Healthy diet associated with better asthma outcomes in elderly women of the French Asthma-E3N study

As recently underlined in the Lancet [1], given the immense societal and individual burden of asthma, there is an urgent need to further develop novel strategies to limit the disease and its consequences. Asthma control, the main objective of asthma management, remains suboptimal in roughly one asthma patient out of two, with even higher rates among women [2, 3]. In this context, investigating the role of modifiable lifestyle factors such as diet is key for the primary and secondary prevention of this highly prevalent disease. While there are promising findings for possible dietary intervention to reduce asthma in children [4], the impact of diet on asthma in adults, and even more so among the elderly, remains largely unknown.

Among adults over 65 years, the prevalence of asthma varied from 4 to 13% [5], a rate likely underestimated as it is frequently underdiagnosed in this age group [6]. The burden of asthma is more significant in the elderly than in their younger counterparts [7, 8] with more hospitalisations [6, 8], and worse health-related quality of life [5, 6, 8] and asthma control [6]. Asthma in the elderly is a phenotype of interest, especially in women because asthma tends to be more prevalent (9.9% vs 6.2%) and more severe in women than in men [9]. Obesity is now an established risk factor for asthma [10, 11], with recent studies supporting the hypothesis that obesity is causally related to asthma [12], and with a higher risk in elderly women as compared to men [13]. Furthermore, multimorbidity is common in asthma patients, especially in the elderly [13]. Besides obesity, allergic rhinitis, chronic obstructive pulmonary disease (COPD), gastroesophageal reflux, and sleep apnoea syndrome that are the most common asthma-related multimorbidities, recent studies have suggested that other chronic conditions such as cardiovascular diseases (CVD), metabolic syndrome, or type 2 diabetes mellitus are also involved [14].

Although published studies on the diet–asthma association have traditionally focused on specific nutrients or foods, it is important in terms of dietary recommendations to emphasize overall dietary patterns rather than specific foods and nutrients to account for synergistic effects on health of foods and nutrients within the overall diet [15]. Several dietary scores have been proposed to globally assess diet quality and among them, the Alternate Healthy Eating Index-2010 (AHEI-2010) score, that reflects a healthy diet, has been associated with a lower risk of chronic diseases [16]. Among middle-aged women and men, a better adherence to a healthy diet evaluated by the AHEI-2010 was associated with a lower asthma symptom score [17, 18] and better asthma control [18]. Being a continuous measure of asthma, the asthma symptom score that has been associated with new onset of asthma, provides more power to detect risk factors for asthma as compared to a dichotomous definition [19]. Up to now, the role of a healthy diet on asthma symptoms, asthma control, and asthma-related metabolic and cardiovascular morbidities, which are commonly associated with both asthma [20] and an unhealthy diet [21], remains unknown among elderly women.

In a large study among elderly women, we aimed to investigate (1) the association between a healthy diet assessed with the AHEI-2010, and the asthma symptom score and, (2) among patients with asthma, the association of this dietary score with uncontrolled asthma and with specific metabolic and cardiovascular multimorbidity-related medications profile.

In this large study of more than 12,000 elderly women, a stronger adherence to a healthy diet evaluated by the AHEI-2010 was associated with a lower asthma symptom score, and among women with asthma, with disease characteristics associated with better asthma prognosis. These findings strengthen evidence supporting the promotion of a healthy diet to target reduction in asthma respiratory symptoms among elderly women, and confirm the importance of considering diet in the relation between asthma and comorbidities, especially cardiovascular diseases.

To our knowledge, five studies have investigated the association between the AHEI-2010 and asthma outcomes, among young [39] to middle-aged adults [17, 18, 40, 41], and using a dichotomous [39‐41] or a continuous [17, 18] definition of asthma, and they reported mixed findings. The three studies that used a dichotomous definition of asthma, namely wheezing (yes/no), incident asthma (yes/no), or at least one out of three current asthma symptoms (yes/no), reported conflicting results. Using the asthma symptom score, two previous studies have investigated its association with the AHEI-2010 and reported that a healthy diet was associated with lower asthma symptoms [17, 18]. Regarding other dietary scores (a priori approach), the most studied are the Mediterranean diet score based on recommended foods or nutrients for disease prevention (such as the AHEI-2010) and the Dietary Inflammatory Index (DII) which relates to pathophysiological processes relevant to asthma (i.e. inflammation). To our knowledge, only one cross-sectional study reported a positive association between the Mediterranean diet and the asthma symptoms score in adults [18]. For the DII, one case–control [42] and two cross-sectional studies [41, 43] investigated associations with asthma in adults, and they all reported that a higher DII was associated with asthma [42] or current asthma [41, 43]. By contrast, using the a posteriori approach to derive data-driven dietary patterns (i.e., statistically derived independently of their relevance to any disease), at least 13 studies have looked at the association between dietary patterns, mostly derived using principal component analysis (PCA), and asthma symptoms or incidence in adults, and they have yielded mixed findings [44‐47]. Our current findings are consistent with those based on younger populations and support the likely impact of a healthy diet in the prevention of asthma in elderly women. They extend results obtained from younger population and provide first evidence for the impact of a healthy diet in the prevention of asthma over the life course.

Regarding the impact of an overall healthy diet as a disease modifier, a healthier diet was associated in our study with a lower risk of uncontrolled asthma but the association did not reach statistical significance. Five studies have been published so far, all among middle-aged adults and overall, they reported mixed findings as follows: three studies reported an association between an overall healthy diet and lower uncontrolled asthma [48‐50], one reported no association [51], and one reported a borderline significant association in women [18]. Published studies are very heterogeneous in term of tools used to evaluate asthma control as follows: two used the ACT [18, 50] and three used the asthma control questionnaire (ACQ) [48, 49, 51]; in terms of tools used to collect dietary data, three used 24-h dietary records [18, 48, 50], and two used a FFQ [49, 51]; and in terms of tools used to estimate the overall diet, one used the Dietary Approaches to Stop Hypertension (DASH) score [48], one used the DII [50], two used the Mediterranean diet score [49, 51], and one used the AHEI-2010 [18]. All five studies used dietary data collected at the same time as asthma control data or in the year before. In our study, diet was evaluated by two validated semi-quantitative food history questionnaires administrated in 1993 and 2005, whereas asthma control was evaluated in 2011. The assessment of diet does not cover the ACT window of exposure and it is likely that participants may have modified their diet between 1993 and 2011. Anyhow, it has been shown that diet remains globally stable in this specific population, namely elderly women with a high level of education [52], and we used the average diet between 1993 and 2005 which is likely more representative of the usual diet as compared to one assessment only. Although diet might be a modifiable risk factor that could be targeted to help reduce asthma exacerbations and reach a good level of asthma control, its potential role on asthma control remains unclear, particularly among elderly women.

Besides allergic-related morbidities, that are common especially in childhood asthma, it has been more recently reported that mechanisms involved in adult-onset asthma also include several metabolic and inflammatory pathways also related to other chronic conditions such as obesity, the metabolic syndrome, type 2 diabetes mellitus, or CVD [53]. Although the underlying mechanisms remain unknown, it has been suggested that asthma and CVD share common risk factors such as smoking, obesity, or more recently reported, air pollution exposure [54], consistent with common etiological pathways. The AHEI-2010 is based on a comprehensive review to identify foods and nutrients that have been consistently associated with lower risks of CVD, cancer, and type 2 diabetes in clinical and epidemiological investigations [16]. The AHEI-2010 also captures additional information on diet quality that may further decrease the risk of metabolic diseases [16]. Our findings provide further evidence for a major role of an unhealthy diet as a common risk factor between asthma and CVD. A better understanding of the role of a healthy diet could lead to risk management strategies for patients with asthma to reduce their CVD risk.

Several hypotheses and mechanisms have been raised to explain the role of diet in asthma, including oxidative stress and inflammation, and more recently, vitamin D, epigenetic regulation, and imbalance in the gut microbiome [55]. Asthma is a chronic inflammatory disease of the airways, and endogenous reactive oxygen species have been implicated in its pathogenesis [56]. It has been extensively reported that a better quality diet (high in fruit, vegetables, whole grains, and legumes as reflected by a high AHEI-2010) is associated with lower inflammatory biomarkers [57], and more recently with a reduction in short-chain fatty acids [58] (produced by bacteria in the gut during fermentation of fibre from dietary plant matter) known to reduce airway inflammation [59]. In this manner, a dietary intervention based on a fibre-rich healthy diet might be relevant for both primary and secondary prevention of asthma.

Diet, physical activity, and body composition are nutritional factors that are not only closely interrelated (at a given time t) but also time-dependent, which makes it difficult to disentangle their separate effects on asthma outcome [60]. Regarding interrelations (at a given time t), although obesity can result from multiple factors (e.g., genetic predispositions, certain disease status, or medication use), it is most often a result of unhealthy lifestyle, including excessive dietary energy intakes and insufficient physical activity [61]. It has been shown that individuals with lower overall diet quality have higher risk of obesity [36]. The established evidence that links diet to obesity and that links obesity to asthma [10] and asthma control [62] can illustrate the role of obesity as a potential mediator in the diet–asthma association [60]. In this context, several studies provided results with and, in addition, without adjustment for BMI [63, 64] but it could lead to biased results [65]. Some novel analyzing approaches, such as the counterfactual approach, provide a new tool to face the above issues in mediation analysis [66]. To our knowledge, only two studies were conducted in the context of nutritional factors and asthma, suggesting that BMI partly mediates the association between high cured meat intake and worsening asthma symptoms over time [67], but does not mediate the association between overall diet quality—assessed using the AHEI-2010—and improved asthma symptoms [17]. From a longitudinal perspective, interrelations between nutritional factors and asthma are also time-dependent. Indeed, in addition to the potential role of each nutritional factor at a given time t on asthma at a time t + 1, asthma at time t − 1 may have modified nutritional factors at time t (e.g., asthma can lead to a decrease in physical activity), and each nutritional factor at time t − 1 may have modified another nutritional factor at time t (e.g., overweight/obesity can lead to modifying dietary and/or physical activity behaviours). To our knowledge, only one study has been conducted in the context of the time-dependent associations between physical activity, BMI and asthma (also using data from the E3N study) and suggested an independent causal deleterious effect of overweight and obesity on current asthma, but no independent causal effect of physical activity on current asthma [30]. Although longitudinal data are warranted to fully address these issues, our models stratified according to BMI yielded to similar associations between diet and asthma outcomes, and estimates obtained from SEM and those obtained from standard models are very close, showing that BMI is unlikely a major modifier or mediator in the association between diet with asthma outcomes.

Our study has several strengths and limitations. First, our analysis was cross-sectional. However, our findings are based on a large sample size, which allows accounting for several potential confounders and performing stratified analyses to address the robustness of the findings. Secondly, diet was evaluated in 1993 and 2005, and asthma in 2011. Although participants may have modified their diet between 1993 and 2011, it has been shown that the diet remains globally stable in this specific population of elderly women with a high level of education [52]. Although a substantial proportion of women did not complete the ACT (28%), analyses with and without imputed data yielded similar results. Besides, associations were consistent in several sub-populations and after adjustment for many potential confounders. In addition, we used validated tools to estimate asthma symptoms [24, 25] and asthma control [26] as well as the diet [68]. Secondly, we acknowledge that the main source of disease misclassification among this population of elderly women is probably misdiagnosis of COPD, and that potential overlap between asthma and COPD may have contributed to the association between the AHEI-2010 and asthma. However, we used the asthma symptom score that measures specific symptoms of asthma (and not of COPD) rather than a dichotomous definition of asthma which is more likely to include COPD patients. Moreover, similar associations were observed whatever the smoking status, especially among never smokers who are less likely to suffer from COPD. We also acknowledge that the association between a healthy diet and the “Predominantly metabolic multimorbidity-related medications” profile might be due in part, to obesity. Indeed, obesity (e.g., BMI ≥ 30 kg/m²) was included in the LCA to identify asthma groups, and whereas the “Few multimorbidity-related medications” profile and the “Predominantly allergic multimorbidity-related medications” include respectively 3.4% and 1.5% of obese women, the “Predominantly metabolic multimorbidity-related medications” profile includes 34% of obese women. Finally, the relative homogeneity of the studied population (e.g., elderly women with mostly high educational levels) actually helps with causal inferences about the relation between healthy diet and asthma outcomes because the comparability of the high and low dietary score groups will be higher than in a more heterogeneous populations (i.e., less potential for residual confounding).

In this elderly female population, we observed that a healthier dietary intake was associated with lower asthma symptoms, and among women with asthma, lower risk to belong to a metabolic multimorbidity-related medication profile. Overall, our findings show that a healthy diet may play an important role in the prevention and management of asthma over the life course. There is a need for longitudinal studies and RCTs to help understand better the role of the AHEI-2010 diet for the primary and secondary prevention of asthma among elderly. More studies are also warranted to better understand the role of a healthy diet as a common risk factor for patients with asthma to better reduce their CVD risk, which could lead to risk management strategies among those patients. Lastly, as the investigation of nutritional factors as a whole (diet, physical activity, obesity) is highly relevant in the etiology of asthma and its control, both in terms of understanding the underlying mechanisms and in terms of guiding efficient multidimensional public health interventions, there is a crucial need for further prospective studies to address this issue.