The association between growth patterns and blood pressure in children and adolescents: A cross‐sectional study of seven provinces in China

Abstract Aimed to investigate the associations between different growth patterns with high blood pressure, and further examine the mediation effect of BMI between growth patterns and high blood pressure among children and adolescents. A total of 31581 children and adolescents aged 7–18 years were selected based on the stratified cluster sampling method. Logistics regression models were used to calculate the odds rations (ORs) and 95% confidence interval (95%CI) of the association between different growth patterns and high blood pressure. Mediation effect analyses were applied to estimate the effect of BMI on the increase of blood pressure levels in different growth patterns. In different sex and ages, compared to reference group of normal growth, blood pressure levels and prevalence of high blood pressure of the catch‐up growth were higher, but that of the catch‐down growth were lower. The prevalence of high blood pressure was 11.69%, 16.06%, and 9.68% in normal growth, catch‐up growth, and catch‐down growth, respectively. In total, compared with the normal growth pattern, the ORs (95%CI) of high blood pressure, high systolic blood pressure and high diastolic blood pressure in the catch‐up growth were 1.171(1.073,1.280), 1.110(1.001,1.230) and 1.141(1.025,1.270) (p < .05), respectively. Additionally, the mediation effect of current BMI existed in the association between blood pressure levels and different growth patterns, particularly in boys. Our findings suggested that different growth patterns after birth could modify blood pressure, and the potential risks of high blood pressure could be increased by catch‐up growth at childhood and adolescence.


INTRODUCTION
High blood pressure (HBP) during childhood and adolescence has become a major public health problem in the world. 1 The global prevalence of HBP in children and adolescents was on a long-term upward trend with obesity pandemic, particularly in Asian countries. 2,3 One study conducted in US reported the prevalence of HBP was 16.3% in 2017 among adolescents aged 10-17 years. 4 Similar in China, the prevalence of HBP in children and adolescents fluctuated between 6.9% and 9.2% from 1995 to 2014 based on a successive national survey. 5,6 The BP levels of children and adolescents should always be a good predictor of adult HBP with the phenomenon of "trajectory", that was often overlooked. [7][8][9] Therefore, HBP that once was considered a rare disease in children is now actually a major public health problem worldwide, 10 due to the tracking of risks with organ damages including coronary artery calcifications and hypertrophy from childhood to adulthood. 11,12 Simultaneously, there were many factors that could affect HBP in children and adolescents, such as family history, physical exercise, diet and body fat. 13 Among the multiple factors, body types, like weight and height, has always been confirmed to be one of the leading drivers to HBP. 14,15 Notably, unlike weight with much attention, height was also a strong and independent indicator in determining the increases in BP, since the definition of HBP among children and adolescents was comprehensively determined by age, sex, and height in the current international criteria. [16][17][18] However, the effect of growth patterns measured by the indicator of height on BP during childhood and adolescence was still unknown.
Some indicators of infants followed a regular growth patterns after birth, such as body length, height, weight, body mass index (BMI), and head circumference. 19 Under various nutrition environments, they were outgrowing, or falling behind, or still following the established trajectory. 20 A study from US divided growth patterns into catchdown growth, normal growth and catch-up growth, 21 which caused widespread. In each stage of growth in children, they might suffer from diseases, malnutrition, psychological stress, and other adverse events, which would cause the phenomenon of catch-down growth. 22,23 Catch-down growth, is often observed in babies born at intrauterine growth restriction and small for gestational age. [24][25][26] Once these factors above were removed, the height growth of children would catch up the original normal trajectory. This phenomenon of accelerated growth was called catch-up growth pattern. 20,27 Thus, catch-up growth was defined as "a height velocity above the statistical limits of normality for age or maturity during a defined period of time, following a transient period of growth inhibition; the effect of catch-up growth is to take the child towards his/her pre-retardation growth curve". 28,29 Many studies reported the adverse events of catch-down growth of children with linear growth failure, including multiple pathological disorders associated with increased morbidity and mortality, loss of physical growth potential, nutritional diseases, and lung diseases in adulthood. 30,31 Previous studies also showed that children with catchup growth could have a higher risk of cardiovascular and metabolic diseases later in life than those with normal growth patterns. 32 In addition, the strong association between BP and BMI had been well documented in children and adolescents, 33 and previous studies had explored potential mechanisms linking growth patterns and cardiovascular diseases. 34 However, few studies explored the associations between different growth patterns and HBP risks among children and adolescents, and given the importance of BMI, the mediation effect of BMI on the increase of BP levels among different growth patterns is still lacking.
Thus, using the data from a national representative survey in Chinese seven provinces, the present study aimed to determine the association between different growth patterns and HBP risks among children and adolescents, and to explore the potential mediation effect of BMI between growth patterns and BP levels.

Design and participants
The data in this study came from 93 primary and secondary schools in seven provinces, including Guangdong, Hunan, Liaoning, Shanghai, Chongqing, Tianjin, and Ningxia (Figure 1). The study adopted the multistage cluster random sampling, which has been published in detail in previous studies. 35 Briefly, one city was selected in each of seven provinces according to their economic level. In each selected city, about 16 primary and secondary schools were selected. Among the selected schools, two classes per grade were randomly selected with all students aged 6-18 years considered eligible. All eligible participants in our study underwent a complete medical examination before data collection and those were excluded if they had one or more of the fol-

Measurements
The measurement of each anthropometric index, including height (cm), weight (kg), BP (mm Hg), in our study followed a standardized procedure by professionals who had passed the training course, details of which have been published previously. 36 The standardized measurement procedure referred to the anthropometry methods in the 2006 Height and weight were calculated the average of two repeated measurements. In rechecking 10% of the students, the error was less than 5%. BMI was calculated as (weight(kg)/height(m) 2 ), and the age was cal- sound. BP was measured twice with a 5 min gap between the two measurements from the right arm. If the difference between two measurements was 10 millimeters of mercury or higher, the third measurement would be conducted until the difference was less than 10 mm Hg, then the two closest values were adopted, and the average of SBP and DBP values were calculated, respectively.
Birth length data was collected by a standard questionnaire survey. Parents needed to record their children's birth length according to their birth certificate or health clinic record. If they did not have it, they were asked to recall their children's data. All of the birth length was measured by trained nurses and recorded to the nearest 0.1 cm when they were born. About 75% of parents recorded the data of their children's birth length based on the birth certificate or health clinic card.
During the entire research process, to ensure the reliability of birth length, parents were asked to repeat the same process of questionnaire survey twice, if the difference was more than 2 cm, we judged the sample as missing.

Definitions of growth patterns
Participants' birth length and height were Z-Scored according to the 2000 Centers for Disease Control and Prevention growth charts for the United States (US CDC standards (2000)), 21 and it was recommended to assess size and growth in infants, children, and adolescents in clinical practice. 38 According to the definition by Robert J. Kuczmarski, and his colleagues, a gain in Z-Score for birth length and current height greater than 0.67 was taken to indicate clinically significant catch-up growth, similarly, a decrease in Z-Score less than -0.67 was considered as catch-down growth, and the others between the two limits were judged as normal growth.

Definitions of HBP
BP levels were divided according to the Clinical Practice Guideline for Screening and Management of High Blood Pressure in Children (2017 Guideline). 16 Normal BP was defined as SBP and DBP that were < 95th percentile for sex, age, and height. HSBP (high systolic blood pressure) and HDBP (high diastolic blood pressure) were defined as SBP or DBP greater than or equal to the 95 th age-, sex-, and height-specific percentile for children younger than 13 years, and greater than or equal to 130/80 mm Hg for children and adolescents aged 13 years or older. 16 Participants with at least HSBP or HDBP were defined as having HBP. between variable X (e.g., Growth patterns) and variable Y (e.g., SBP, DBP). Thereafter, we used same methods to analyze associations (a coefficient) between X (e.g., Growth patterns) and potential mediators Z (e.g., BMI). Then, associations between potential mediators Z (e.g., BMI) and dependent variable Y (e.g., SBP, DBP) were examined (b coefficient). Finally, if c, a and b coefficients were all statistically significant, the mediation effect of potential mediator Z on the association between X and Y was proved. Additionally, the proportion of mediation was calculated. Statistical analysis was performed using Stata version 14.0, we considered the associations to be significant when the twosided p value was less than .05.

General demographic characteristics
Characteristics of participants in different growth patterns were shown in Note: Continuous variables were expressed by mean values and standard deviations (Mean ± SD), and categorical variables were expressed by numbers and percentages (n (%)).

Associations of BP Z-Scores with birth length, height, and the difference Z-Scores
A scatter diagram and its fitted lines showed the visual association between the SBP and DBP Z-Scores with birth length, height, and the difference Z-Scores among boys and girls (Figure 2). We found that the DBP Z-Scores was positively associated with birth length Z-Scores (r = 0.022, p = .006). Apart from this, a significant U-shaped relationship between both SBP and DBP Z-Scores with the current height Z-Scores was observed, and the same results also existed between both SBP and DBP Z-Scores and the difference of current height and previous birth length Z-Scores.

The levels of BP by sex and age group
BP levels of catch-up growth were higher than that in other two growth patterns groups (Figure 3)      Abbreviations: HBP, high blood pressure; HSBP, high systolic blood pressure; HDBP, high diastolic blood pressure. Model1: represents there were age, city, birthweight, first menstrual, gonacrat, father education, mother education, father BMI, mother BMI, family history of hypertension, breast feeding, daily moderate physical activity.; Model2: represents there were current BMI based on Model1.

DISCUSSION
Our study mainly illustrated the associations between different growth patterns and BP levels among children and adolescents in China. The BP levels of catch-up growth were higher than that of normal growth.
After adjusted for confounders, compared with normal growth group, catch-up growth pattern was positively associated with the risks of HBP, HSBP and HDBP, but the catch-down growth pattern failed to reach significance. Furthermore, current BMI value could act as a mediating indicator in the relationship between the catch-up growth patterns and HBP risks, especially among boys. This study with national large data of seven provinces provided an obvious evidence for guiding the scientific and reasonable growth and development of children and adolescents after birth with different growth patterns.
Compared to reference group of normal growth, BP levels and prevalence of HBP of the catch-up growth were higher, but the that of the catch-down growth were lower. Currently, catch-up growth patterns are usually recommended, which can promote neurodevelopment and height growth. However, massive studies demonstrated that catch-up growth patterns might lead to the high risk of obesity, insulin resistance, cardiovascular diseases and other adverse health outcomes. [40][41][42] Besides, one study found that adults who were born with small for gestational age but experienced accelerated growth during growth and development had been considered to be at an increased risk of HBP, 43 which was similar to our results. It was also found that insulin resistance and signal impairment in skeletal muscle inhibited thermogenesis during catch-up growth, which could be the potential mechanisms of catch-up growth with increased risk of HBP in later life. 44,45 There were some evidences that the state of insulin resistance in catch-up growth would increase preferentially body fat content, which could induce metabolic diseases. 44,46 The possible mechanisms for this result were that the process of catch-up growth could excessively consume the function of the early blood vessel wall, which affected the power of blood flow in the blood vessel and leads to increase in BP levels. 47 The finding of this study manifested that cutch-down growth had low BP levels, but had no substantial effect.
Until it was not clear that enabling catch-up growth was of no cognitive benefit, 48 specially, our study conducted that catch-up growth could be a risk factor for HBP, HSBP and HDBP. The studies that investigated the association between catch-up growth and SBP in adulthood all reported similar positive associations, 49,50 which supported the suggestion that small-for-gestational-age infants, who tended to exhibit catch-up growth, were at an increased risk of HBP. 51 Instead, a more comprehensive guidance of HBP and catch-up growth, which kept children and adolescents healthier after birth, should be more stressed. Noticeably, in the process of adjusting for current BMI, no statistically significant differences were observed between boys and girls.
In addition, the mediation effect of BMI in different growth patterns on BP levels was only found in boys. The difference by sex might be attributed to the hormones. Considering that the serum estradiol and testosterone concentrations of prepubertal girls and boys were different, the estradiol level of prepubertal girls was much higher than that of prepubertal boys, and the serum testosterone concentration was the opposite. 52 Our study had obvious strengths that the national representative large sample size from seven provinces in China reflected the associations between different growth patterns and HBP risks in children and adolescents. However, several limitations should be noted. Firstly, birth weight and length were the core indicators of this study, therefore, we chose the information on the birth certificate as much as possible. Indeed, the results obtained in this way have a certain degree of bias, because there are measurement biases among nurses in different regions of the country. Another way collected by parents record according to their birth certificate or health clinic record, also existed recall bias. However, on this process we carried out strict quality control to ensure the reliability. Secondly, different growth patterns were defined by the Z-Score of birth length and current height during a long interval, thus, the assessment of the growth pattern could be a bit rough, and it might neglect the detailed patterns in the mid-dle of growth. In order to overcome the information of lacking of the Tana stage, we included these factors (first steatorrhea(yes/no), menarche(yes/no)) to replace the pubertal development status. Thirdly, our study was a cross-sectional design, so we cannot conclude into a causal relationship, however, the infant birth length existed before the measurement of BP, an inferred cause association between growth patterns and BP levels should be reasonable. In addition, regarding the diagnosis of HBP, the standard we adopted was only a preliminary screening, but to some extent, our research was based on a large-scale epidemiological survey, which could reduce the bias partially. Besides, we defined BP > 95% as HBP, which would reduce the strength of the correlation to a certain extent. A final limitation was that some potential confounders in our study were adjusted, but the risks of HBP could also be influenced by genetics and environment factors, which needed to be further explored and vitrificated in future study.

CONCLUSIONS
Catch-up growth pattern was positively associated with higher risks of HBP, HSBP and HDBP in children and adolescents, but catch-down growth patterns failed to reach significances. Current BMI value could act as a mediating indicator in the relationship between the catch-up growth patterns and HBP risks, particularly in boys. Our study provided new theoretical basis and research ideas for the future identification and intervention of HBP among children and adolescents, and helped to assess potential risks of HBP and related metabolic diseases caused by catch-up growth pattern after birth and at adulthood.