Risk factors during first 1,000 days of life for carotid intima-media thickness in infants, children, and adolescents: A systematic review with meta-analyses.

BACKGROUND: The first 1,000 days of life, i.e., from conception to age 2 years, could be a critical period for cardiovascular health. Increased carotid intima-media thickness (CIMT) is a surrogate marker of atherosclerosis. We performed a systematic review with meta-analyses to assess (1) the relationship between exposures or interventions in the first 1,000 days of life and CIMT in infants, children, and adolescents; and (2) the CIMT measurement methods. METHODS AND
FINDINGS: Systematic searches of Medical Literature Analysis and Retrieval System Online (MEDLINE), Excerpta Medica database (EMBASE), and Cochrane Central Register of Controlled Trials (CENTRAL) were performed from inception to March 2019. Observational and interventional studies evaluating factors at the individual, familial, or environmental levels, for instance, size at birth, gestational age, breastfeeding, mode of conception, gestational diabetes, or smoking, were included. Quality was evaluated based on study methodological validity (adjusted Newcastle-Ottawa Scale if observational; Cochrane collaboration risk of bias tool if interventional) and CIMT measurement reliability. Estimates from bivariate or partial associations that were least adjusted for sex were used for pooling data across studies, when appropriate, using random-effects meta-analyses. The research protocol was published and registered on the International Prospective Register of Systematic Reviews (PROSPERO; CRD42017075169). Of 6,221 reports screened, 50 full-text articles from 36 studies (34 observational, 2 interventional) totaling 7,977 participants (0 to 18 years at CIMT assessment) were retained. Children born small for gestational age had increased CIMT (16 studies, 2,570 participants, pooled standardized mean difference (SMD): 0.40 (95% confidence interval (CI): 0.15 to 0.64, p: 0.001), I2: 83%). When restricted to studies of higher quality of CIMT measurement, this relationship was stronger (3 studies, 461 participants, pooled SMD: 0.64 (95% CI: 0.09 to 1.19, p: 0.024), I2: 86%). Only 1 study evaluating small size for gestational age was rated as high quality for all methodological domains. Children conceived through assisted reproductive technologies (ART) (3 studies, 323 participants, pooled SMD: 0.78 (95% CI: -0.20 to 1.75, p: 0.120), I2: 94%) or exposed to maternal smoking during pregnancy (3 studies, 909 participants, pooled SMD: 0.12 (95% CI: -0.06 to 0.30, p: 0.205), I2: 0%) had increased CIMT, but the imprecision around the estimates was high. None of the studies evaluating these 2 factors was rated as high quality for all methodological domains. Two studies evaluating the effect of nutritional interventions starting at birth did not show an effect on CIMT. Only 12 (33%) studies were at higher quality across all domains of CIMT reliability. The degree of confidence in results is limited by the low number of high-quality studies, the relatively small sample sizes, and the high between-study heterogeneity.
CONCLUSIONS: In our meta-analyses, we found several risk factors in the first 1,000 days of life that may be associated with increased CIMT during childhood. Small size for gestational age had the most consistent relationship with increased CIMT. The associations with conception through ART or with smoking during pregnancy were not statistically significant, with a high imprecision around the estimates. Due to the large uncertainty in effect sizes and the limited quality of CIMT measurements, further high-quality studies are needed to justify intervention for primordial prevention of cardiovascular disease (CVD).

Introduction

Background

Within a Developmental Origins of Health and Disease (DOHaD) framework, the first 1,000 days of life, i.e., the period from conception to age 2 years, is regarded as a critical period for long-term cardiovascular health [1]. Evidence suggests that adaptations in body structure and function during development, as a response to cues from the environment, for instance, assisted reproductive technologies (ART) for conception [2], undernutrition, or environmental pollutants [1], could increase the risk for cardiovascular disease (CVD). Indeed, several studies showed that a low birth weight is associated with higher blood pressure [3] or diabetes risk [4], as well as with cardiovascular mortality in adulthood [5].

Carotid intima-media thickness (CIMT) is a marker of CVD risk linked to pathways that originate in the first 1,000 days of life. In adults, CIMT is predictive of heart attack and stroke risk [6,7] and is also associated with prenatal risk factors, such as impaired fetal growth and preterm birth [8,9]. Within a DOHaD perspective, these associations may be explained by early life adaptative changes in the vascular phenotype [10], with lifelong effects on the CVD risk. Nonetheless, in children, the evidence on increased CIMT after impaired fetal growth or preterm birth is inconsistent [11-14]. Further, CVD risk is determined by a combination of factors, at multiple levels, rather than a single factor [15]. A large range of early life factors at the child, familial, or environmental levels needs to be explored to better understand CIMT changes in children. To our knowledge, such a comprehensive overview is lacking.

CIMT assessment in children can be challenging [16,17]. Several aspects related to the ultrasound equipment [18-20], the site of measurement, or the edge detection approach may influence the reliability of measurements [17,21]. Further, measurements in infants and young children are challenging due to limited compliance, anatomic particularities, and a lack of recommendations tailored to this age group [16,22]. The comparability of studies is therefore limited, and the inconsistency in findings may be attributed, at least in part, to the heterogeneity in CIMT measurement. An appraisal of the CIMT measurement methods in children is needed to shed light on these limitations.

We therefore aimed to perform a systematic review with meta-analyses to (1) assess the relationship between exposures or interventions in the first 1,000 days of life and CIMT in infants, children, and adolescents; and (2) critically appraise the CIMT measurement methods.

Methods

Protocol development and reporting

The protocol for this systematic review was registered on the International Prospective Register of Systematic Reviews (PROSPERO; CRD42017075169) and published in full in a peer-reviewed journal [23]. The reporting of this paper complies with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [24] (S1 PRISMA Checklist) and the Meta-analysis of Observational Studies in Epidemiology (MOOSE) guidelines [25].

Eligibility criteria

Study designs

Observational and interventional studies with the following designs were considered for inclusion: cohort, case–control, cross-sectional studies, randomized trials, nonrandomized controlled trials, and noncontrolled trials. Case reports, case series, opinion papers, letters to the editor, comments, conference abstracts, policy papers, reviews and meta-analyses, study protocols without baseline data, and animal studies were excluded.

Participants

We considered for inclusion studies in children from birth up to 18 years old, including studies where participants were recruited pre-birth. Apparently, healthy children and participants with clinical conditions relating to the first 1,000 days of life (e.g., prematurity) were included. Studies in children with rare or special conditions, such as congenital heart diseases, or postnatal conditions conferring an increased risk of CVD, such as hypertension or diabetes, were not considered for inclusion. We aimed to make inferences to the general population. Further, we assumed that there may be a different effect of DOHAD exposures in these groups than in the general population, which should be evaluated separately.

Exposures/Interventions

We considered for inclusion prenatal and postnatal exposures, at the individual, familial, or environmental level, which reflected the developmental milieu of the child and occurred between conception and age 24 months (i.e., first 1,000 days of life). These included biological factors related to physiology, physiopathology, or epigenetics and contextual factors related to socioeconomic status, behaviors, or environmental toxins. The selection of exposures was informed by the review of Hanson and Gluckman [1] and Moore and colleagues [26] and was conditioned by the timing of occurrence or ascertainment of exposure, as appropriate. Exposures that started in the first 1,000 days of life, but continued in later life, were included if they spanned up to 5 years of age. Likewise, periconceptional factors that reflected the environment in which the baby was conceived and grew were included. For interventional studies, we considered both pharmacological and non-pharmacological interventions, at the individual, familial, or environmental level, conducted in the first 1,000 days of life. Specifically, eligible interventions had to start and end in the first 1,000 days of life or start in the first 1,000 days of life and continue in later life no longer than child age 5 years.

Comparators

Where applicable, the use of a reference group without the exposure or intervention of interest was sufficient for inclusion.

Outcome measures

The outcome was the intima-media thickness of the carotid artery measured by ultrasonography from birth to 18 years of age.

Time frame and setting

There was no restriction by length of follow-up. There was no restriction by type of setting.

Language

Studies in English and French were considered for inclusion.

Search strategy

Systematic searches were conducted in the Medical Literature Analysis and Retrieval System Online (MEDLINE) database, Excerpta Medica database (EMBASE), and Cochrane Central Register of Controlled Trials (CENTRAL) from inception to 18 March 2019. The strategies for the systematic searches included the 2 main concepts of this systematic review: (1) children and adolescents; and (2) CIMT (detailed search strategies are available in S1 Table). Supplementary searches consisted of (1) a manual search of reference lists and other reviews on the topic; (2) forward citation tracking on Web of Science based on retrieved eligible reports; and (3) personalized search queries in Google Scholar and trial registers. The full strategies are available in S2 Table.

Study selection process

Following systematic searches in the databases, all retrieved reports were imported into Endnote (version X8.1, Clarivate Analytics, London, United Kingdom), and duplicates were removed according to the method of Bramer and colleagues [27]. Subsequently, reports were uploaded to Covidence (www.covidence.org), a systematic review management platform. Two reviewers performed study eligibility screening independently and in duplicate. Each reviewer screened the reports in a 2-step approach: (1) based on titles and abstracts; and (2) based on full-text of reports retained in the first step. The reviewers agreed on 94% of the titles and abstracts and on 84% of the full texts. Disagreements between reviewers were resolved by discussion, or, if necessary, by a third reviewer. The investigator of 1 completed study (ClinicalTrials.gov identifier: NCT02147457), with unpublished data on CIMT, was contacted by e-mail, but data could not be made available due to technical and logistical issues when performing measurements.

Data extraction

Data were extracted independently and in duplicate by 2 reviewers using a standardized electronic form in Microsoft Excel (version 2016, Microsoft Corporation, Redmond, United States of America). The data extraction process was piloted on a subset of articles, and a guide for extraction was designed to minimize errors and enhance consistency between reviewers. We extracted information about (1) study characteristics; (2) population characteristics; (3) CIMT definition and additional characteristics of the CIMT measurement methods; (4) characteristics of each exposure or intervention type; and (5) adjusted and unadjusted association or effect estimates.

The quality of each study was evaluated independently by 2 reviewers using an adjusted version of the Newcastle–Ottawa Scale for observational studies [28] and Cochrane collaboration risk of bias tool for interventional studies [29,30] (S7 Table). The quality of the CIMT measurement method was examined according to a predefined tool with 3 levels (higher, lower, and unclear reliability). The complete tool and the algorithm of judgment are available in S3 Table.

Disagreements in extracted data between the 2 reviewers were resolved by discussion or with the arbitration of a third reviewer. Essential missing information was searched by checking additional references related to that study, such as the published research protocol. We also attempted to contact authors of 2 studies by e-mail.

Data analysis

Analyses were conducted as prespecified in the study protocol [23]. Data transformations and estimations were done according to recommendations and formulae provided in Borenstein and colleagues [31,32], the Cochrane Handbook for Systematic Reviews [29], Lipsey and colleagues [33], Rücker and colleagues [34], Wan and colleagues [35], Aloe and colleagues [36], and Nieminen and colleagues [37]. The following data transformations were performed: (1) when the mean was not available, it was estimated from the median [35]; (2) when standard deviations were not available, they were estimated from the range or interquartile range [36], standard errors [33], and confidence intervals (CIs) [29,33]; and (3) when standard errors were not provided, they were calculated from CIs, p-values, and t-values [29,33]. To perform meta-analyses, additional data simplifications were performed as follows: (1) if a study reported on the same exposure or outcome type at multiple time points, the latest time point was included in the analyses; (2) if a study reported on multiple exposed or reference subgroups, they were combined into 1 using formulae provided by the Cochrane collaboration in case of means and standard deviations [29] or fixed-effect meta-analysis in case of association estimates for each subgroup [32]; (3) if a study provided association estimates separately for left and right CIMT or multiple sample groups, they were combined using fixed-effect meta-analysis [32,34]; (4) if a study reported on both mean and maximum CIMT, mean CIMT was included in the analyses; and (5) if multiple publications from the same study reported on the same exposure type, the publication providing data on the longest follow-up, left and right and/or mean thickness CIMT, or with the clearest reporting or highest sample size was used.

Standardized association measures were used to pool results across studies so that the comparison of different exposure types would be possible on the same scale. The standardized mean difference (SMD) in CIMT between exposed and reference groups was used as the common association measure for binary exposure variables [31-33]. When means and standard deviations for exposed and reference were not available or could not be estimated from other statistics as reported above, SMD were estimated from the unstandardized regression coefficient and the full-sample standard deviation [33,38]. The correlation coefficient was used as the common association measure for continuous exposure variables [32,33,36,37]. When the correlation was not available, it was estimated from the regression coefficient and its standard error or the unstandardized regression coefficient and the standard deviations of the exposure and outcome variables [36,37]. If the reported association measures could not be transformed to SMD or correlation coefficient, they were analyzed and reported separately. Estimates from bivariate associations or partial associations that were least adjusted for sex were used for pooling data across studies [39]. If at least 3 studies reported on the same exposure type, with a similar definition, association estimates were pooled using the DerSimonian–Laird random-effects model. As recommended by Amrhein and colleagues [40], together with more than 800 cosignatories of the initiative “Retire statistical significance,” and because we assessed multiple associations, we interpreted the point estimate as the most compatible value of the observed association. The uncertainty around the point estimate is comprised in the accompanying 95%CI, which is reported along the corresponding p-value.

The heterogeneity was assessed using the Cochran Q test and I2 and tau2 statistics [29,32,41]. I2 and tau2 are standard measures of heterogeneity in random-effect meta-analyses that reflect the amount of true heterogeneity between studies, i.e., the variation in effect sizes that is not due to sampling error. I2 is the percentage of variation in effect sizes attributable to heterogeneity, ranging from 0% to 100%, with 0% indicating that all variation in effect sizes is due to sampling error within studies and 100% indicating that all variation is due to true heterogeneity between studies [32]. Tau2 is used to assign study weights under the random-effects model [32]. Sources of heterogeneity were explored using subgroup meta-analyses, meta-regression, and sensitivity analyses. Publication bias was evaluated by the visual inspection of funnel plots and, where appropriate, by Egger test. All statistical analyses were conducted with Stata (version 16, StataCorp, Texas, USA) and Microsoft Excel (version 2019).

Results

Study characteristics

The searches retrieved 8,338 reports (Fig 1). Following deduplication, titles and abstracts of 6,221 reports were screened, and 154 were selected for full-text screening. A total of 50 full-text articles, pertaining to 36 studies, were eventually included. Details on study characteristics and associated reports are presented in Table 1. Studies were conducted in Europe (n = 28), Asia (n = 3), North America (n = 3), and Australia (n = 2). Some 23 studies were conducted in healthcare facilities, 6 were community/population based, 4 included a mix of the aforementioned 2 settings, and 3 did not report any information about their participant recruitment sites or sources. Exposures were evaluated in studies with a cohort (n = 22) or cross-sectional (n = 14) design. Interventions were studied in 2 randomized controlled trials. All studies comprised girls and boys. Overall, the minimum age at CIMT assessment was 5 hours, and the maximum age was 18 years.

Fig 1

Flow diagram showing the study selection process.

CENTRAL, Cochrane Central Register of Controlled Trials; EMBASE, Excerpta Medica database; MEDLINE, Medical Literature Analysis and Retrieval System Online.

Table 1

Characteristics of included studies.

Author, publication year	Country	Design	Setting	N* participants	Age at CIMT assessment	Mean (SD), CIMT, mm	Exposure/intervention categories at child (1), family (2), or environmental (3) levels
Observational studies
Gale and colleagues[42], 2006	UK	cohort	healthcare facility	216	9 years	right: 0.34 (0.06)	Diet and feeding practices (1), Fetal growth (1), Postnatal growth (1), Maternal weight and growth, nutrition, and physical activity (2), Socioeconomic status (3), Tobacco exposure (3)
Gale and colleagues [43], 2008	UK	cohort	healthcare facility	178	9 years	right: 0.34 (0.06)	Maternal weight and growth, nutrition, and physical activity (2)
Ayer and colleagues [44], 2009a	Australia	cohort	healthcare facility	405	8 years	left and right: 0.59 (0.07)	Fetal growth (1), Pregnancy-specific factors (2), Tobacco exposure (3)
Ayer and colleagues [45], 2009b	Australia	cohort	healthcare facility	405	8 years	left and right: 0.59 (0.06)	Diet and feeding practices (1)
Ayer and colleagues [46], 2011	Australia	cohort	healthcare facility	405	8 years	left and right: 0.59 (0.1)	Tobacco exposure (3)
Skilton and colleagues [47], 2012	Australia	cohort	healthcare facility	363	8 years	left and right: 0.59 (0.06)	Fetal growth (1), Gestational age (1)
Skilton and colleagues [48], 2013	Australia	cohort	healthcare facility	395	8 years	left or right: 0.77 (0.08)	Diet and feeding practices (1), Fetal growth (1), Gestational age (1), Postnatal growth (1), Socioeconomic status (3)
Crispi and colleagues [11], 2010	Spain	cohort	healthcare facility	200	2 to 6 years	left and right: 0.38 (0.11)	Fetal growth (1), Gestational age (1)
Rodriguez-Lopez and colleagues [49], 2016	Spain	cohort	healthcare facility	202	4 to 5 years	left and right: 0.38 (0.03)	Diet and feeding practices (1), Fetal growth (1), Gestational age, Other (1), Postnatal growth (1), Pregnancy-specific factors (2), Socioeconomic status (3), Tobacco exposure (3)
Trevisanuto and colleagues [50], 2010	Italy	cohort	unclear	38	3 to 5 years	left: 0.51 (0.08); right: 0.51 (0.07)	Fetal growth (1)
Evelein and colleagues [51], 2011	the Netherlands	cohort	community/population based	296	5 years	right: 0.39 (0.04)	Diet and feeding practices (1)
Geerts and colleagues [52], 2012	the Netherlands	cohort	community/population based	258	5 years	right: 0.38 (0.03)	Tobacco exposure (3)
Evelein and colleagues [53], 2013	the Netherlands	cohort	community/population based	323	5 years	right: 0.39 (0.69)	Fetal growth (1), Postnatal growth (1)
Pluymen and colleagues [54], 2017	the Netherlands	cohort	community/population based	413	5 years	right: 0.39 (0.04)	Diet and feeding practices (1)
Atabek and colleagues [55], 2011	Turkey	cohort	healthcare facility	55	48 to 72 hours	right: 0.31 (0.03)	Cardio-metabolic and inflammatory factors (1), Fetal growth (1), Pregnancy-specific factors (2)
Dratva and colleagues [12], 2013	USA	cohort	community/population based	670	mean age: 11 years	right: 0.57 (0.04)	Fetal growth (1), Gestational age (1), Pregnancy-specific factors (2)
Breton and colleagues [56], 2016b	USA	cohort	community/population based	240	mean age: 11 years	N/S: 0.55 (0.05)	Epigenetics (1)
Breton and colleagues [57], 2016a	USA	cohort	community/population based	392	mean age: 11 years	left: 0.56 (0.05); right: 0.57 (0.04)	Epigenetics (1), Air pollution (3)
Schubert and colleagues [58], 2013	Sweden	cohort	healthcare facility	50	3 months	N/S: 0.39 (0.12)	Gestational age (1)
Valenzuela-Alcaraz and colleagues [59], 2019	Spain	cohort	healthcare facility	160	3 years	left and right: 0.49 (0.08)	Pregnancy-specific factors (2)
Lee and colleagues [60], 2014	Germany	cohort	healthcare facility	47	10 to 14 years	left and right: 0.45 (0.03)	Gestational age (1)
Morsing and colleagues [61], 2014	Sweden	cohort	healthcare facility	93	6 to 9 years	N/S: 0.34 (0.06)	Fetal growth (1), Gestational age (1)
Stergiotou and colleagues [62], 2014	Spain	cohort	healthcare facility	201	0 to 1 week	left and right: 0.24 (0.05)	Cardio-metabolic and inflammatory factors (1), Fetal growth (1)
Gruszfeld and colleagues [63], 2015	Belgium, Germany, Italy, Poland, and Spain	cohort	unclear	383	5 years	left and right: 0.4 (0.11)	Diet and feeding practices (1), Fetal growth (1), Maternal weight and growth, nutrition, and physical activity (2), Paternal factors (2), Tobacco exposure (3)
Sebastiani and colleagues [64], 2016	Spain	cohort	healthcare facility	46	6 years	left and right: 0.35 (0.08)	Fetal growth (1)
Liu and colleagues [65], 2017	Australia	cohort	community/population based	1,477	11 to 12 years	right: 0.58 (0.05)	Socioeconomic status (3)
Mohlkert and colleagues [66], 2017	Sweden	cohort	community/population based	235	6 years	left: 0.38 (0.04); right: 0.38 (0.04)	Gestational age (1)
Tzschoppe and colleagues [67], 2017	Germany	cohort	healthcare facility	20	6 years	left and right: 0.38 (0.08)	Fetal growth (1)
Carreras-Badosa and colleagues [68], 2018	Spain	cohort	healthcare facility	66	5 to 6 years	right: 0.37 (0.03)	Maternal weight and growth, nutrition, and physical activity (2)
Chen and colleagues [69], 2019	Canada	cohort	healthcare facility	119	1 year	left and right: 0.56 (0.06)	Cardio-metabolic and inflammatory factors (1), Cardio-metabolic and inflammatory factors (2)
Prins-Van Ginkel and colleagues [70], 2019	the Netherlands	cohort	other†	221	16 years	left and right: 0.47 (0.04)	Cardio-metabolic and inflammatory factors (1)
Sebastiani and colleagues [71], 2019	Spain	cohort	healthcare facility	68	2 years	left and right: 0.3 (0.02)	Fetal growth (1)
Sundholm and colleagues [72], 2019	Finland	cohort	other†	201	mean age: 6 years	left and right: 0.36 (0.04)	Maternal weight and growth, nutrition, and physical activity (2), Pregnancy-specific factors (2)
Jouret and colleagues [13], 2011	France	cross-sectional	healthcare facility	113	mean age: 11 years	right: 0.46 (0.05)	Fetal growth (1)
Scherrer and colleagues [73], 2012	Switzerland	cross-sectional	other†	39	mean age: 11 years	left and right: 0.39 (0.03)	Pregnancy-specific factors (2)
de Arriba and colleagues [74], 2013	Spain‡	cross-sectional	unclear	269	4 to 15 years	N/S: 0.36 (0.08)	Fetal growth (1)
Maurice and colleagues [75], 2014	Canada	cross-sectional	healthcare facility	14	13 to 15 years	right: 0.48 (0.05)	Fetal growth (1)
Xu and colleagues [76], 2014	China	cross-sectional	healthcare facility	124	3 to 7 years	left: 0.39 (0.05)	Pregnancy-specific factors (2)
Putra and colleagues [77], 2015	Indonesia	cross-sectional	community/population based	285	15 to 18 years	N/S: 0.44 (0.07)	Diet and feeding practices (1)
Sodhi and colleagues [78], 2015	India	cross-sectional	healthcare facility	100	0 to 4 days	left and right: 0.42 (0.05)	Cardio-metabolic and inflammatory factors (1), Fetal growth (1)
Ciccone and colleagues [79], 2016	Italy	cross-sectional	healthcare facility	30	3 to 5 years	left: 0.51 (0.06); right: 0.52 (0.06)	Fetal growth (1), Gestational age (1)
Faienza and colleagues [80], 2016	Italy	cross-sectional	healthcare facility	52	mean age: 10 years	left and right: 0.48 (0.06)	Fetal growth (1)
Olander and colleagues [81], 2016	Finland	cross-sectional	healthcare facility	174	5 to 184 hours	left and right: 0.17 (0.03)	Cardio-metabolic and inflammatory factors (1), Fetal growth (1), Gestational age (1)
Dilli and colleagues [82], 2017	Turkey	cross-sectional	healthcare facility	60	0 to 48 hours	left and right: 0.32 (0.03)	Cardio-metabolic and inflammatory factors (1), Fetal growth (1)
Stock and colleagues [14] 2018	Austria and Italy	cross-sectional	community/population based	930	15 to 16 years	left and right: 0.38 (0.05)	Fetal growth (1), Gestational age (1)
Wilde and colleagues [83], 2018	the Netherlands	cross-sectional	other†	162	6 to 8 years	right: 0.38 (0.04)	Other (2)
Muñiz Fontán and colleagues [84], 2019	Spain	cross-sectional	healthcare facility	239	6 to 8 years	left and right: 0.43 (0.04)	Fetal growth (1)
Interventional studies
Ayer and colleagues [45], 2009b	Australia	randomized controlled trial	healthcare facility	405	8 years	left and right: 0.59 (0.06)	Diet and feeding practices (1)
Gruszfeld and colleagues [63], 2015	Belgium, Germany, Italy, Poland, and Spain	randomized controlled trial	unclear	383	5 years	left and right: 0.4 (0.11)	Diet and feeding practices (1)

Note: Studies reporting on both interventions and exposures were analyzed as randomized controlled trial for the data on interventions and as observational cohort for the data on exposures (N = 2, namely Ayer and colleagues [45], 2009b and Gruszfeld and colleagues [63], 2015).

*N = number of participants with CIMT measurements introduced in the analyses.

†Mix of healthcare facility and community based.

‡Country of the first author.

CIMT, carotid intima-media thickness; N/S, not specified; SD, standard deviation; UK, United Kingdom; USA, United States of America.

CIMT measurement methods

Image acquisition methods were relatively uniform across studies, but image analysis methods varied greatly (Tables 2 and S4). Measurements were performed on the common carotid artery (CCA) in the majority of studies (92%). However, poor or missing information on key characteristics of the CIMT measurement, namely the carotid wall, the method for edge detection and analysis of the distance between interfaces, or quality control procedures were relatively common. The segmental thickness (mean or maximum) and the type and number of sides (right, left, or the mean of both) included in the CIMT calculation were among the main sources of heterogeneity between studies. Several studies used multiple approaches for analysis; thus, up to 3 different CIMT outcomes were reported in each publication. Only 12 (33.3%) studies were at higher quality across all domains of CIMT reliability.

Table 2

CIMT measurement characteristics.

Author, publication year	N outcomes	Image acquisition			Image analysis			Reliability
		Side	Segment	Wall	Edge detection, analysis of the distance between interfaces	Cardiac cycle phase	Wall thickness	Acquisition site	Analysis	Reproducibility assessment
Gale and colleagues [41,43], 2006, 2008	1	right	CCA	far	N/S, automatic over a specific length	end diastole	N/S	higher	higher	higher
Ayer and colleagues [44–46], 2009a, 2009b, 2011;Skilton and colleagues[47], 2012	1	left and right	CCA	far	automatic/semiautomatic, automatic over a specific length	end diastole	mean	higher	higher	higher
Skilton and colleagues [85], 2013	1	left or right	CCA	far	automatic/semiautomatic, automatic over a specific length	end diastole	maximum	higher	higher	higher
Crispi and colleagues [11,85], 2010, 2012	1	left and right	CCA	far	automatic/semiautomatic, automatic over a specific length	end diastole	mean	higher	higher	higher
Rodriguez-Lopez and colleagues [49], 2016	1	left and right	CCA	far	automatic/semiautomatic, automatic over a specific length	end diastole	N/S	higher	higher	higher
Trevisanuto and colleagues [50], 2010	2	left; right	CCA	far	manual, point-to-point measurements	N/S	maximum	higher	lower	higher
Evelein and colleagues [51,53], 2011, 2013Geerts and colleagues [52], 2012	1	right	CCA	far	automatic/semiautomatic, automatic over a specific length	end diastole	N/S	higher	higher	higher
Pluymen and colleagues [54], 2017	1	right	CCA	far	automatic/semiautomatic, automatic over a specific length	end diastole	maximum	higher	higher	higher
Atabek and colleagues [55], 2011	1	right	CCA	far	manual, point-to-point measurements	N/S	maximum	higher	lower	unclear
Dratva and colleagues [12], 2013	1	right	CCA	far	automatic/semiautomatic, automatic over a specific length	end diastole	mean	higher	higher	higher
Breton and colleagues [57], 2016a	2	left; right	CCA	far	automatic/semiautomatic, automatic over a specific length	end diastole	N/S	higher	higher	higher
Breton and colleagues [56], 2016b	1	N/S	CCA	far	automatic/semiautomatic, automatic over a specific length	end diastole	N/S	higher	higher	higher
Schubert and colleagues [58], 2013	1	N/S	CCA	far	N/S, automatic over a specific length	end diastole	mean	higher	higher	higher
Valenzuela-Alcaraz and colleagues [59,86], 2013, 2019	2	left and right	CCA	far	automatic/semiautomatic, automatic over a specific length	end diastole	mean; maximum	higher	higher	higher
Lee and colleagues [60], 2014	1	left and right	CCA	far	automatic/semiautomatic, automatic over a specific length	N/S	mean	higher	higher	higher
Morsing and colleagues [61], 2014	1	N/S	CCA	N/S	manual, point-to-point measurements	end diastole	mean	unclear	lower	unclear
Stergiotou and colleagues [62], 2014	2	left and right	CCA	far	automatic/semiautomatic, automatic over a specific length	end diastole	mean; maximum	higher	higher	higher
Gruszfeld and colleagues [63], 2015	1	left and right	CCA	far	manual, point-to-point measurements	end diastole	mean	higher	lower	lower
Sebastiani and colleagues [64], 2016	1	left and right	CCA	far	N/S, N/S	end diastole	N/S	higher	unclear	higher
Liu and colleagues [65], 2017	1	right	CCA	far	automatic/semiautomatic, automatic over a specific length	end diastole	maximum	higher	higher	higher
Mohlkert and colleagues [66], 2017	2	left; right	CCA	far	automatic/semiautomatic, automatic over a specific length	end diastole	mean	higher	higher	unclear
Tzschoppe and colleagues [67], 2017	1	left and right	CCA	N/S	N/S, N/S	N/S	maximum	unclear	unclear	unclear
Carreras-Badosa and colleagues [68], 2018	1	right	CCA	N/S	N/S, N/S	N/S	N/S	unclear	unclear	higher
Chen and colleagues [69], 2019	2	left and right; left or right	other*	far	N/S, automatic over a specific length	N/S	N/S; maximum	unclear	unclear	unclear
Prins-Van Ginkel and colleagues [70], 2019	1	left and right	CCA	N/S	automatic/semiautomatic, automatic over a specific length	end distole	N/S	unclear	higher	unclear
Sebastiani and colleagues [71], 2019	1	left and right	CCA	far	N/S, N/S	end diastole	N/S	higher	unclear	higher
Sundholm and colleagues [72], 2019	1	left and right	CCA	far	manual, N/S	end diastole	N/S	higher	unclear	higher
Jouret and colleagues [13], 2011	1	right	CCA	N/S	N/S, automatic over a specific length	N/S	N/S	unclear	unclear	unclear
Scherrer and colleagues [73,87], 2012	1	left and right	CCA	N/S	automatic/semiautomatic, automatic over a specific length	end diastole	mean	unclear	higher	higher
de Arriba and colleagues [74], 2013	1	N/S	CCA	far	N/S, N/S	N/S	maximum	higher	unclear	unclear
Maurice and colleagues [75], 2014	1	right	CCA	far	automatic/semiautomatic, automatic over a specific length	end diastole	N/S	higher	higher	unclear
Xu and colleagues [76], 2014	1	left	CCA	N/S	N/S, N/S	N/S	N/S	unclear	unclear	unclear
Putra and colleagues [77], 2015	1	N/S	N/S	N/S	N/S, N/S	N/S	N/S	unclear	unclear	unclear
Sodhi and colleagues [78], 2015	2	left and right	CCA	N/S	manual, point-to-point measurements	N/S	mean; maximum	unclear	lower	lower
Ciccone and colleagues [79], 2016	2	left; right	CCA	far	N/S, N/S	end diastole	N/S	higher	unclear	unclear
Faienza and colleagues [80], 2016	1	left and right	CCA	N/S	manual, point-to-point measurements	end diastole	N/S	unclear	lower	higher
Olander and colleagues [81], 2016	1	left and right	CCA	far	manual, point-to-point measurements	end diastole	mean	higher	lower	higher
Dilli and colleagues [82], 2017	3	left; right; left and right	CCA	far	N/S, automatic over a specific length	end diastole	N/S	higher	higher	higher
Stock and colleagues [14], 2018	2	left and right; left or right	CCA	far	manual, point-to-point measurements	N/S	mean; maximum	higher	lower	unclear
Wilde and colleagues [83], 2018	1	right	CCA	far	automatic/semiautomatic, automatic over a specific length	diastole	N/S	higher	higher	higher
Muñiz Fontán and colleagues [84], 2019	2	left and right	CCA, ICA, and CB	far	automatic/semiautomatic, automatic over a specific length	end diastole	mean; maximum	higher	higher	unclear

*Insufficient information to select a specific segment or combination of segments. Place of measurement was described as follows: “the carotid bulb, internal carotid artery, and carotid bifurcation, and at least 2 cm below the bifurcation (i.e., into the common carotid artery) were imaged. CIMT was measured over a length of 1.5 cm on each carotid artery.”

CB, carotid bulb; CCA, common carotid artery; ICA, internal carotid artery; N, number; N/S, not specified.

Risk factors at the child level

Child risk factors evaluated across studies included (1) small size for gestational age and other fetal growth indicators, such as birth weight, length, head circumference; (2) prematurity; or (3) duration of breastfeeding. The effect of interventions targeting the dietary intake of proteins and polyunsaturated fatty acids after birth was also assessed.

CIMT was higher in children born small for gestational age (16 studies, 2,570 participants, pooled SMD 0.40 (95% CI: 0.15 to 0.64, p: 0.001), I2: 83%) compared with those with an appropriate size at birth (Table 3; Fig 2A). The same relationship was found when analyses were restricted to children born small and with a documented prenatal diagnosis of intrauterine growth restriction (IUGR; 8 studies, 623 participants, pooled SMD 0.35 (95% CI: −0.06 to 0.77, p: 0.097), I2: 77%). There was no clear association of prematurity with CIMT when data were pooled across all studies (7 studies, 2,024 participants, pooled SMD 0.03 (95% CI: −0.17 to 0.22, p: 0.805), I2: 50%) (Table 3; Fig 3A). However, small size for gestational age (3 studies, 461 participants, pooled SMD: 0.64 (95% CI: 0.09 to 1.19, p: 0.024), I2: 86%) (Fig 2B) and prematurity (3 studies, 736 participants, pooled SMD: 0.25 (95% CI: 0.02 to 0.48, p: 0.032), I2: 0%) (Fig 3B) were consistently associated with higher CIMT, and the associations increased in magnitude when restricted to studies of higher CIMT quality. Additionally, subgroup analyses showed that CIMT was higher in infants than in older children exposed to these risk factors (Table 4). Meta-analyses of other fetal growth indicators showed that CIMT tended to be lower with higher values of birth weight or length, but the magnitude of these relationships was low (Table 3).

Table 3

Summary of findings for each exposure type included in meta-analyses.

Category of exposure	Type of exposure	Level of comparison	N studies	N participants	Association measure	Association estimate	Heterogeneity
Child level
Fetal growth	Birth size for gestational age	small* vs appropriate	16	2,570 (848 vs 1,722)	SMD	0.40 (95% CI: 0.15 to 0.64, p: 0.001)	I2: 83%, tau2: 0.18, p: <0.001
		small with IUGR diagnosis vs appropriate	8	623 (176 vs 447)	SMD	0.35 (95% CI: −0.06 to 0.77, p: 0.097)	I2: 77%, tau2: 0.26, p: <0.001
	Birth weight	range of values	7	1,445	Correlation	−0.06 (95% CI: −0.19 to 0.08, p: 0.404)	I2: 82%, tau2: 0.02, p: <0.001
	Birth length	range of values	3	424	Correlation	−0.18 (95% CI: −0.36 to 0.00, p: 0.052)	I2: 64%, tau2: 0.02, p: 0.062
	Birth head circumference	range of values	3	329	Correlation	0.01 (95% CI: −0.40 to 0.42, p: 0.960)	I2: 93%, tau2: 0.14, p: <0.001
Gestational age	Gestational age	preterm† vs term	7	2,024 (369 vs 1,655)	SMD	0.03 (95% CI: −0.17 to 0.22, p: 0.805)	I2: 50%, tau2: 0.03, p: 0.062
Family level
Pregnancy-specific factors	Mode of conception	ART‡ vs natural	3	323 (177 vs 146)	SMD	0.78 (95% CI: −0.20 to 1.75, p: 0.120)	I2: 94%, tau2: 0.69, p: <0.001
	Maternal diabetes in pregnancy	yes§ vs no	3	658 (150 vs 508)	SMD	0.08 (95% CI: −0.16 to 0.33, p: 0.495)	I2: 17%, tau2: 0.01, p: 0.299
Environmental level
Tobacco exposure	Maternal smoking in pregnancy	yes vs no	3	909 (145 vs 734)	SMD	0.12 (95% CI: −0.06 to 0.30, p: 0.205)	I2: 0%, tau2: 0, p: 0.380

*Birth weight and/or length below the 10th percentile or 2 SDs below the mean, with or without a documented prenatal diagnosis of IUGR (fetal biometry or Doppler velocimetry).

†Below 37 weeks of gestation, where specified.

‡In vitro fertilization or intracytoplasmic sperm injection.

§Gestational diabetes, where specified.

ART, assisted reproductive technologies; CI, confidence interval; IUGR, intrauterine growth restriction; N, number; p, p-value; SD, standard deviation; SMD, standardized mean difference; vs, versus.

Fig 2

Association of small size for gestational age with CIMT in children.

SMD in CIMT between children born with a small size for gestational age (exposed) and those born with an appropriate size for gestational age (reference) in (A) all studies or (B) studies at higher CIMT reliability. Weights are from random-effects model. A positive SMD corresponds to a higher CIMT in the exposed as opposed to reference. CI, confidence interval; CIMT, carotid-intima media thickness; N, sample size; p, p-value; SMD, standardized mean difference.

Fig 3

Association of prematurity with CIMT in children.

SMD in CIMT between children born preterm (exposed) and those born at term (reference) in (A) all studies or (B) studies at higher CIMT reliability. Weights are from random-effects model. A positive SMD corresponds to a higher CIMT in the exposed as opposed to reference. CI, confidence interval; CIMT, carotid-intima media thickness; N, sample size; p, p-value; SMD, standardized mean difference.

Table 4

Subgroup meta-analyses of SMD for the association of small size for gestational age or prematurity with CIMT.

	Birth size for gestational age (small vs appropriate)				Gestational age (preterm vs term)
	N studies	SMD	I2, tau2*	p†	N studies	SMD	I2, tau2*	p†
All	16	0.40 (95% CI: 0.15 to 0.64, p: 0.001)	83%, 0.18	-	7	0.03 (95% CI: −0.17 to 0.22, p: 0.805)	50%, 0.03	-
Study design
Cohort	7	0.39 (95% CI: 0 to 0.78, p: 0.049)	79%, 0.2	0.528	5	0.1 (95% CI: −0.13 to 0.33, p: 0.383)	45%, 0.03	0.030
Cross-sectional	9	0.41 (95% CI: 0.07 to 0.74, p: 0.017)	87%, 0.21		2	−0.18 (95% CI: −0.39 to 0.04, p: 0.104)	0%, 0
Study setting
Community or population based	1	−0.09 (95% CI: −0.30 to 0.13, p: 0.440)	-	<0.001	3	0.08 (95% CI: −0.23 to 0.38, p: 0.626)	73%, 0.05	0.716
Healthcare facility	13	0.43 (95% CI: 0.14 to 0.72, p: 0.004)	83%, 0.21		4	−0.04 (95% CI: −0.34 to 0.27, p: 0.816)	32%, 0.03
N/S	2	0.60 (95% CI: 0.36 to 0.85, p: <0.001)	1%, 0.00		-	-	-
CIMT wall
Far	11	0.38 (95% CI: 0.11 to 0.64, p: 0.005)	82%, 0.15	0.756	6	0.08 (95% CI: −0.11 to 0.27, p: 0.425)	42%, 0.02	0.066
N/S	5	0.47 (95% CI: −0.17 to 1.12, p: 0.149)	88%, 0.46		1	−0.37 (95% CI: −0.81 to 0.06, p: 0.090)	-
CIMT sides
Left and right	12	0.5 (95% CI: 0.22 to 0.79, p: 0.001)	84%, 0.19	0.196	4	0.02 (95% CI: −0.2 to 0.23, p: 0.889)	41%, 0.02	0.059
Right	2	0 (95% CI: −0.35 to 0.35, p: 1.000)	0%, 0.00		1	0.35 (95% CI: 0.01 to 0.69, p: 0.045)	-
N/S	2	0.11 (95% CI: −1.00 to 1.21, p: 0.849)	95%, 0.60		2	−0.22 (95% CI: −0.6 to 0.17, p: 0.272)	19%, 0.02
CIMT edge detection method
Automatic or semiautomatic	4	0.35 (95% CI: 0.04 to 0.67, p: 0.029)	69%, 0.06	<0.001	3	0.21 (95% CI: 0.03 to 0.39, p: 0.021)	0%, 0	0.006
Manual	6	0.22 (95% CI: −0.26 to 0.71, p: 0.367)	88%, 0.31		2	−0.23 (95% CI: −0.43 to −0.03, p: 0.025)	0%, 0
N/S	6	0.64 (95% CI: 0.27 to 1.01, p: 0.001)	71%, 0.14		2	−0.02 (95% CI: −0.46 to 0.43, p: 0.949)	0%, 0
Ultrasound transducer frequency
≤12 MHz	10	0.69 (95% CI: 0.45 to 0.93, p: <0.001)	57%, 0.07	<0.001	3	0.21 (95% CI: −0.06 to 0.48, p: 0.123)	0%, 0	0.253
>12 MHz	4	0.03 (95% CI: −0.5 to 0.56, p: 0.904)	90%, 0.26		2	−0.06 (95% CI: −0.66 to 0.54, p: 0.846)	77%, 0.14
N/S	2	−0.06 (95% CI: −0.25 to 0.12, p: 0.505)	0%, 0.00		2	−0.04 (95% CI: −0.34 to 0.27, p: 0.808)	69%, 0.03
Age
0 to 1 years	4	0.63 (95% CI: −0.02 to 1.27, p: 0.057)	91%, 0.39	0.002	1	0.03 (95% CI: −0.53 to 0.59, p: 0.924)	-	0.957
2 to 16 years	12	0.31 (95% CI: 0.06 to 0.55, p: 0.014)	77%, 0.13		6	0.02 (95% CI: −0.2 to 0.25, p: 0.828)	58%, 0.04
N participants
<200	11	0.44 (95% CI: 0.05 to 0.84, p: 0.027)	83%, 0.35	0.699	4	−0.04 (95% CI: −0.34 to 0.27, p: 0.816)	32%, 0.03	0.716
≥200	5	0.34 (95% CI: 0.03 to 0.65, p: 0.034)	86%, 0.11		3	0.08 (95% CI: −0.23 to 0.38, p: 0.626)	73%, 0.05

*I2 and tau2 assess within-group heterogeneity.

†p-value assesses between-group heterogeneity.

CI, confidence interval; CIMT, carotid intima-media thickness; N, number; N/S, not specified; p, p-value; SMD, standardized mean difference.

Inconclusive evidence was found for breastfeeding. Four studies identified a lower CIMT among children with a higher duration of breastfeeding (S5 Table). Two studies identified a higher CIMT among children exclusively breastfed compared to exclusively formula fed (S5 Table). Meta-analysis was not possible due to very heterogenous comparisons across studies. Two interventional studies on dietary interventions at the child level involved a total of 787 participants. In 1 study, a diet comprising a higher n-3 to n-6 fatty acids ratio from birth to age 5 years had no effect on CIMT (SMD 0 (95% CI: −0.20 to 0.20)) [45]. In the other study, non-breastfed children on formulas with higher protein content during their first year of life had a lower CIMT (SMD −0.21 (95% CI: −0.45 to 0.04)) compared to children on lower protein formulas [63]. Scarce evidence, reported in 1 study, existed for other factors, such as cardio-metabolic and inflammatory factors, for instance, blood pressure [81], cord blood cholesterol [55], cortisol [69], and epigenetic changes, such as DNA methylation levels following exposure to air pollution [56,57] (Tables 1 and S6).

Risk factors at the family level

Family risk factors evaluated included (1) mode of conception; and (2) pregnancy complications, such as maternal diabetes. Children conceived through ART had, on average, an increased CIMT (3 studies, 323 participants, pooled SMD 0.78 (95% CI: −0.20 to 1.75, p: 0.120), I2: 94%), but the imprecision around the estimate was high (Table 3; Fig 4A). There was no clear association between maternal diabetes during pregnancy and offspring’s CIMT (3 studies, 658 participants, pooled SMD 0.08 (95% CI: −0.16 to 0.33, p: 0.495), I2: 17.1%) (Table 3; Fig 4B). Scarce evidence, reported in 1 or 2 studies, existed for other risk factors, such as preeclampsia or gestational hypertension [44,49], maternal physical activity level [42], vitamin D levels [43,68], energy and macronutrient intakes during pregnancy [42], or paternal BMI [63] (Tables 1 and S5 and S6).

Fig 4

Association of (A) ART conception, (B) maternal diabetes in pregnancy, and (C) maternal smoking in pregnancy with CIMT in children. SMD in CIMT between children: (A) conceived through ART (exposed) or naturally (reference); (B) exposed to maternal diabetes during pregnancy (exposed) or not exposed (reference); and (C) exposed to maternal smoking during pregnancy (exposed) or not exposed (reference). Weights are from random-effects model. A positive SMD corresponds to a higher CIMT in the exposed as opposed to reference. ART, assisted reproductive technologies; CI, confidence interval; CIMT, carotid-intima media thickness; N, sample size; p, p-value; SMD, standardized mean difference.

Risk factors at the environmental level

Environmental risk factors evaluated included (1) exposure to tobacco (maternal, paternal, and household); and (2) socioeconomic status (maternal, family, and neighborhood). Children exposed to maternal smoking during pregnancy had, on average, an increased CIMT, but the imprecision around the estimate was high (3 studies, 909 participants, pooled SMD 0.12 (95% CI: −0.06 to 0.30, p: 0.205), I2: 0%) (Table 3; Fig 4C). Paternal smoking during pregnancy was linked with increased CIMT in 1 study [49]. No difference in CIMT after exposure to passive smoking in the first year of life was reported in 1 study [46]. Children with a lower socioeconomic status tended to have a higher CIMT (S5 Table). Scarce evidence, reported in 1 study, existed for ambient air pollutants [57], such as nitrogen dioxide (NO2), ozone (O3), or particulate matter (PM10 and PM2.5) (Tables 1 and S6).

Discussion

Main findings

In this systematic review of 36 studies, involving 7,977 participants, multiple exposures at the child, family, or environmental levels were evaluated. Small size for gestational age had the most consistent association with increased CIMT. The magnitude of this association was higher when restricted to studies with a higher quality of CIMT measurement. The associations with conception through ART or with maternal smoking during pregnancy were not statistically significant. Evidence from interventional studies was scarce and focused exclusively on dietary interventions in the child, without showing an effect. CIMT measurement methods varied across studies, and they were frequently poorly reported, with only 33% of studies rated at higher quality across all domains of CIMT reliability.

Comparison with other studies

To the best of our knowledge, this is the first systematic review with such a multidimensional view on risk factors in the first 1,000 days of life and CIMT in children. The underlying mechanisms of CVD programming and atherosclerosis are multifactorial. Exposure to risk factors in critical periods of development may influence life course cardiovascular health trajectories through epigenetic changes [88] and cardio-metabolic factors [89] (e.g., growth patterns, blood pressure, total cholesterol, or glucose levels), but also via direct effects on the vessels structure and function. Assessing the relationship with carotid remodeling already in children was needed to shed light on the consistency and magnitude of associations, their clinical and public health importance, as well as to provide suggestions for mechanisms that need to be addressed in subsequent high-quality studies. Given that atherosclerosis is a disease affecting the whole arterial system, similar research efforts focusing on the aortic bed are currently underway (PROSPERO; CRD42019137559).

The relationship of impaired fetal growth with a higher CVD risk in later life was among the first one to be reported in the DOHaD literature. Systematic reviews and meta-analyses in adults showed that a low birth weight is associated with an increased risk of elevated blood pressure and coronary heart disease [90]. Other studies found evidence of a thicker CIMT in young adults with a small size at birth [8,9]. We found a thicker CIMT in children that were born small for gestational age, with a stronger relationship in infants than in older children. This finding is in line with previous studies and indicates that fetal growth restriction may place the individual on a higher risk trajectory since birth. It has been argued that the definition of small for gestational age may not distinguish infants that are born small due to an intrauterine pathologic process from those that have reached their genetic potential but are constitutionally small [91]. Further, fetal growth restriction was shown to be associated with cardiac remodeling and reduced arterial compliance already in utero [92]; therefore, the severity of the growth impairment might play a key role in the vascular abnormalities of these children. Following this reasoning, we restricted our analyses to children that were born small for gestational age and had a documented prenatal diagnosis of IUGR. CIMT remained thicker in IUGR children compared with those born appropriate for gestational age (Table 3). Also, the vast majority of included studies used Doppler velocimetry in the maternal or fetal vessels to diagnose IUGR; thus, these findings support the role of a well-characterized mechanism of vascular remodeling after poor fetal growth, namely the placental dysfunction [10].

The duration of gestation, maternal diabetes, or parental smoking during pregnancy are known determinants of the size at birth [1] and may also directly affect the vascular function and structure. Prematurity was shown to increase the risk of hypertension in adolescents and adults [93], probably through impaired renal development. Young adults that were born preterm were shown to have a slightly higher CIMT, noted in 3 of 4 studies of a recent systematic review and meta-analysis [94]. We did not find consistent evidence of a relationship between prematurity and higher CIMT in children. This may be because prematurity is more likely to be associated with a higher cardiovascular risk only if accompanied by fetal growth restriction or higher blood pressure in later life [8,95]. Also, it may be due to the measurement error in CIMT. In our case, it is plausible to be a combination of these mechanisms, as studies meta-analyzed involved former preterm children, with or without poor fetal growth; therefore, a distortion of this association cannot be excluded, and CIMT was higher when restricted to studies with more reliable measurements. Of note, the relationship of CIMT with prematurity was weaker than that with small for gestational age, which pleads for a greater impact of the latter, and potentially, a more pronounced impact when the 2 risk factors are present concomitantly. Interestingly, we found inverse, but very weak, relationships of CIMT with other growth indicators, such as birth length, potentially because they do not account for the duration of gestation. We also found no clear evidence of a higher CIMT after exposure to maternal diabetes during pregnancy and a consistently higher CIMT, although of very low magnitude, in children exposed to maternal smoking during pregnancy. It is known that gestational diabetes is associated with macrosomia in the offspring, whereas maternal smoking during pregnancy is associated with poor fetal growth [1]. Our data indicate that CIMT was consistently increased in children with abnormal growth, either large or small. The effect of maternal gestational diabetes or smoking may thus be mainly indirect, through the size at birth. However, adequately powered studies that formally assess the mediator effect of size at birth on these relationships are needed to shed light on this hypothesis [96]. Finally, our finding of a higher CIMT in children conceived through ART draws attention on the highly sensitive periconceptional period, with potentially new underlying mechanisms that need to be disentangled.

Heterogeneity and poorly reported or missing information on key characteristics of the CIMT measurement methods were among the main issues that we encountered when analyzing the evidence presented herein. Similar conclusions were drawn in systematic reviews on CIMT in adults that carefully considered measurement methods when interpreting their results [6,97]. In our case, the sensitivity analyses performed for small size for gestational age and prematurity showed that the quality of CIMT measurement methods had a significant impact on the results. Additionally, CIMT values may differ with the edge detection method and the specificities of the ultrasound equipment used, such as the transducer frequency. We could notice large decreases in the magnitude of association estimates across studies using higher transducer frequencies. Our finding is in line with other evidence showing that conventional high frequencies (≤12 MHz) may have insufficient ultrasound resolution and result in an overestimation of the true thickness in children under 12 years of age [18,19]. Novel techniques that measure CIMT with very high-resolution ultrasound (25 to 55 MHz) seem to be more precise in infants and young children and at very early stages of atherosclerosis [98,99].

Strengths and limitations

The strength of our systematic review lies primarily in the fact that it was performed according to a detailed protocol that was prospectively registered on PROSPERO and published in full in a peer-reviewed journal. Secondly, we collected extensive data on the CIMT measurement characteristics, which permitted interpretation of results in the light of the quality of measurement. Thirdly, we included both observational and interventional studies reporting on exposures or interventions at the child, family, or environmental levels. This helped us have a comprehensive overview on the current status of the evidence and highlight needed areas of future research.

The main limitation is that the body of evidence for this systematic review is largely observational, which is of low certainty by design and restricts the ability to draw conclusions about causality. Additionally, the degree of confidence in results is limited by the high between-study heterogeneity. Studies varied on several aspects, such as design, exposure metric, age at CIMT assessment, or CIMT image analysis. The variability in age at CIMT assessment was large, from 0 to 18 years. Our subgroup analysis was mainly based on the age range reported in each study and, thus, could only distinguish the infants from the other pediatric age groups combined. Meta-analyses were performed mainly using estimates from bivariate associations. Data transformation and simplification were needed, and associations were estimated from various statistics so that relationships could be summarized in a consistent way across studies. Therefore, differences in body size between exposed and nonexposed could not be taken into account in the analyses, confounding bias cannot be excluded, and our association estimates may be distorted, probably underestimated, due to measurement error in the CIMT. Furthermore, CIMT was evaluated in youth with tiny, submillimetric vessel structures, and many studies had relatively small samples and few individuals exposed to the risk factors of interest, which resulted in large CIs for the estimates. The low number of studies meta-analyzed often impeded conduction of subgroup analyses to explore the sources of heterogeneity predefined in the protocol. However, to increase our confidence in the overall estimates, whenever possible, we conducted analyses restricted to studies at higher quality for all CIMT reliability domains. Other limitations are represented by exclusion of studies not reported in English or French [100-102], publication bias, and selective exposure or analysis reporting. Usable data for the majority of exposures or interventions were reported by a low to a very low number of studies. However, for small size for gestational age, the risk factor most studied, the funnel plot inspection, and Egger test were not indicative of publication bias (S2 Fig).

Implications for research and practice

We highlight the need of a standardized ultrasound protocol for measuring CIMT in children. Due to the paucity of interventional studies and the limited quality of CIMT measurements, further high-quality studies are needed to justify the use of CIMT for child CVD risk assessment in clinical practice. However, promotion of a healthy lifestyle is important at any age, and screening of postnatal CVD factors, such as hypertension or obesity, targeted at children exposed to risk factors in the first 1,000 days of life, may be warranted.

Further, our study indicated that children that were small for gestational age had a higher CIMT in early life. Prevention of small for gestational age could have a big impact from a public health perspective. Its prevalence varies among populations and rises with decreasing gestational age and availability of resources, reaching about 20% in low and middle-income countries [103]. Moreover, modifiable factors, including gestational smoking or poor maternal nutrition, account for a large share of small for gestational age infants [104]. In fact, smoking cessation interventions during pregnancy were shown to result in a 17% reduction in low birth weight cases [105]. Public health strategies to reduce smoking and improve the nutritional status in women of reproductive age should be advocated and may improve the cardiovascular health too [105,106].

Conclusions

In our meta-analyses, we found several risk factors in the first 1,000 days of life that may be associated with increased CIMT in infants, children, and adolescents. Poor fetal growth had the most consistent and strongest association with increased CIMT. The associations with ART conception or with maternal smoking during pregnancy were not statistically significant. From a DOHaD perspective, exposure to adverse experiences in early life determines adaptative responses in the organs’ structure and function with lifelong effects on cardiovascular health and disease, ranging from elevated blood pressure, early vascular aging and increased CIMT, and premature CVD morbidity and mortality [1,107]. Therefore, assessing CIMT early in life is important to distinguish between changes occurring in childhood and adulthood, with the goal of better characterizing lifetime risk trajectories. Provided this hypothesis will stand the test of time and effective intervention, the interplay of risk factors in the first 1,000 days of life and vascular remodeling will offer a great opportunity for primordial prevention of CVD [108].

Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

(PDF)

Click here for additional data file.

Assessment of study quality for each exposure type included in meta-analyses.

(PDF)

Click here for additional data file.

Assessment of small-study effects, including publication bias, for each exposure type included in meta-analyses.

(PDF)

Click here for additional data file.

Random-effects meta-regression examining the influence of sample size on the association of small size for gestational age with CIMT.

(PDF)

Click here for additional data file.

Strategies for systematic searches.

(PDF)

Click here for additional data file.

Strategies for supplementary searches.

(PDF)

Click here for additional data file.

Criteria for CIMT quality assessment.

(PDF)

Click here for additional data file.

CIMT equipment and operator.

(PDF)

Click here for additional data file.

Association of CIMT with main exposures or interventions types in the first 1,000 days of life.

(PDF)

Click here for additional data file.

Other exposure types in the first 1,000 days of life by level and category of exposure (reported in a single study).

(PDF)

Click here for additional data file.

Criteria for study quality assessment.

(PDF)

Click here for additional data file.

Assessment of study quality for each intervention type in interventional studies.

(PDF)

Click here for additional data file.

Assessment of study quality for observational studies included in meta-analyses.

(PDF)

Click here for additional data file.

11 May 2020

Dear Dr Epure,

Thank you for submitting your manuscript entitled "First 1000 days risk factors for carotid intima-media thickness in childhood: a systematic review with meta-analyses" for consideration by PLOS Medicine.

Your manuscript has now been evaluated by the PLOS Medicine editorial staff and I am writing to let you know that we would like to send your submission out for external peer review.

However, before we can send your manuscript to reviewers, we need you to complete your submission by providing the metadata that is required for full assessment. To this end, please login to Editorial Manager where you will find the paper in the 'Submissions Needing Revisions' folder on your homepage. Please click 'Revise Submission' from the Action Links and complete all additional questions in the submission questionnaire.

Please re-submit your manuscript within two working days, i.e. by .

Login to Editorial Manager here: https://www.editorialmanager.com/pmedicine

Once your full submission is complete, your paper will undergo a series of checks in preparation for peer review. Once your manuscript has passed all checks it will be sent out for review.

Feel free to email us at plosmedicine@plos.org if you have any queries relating to your submission.

Kind regards,

Artur Arikainen,

Associate Editor

PLOS Medicine

20 Jun 2020

Dear Dr. Epure,

Thank you very much for submitting your manuscript "First 1000 days risk factors for carotid intima-media thickness in childhood: a systematic review with meta-analyses" (PMEDICINE-D-20-01898R1) for consideration at PLOS Medicine.

Your paper was evaluated by a senior editor and discussed among all the editors here. It was also discussed with an academic editor with relevant expertise, and sent to independent reviewers, including a statistical reviewer. The reviews are appended at the bottom of this email and any accompanying reviewer attachments can be seen via the link below:

[LINK]

In light of these reviews, I am afraid that we will not be able to accept the manuscript for publication in the journal in its current form, but we would like to consider a revised version that addresses the reviewers' and editors' comments. Obviously we cannot make any decision about publication until we have seen the revised manuscript and your response, and we plan to seek re-review by one or more of the reviewers.

In revising the manuscript for further consideration, your revisions should address the specific points made by each reviewer and the editors. Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments, the changes you have made in the manuscript, and include either an excerpt of the revised text or the location (eg: page and line number) where each change can be found. Please submit a clean version of the paper as the main article file; a version with changes marked should be uploaded as a marked up manuscript.

In addition, we request that you upload any figures associated with your paper as individual TIF or EPS files with 300dpi resolution at resubmission; please read our figure guidelines for more information on our requirements: http://journals.plos.org/plosmedicine/s/figures. While revising your submission, please upload your figure files to the PACE digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at PLOSMedicine@plos.org.

We expect to receive your revised manuscript by Jul 13 2020 11:59PM. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

We ask every co-author listed on the manuscript to fill in a contributing author statement, making sure to declare all competing interests. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. If new competing interests are declared later in the revision process, this may also hold up the submission. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT. You can see our competing interests policy here: http://journals.plos.org/plosmedicine/s/competing-interests.

Please use the following link to submit the revised manuscript:

https://www.editorialmanager.com/pmedicine/

Your article can be found in the "Submissions Needing Revision" folder.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see http://journals.plos.org/plosmedicine/s/submission-guidelines#loc-methods.

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

We look forward to receiving your revised manuscript.

Sincerely,

Emma Veitch, PhD

PLOS Medicine

On behalf of Clare Stone, PhD, Acting Chief Editor,

PLOS Medicine

plosmedicine.org

-----------------------------------------------------------

Requests from the editors:

*In the last sentence of the Abstract Methods and Findings section, please add a brief note commenting on any key limitation(s) of the study's methodology.

-----------------------------------------------------------

Comments from the reviewers:

Reviewer #1: See attachment

Michael Dewey

-----------------------------------------------------------

Reviewer #2: This is a systematic review and meta-analysis titled "First 1000 days risk factors for carotid intima media thickness in childhood: a systematic review and meta-analysis" assessing different DOHaD variables associations with CIMT mainly during childhood (<10y). The protocol has been published previously (Epure AM, Leyvraz M,Mivelaz Y, et al. Risk factors and determinants of carotid intimamedia thickness in children: protocol for a systematic review and meta-analysis. BMJ Open 2018;0:e019644). The focus from the original protocol (pediatric age group 0-18; postnatal child CVD risk factors obesity, hypertension, smoking, disease and treatments) has changed in the current review manuscript to focus on younger pediatric age and the role of early life DOHaD predictors.

The title uses the term "risk factors for CIMT in childhood", but as there is currently no evidence linking CIMT in infancy/childhood to CVD disease later in life, I would certainly propose a more comprehensive vascular phenotype (not only CIMT) in children. Furthermore, if the main focus and aim is to explore relations between DOHaD and vascular changes in children, one would perhaps expect a more comprehensive set of vascular variables included (other sites of intima-media thickness, different measures of arterial function such as regional and local stiffness, blood pressure and heart rate, perhaps also endothelial function for older children) and a more comprehensive set of covariates on postnatal anthropometrics, growth, body composition, and different traditional aspects of CVR (diet, activity etc.). The present manuscript is, thus, limited to child CIMT only, and mainly prenatal, perinatal factors and early life DOHaD factors. In addition, the high heterogeneity and apparent problems in the reported CIMT measurement methodology in infancy/childhood makes the analysis with DOHaD variables and conclusions thereof challenging and uncertain indeed.

The CIMT methodology issues warranted the investigators to critically evaluate and explore explanations for this variance, as mentioned in the published protocol as well. This is warmly endorsed by this reviewer, and forms a major strength of this manuscript. The manuscript provides in-depth analyses to shed light on factors related with this heterogeneity and how this might influence the DOHaD analyses. The current guidelines on vascular assessments and reports of population and healthy CIMT references in the pediatric age group have been made to address mainly adolescent populations with CVR and different disease states and treatments affecting the vasculature. This review is welcome in attempting to provide a summary of the very diverse current literature of the ultrasound CIMT during infancy and childhood.

The manuscript reads well, many tables and figures are included and the authors have made good attempts in trying to address the objectives. The review also seems to include the most relevant articles in the field. There are, however, some concerns related with method presentation and the analyses and the conclusions that warrant commenting and modifications to the presentation, as outlined in the comments to the authors.

1. Title and abstract. See general commenting and suggestion above for revising wording in title. Mean difference 0.40. Please provide information on adjustments for body size, HC size, gestational age, child age made in this main result, as appropriate. The first sentence in the conclusion: "Several risk factors in the first 1000 days of life may be associated with increased CIMT during childhood", comments on ART and smoking (tendencies only) also seem poorly substantiated. What intervention?

2. Introduction. Please see general commenting above. The intro could focus more on the objective, i.e. 1) relations with DOHad and 2) challenges with ultrasound derived pediatric IMT in order to provide a rationale for the study. The second paragraph could be more focused on the main objective. What is known about the role of early life factors, early child vascular phenotype assessment, and later CVD, and why should the vascular phenotype be assessed at this stage? The last sentence relating to the guidelines could emphasize the difference in age group addressed in the guidelines compared with the focus of the present manuscript.

3. Methods. Methods/Eligibility criteria/Participants. Definition of child population. "Apparently healthy children" vs children with different disease states? On the other hand "Both apparently healthy children and individuals with clinical conditions will be included." (information from PROSPERO-registry). And furthermore in published study protocol: "We will include studies in children from birth up to 18 years. Studies with both children and adults will be included if the data for children can be extracted separately. Both apparently healthy children and subjects with clinical conditions will be included" The manuscript, however, states, "Studies in children with postnatal conditions that confer an increased risk for CVD, e.g. hypertension or diabetes, were excluded. Additionally, children with rare conditions (e.g. congenital heart disease) were not included as they are not representative of the general population". Both these exclusion criteria seem irrational as future cardiovascular disease is of major interest in and early development and interventions performed could well alter both. In addition, the study includes analyses related with maternal gestational risk (egg gestational diabetes), premature birth, abnormal fetal growth, that are per se risk factors for later disease development in offspring. What is the rationale for these exclusion statements? Figure 1 shows that adult populations were excluded. Study population could then be better defined in methods and mentioned in title/abstract as well to avoid confusion.

4. Childhood is per se not defined as 0-18y, but as the age span ranging from birth to adolescence. On the other hand although the study states that they included children aged 0-18. Table 3 shows that a majority of the study subjects included belong to the younger age groups (36/50 studies report age less than 9y), so the focus is on childhood. However, the use of the term childhood should be consistent with common definitions. Please check this the use of this terminology throughout manuscript including title, abstract, manuscript text, figure and tables, supplements.

5. Comparators. No eligibility criteria is not provided although this seems key in the comparison. Published study protocol states: "Studies assessing specific clinical conditions in relationship with CIMT will be included provided they use a control group without the clinical condition of interest". Age, body size, puberty, CVR-profile including BP, race etc. are well known important key predictors of CIMT in children. Do the authors mean to say that no requirements for comparator groups needs to be specified in order to assess differences in CIMT in relation to DOHaD variables? Differences in vascular dimensions reported in the different studies should be adjusted for differences in key covariates. This is not clearly presented in Data Analysis. Please do elaborate and revise accordingly.

6. Exposures/interventions. "For observational studies, we considered for inclusion prenatal and postnatal exposures, at the individual, familial, or environmental level, which reflected the developmental milieu of the child and occurred between conception and age 24 months" This relates to the DOHaD Figure 1 states that 20 articles were excluded due to missing exposure or intervention of interest. Please specify in more clear terms what prenatal and postnatal exposures were considered here? Birth weight, gestational age at birth, maternal disease, gestational disease or complication, infections, medications etc. Examples could be provided.

7. Disagreement between reviewers were solved by discussion. Could this be presented with numbers somehow?

8. Data extraction. Quality of study assessment. Supplemental Table 3. Criteria for study quality and study reliability tool criteria and algorithm of judgement seems arbitrary and poorly justified. Why is at least 2 images/frames needed in addition to manual tracings to indicate improved reliability? How does assessing more frames improve CIMT measurement reliability? No data regarding this is presented. Intra and interobserver variabilities are commonly performed in standardized vascular laboratories as quality assurance measures, and may not always be reported in articles methods section. The non-reporting of these (or only referring to a reference) is then by no means evidence of a lower technical operator performance or lower measurement reproducibility. There other more important factors related with this that could be included: standardized imaging and measurement protocols, operator experience and skill, image quality, ultrasound equipment and transducer frequency, imaging feasibility (child preparation with measurements performed at rest; sedation as appropriate), site of carotid intima media thickness, just to mention a few. Reading Box 2 in the published study protocol in BMJ Open suggests a different approach for this assessment. Please justify the choice of different factors included in this manuscript Supplemental table 3.

Moreover, "The quality of each study was evaluated using an adjusted version of the Newcastle-Ottawa Scale for observational studies and Cochrane's collaboration risk of bias tool for experimental studies." How this was done remain largely unclear. This also applies to the information provided in Supplemental Table S7, Supplementary Figure 1S. How was study quality data generated? What defines a low quality study and a high quality study addressed in subanalyses and mentioned in the discussion? The study quality assessment should be clearly presented and outlined.

9. Data analyses. The SMD measure adjusts for methods differences in scale and precision. However, there may be differences between the exposure and comparator that are explained by covariates such as body size (weight, lean body mass or head circumference) that is not accounted for in this meta-analysis metric. The main stated results relates to being born small for gestational age. How differences in body size between exposure and comparator was (reference) groups taken into account in the generation of the SMD and this conclusion? How does measurement precision affect SMD (influence of the SD on the SMD metric?) and how does this influence the reported outcome SMD and conclusions thereof?

10. The reported I2 50-83% (child level) in conjunction with the SMD in Table 3 indicate a high heterogeneity and inconsistency between the studies included in the meta-analysis (similarly high in Table 4 subgroup analyses). Could the high heterogeneity be explained by methodological differences between the studies? Could this be explored in more detail. Moreover, how does the high heterogeneity between the studies impact on the conclusion of a higher CIMT among children born small for gestational age? Please explain the use of I2 and tau2 metrics of heterogeneity in Data Analysis and, in particular, include interpretations on these as well. In addition, please elaborate on this in the manuscript discussion.

11. Results. Table 4, footnote. "P-value reported for between group heterogeneity". What groups? Is this p-value in conjunction with I2 and tau2 metrics (similar to table 3) between small vs appropriate birth size (weight?) or is it some kind of comparison of the heterogeneity related with subgroup factor such as transducer frequency or age? Please clarify.

12. Table 4 shows that the higher CIMT in children born small for gestational age (SMD) disappears when assessing studies using higher ultrasound frequencies. This finding should be highlighted, as the measurement accuracy is likely to be much better younger populations with thinner CIMTs.

13. Exposure to maternal smoking had a consistent association with higher CIMT (SMD 0.12, CI95% -0.06-0.30), is this statistically significant? Is this mediated by the effect of gestational smoking on birth weight?

14. Figure 2. Small size for gestational age. What was the criteria for SGA applied? What was the age of the population at CIMT assessment? Are these newborns or children of higher age? What is the effect of postnatal growth? Are the exposure and reference groups comparable with respect to age and body size?

15. What is number in brackets after first author name in Supplemental Table S4?

16. Supplemental Table 5. Birth size (weight). How comparable are the studies concerning birth size? Has the mean difference been adjusted for body size? Covariates? Is the CIMT in accordance with body size (lean body mass or HC size)?

17. Discussion. Discussion. … followed by ART and maternal smoking during pregnancy consistent effect on CIMT? Was the effect of ART statistically significant? Reported as a tendency in results (SMD 0.78, CI95% -0.20 to 1.75).

18. Please address CIMT associations with race and sex, and body composition (e.g. adiposity)?

19. What does the SMD 0.40 correspond to in mm? What is the clinical relevance of this from a CVR perspective?

20. Overall, the discussion is too long and speculative. Please focus on discussing in relation to objective. The recommendations for CIMT measurements presented in the conclusions are not warranted by the results, they are too strong and very subjective (e.g. no data provided suggesting superiority of border detection software in pediatric CIMT or the analysis of 2 image frames using calipers), the line of reasoning is missing here.

Minor comment from reviewer related with CIMT methodology

The authors might be interested in studies addressing non-invasive ultrasound-derived pediatric (and adult) arterial layer method validation using histology (Sarkola et al Atherosclerosis. 2010 Oct;212(2):516-23; Sarkola et al Atherosclerosis. 2012 Sep;224(1):102-7; Sundholm Atherosclerosis. 2015 Apr;239(2):523-7; Sundholm et al Ultrasound Med Biol. 2019 Aug;45(8):2010-2018). These studies illustrate important caveats related with the ultrasound CIMT measurement during infancy/childhood. In brief, these are related with 1) the true histological thinness of the arterial wall layers during infancy/childhood exemplified in the present review by references reporting ultrasound IMTs in infants (e.g.Trevisanuto et al 2010, Schubert et al 2013, Sodhi et al 2015, Chen et al 2019) similar to commonly reported in young adult age (e.g. Engelen et al https://doi.org/10.1093/eurheartj/ehs380), and references reporting IMTs childhood (e.g. multiple studies by Ayer and Skilton, Ciccone et al 2016, Valenzuela-Alcaraz et al 2019) similar to commonly reported in older adult age which is biologically impossible, 2) frequency dependence of ultrasound axial resolution (i.e. lower frequency providing higher IMT) and ultrasound frequency effect on IMT measurement accuracy (conventional frequency axial resolution limit >0.25 mm higher than infancy/childhood IMT <0.25). I hope that these could guide the authors in their revision.

Review provided by Dr Taisto Sarkola.

-----------------------------------------------------------

Reviewer #3: This meta-analysis summarizes the relevance of perinatal life factors for carotid-intima thickness in children and adolescents. In view of the fact that the first 1000 days are regarded as a potentially critical period for shaping later cardiometabolic risk this is a topic of substantial public health relevance. In addition, evidence from several studies suggests that associations of factors acting in this window of opportunity may already impact carotid intima-media thickness, i.e. an established marker of subclinical atherosclerosis.

Nonetheless, a number of issues arise with this manuscript

Major

* Main results. Only one of the three "main" results (born small for gestational age) is statistically significant. This is not correctly presented and discussed by the authors.

* Aim 2:

o The necessity for cIMT method appraisal remains elusive (I do not question them, but they should be explained in the introduction). Why should the between-study heterogeneity matter? What are the main problems associated with this and could this affect the associations under investigation?

o The authors provide very little information on their cIMT measurement quality assessment.

o Generally, the term "reliability" appears incorrect. Many of the aspects assessed could also affect the validity of the measurements.

* Intervention studies:

o These can only be considered as such if the primary endpoint of the study was cIMT. This was only the case for the study of Ayer et al. (Ref. 54) The second study had body composition as a major outcome. Hence, the post-hoc analysis of this study for an association with later cIMT should be regarded as an observational cohort study.

o The study of Ayer et al. (Ref. 54) is mentioned both as a cohort and an intervention study. Please, clarify.

* Exposures

o it remains elusive how the authors arrived at their set of exposures. Were there any a priori criteria (they were not part of the search strategy)? Were all exposures analysed in a relevant publication included?

o Many exposures could have multiple indicators, e.g. duration of breastfeeding would certainly be reported more specifically (exclusive vs full vs any), yet it is not clear how the authors attempted to address/combine these.

o A Table on all early life risk factors and their "definitions" would be helpful, e.g. clarify "diet and feeding practices"

o Some of the exposures appear questionable: epigenetic changes, cortisol and cardiometabolic factors in childhood are presumably markers of mechanisms by which perinatal risk factors could affect the cardiometabolic outcome cIMT rather than exposures

* Age at outcome

o One of the main limitations is the large variability in the age at outcome (0-18 years). This is hardly mentioned and the sensitivity analysis only distinguishes 0-1 from 2-18 years.

* Search

o End date March 2019 - this is outdated. Please, update.

o Were searches performed in duplicate?

* Risk of bias / Grading of evidence

o The authors use the Newcastle Ottawa scale for observational studies, but do not report findings per study. They only provide a summary figure by exposure, i.e. the reader cannot confirm their judgement

o Some aspects in this figure would need explanation (i.e. "same method of ascertainment or "comparability")

o The Assessment of study quality for interventions studies given in Supplementary Table 7 is not referred to in the text and the table is not intelligible. Does "high" indicate high quality or high risk of bias?

o The quality of meta-evidence needs to be graded

* Discussion of the findings

o An appraisal of the "effect sizes" and their public health relevance is missing

o Heterogeneity due to the age-range at outcome measurement (only first year vs 2-18 yrs) longitudinal vs cross-sectional needs to be discussed

o Lack in focus: The authors discuss a number of potential mechanisms driving their observations, even for small sensitivity analyses for which data are not shown (and are perhaps not significant). In view of this reviewer, this should be focussing on the main findings of the meta-analysis only.

Minor

* Abstract: CIMT methods cannot be appraised with a meta-analysis

* Number for observational studies should be given separately for cross-sectional and longitudinal studies

* Why was this study done: "It is unclear whether …" is not a convincing argument to perform such a study. Preliminary (mechanistic) evidence suggesting that this could be relevant should be given.

* Funding information should not be mentioned in methods and findings

* What does the age range in the abstract refer to? Age at cIMT measurement?

* The authors should already introduce the concept of perinatal programming and the mechanisms discussed in this context in the background. Why should we expect that early life factors could already affect cIMT in childhood/adolescence?

* The conclusion on CIMT method "cIMT measurement protocols are heterogeneous and often poorly reported" is vague. What does this refer to? Validity or reliability of the method? Lack of data on intra- and inter-observer variability? Potential for investigator-associated bias (unstandardized measurement procedure)?

* Purpose of a meta-analysis. In the discussion the authors state that their review is needed to better understand DOHaD mechanisms. In the view of this reviewer a meta-analysis based primarily on observational studies can only provide suggestions for potential mechanisms which need to be addressed in other subsequent studies. Instead, quantification of associations, evidence on public health relevance etc. are most likely more important contributions of a meta-analysis.

* Experimental studies should be termed intervention studies

* Participants cannot be recruited pre-birth, whereas their mothers can

* Please check language, e.g. lines 148-150, does not appear to be a sentence

* Results: When referring to specific studies the reference should be given

* Line 293: this relation is no longer significant!

* Strength: performing a meta-analysis according to a protocol is not a strength.

* How was the association between study quality (RoB) and method quality?

* Strength of relationships (e.g. line 301) are an interpretation of the results, not a result per se

* Fixed-effects meta-analysis was used for data simplification, yet random-effects meta-analysis for the main analysis. Please explain.

* The discussion of the breastfeeding non-finding is misleading and should be shortened. The findings in this meta-analysis do not permit any public health statements relating to breastfeeding.

-----------------------------------------------------------

Any attachments provided with reviews can be seen via the following link:

[LINK]

Submitted filename: epure.pdf

Click here for additional data file.

10 Aug 2020

Submitted filename: response_to_editors_and_reviewers.docx

Click here for additional data file.

17 Sep 2020

Dear Dr. Epure,

Thank you very much for re-submitting your manuscript "First 1,000 days risk factors for carotid intima-media thickness in infants, children, and adolescents: a systematic review with meta-analyses" (PMEDICINE-D-20-01898R2) for review by PLOS Medicine.

I have discussed the paper with my colleagues and the academic editor and it was also seen again by three reviewers. I am pleased to say that provided the remaining editorial and production issues are dealt with we are planning to accept the paper for publication in the journal.

The remaining issues that need to be addressed are listed at the end of this email. Any accompanying reviewer attachments can be seen via the link below. Please take these into account before resubmitting your manuscript:

[LINK]

Our publications team (plosmedicine@plos.org) will be in touch shortly about the production requirements for your paper, and the link and deadline for resubmission. DO NOT RESUBMIT BEFORE YOU'VE RECEIVED THE PRODUCTION REQUIREMENTS.

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

In revising the manuscript for further consideration here, please ensure you address the specific points made by each reviewer and the editors. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments and the changes you have made in the manuscript. Please submit a clean version of the paper as the main article file. A version with changes marked must also be uploaded as a marked up manuscript file.

Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. If you haven't already, we ask that you provide a short, non-technical Author Summary of your research to make findings accessible to a wide audience that includes both scientists and non-scientists. The Author Summary should immediately follow the Abstract in your revised manuscript. This text is subject to editorial change and should be distinct from the scientific abstract.

We expect to receive your revised manuscript within 1 week. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

We ask every co-author listed on the manuscript to fill in a contributing author statement. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT.

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

If you have any questions in the meantime, please contact me or the journal staff on plosmedicine@plos.org.

We look forward to receiving the revised manuscript by Sep 24 2020 11:59PM.

Sincerely,

Artur Arikainen,

Associate Editor

PLOS Medicine

plosmedicine.org

------------------------------------------------------------

Requests from Editors:

1. Please address the reviewers’ final comments below.

2. Title: Please amend to: “Risk factors for carotid intima-media thickness during first 1,000 days of life in infants, children, and adolescents: a systematic review with meta-analyses”

3. Abstract:

a. Line 33: Please include the search date ranges.

b. Please include the measure of study quality/bias and the overall quality or risk of bias of the included studies.

c. Please include a list of the risk factors/exposures tested.

d. Line 44: For ART and maternal smoking, we ask that you rephrase this to something like: "...non-significantly increased CIMT…" for these associations. Similarly, it may be necessary to rephrase lines 52-54 and 71-74, which currently refer to associations with several risk factors:

i. “In our meta-analyses, we found that several risk factors in the first 1,000 days of life may be associated with increased CIMT during childhood. Poor fetal growth had the most consistent relationship with increased CIMT, while for ART conception and smoking during pregnancy the increase was not significant.”

ii. “Risk factors in the first 1,000 days of life relating to poor fetal growth are associated with increased CIMT in infants, children, and adolescents. Conception through assisted reproductive technologies and maternal smoking during pregnancy also increased CIMT, but non-significantly.”

e. Please quantify all results with p values as well as 95% CI. Please include data for the following: “When restricted to studies of higher quality of CIMT measurement, the relationships were stronger.”

f. Please mention the primary funding source.

4. Author Summary:

a. Lines 65-66: Please clarify for a lay audience the significance of “vascular morphology” and “increased carotid intima-media thickness”. Perhaps also define “vascular”.

b. Line 69: Please clarify “…systematic review of published literature…”

c. Line 70: CIMT is not defined earlier.

d. Line 83: Please clarify or define “primordial” in this context.

5. Line 120: We recommend removing this subheading.

6. When completing the PRISMA checklist, please use section and paragraph numbers, rather than page numbers.

7. Lines 131-132: Please cite the included PRISMA Checklist as “S1 Checklist” here.

8. Please remove spaces from within citation callouts, eg. (line 535) “…health too [103,104].”

------------------

Comments from Reviewers:

Reviewer #1: The authors have addressed my points.

I still find it regrettable that studies published in languages other than English or French were deliberately excluded and I suggest that the references to the three that were found should be included somewhere for the benefit of future researchers who can understand them.

Michael Dewey

Reviewer #2: To the editor and authors,

For this reviewer the comments have been very well addressed in the responses and the manuscript overall. The authors have made an extraordinary effort in the systematic analyses and meta-analyses of this review. In this reviewers view, the responses to the comments provide essential information to the manuscript and, therefore, the publication of this document is warmly recommended in conjunction with the manuscript.

This reviewer has only on additional comment to provide for the revised version. This relates to the main conclusion of CIMT being increased due to poor fetal growth (or born small for gestational age) with analyses provided in Table 4, Table 5S, Results lines 315-327 and Fig 2.

Differences in CIMT attributed to differences in body size/anthropometrics has not adequately been taken into account in the analyses in response the reviewer comments 9 and 16. The following responses are provided by the authors in relation to this issues: Response 9 (differences attributed to body size/anthropometrics ignored in the response), and Response 16 "…CIMT was measured in mm, thus, it was not normalized according to body size. "

It is well known that the most important determinant of cardiovascular dimensions in the growing fetus and child is body size and anthropometry. Differences related with e.g. brain sparing in the setting of IUGR and differences in e.g. head circumference (HC) has not been adequately addressed in this review (although partly for HC in Table 5S). This review does not then answer the question: Is CIMT proportional to head/brain anthropometrics in poor fetal growth/SGA?

The lack of analyses accounting for differences in HC/brain in assessing CIMT in relation to poor fetal growth/SGA is then a limitation (may not be adequately reported in the original references too precluding these analyses) in the analyses and this should briefly be acknowledged in the Discussion/limitations section.

Taisto Sarkola

Reviewer #3: The authors have responded to my comments in detail

Any attachments provided with reviews can be seen via the following link:

[LINK]

7 Oct 2020

Submitted filename: response_to_editors_and_reviewers_corrections.docx

Click here for additional data file.

19 Oct 2020

Dear Dr Epure,

On behalf of my colleagues and the academic editor, Dr. Sanjay Basu, I am delighted to inform you that your manuscript entitled "Risk factors during first 1,000 days of life for carotid intima-media thickness in infants, children, and adolescents: a systematic review with meta-analyses" (PMEDICINE-D-20-01898R3) has been accepted for publication in PLOS Medicine.

PRODUCTION PROCESS

Before publication you will see the copyedited word document (within 5 business days) and a PDF proof shortly after that. The copyeditor will be in touch shortly before sending you the copyedited Word document. We will make some revisions at copyediting stage to conform to our general style, and for clarification. When you receive this version you should check and revise it very carefully, including figures, tables, references, and supporting information, because corrections at the next stage (proofs) will be strictly limited to (1) errors in author names or affiliations, (2) errors of scientific fact that would cause misunderstandings to readers, and (3) printer's (introduced) errors. Please return the copyedited file within 2 business days in order to ensure timely delivery of the PDF proof.

If you are likely to be away when either this document or the proof is sent, please ensure we have contact information of a second person, as we will need you to respond quickly at each point. Given the disruptions resulting from the ongoing COVID-19 pandemic, there may be delays in the production process. We apologise in advance for any inconvenience caused and will do our best to minimize impact as far as possible.

PRESS

A selection of our articles each week are press released by the journal. You will be contacted nearer the time if we are press releasing your article in order to approve the content and check the contact information for journalists is correct. If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact.

PROFILE INFORMATION

Now that your manuscript has been accepted, please log into EM and update your profile. Go to https://www.editorialmanager.com/pmedicine, log in, and click on the "Update My Information" link at the top of the page. Please update your user information to ensure an efficient production and billing process.

Thank you again for submitting the manuscript to PLOS Medicine. We look forward to publishing it.

Best wishes,

Artur Arikainen,

Associate Editor

PLOS Medicine

plosmedicine.org