally beneficial in treating African American and European American individuals.

• Easily obtained measures of lung function and asthma control may be useful in assessing the likelihood of response.

Abbreviations

ICS

inhaled corticosteroid

ACT

Asthma Control Test

SAPPHIRE

Study for Asthma Phenotypes and Pharmacogenomic Interactions by Race-ethnicity

FEV₁

forced expiratory volume at one second

MDI

metered dose inhaler

BMI

body mass index

SD

standard deviation