OVERWEIGHT/OBESITY ENHANCES ASSOCIATIONS BETWEEN SECONDHAND SMOKE EXPOSURE AND ASTHMA MORBIDITY IN CHILDREN

Summary

In two observational cohorts of inner-city children with asthma, secondhand smoke exposure was associated with worsened symptoms in overweight/obese compared to normal weight children, suggesting that elevated BMI may increase susceptibility to secondhand smoke, a common indoor pollutant.

TO THE EDITOR —

Asthma in overweight and obese children has been associated with worsened morbidity. Overweight and obese children may also be more susceptible to air pollution. Elevations in particulate matter, sulfur dioxide, and ozone have been associated with greater respiratory symptoms in overweight or obese children compared to normal weight children(1, 2).

Secondhand smoke (SHS) is a common indoor air pollutant. Approximately 37% of children in the United States were still exposed in 2011-2012, amid encouraging secular declines(3). If SHS exposure is associated with worsened asthma control in overweight and obese children, this may offer additional insight into the increased asthma morbidity observed in these children, and it would raise important questions about pathways that determine susceptibility. While public health initiatives to reduce obesity and SHS exposure are underway, such a finding would also have clinical relevance.

To investigate possible effect modification of the effects of SHS by overweight/obesity, we analyzed data from two observational cohorts of inner-city, predominantly African American children with asthma, the Mouse Allergen and Asthma Cohort Study (MAACS) and the Discover Study. In MAACS, 150 children were followed every three months for one year. SHS exposure was assessed by urine cotinine and asthma morbidity by caregiver recall of the number of days of symptoms in the prior two weeks. Symptoms included: days with cough, wheeze, chest tightness; slowed activity due to asthma; exercise-related symptoms; trouble speaking due to wheeze; cough without cold; nocturnal awakening; and a composite maximum asthma-days (days with slowed activity, exercise-related symptoms, or nocturnal symptoms due to asthma). In Discover, 162 children were followed every three months for nine months. SHS exposure was assessed by in-home airborne nicotine, and asthma morbidity by daily diary that asked if the participant had symptoms, used a rescue inhaler, had an unanticipated acute care need, or missed school due to asthma(4). Symptoms included trouble breathing; limited activities due to asthma; bother due to asthma; and nocturnal awakening. Full details are in the online repository.

Normal weight, overweight, and obesity were defined as BMI ≥5th, ≥85th, and ≥95th percentile by U.S. Centers for Disease Control criteria. Underweight participants were excluded. Baseline characteristics were statistically compared across BMI categories with Fisher’s exact and ANOVA/Kruskal-Wallis tests. Binomial generalized estimating equation regression models were constructed to estimate the association between SHS exposure and the daily odds for each outcome, utilizing an exchangeable correlation structure and robust variance estimator. Models controlled for age, gender, race, inhaled corticosteroid use, atopy, caregiver education level, and season of visit. A Wald-type test for interaction by BMI category was performed. Statistical significance was accepted at p<0.05 for main effects and at p<0.1 for interactions as suggested by Selvin and applied previously in modest sample-size environmental studies(5). For MAACS, overweight and obese participants were combined for the outcome of trouble speaking due to low event rates. Cotinine and nicotine were log₁₀ transformed to achieve normality; values below the limit of detection were multiply imputed from a log-normal distribution fit by maximum likelihood estimation(6) and were resampled at the participant level due to its longitudinal structure. Analysis was performed in Stata 13 (StataCorp; College Station, TX). Studies were approved by the Johns Hopkins Institutional Review Board. Assent from participants and informed consent from caregivers were obtained.

Of 289 participants with complete information, 48% were overweight/obese (Table E1). Participant were ages 5-17 years, 92% African American, and 52% male; 80% were atopic by skin test or blood testing, and 53% reported inhaled corticosteroid use. Half (49%) of urine cotinine and the majority (71%) of airborne nicotine measurements were above the limit of detection. With exception of there being fewer obese males within MAACS, there were no significant differences in baseline characteristics, proportion of cotinine or nicotine values above limit of detection, or their measured concentration between BMI categories (Tables E2, E3). In Discover, participants spent on average 15.6 hours/day at home, which was also similar between BMI categories.

There was evidence that overweight and obesity enhanced associations between SHS exposure and asthma outcomes (Table 1). For example, SHS exposure in MAACS was associated with greater odds of a maximum asthma-symptom day among obese (aOR 2.25; 95% CI 1.12-4.50) and overweight (3.40 [1.32-8.75]) compared to normal weight children (1.21 [0.73-2.00]) and in Discover with greater trouble breathing among obese (1.43 [1.02-2.01]) compared to overweight (1.34 [0.81-2.23]) or normal weight children (0.86 [CI 0.61-1.21]) (p-interaction<0.1 for all comparisons, Figure 1).

Table 1.

Association Between Secondhand Smoke Exposure and Asthma Outcomes^*

Study Outcome	BMI Category			P-value†
	Normal Weight	Overweight	Obese
Mouse Allergen and Asthma Cohort Study
Maximum asthma symptom-day‡	1.25 (0.81-1.92)	3.40 (1.32-8.75)	2.25 (1.12-4.50)	0.096
Cough, wheeze, or chest tightness	1.34 (0.88-2.03)	2.76 (1.09-7.03)	1.92 (1.01-3.65)	0.331
Slowed activity due to asthma	1.21 (0.73-2.00)	3.22 (1.08-9.64)	2.72 (1.24-5.95)	0.090
Trouble speaking due to wheeze§	1.06 (0.48-2.31)	14.81 (4.09-53.66)		<0.01
Nocturnal symptoms	0.81 (0.46-1.41)	5.66 (1.12-28.68)	1.68 (0.81-3.49)	0.037
Exercise-related symptoms	1.24 (0.73-2.12)	4.28 (1.42-12.96)	1.40 (0.72-2.73)	0.169
Cough without cold	1.00 (0.47-2.13)	3.93 (1.68-9.22)	1.04 (0.43-2.54)	0.046
Rescue inhaler use	1.01 (0.64-1.60)	1.27 (0.62-2.60)	1.09 (0.46-2.59)	0.848
Acute care need	0.77 (0.50-1.18)	1.40 (0.68-2.87)	0.92 (0.55-1.54)	0.367
Discover Study
Trouble breathing	0.86 (0.61-1.21)	1.34 (0.81-2.23)	1.43 (1.02-2.01)	0.093
Limiting activities due to asthma	0.85 (0.64-1.11)	1.20 (0.54-2.65)	1.18 (0.84-1.64)	0.266
Bother due to asthma	0.81 (0.60-1.10)	1.06 (0.49-2.32)	1.49 (1.08-2.04)	0.019
Woken due to asthma	0.79 (0.56-1.13)	1.26 (0.91-1.74)	0.97 (0.59-1.59)	0.141
Rescue inhaler use	0.76 (0.55-1.04)	1.86 (0.71-4.87)	0.79 (0.47-1.32)	0.231
Acute care need	0.68 (0.37-1.25)	1.07 (0.34-3.39)	2.05 (0.94-4.45)	0.046
Absent from school	0.86 (0.58-1.28)	1.03 (0.35-3.07)	1.72 (0.89-3.31)	0.216

^*Odds ratio per 10-fold increase in cotinine/nicotine, adjusted for age, sex, race, inhaled corticosteroid use, atopic status, caregiver education level, and season.

^†P-value for overall interaction between BMI category and cotinine/nicotine.

^‡Highest number of days of slowed activity, nocturnal symptoms, or exercise-related symptoms

^§Due to low event rate, overweight and obese categories combined

Figure 1.

The finding that overweight and obesity may increase susceptibility to SHS has potential clinical and public health relevance. It also adds to existing evidence that elevated BMI increases susceptibility to air pollution(1, 2, 7). Confirmatory and mechanistic investigation of these findings is warranted. Potential pathways may involve interdependent effects of obesity and SHS exposure on airways and systemic oxidative stress and inflammation. Case-control studies have identified additive interactions between asthma and obesity for elevations in inflammatory cytokines and propensity toward non-atopic systemic inflammation(8). Alternatively, pathways may involve differences in respiratory mechanics; overweight and obese children may heightened exposure to SHS due to higher resting minute ventilation(9) or exaggerated increases in minute ventilation as a result of pollutant exposure(10).

The finding in the MAACS study that the overweight group was most affected has been reported previously in studies of outdoor and indoor air pollution(1,2). This could reflect the condition that obese children are impacted by other factors that worsen asthma symptoms beyond air pollution, including more extreme impact of body mass on respiratory mechanics and greater influence of comorbidities. However, this pattern was not seen in Discover. Potential reasons for this heterogeneity between studies include differences in age, atopy prevalence and use of ICS, as well as unmeasured differences in adherence to ICS. Further, SHS exposure in MAACS was ascertained by urine cotinine and in Discover by airborne nicotine; differential accuracy in exposure measurement may contribute to differences between study findings. Finally, studied outcomes were all patient-reported, and our findings may be confounded by differential symptom reporting between BMI categories. While inclusion of two studies addressing the same fundamental question is a strength, we were unable to combine outcomes systematically because of these differences in study design and population. Larger studies that provide greater statistical power are needed to confirm the effects we have observed.

In two well-characterized cohorts, SHS exposure was associated with enhanced asthma symptoms in overweight and obese children compared to normal weight children. This suggests that overweight and obese children are more susceptible to SHS and adds to evidence that elevated BMI increases vulnerability to air pollution. These findings offer new insight into why asthma in overweight and obese children may be more severe.