Childhood Developmental Milestones and Risk of Adult Cerebrovascular Disease: The Northern Finland Birth Cohort 1966 Open Access

Incidence of ischemic cerebrovascular disease (CeVD) among young adults is rising globally [1], posing an enormous burden. Unlike older adults, CeVDs among young adults are more heterogenous due to various risk factors [2]. Early developmental factors play a crucial role in later susceptibility to chronic diseases, including CeVD [3, 4].

Developmental milestones are divided into five main domains: fine motor, gross motor, language, cognitive, and social-emotional [5]. The World Health Organization has classified the age ranges for these milestones [6]. Childhood neurodevelopment is influenced by factors before and after birth, some of which are also risk factors for adult CeVD. For instance, early growth is linked to developmental milestones and overall health, including brain growth [7]. On the other hand, poor early childhood growth has been linked to adult CeVD [8‒10].

It has previously been established in this cohort that delayed motor development in childhood is associated with lower sports participation in adolescence and raised blood pressure levels in adulthood [11, 12] but higher light physical activity at middle age [13]. We found no previous literature examining childhood developmental milestones and later risk of hemorrhagic or ischemic CeVD. Our aim in this study was to address this evidence gap and investigate relationships between childhood achievement of fine motor, gross motor, and language milestones and adult CeVDs. Further, we investigate sex differences in these relationships as previous literature shows early life risk factors for CeVD differ between sexes [14].

This prospective cohort study uses data from the Northern Finland Birth Cohort 1966 (NFBC1966), a general population-based birth cohort containing data on 12,068 pregnant women and their 12,058 live-born children in two northernmost provinces of Finland with an expected date of birth in 1966 [15, 16]. Data collection started antenatally has been regularly followed up ever since and linked to nationwide registers.

This study included 11,688 subjects who were followed from birth to their first stroke, death, moving abroad, or until the end of year 2020. Excluded from the study were the subjects for whom informed consent was not obtained (n = 133), persons who had suffered a cerebrovascular event under/before the age of 15 years (n = 8) and those with a diagnosis of intellectual disability until the age of 14 years (n = 229). Detailed information on intellectual disability covering the age group 0–14 years was obtained from school health authorities, the national hospital discharge register, and other regional or national registers [17].

Developmental data were obtained from the records of child welfare clinics in a standard checklist format. All Finnish children are offered regular free-of-charge visits at child welfare clinics, and nearly everyone attends. During these visits, a public health nurse records any new milestones the child may have attained after the previous visit. Data were supplemented during with the information obtained in a special examination performed at the age of 1 year by the public health nurses. Age of achievement of motor milestones in the first year of life (in months) and age of achievement of milestones addressing hearing and speech (in months or years) from age 1 month to 4 years were used. Following gross motor milestones were addressed: being able to hold the head up, to turn over from back to tummy, to sit without support, to stand up, to stand without support, and to walk without support. Following fine motor milestones were addressed: to grab an object, to touch the thumb with the index finger. After separately studying each motor milestone variable, we further investigated overall motor development by using a principal component analysis (PCA) with a one-component model. The following language milestones were addressed: age at making sounds, speaks words at 1 year, speaks sentences at 2 years, and pronounces letters at 4 years.

Strokes and transient ischemic attacks (TIA) were identified from the Care Register for Health Care and Causes of Death Register based on medical records. CeVDs were classified by first primary diagnosis. Detailed ICD coding information is provided in Supplement. The linkage to other data with pseudonymization was fully complete for stroke diagnoses. Ischemic strokes and TIA’s were defined as ischemic strokes, and subarachnoid hemorrhages and intracerebral hemorrhages as hemorrhagic strokes. For analyses of any CeVD, ischemic strokes, hemorrhagic strokes, and other CeVD were combined.

Covariates were socioeconomic status (SES) of family and birth weight for gestational age. The SES of the family was asked in the pregnancy questionnaire and was defined as the highest occupational status of the mother or father, and categorized as unemployed or unskilled worker, skilled worker, and professional. Birth weight was obtained from delivery reports and gestational age at birth was calculated from the mother’s last menstrual period. Birth weight for gestational age was defined from the mean for sex and each gestational week and based on national birth weight standards [18]. Sex was categorized binary as women and men and assigned at birth based on biological attributes. We do not distinguish the biological sex differences from socially constructed gender differences. We refer to all differences as “sex differences” for clarity.

A total of 92.4% of cohort members missed some information on developmental milestones. All variables had missing information. The amount of missing information per variable in the total cohort and in those with CeVD is presented in the online supplementary Table 1 (for all online suppl. material, see https://doi.org/10.1159/000541702). A total of 50.2% of the values were missing. The missing data were assumed to be missing at random. Multiple imputation was not performed.

Baseline characteristics of women and men are shown as % or mean (standard deviation [SD]). Cox proportional hazards model was used to estimate the associations between age of attaining childhood developmental milestones (in months) and CeVD separately in women and men during follow-up. Results are presented with adjusted hazard ratios (aHR) with 95% confidence intervals (CIs).

Each milestone (per 1 month delay in achievement), sex, covariates, and a product term of the milestone with sex was included to obtain the women-to-men relative hazard ratio (RHR) for that milestone. For interpretability of the results, each milestone was studied in a separate model. Models also included sex-interaction terms with covariates as both covariates were assumed to have different associations in women and men. After investigating each motor milestone variable separately, we further examined motor development in respect to CeVD by using a PCA. PCA is a method used to reduce dimensionality of a large dataset by finding new variables (principal components) that are linear functions of those in the original dataset [19]. We tested 1–3 component PCA models and calculated their covariance matrix, eigenvalues, and eigenvectors. The best model based on eigenvalues and explaining most of the variance was selected. All statistical analyses were stratified by sex and adjusted for SES of family and birth weight for gestational age. In sensitivity analyses, we excluded TIA from the outcomes to account for possible diagnosing differences, as earlier definitions for TIA were established prior to largely available advanced imaging [20]. The results were similar to original analyses (online suppl. File).

Statistical analyses were carried out using IBM SPSS Statistics 27 except for sex-interaction analyses which were carried out using SPSS version 29. R version 4.3.1 was used to create data visualization.

In total, 11,688 people (48.8% women) were followed up for 575,723 person years (women 282,265 person years and men 293,458 person years) (Table 1). The mean follow-up time per participant was 49.3 years (SD 13.8 years) and 10,775 persons were alive at the end of follow-up. Altogether 498 (4.3%) CeVDs occurred during follow-up. Of these, 155 (31.1%) were ischemic strokes, 188 (37.8%) TIAs, 62 (12.4%) subarachnoid hemorrhages, 37 (7.4%) intracerebral hemorrhages, and 56 (11.2%) other cerebrovascular events. Median age at onset was 45.1 (SD 7.4) years for ischemic stroke, 47.5 (SD 5.2) years for TIA, 40.6 (SD 10.1) years for subarachnoid hemorrhage, and 44.4 (SD 8.8) years for intracerebral hemorrhage.

Among men, delayed turning from back to tummy related to risk for any CeVD with an aHR: 1.21 (95% CI: 1.05–1.39) per 1 month’s delay and an aHR of 1.25 (1.06–1.46) per 1 month’s delay for ischemic CeVD (Table 2). Among women, later turning from back to tummy was associated with increased risk of any CeVD with an aHR: 1.18 (95% CI: 1.02–1.38) per 1 month’s delay and aHR: 1.20 (95% CI: 1.02–1.42) per 1 month’s delay for ischemic CeVD. Later achievement of fine motor milestone “gripping an object” was related to increased risk of hemorrhagic CeVD among women with an aHR of 1.80 (95% CI: 1.01–3.19) per 1 month’s delay (Table 3). Other individual motor milestones were not associated with later CeVDs among either sex. Women-to-men RHR’s of motor milestones showed no significant findings or evidence for sex differences in sex-specific findings (Table 4; online suppl. File).

A one-component model provided the most appropriate interpretation of the data showing better fit than models with 2–3 classes. The one principal component model explained 41.5% of the variance in the motor milestones, eigenvalue 3.3, communalities of the milestones varied between 0.34 and 0.82.

Among men, delayed overall motor developmental reflected by a higher principal component score of motor milestones was associated with ischemic CeVD with an aHR of 1.50 (95% CI: 1.03–2.19) (Table 2). Among women, no association was found between motor milestone principal component score and later CeVD (Table 3). All results were similar in unadjusted models and models excluding TIA’s (online suppl. Tables 2–9). Women-to-men RHR’s showed no evidence for sex differences (Table 4; online suppl. File).

Among men, slow language development at two timepoints associated with increased risk of adult CeVDs. Later achievement of making sounds (at 3 months vs. at 1 or 2 months) was associated with any CeVD with an aHR of 2.74 (95% CI: 1.39–5.40) and ischemic CeVD with an aHR of 3.41 (95% CI: 1.65–7.06) (Table 5). Not speaking sentences at 2 years was associated with ischemic CeVD with an aHR of 1.89 (95% CI: 1.10–3.24). Among women, no associations were found between language milestones and later CeVD (Table 6). All results were similar in unadjusted models and models excluding TIA’s (online suppl. Tables 2–9). Later achievement of making sounds showed a women-to-men RHR of 0.17 (95% CI: 0.04–0.81) for any CeVD and RHR: 0.18 (95% CI: 0.04–0.89) for ischemic stroke indicating its association with risk to be lower in women compared to men (Table 4; online suppl. File). Women-to-men RHR’s of other language milestones showed no evidence for sex differences.

We present the first study observing associations between childhood achievement of motor and language milestones and later CeVD. Later gross motor development was associated with adulthood CeVD risk among both sexes. Delayed overall motor milestone development (modeled by motor milestone principal component score) was related to increased risk of ischemic CeVD among men. Among women, later acquirement of fine motor milestone “gripping an object” was associated with increased risk for hemorrhagic CeVD in adulthood. Delayed language development also showed to be associated with later risk of CeVD in the male study population. These observations were independent of family SES and birth weight for gestational age and analyses did not include persons with a diagnosis of intellectual disability. Sex differences were tested for and found that later making sounds was associated with increased CeVD risk in men. For other developmental milestones, no evidence of a sex disparity was observed.

Examining underlying mechanisms and proof of causality for the found associations was outside the scope of this study, but we will discuss some possible mechanistic explanations behind the reported findings. The first possible mechanistic explanation between delayed motor and language development and risk of adult CeVD is suboptimal brain development during the fetal period and early childhood. Human brain development begins in the third gestational week and brain size increases rapidly during childhood, reaching approximately 90% of adult volume by age 6 [21]. Especially, the first 1,000 days after conception have been found to be an important period in neurodevelopment [22]. Both reduced early growth and later achievement of motor and language milestones in childhood have been shown to correlate with lower cognitive abilities in later life [23, 24]. Children born with low birth weight are at increased risk for neurodevelopmental delays and adult CeVD, but adjustments for birth weight did not change the results of our study, suggesting other underlying mediators. A recent study using data from four prospective cohort studies showed that a risk for cerebral small vessel disease originates in early life and could provide a mechanistic connection between early life factors and risk of CeVD [25]. Current evidence suggests that favorable prenatal and early life factors related to nutrition, SES and IQ can decrease the risk of small vessel disease and stroke [26]. Education and early life cognition could reduce susceptibility to brain pathology through reducing risk factors for disease, but associations between childhood IQ and small vessel disease have been found to be independent of vascular risk factors and adult SES, suggesting an effect independent of adult vascular risk factors [25].

Delayed neurodevelopment during childhood predisposes to lower performance in school and work life, lower SES and possibly thereby increased risk for later CeVD. Previous studies in our cohort have established that earlier standing, walking, and talking are each associated with higher educational attainment in adulthood, and age of walking is linked to sports participation later in life [11, 27]. Higher cognitive ability in childhood, better SES, and longer duration of education have been shown to predict lower risk of CeVD in later life [28]. However, in our current study, results persisted even after adjustments for family SES.

Concerning motor development, one possible mechanistic explanation is that delayed childhood motor development and adult CeVDs are mediated through stroke risk factors emerging in later life, including obesity and high blood pressure. Supporting the motor milestone findings of this study, a previous study has reported that later turning from back to tummy is related to greater central adiposity age 3 years, which predisposes to later obesity [29‒31]. Hand control and coordination in childhood have also been associated with adult obesity and a delayed achievement of motor milestones with disturbances in lipid metabolic pathways [32, 33]. It is possible that children with slower motor development develop poorer motor skills, may not find physical activity rewarding and adopt a more sedentary lifestyle [34, 35]. In our cohort, it has also been established that later motor milestone development during infancy is related to higher systolic and diastolic blood pressure levels at age 31 [12].

We found delayed language development to be associated with an increased risk of later CeVDs among men. According to epidemiological studies, risk factors for language delays and disorders include male sex, low SES, low birth weight or preterm birth and family history of language delays [36‒39]. In support of these previously established risk factors, our findings on delayed language milestone achievement did not replicate among women. They did, however, persist after adjustment for family SES and small birth weight for gestational age, suggesting a common developmental background between language development and later CeVD. Studies performed on children with history of late talking have revealed a variety of structural and functional differences of the brain compared to peers with typical language development, supporting the hypothesis of developmental delays occurring due to underlying neurobiological defects [40, 41].

The present study had some limitations, the main one being the small sample sizes of CeVD groups. Even though we had a large-scale cohort with comprehensive follow-up data, only 498 CeVDs were recorded. Some of the developmental milestone variables might not reflect delays in development, but rather average group development. Missing data were not imputed. We performed multiple analyses so there is a possibility of chance findings. Also, as with observational population-based research in general, unidentified, or residual confounders are a possible limitation.

The population-based and longitudinal design of the study also provides several important strengths, one being the possibility to use a large, unselected birth cohort with almost 12,000 participants and nearly 580,000 person years of follow-up. Data collection started from the antenatal period and cohort members have been followed up regularly ever since. Register data were complete for the whole cohort with no loss to follow-up. Information on childhood milestones was registered prospectively and we were able to perform analyzes separately for both sexes. Very few stroke patients had diagnoses of other possible comorbidities that might have affected neurodevelopment, for example, congenital heart disease (n = 2). We believe the results of these dataset to be generalizable to most high-resource settings.

In conclusion, childhood motor milestone achievement among both sexes and language development among men showed to be related to adulthood CeVDs. These new findings point toward a common neurodevelopmental background and could in part explain lifetime CeVD risk accumulation. The results highlight the need for elucidating the pathways which may explain these novel associations. Better understanding of the relationship between childhood development and adult CeVD risk is an important research target for public health policy as it provides tools for targeting primary preventative measures to individuals at increased risk.

We thank all cohort members who participated in the study and the researchers of NFBC project center.

Permission to gather register data was obtained from the Ministry of Social Affairs and Health, and the study was approved by the Ethical Committee of Northern Ostrobothnia Hospital District in Oulu, Finland, Approval No. EETTMK 94/11, September 17, 2012. Data protection was scrutinized by the Office of Data Protection Ombudsman of Finland. Written informed consent was inquired from all the participants.

None to be reported.

NFBC1966 received financial support from University of Oulu (65354, 24000692); Oulu University Hospital (2/97, 8/97, 24301140); Ministry of Health and Social Affairs (23/251/97, 160/97, 190/97); National Institute for Health and Welfare (54121); and European Regional Development Fund (539/2010, A31592). This work was supported by Orion Research Foundation, Maire Taponen foundation, Päivikki & Sakari Sohlberg foundation, Paavo Ilmari Ahvenainen foundation, Finnish-Norwegian Medical foundation and Finnish Medical foundation. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Milja Kivelä: conceptualization, methodology, formal analysis, data curation, and writing – original draft. Ina Rissanen: conceptualization, methodology, data curation, writing – review and editing, and supervision. Eero Kajantie, Harri Rusanen, and Marja Ojaniemi: conceptualization, writing – review and editing. Jouko Miettunen: conceptualization, methodology, resources, and writing – review and editing. Markus Paananen: conceptualization, methodology, writing – review and editing, and supervision.

Northern Finland Birth Cohort 1966 (NFBC1966) data are available from the Infrastructure for Population Studies at University of Oulu. Permission to use the data can be applied for via electronic material request portal. EU general data protection regulation (679/2016) and Finnish Data Protection Act are followed in the use of data. Cohort participants’ written informed consent at latest follow-up study is used as base for personal data usage, which may cause some limitations to its use. For more information, contact NFBC1966 project center ([email protected]) and visit the cohort website for more information.