Body mass index trajectories may represent modifiable targets in the promotion of respiratory health

Lung function and development are critical to long-term respiratory health [1, 2]. Identifying modifiable factors, particularly within the early life period, that can be targeted to optimise lung function is therefore a critical priority. Body mass index (BMI) is a well-established risk factor for poor lung function [3], and given the inherently variable nature of BMI, could offer such a target. However, conflicting reports [4–6] from the literature have limited our understanding of the impact of BMI on lung health, specifically with regards to changes in an individual’s BMI trajectory over time, which does not necessarily follow a linear or predictable course [7]. Furthermore, the biochemical underpinnings of the purported relationship between BMI and lung health have not been well explored. In this issue of the European Respiratory Journal, Wang et al. [8] assessed the impact of BMI trajectories from birth to age 24 years on lung function metrics within a deeply phenotyped cohort. The authors then performed metabolomic characterisation of urine samples collected at the conclusion of follow-up to identify biomarkers of impaired lung function within the BMI trajectory groups.

A unique strength of this study was the dense and longitudinal data on BMI measures in the first 24 years of life across a large number of participants, allowing the authors to construct BMI trajectories encompassing critical periods of lung development and maturation [9]. The study included 3204 participants from the Swedish BAMSE (Barn, Allergy, Milieu, Stockholm, Epidemiological) cohort, a population based prospective cohort which recruited children born between 1994 and 1996 in inner-city, urban and suburban areas of Stockholm, with follow-up ongoing [10]. By 24 years of age, the response rate remained above 75%, allowing extensive and robust inquiry into the relationship between longitudinal BMI and lung function. The authors applied latent class mixture modelling in individuals with at least four and as many as 14 BMI z-score measures to derive six BMI trajectory groups incorporating sex as a covariate, then compared lung function at age 24 years between the groups. In agreement with previous literature [4, 11], the authors found that BMI trajectories began to diverge as early as in infancy and accelerated BMI gain during childhood was associated with poorer lung function. Prior studies have reported that a persistently high BMI was associated with obstructive lung function patterns [4] and increased risk of asthma development [4, 11], while BMI trajectories measured into adulthood have demonstrated consistently harmful impacts on lung function [12]. However, leveraging the extended follow-up period of the BAMSE cohort, Wang et al. [8] further subdivided those with a high BMI in childhood into “persistently high” and “accelerated resolving” groups. The accelerated resolving group followed a similar BMI trajectory as the persistent high group until approximately age 8–10 years, at which time these individuals experienced a “resolving” phase when their BMI values approached BMI z-scores of 0. Crucially, this group showed improved lung function metrics as compared to the persistent high group. These findings have important public health implications for the identification of critical windows for intervention, as lung maturation is commonly understood to be complete around 8 years old [13]. They also have biological relevance, as this suggests a possible relationship between BMI and lung function beyond the changes in the mechanical properties of the lungs and chest wall that occur with an increasing BMI. Finally, these results demonstrate the importance of considering an individual’s phenotype trajectory as their exposure, rather than a single timepoint across that trajectory – as is often done in epidemiological research – which can lead to spurious associations. It is of course noteworthy, however, that these findings are based on lung function measured at only a single timepoint. The most complete understanding of the relationship between BMI and lung function can only be achieved by considering and appropriately modelling both as longitudinal and dynamic variables.

The authors added mechanistic insight though metabolomic profiling of urine samples collected at age 24 years. They assessed interactions between individual metabolites and BMI trajectory groups on lung function measures, reporting an enrichment of histidine metabolism among the metabolites that interacted with trajectories for pre- and post-bronchodilator forced expiratory volume in 1 s and forced vital capacity. Histidine is an essential amino acid necessary for the synthesis of proteins, with particularly important roles in the active site of enzymes [14], affecting anti-oxidant and anti-inflammatory processes [15], and has previously been associated with obstructive lung disease [15]. Obesity itself is considered to be a condition of chronic inflammation [16], and histidine supplementation has been shown to both improve insulin resistance and reduce BMI [17]. Therefore, it is feasible that a disruption to histidine metabolism may link BMI trajectories and lung function. A limitation in interpreting these findings is that urine metabolomic profiling was conducted only at age 24 years, which was the end of BMI trajectory monitoring. The metabolome is a dynamic ‘omic tool that reflects time-sensitive biochemical processes [18], so it is possible important developmental processes relevant to the lung and BMI were missed by not including additional time periods. Therefore, disentangling the relationships between histidine metabolites and BMI trajectory group interactions, and validating the potential of histidine metabolites as biomarkers, requires further work.

While the longitudinal BAMSE cohort offers many advantages, generalisability must be considered; BAMSE is an ethnically homogeneous cohort with >95% European ancestry. Previous studies of geographically diverse cohorts have supported connections between resolving BMI over time and improvements in lung function [3], but future analyses should consider ethnic and social diversity. Socioeconomic factors, such as diet, and lifestyle are particularly important for BMI trajectories [19]. Further, underrepresented groups often include high-risk individuals for respiratory diseases who may suffer from poor access to healthcare or suboptimal living conditions [20], and it is imperative that future public health strategies focus on such populations to consider the underlying causes of changes to an individual’s BMI over time. The authors also did not consider genetics in their investigation, despite the well-established heritable component to an individual’s BMI [21, 22]. There are shared genetic effects between BMI and lung function [21], which – if considered in future work – could further enlighten whether the differences in lung function metrics between the trajectory groups were driven, at least in part, by genetics. In paediatric cohorts, BMI polygenic risk scores have been associated with severe wheeze and increased incidence of lower respiratory tract infections [22], suggesting that the genetic component of BMI is relevant during the early life developmental period that was the focus of this current study. A major challenge in both validating the current findings and extending upon them in additional populations is finding suitable cohorts with equivalent data structures. The BAMSE study collected extensive data over a long period of follow-up, which is time-intensive, immensely costly and practically unfeasible in the majority of cases. An alternative to the traditional longitudinal cohort design that researchers are increasingly taking advantage of, is leveraging biobanks linked to electronic medical records (EMR) [23] to conduct human health studies across a large number of individuals. EMRs contain a wealth of extensive longitudinal data, including diagnoses, phenotypic measures, medical notes, laboratory tests, medications and other data, to provide a more complete picture of exposures and outcomes over long timespans. By linking these data to biobank samples, there are innumerable opportunities to search for ‘omic and other biomarkers along these health trajectories (figure 1). Consequently, EMR and biobank data have been increasingly implemented towards precision medicine aims [24]. However, extracting, cleaning and analysing such data is exceptionally complex [25], and novel computational methodologies are still in development, and ethnic and social diversity issues are not entirely resolved through application of EMR-linked biobanks [26]. Nevertheless, they represent a unique and exciting resource with unprecedented potential, particularly in the consideration of longitudinal exposure trajectories.

FIGURE 1.

Electronic medical records and biobank samples hold immense promise to improve understanding of the connection between body mass index (BMI) and lung function. Longitudinal monitoring of BMI and lung function, in conjunction with biosampling, could enhance insights into the connection between BMI trajectories and lung function outcomes, including identification of the optimal window for intervention to improve BMI. Biosampling throughout this time period could enhance insights into underlying biochemistry impacted by BMI and support the move towards biomarker identification.

In conclusion, the study by Wang et al. [8] provides novel insights into the connection between BMI trajectories and lung function and development over time while providing biochemical context through metabolomic investigation. The authors report encouraging results regarding the ability to intervene using modifiable lifestyle factors to improve lung function, with a window of opportunity extending to at least the pre-adolescent period. Furthermore, the authors propose histidine metabolites as potential urinary biomarkers of obesity-associated lung function changes, which provides a cost-effective method to clinically monitor patients in an easily accessible biofluid. Future studies leveraging a similar design, but considering both exposure and outcome as trajectories and incorporating additional longitudinal ‘omic measurements, particularly genetics and additional metabolomics time points, could capture the full spectrum of changes in underlying biology and greatly benefit our understanding of the role of BMI trajectories in lung function over time. Increasing our understanding will allow higher translation potential for early life and childhood interventional strategies that can mitigate the risks of poor lung function across the life course.