Ethnicity and tracking blood pressure in children

Abstract

There is growing evidence¹ that hypertension, one of the major modifiable risk factors for cardiovascular disease (CVD), is established early in life. Given this, it is important to discover when hypertension first becomes apparent. Further, it is of particular importance to examine the ethnic differences in blood pressure (BP) in children, given the variation in rates of CVD morbidity and mortality among adults from different ethnic groups.²

Prevalence of hypertension, mean BP levels and CVD mortality in adults

The majority of American adult population-based BP studies have reported an increased prevalence of hypertension, higher mean BP levels, and higher morbidity and mortality rates, attributable to hypertension among African-Americans compared to white-Americans. A similar picture for mortality emerges from the United Kingdom (UK). There is considerable variation in the mortality rates from CVD among the main ethnic groups in the UK.^2,3,4 As in the USA, black migrants (from the Caribbean and West Africa) have an increased risk of stroke³ and end-stage renal failure,⁴ whereas coronary heart disease (CHD) is less common (in the UK). South-Asians (people originating from the Indian subcontinent and East Africa), the other main ethnic group in the UK, have a greater incidence of CHD compared to the white group,³ but also have high mortality rates attributable to stroke and end-stage renal failure.⁵ These differences are due, in part, to the high prevalence of diabetes and hypertension.⁵

However, data regarding higher mean BP and a greater prevalence of hypertension among adults from ethnic groups in the UK are not as consistent as data from the USA. Indeed, three recent reviews of population-based studies from the UK^6,7,8 reveal that the majority of these studies report an increased prevalence of hypertension among both Afro-Caribbeans^7,8 and South-Asians,^6,7,8 and/or higher mean BP among Afro-Caribbeans,^7,8 but similar mean BP levels among South-Asians^6,8 when both groups were compared to the white population, although some counter-examples exist.⁸

Mean BPs in children in the USA and UK

During the 1970s and early 1980s, many studies were conducted, mainly in the USA, to assess BP in children and adolescents. A task force on BP control in children⁹ was established, which produced guidelines on BP standards for children, and now includes height percentiles, age and gender, but does not mention ethnic differences in BP.¹⁰

Previous cross-sectional studies to assess the ethnic differences in BP among children and adolescents in the USA have produced conflicting results. The majority of these studies have assessed BP among African-American and white-American children and adolescents, and report higher BP levels among African-Americans, although these differences are often minimised or abolished once BMI, age and socio-economic status (SES) are controlled for.^11,12 Many other US studies do not report significant differences in BP between these two ethnic groups¹³ or lower BP among African-Americans.¹⁴ More recent studies investigating BP in children from other ethnic minority groups, such as Mexican-Americans, do not report higher BP levels in these children.¹⁵ These conflicting results are likely to be due to methodological differences and the groups sampled.

Agyemang et al,¹⁶ in this issue of the Journal, reviewed UK studies which assessed BP in children (⩽16 years of age) from minority ethnic groups and identified five such studies, three of which included children of African origin, four included children of South-Asian origin, and one study included children of Chinese origin. However, the results of these studies are also inconsistent, and are probably also due to methodological differences between the studies. Further, the studies reviewed were cross-sectional and did not track BP from childhood to adolescence or adulthood. This paper highlights several important points. First, the paucity of UK studies investigating ethnic differences in BP in children and adolescents, given the high CVD mortality rates attributable, in part, to hypertension, among minority ethnic groups in the UK. Second, the heterogeneity of the ethnic groups studied particularly among South-Asians, and the relatively small numbers of children in each study. Future epidemiological research should examine South-Asian subgroups separately, as the mean BP levels and the prevalence of hypertension and other CVD risk factors may vary considerably among these groups. Third, those ethnic differences in BP in children do not correspond to the pattern of BP seen in adults from those ethnic groups in the UK. Fourth, the importance of evaluating BP in all minority ethnic groups reflecting the changing diversity of the UK population.² And, finally, it calls for the need to assess the persistence of BP from childhood to adulthood, to identify which children and adolescents will become hypertensive later in life, so that preventive measures can be undertaken.

Tracking of BP in Western populations

Compared to the plethora of studies providing cross-sectional data on the relationship between BP and age in both childhood and adolescence, and adulthood, relatively few have investigated this association longitudinally starting in childhood and continuing into adulthood. This phenomena, known as ‘tracking’, refers to the stability over time of an individual's level of BP relative to that of his/her peers,¹⁷ and has been quantified by correlating the rank order of BP at different points in time.¹⁸ The correlation coefficients vary considerably (0.1–0.7 for SBP and 0.02–0.5 for DBP),¹⁸ and are higher when measurements are separated by shorter durations and in older children.¹⁸ Another method used to quantify the persistence of BP has been to split the BP distribution into percentiles or quintiles and observe how many children remain within these parameters over time. Once again, the results vary considerably, but the Bogalusa Heart Study showed that 40% of individuals with SBP levels above the 80th percentile at baseline had levels above this percentile 15 years later. The analogous figure for persistence of DBP in the top quintile was 37%.¹⁹

The ability to predict BP levels in adulthood from measurements taken over time in childhood would provide the opportunity to intervene before hypertension is firmly established, thereby reducing CVD risk. Most cohort studies tracking BP in childhood were conducted during the late 1970s and 1980s,^20,21 and only the Bogalusa Heart Study,²² to our knowledge, investigated ethnic differences. Since then, two further studies have tracked BP in children from different ethnic groups in the USA^12,23 and a further re-analysis of the Bogalusa Heart Study cohort has been published.¹⁹

Tracking of BP from childhood to early adulthood in different ethnic groups

Table 1 summarises three studies, all conducted in the USA, which have tracked BP in African-Americans and white-Americans, one from childhood to adolescence¹² and two from childhood to early adulthood.^19,23 The results of these studies are inconsistent. All three studies reveal that black children have significantly higher mean BP than white children. This finding survived the adjustment for potential confounders such as weight gain or increase in BMI^19,23 and SES²³ in two studies, but only the ethnic difference for SBP in girls remained significant after controlling for age, weight and height in the study by Manatunga et al.¹² The conflicting results may be due to population differences as this study tracked children to adolescence only,¹² whereas the other two studies tracked BP from childhood to early adulthood.^19,23 Black children had significantly greater increases over time for SBP and DBP than white children,¹² and black females exhibited a greater increase in SBP over time than white females.²³ Bao et al¹⁹ demonstrated similar tracking of BP in the two ethnic groups, with correlations ranging from 0.36 to 0.50, but no significant differences between the ethnic groups. Adult onset hypertension was also more prevalent in African-Americans and in individuals who had higher BP or BMI in childhood, or had gained more weight (with respect to height) from childhood to adulthood. However, BP was only measured twice in this study,¹⁹ 15 years apart, which may be insufficient to describe the development of BP.²³

Table 1 Studies which have tracked blood pressure in longitudinal cohorts of children from different ethnic groups

Despite the inconsistency in results, these studies revealed that raised childhood BP was predictive of subsequent BP^19,23 and hypertension¹⁹ in early adulthood, and that BP trajectories varied between ethnic groups when repeated BP measurements were recorded from childhood to adolescence¹² and adulthood.²³ Ethnic differences in SBP become evident in childhood in girls, in early adolescence in boys, and tend to increase with age. For both sexes, ethnic differences in DBP are evident in childhood and remain stable over time, and these ethnic differences in BP cannot be entirely explained by individual differences in adiposity, growth or SES.²³

Parental history of disease and clustering of CVD risk factors

Children of parents with a positive history of CVD tend to display an adverse CVD risk profile, namely being overweight, and having elevated levels of BP and cholesterol, after adjustment for age and weight,^17,24 due to a combination of genetics and a shared environment. This clustering of CVD risk factors has been termed syndrome X or the insulin-resistance syndrome. Investigation of such CVD risk factors in asymptomatic young people may help explain how the precursors of adult CVD develop. Studies, including black and white children, have shown that CVD risk factors cluster even in childhood, that clustering tracks over time, and that the persistence and prevalence of multiple risk factor clustering is increased by obesity.^25,26 Further, recent follow-ups of young adults from the Muscatine,²⁷ Bogalusa,²⁸ and Young Finns²⁹ studies have reported that carotid intima-media thickness, a marker of preclinical atherosclerosis, was associated with childhood levels of CVD risk factors, particularly LDL cholesterol,^28,29 total cholesterol²⁷ and BMI.^27,28,29

Birth weight and subsequent BP

Retrospective cohort studies have reported that lower birth weight is associated with higher subsequent mean SBP in children³⁰ and adults,³¹ although the data are not consistent among adolescents,³² and that this relationship becomes amplified with increasing age. Lucas³³ has postulated that this association is an example of ‘programming’, a view that is partly supported by the existence of tracking of BP in children. These findings suggest that prenatal as well as postnatal factors may be implicated in the aetiology of hypertension.³⁴

Implications for research and policy

On the whole, studies from the USA have shown that elevated BP in childhood is a good predictor of elevated BP^19,23 in adulthood, and that there appears to be ethnic differences in BP trajectories, especially when multiple observations are available in both adolescence¹² and early adulthood.²³ Given the findings reported in this Journal,¹⁶ that childhood BP patterns do not mirror those of adult BP in four UK ethnic groups, there is a need for a cohort study of children, including data on birth weight, to investigate whether childhood BP trajectories predict hypertension in adulthood among the different ethnic groups in the UK.

CVD is largely preventable by reduction of modifiable risk factors. However, effective primary and secondary prevention requires total risk factor identification, treatment and management rather the intervention of single risk factors. Therefore, it is extremely important to identify at what age CVD risk factors, not just hypertension, become apparent, given that studies from the USA have shown that CVD risk factors cluster in childhood and track over time.^25,26 Future epidemiological studies should focus on the heterogeneity of different ethnic groups and examine the subgroups separately to identify CVD risks specific to that group. Since a growing proportion of people from ethnic minorities are UK-born, epidemiological research among second- and third-generation South-Asians and Afro-Caribbeans is necessary to re-examine the prevalence of CVD risk factors and outcomes, to identify how disease patterns may be changing, so that appropriate interventions can be introduced early in life before unhealthy lifestyles become firmly established.