Life‐course trajectories of body mass index and cardiovascular disease risks and health outcomes in adulthood: Systematic review and meta‐analysis

This systematic review aimed to assess the effect of life course body mass index (BMI) trajectories (childhood to adulthood) on cardiovascular disease (CVD) risk factors and outcomes.

and obesity, respectively. 2[4][5][6] Overweight is a major contributor to the global burden of disease. 1 Global deaths and disability-adjusted life years (DALYs) attributable to high body mass index (BMI) more than doubled between 1990 and 2017, 1 and high BMI was one of the leading risks in 2019. 710][11] CVDs are the leading cause of mortality worldwide 12 and are a major contributor to disability. 13The recent Global Burden of Disease study revealed that the prevalence of CVDs almost doubled from 271 million in 1990 to 523 million in 2019, and CVD deaths increased from 12.1 million in 1990 to 18.6 million in 2019. 13High BMI is one of the most important risk factors for CVDs. 13,14High BMI levels during childhood strongly track (persist) into adulthood, leading to a higher risk of CVD, 15 which highlights the importance of promoting a healthy weight in early childhood and across the life course.
The lifetime health consequences of high BMI are dependent on both the severity and duration of overweight throughout the life course. 15,16The effect of high BMI on health needs to be investigated through a life course research approach.The life course research framework provides the capacity to examine the relationships between individual's life exposure to different factors and subsequent effects on health outcomes across the life span. 17,180][21][22] Trajectory analysis can illustrate the impact of an exposure across the life course, allowing the identification of critical time points and subgroups to inform prevention and control practices. 23wever, BMI trajectories are typically not consistently defined and their effect on CVD risk and outcomes not well synthesized.Some studies have focused on childhood BMI trajectories, 19,21,22 whereas others have assessed the BMI trajectories from childhood to adulthood 14,16,20 and their effect on CVD risks.As a result, it is difficult to understand the full impact of BMI trajectories on health outcomes across the life span.Thus, this study aims to identify common life course BMI trajectories and their impact on adulthood CVD risks and outcomes.Specifically, the following questions will be addressed.
1. What are the most common BMI trajectories across the life span from childhood (5-18 years) to adulthood (over 19)?
2. What are the effects of different BMI trajectories during childhood (5-18 years), or from childhood to adulthood, on CVD risk factors (hypertension, diabetes, and dyslipidaemia) and CVD outcomes (morbidity and mortality from CVD) in adulthood (over 19)?

| Study design
This systematic review and meta-analysis is reported in line with the 2020 preferred reporting items for systematic reviews and metaanalysis (PRISMA) guidelines. 24The protocol of this review was registered with the International Prospective Register of Systematic Reviews (PROSPERO; https://www.crd.york.ac.uk/prospero/display_ record.php?RecordID=336866).

| Search strategy and data source
Seven electronic databases were searched (MEDLINE, EMBASE, Scopus, CINAHL, Cochrane Library, Web of Science and Global Health) from inception to September 2022 and updated in September 2023.
In addition, we manually searched reference lists of relevant articles, including backward and forward citation searches.The final literature search approach was developed in a discussion with a senior librarian at Deakin University.The following search terms/keywords were used to identify relevant studies and titles, abstracts and keywords were searched: ("body mass index" OR bmi OR "overweight*" OR "over weight*" OR "body weight gain*" OR "weight gain*" or obes* OR anthropometr*) AND (trajector* OR "life course*" OR "life long" OR "lifelong" OR "life course trajector*" OR "life-course trajector*" OR "growth mixture model*" OR "group based trajectory model*" OR "latent class model*") AND ("cardiovascular risks" OR "cardiovascular risk factors" OR "cardiovascular disease risks" OR "cardiovascular" OR "cardiovascular disease*" OR CVD OR hypertension OR "raised blood pressure" OR "high blood pressure" OR "diabetes mellitus, type 2" OR diabet* OR dyslipidaemia OR "metabolic syndrome*" OR "health outcome*" OR "health impact*" OR "health effect*" OR mortalit* OR death*).

| Study selection
All identified studies were exported into Covidence 25 for duplicate removal, screening, full-text review, and data extraction.Three independent reviewers conducted title and abstract screening (KTK screened all papers; CS and TKT screened 50% each): full-text review was completed by KTK (all), and CS and TKT (50% each) against the inclusion and exclusion criteria (see below).Disagreements or conflicts were resolved through discussion and consensus.

| Inclusion criteria
Studies that met the following criteria were included in this review: • Exposure/BMI trajectories: • used BMI measure to assess body weight trajectory.We included studies that used either objective (measured) or subjective (selfreported or recalled) body weight and height measures to calculate BMI; • studies were eligible for inclusion if they evaluated BMI trajectories either within childhood (ages 5-18) or from childhood to adulthood (beyond age 18), with the outcomes specifically measured during adulthood (ages 18 and older); • used a longitudinal (cohort) study design that assessed BMI trajectories across life course; • the first time point for BMI value was collected in childhood (5-18 years; the trajectory should start from children aged 5 and over); • BMI value at a minimum of three time points, to identify trajectories in longitudinal data. 23Outcomes: cardiovascular health outcomes to be included were hypertension, diabetes, dyslipidemia, and morbidity or mortality due to CVDs, measured during adulthood (ages 18 and older).
Studies were included if outcomes were measured by self-report or objectively, based on ICD-10 Chapter I (I00-I99).
• Others: peer reviewed publications of original research published in English.

| Exclusion criteria
• Studies that focused on disease-specific populations, for instance, only people with diabetes or hypertension; • studies published in languages other than English, and non-human studies; • non-primary research articles (brief communications, commentary, editorials, and reviews).

| Data extraction
After identifying eligible studies through full-text review, one reviewer (KTK) extracted data using the Covidence data extraction template.The second reviewer (TKT) double-checked 15% subsample of the extracted information to check the consistency.General information that was extracted for each study included: the first author, year of publication, sample size, country, and study population.Information on method and trajectory-specific data such as statistical modelling methods, number of trajectories identified, names and nature of trajectories, the proportion of people in each trajectory, number of body weight measures, period of life (age range), duration/follow up of the measurement, and adjusted variables were also extracted.Finally, data on the association between BMI trajectories and the outcomes of interest, such as CVD risks (hypertension and diabetes), CVDs, mortality, measure of outcomes (objective or self-report), and ages at which outcome measured, were also extracted.

| Quality assessment
The quality of the included studies was assessed using the Newcastle-Ottawa Scale (NOS) for cohort studies. 26The NOS comprises eight items grouped into three domains: study population/ group selection, study group comparability, and outcome or exposure determination.This NOS tool semi-quantitatively appraises the methodological quality of the studies using a star system.If the study has the highest quality, a maximum of four stars, two stars, and three stars are awarded for selection domain, comparability, and outcome domain items, respectively.The minimum NOS score is zero, and the maximum is nine stars.Using the stars, the quality of each study is classified as good (seven to nine stars), fair (four to six stars), or poor quality (zero to three stars). 26Quality assessment was completed by one reviewer (KTK) and 15% of the articles double assessed by the second reviewer (TKT) to check the consistency.

| Synthesis of results/data analysis
A narrative synthesis was performed for the results of studies not included in the meta-analysis to summarize the types and nature of BMI trajectories and their association with CVD risk factors and outcomes (mortality and incidence).8][29] When studies included in the meta-analysis reported only a regression coefficient (β), exponentiation of coefficients was used to estimate OR/RR. 30Certain trajectories were consistently reported across multiple studies.Based on the characteristics of consistently identified BMI trajectories, the following four classifications of trajectories were defined for meta-analysis: (1) "stable normal weight" trajectory (normal BMI throughout the life course/ follow-up [childhood to adulthood]); (2) "normal-to-overweight trajectory (normal BMI during childhood and overweight in adulthood); (3) "overweight-to-normal" trajectory (overweight in childhood and normal in adulthood); and (4) "persistently overweight" trajectory (overweight/obesity during both childhood and adulthood).The trajectory descriptions reflect how study participants moved between weight status classifications over the study period (e.g., staying within the normal weight category or moving from normal weight to overweight classification).Note that these categories do not include all trajectories identified in the included studies.So that the results of the studies that were not included in the pooled estimates of the meta-analysis were presented as a narrative summary.The pooled effects of BMI trajectories from childhood to adulthood on the RR for type 2 diabetes, hypertension, and dyslipidemia were analyzed using the random effects model with a 95% confidence interval (CI).We could not calculate the pooled estimate for other outcomes due to insufficient data.
Heterogeneity was assessed visually using a forest plot 31 and statistically using tau square, Cochran's Q test, and I 2 statistics. 32nsitivity analyses, including leave-one-out meta-analysis, were conducted to assess the stability of the pooled estimates in the metaanalysis.Leave-one-out meta-analysis is a statistical technique employed to evaluate the impact of each study on the overall results.
In this method, each study is removed from the meta-analysis in turn, and the analysis is performed without that study.This allows evaluation of how sensitive or robust the overall findings are to the exclusion of any single study.For the outcomes with sufficient studies available (hypertension and type 2 diabetes), additional sensitivity analyses were performed in which the meta-analyses were restricted to studies using objectively measured BMI, to studies with length of follow-up ≥20 years, and to studies with a quality rating of "good." We conducted all analyses using Stata version 17 (StataCorp LP, College Station, Texas, USA).

| Characteristics of included studies and quality
The database searches yielded a total of 15,875 results.After the removal of duplicates, 7541 articles remained and were screened on title and abstract.We identified 80 articles for full-text screening, after which 17 were found to be eligible for inclusion in the review (Figure 1).
The included studies were published between 2013 and 2022 and included a total of 370,689 participants.The sample size of individual studies ranged from 336 to 116,888.Twelve of the included studies were conducted in high-income countries: two each in the USA, 33,34 Denmark, 19,35 the UK, 36,37 and Spain, 38,39 and one each in the Netherlands, 40 France, 41 Sweden, 20 and Finland. 16Of the remaining five studies, four were conducted in China 14,21,42,43 and one in Iran. 22The proportion of women among the study samples ranged from 40% to 67%, except one study, 41 which recruited only women (Table 1).
Most studies used a prospective cohort design (n = 11), with six using a retrospective cohort design.Eleven studies assessed the effect of BMI trajectories on hypertension or high blood pressure; 10 studies assessed the effect on type 2 diabetes or high blood glucose, six reported on dyslipidemia, and four reported on morbidity and mortality of CVDs (Table 1).Six studies used a group-based trajectory model to identify BMI trajectories, four used a latent class growth model, three used a latent class trajectory model, and two used a latent growth mixture model, in two studies not described (Table S1).All studies were assessed to be of good (n = 6) or fair (n = 11) quality, according to the NOS assessment (Table S1).
T A B L E 1 Characteristics of studies included on life-course trajectories of body mass index and cardiovascular disease risks and outcomes in adulthood, by outcome type.S1).The intermittent overweight trajectory, characterized by normal BMI during childhood and overweight BMI in adulthood or high BMI during childhood and normal BMI in adulthood, was identified in some studies.

| BMI trajectories and hypertension
The results of studies that were not included in the meta-analysis, due to the absence of trajectories that could be classified into the defined trajectory categories, are presented here as a narrative summary.A study from the Netherlands 40 reported that an overweight-to-normal trajectory was not significantly associated with an increased mean arterial pressure.Another retrospective study from China 21 showed that compared with the normal-stable BMI trajectory during childhood (6-18 years), the risk of hypertension in adulthood was higher in low Table 1).

| BMI trajectories and type 2 diabetes
Ten studies 14,16,19,20,22,33,35,37,41,42 reported results on the effect of different BMI trajectories from childhood to adulthood on type 2 diabetes mellitus risk.Three studies were not eligible for meta-analysis because they did not include any trajectories that could be classified into the most common trajectory categories that we defined.In a prospective cohort study from Denmark, 19 it was observed that a steep or steeper increasing childhood BMI trajectory was not associated with a higher risk of adulthood type 2 diabetes compared to a stable normal weight trajectory.Similarly, a population-based prospective cohort study from Iran reported that overweight trajectory in both females and males was not significantly associated with a higher risk of type 2 debates. 22However, in another cohort study, it was reported that a childhood with persistently overweight trajectory in both females and males was associated with a higher incidence rate of adult-onset type 2 diabetes compared with a stable normal weight trajectory 35 (Table 1).
The persistently overweight trajectory was reported by three studies 14,16,36 to be associated with an increased risk of dyslipidemia compared to the normal-stable weight trajectory.A prospective cohort study 36 in the United Kingdom with 3549 participants showed that participants with persistent overweight from childhood to adulthood were at a higher risk of having high-risk total cholesterol (β = 0.20; 95% CI: 0.03, 0.37, p = 0.023), and high LDL-C (β = 0.17; 95% CI: 0.024, 0.32, p = 0.023) during adulthood compared with participants with stable normal weight trajectory from childhood to young adulthood.Similarly, a study from China 14 reported significantly higher odds of dyslipidemia in the constant high BMI trajectory (OR: 5.9; 95% CI: 3.9, 8.8).Another prospective study 16 1).

| BMI trajectories and CVDs and mortality
Four studies 33,34,37,38 assessed the association between BMI trajectories and CVDs and mortality.Meta-analysis was not conducted for these outcomes because these studies did not analyze trajectories that could be classified into the major groupings used for the metaanalyses.Two studies assessed the association between BMI trajectories and CVDs.A study conducted in the United States 33 reported that compared to a stable normal trajectory, the risk of CVD was higher in a normal-to-overweight trajectory in females (HR:  1).

| DISCUSSION
This systematic review and meta-analysis synthesized the evidence on the effect of BMI trajectories across the life course (from childhood to adulthood) on CVD risks and outcomes.The findings suggest that a persistently overweight trajectory from childhood to adulthood is associated with a higher risk of hypertension, type 2 diabetes, dyslipidemia, and CVD morbidity and mortality.Findings were generally robust in sensitivity analyses restricted to objectively measured BMI, the follow-up of ≥20 years, or to good quality studies.The results of this systematic review underscore the significance of overweight prevention and intervention from childhood and throughout life to prevent CVD risks and outcomes in later adult life.
Previous studies have shown that children with overweight or obesity during childhood are at higher risk of CVD in adulthood. 44ndelian randomization study revealed that childhood obesity is associated with increased risks of type 2 diabetes and CVD in adults. 45However, adulthood BMI status may mediate these relationships that need to be traced over the life span. 44,46,47In line with this, the results of the current study showed that the risk of hypertension and CVD mortality was not significantly higher in those who were overweight as children, but have a normal BMI as adults, compared to those with normal BMI throughout their life span.The evidence has also indicated that children with overweight or obesity who achieve a normal BMI in adulthood may have similar cardio-metabolic profiles as those who have had a normal BMI at both stages. 48,49This result may suggest that CVD risks in adulthood can be avoided if childhood overweight is successfully treated.However, since our results are derived from a limited number of studies with considerable heterogeneity, it is essential to interpret these findings cautiously.Although the results were consistent in sensitivity analyses restricting the metaanalysis to studies with objectively measured BMI, with longer followup, and higher quality, further investigation is necessary to validate and confirm the results.
The findings of the current study showed that adult CVD risk factors were significantly higher among those with a persistently overweight trajectory from childhood to adulthood compared to those who had a normal BMI trajectory throughout the life course.This is consistent with the evidence that greater overweight/obesity duration increases the risk of cardiovascular risk factors. 50A doseresponse relationship between the duration of overweight/obesity and type 2 diabetes, CVD and mortality has also been reported. 51,52us, the association between persistently overweight trajectory in childhood and adulthood could be due to tracking of BMI across the life course or reflects the duration of overweight.Similarly, this result also aligned with the life course risk theory of the cumulative risk effect of exposure (high BMI) across the life span. 53These findings highlight the significance of devising prevention strategies, starting from childhood across the life span, to prevent cardiovascular risk factors and outcomes in adulthood.
Various pathophysiological and metabolic mechanisms could explain the association between high BMI trajectories from childhood to adulthood.Persistently overweight across life span leads to high adipose tissue distribution, which increases free fatty acid.Adipose tissue releases adipocytokines (interleukin-6, leptin, and resistin), which stimulate hypercoagulation, insulin resistance, and inflammation, which speeds up the atherosclerosis process. 54The inflammatory process leads to endothelial dysfunction, atherosclerotic plaques, insulin resistance, type 2 diabetes, and hypertension. 55,56In addition, C-reactive proteins are increased in obesity because of inflammation, which is associated with an increased risk of myocardial infarction, peripheral vascular disease, and diabetes. 54Therefore, overweight/ obesity from childhood to adulthood may influence adult CVD risk and outcomes through such a pathophysiological and metabolic process.
Several factors, including socioeconomic or sociodemographic and behavioural factors, may also be related to high BMI or to CVD risks and outcomes. 53,57Some studies included in this review adjusted for various variables, including sex, age, physical activity, alcohol, smoking, diet, and alcohol consumption.
The strengths of this review include that it used a systematic approach to assess the effect of long-term BMI trajectory (from childhood to adulthood) on CVD risks and outcomes and quantitatively synthesized the available evidence.In addition, BMI trajectories were determined using multiple measurements (minimum three-time points) across the life span.As a result, various patterns of change in BMI were identified, and subsequent cardiovascular risk was assessed, making it possible to identify high-risk groups.All the included studies were prospective or retrospective cohorts with large sample sizes and long-term follow-ups.The limitations of this review include that some of the included studies relied on self-reported BMI values, but the majority used objectively derived values.Most of the included studies adjusted for relevant covariates, including sociodemographic and behavioural factors; however, the potential role of confounding effect could not completely be removed as all the included articles are observational studies and all studies could not adjust all potential confounders.The causal association between BMI trajectories and CVD risks could not be established, because all the included studies are observational.However, the results demonstrate temporal relationships, consistency, and strong associations.Another limitation of this review is that BMI trajectories were not consistently defined across studies.Consequently, it is challenging to synthesize their effect on outcomes, and we did not include them in the meta-analysis.We only conducted a synthesis for some common trajectories (i.e., stable

F I G U R E 2
Forest plots for the pooled risk ratio (RR) of the association between BMI trajectories and hypertension, compared to stable normal weight trajectory: (A) RR of hypertension for normal-to-overweight trajectory; (B) RR of hypertension for persistently overweight trajectory.moderate increase trajectory (the normal BMI range during childhood with moderate increase; OR: 2.48; 95% CI: 1.39, 4.42), low rapid increase trajectory (OR: 3.24; 95% CI: 1.66, 6.31), and moderate increasing trajectory during childhood (OR: 3.28; 95% CI: 1.19-9.08;

F I G U R E 3
Forest plots for the pooled risk ratio (RR) of type 2 diabetes for different BMI trajectories compared to stable normal weight trajectory: (A) RR of type 2 diabetes for normalto-overweight trajectory; (B) RR for type 2 diabetes for persistently overweight trajectory.