Viviane Cunha Cardoso Size at Birth and Blood Pressure in Young Adults: Fi Ndings from a Brazilian Birth Cohort Study

Tamanho ao nascer e pressão arterial em adultos jovens: achados de uma coorte de nascimentos brasileira ABSTRACT OBJECTIVE: To describe the prevalence of borderline blood pressure (BBP) and hypertension (HT) among young adults and to assess the association between size at birth and BBP/HT. METHODS: Data were collected from the fi rst Ribeirão Preto Birth Cohort Study conducted in the city of Ribeirão Preto, southeastern Brazil, that started in 1978-1979. Of the 6,827 singletons born in hospitals, 2,060 were assessed at 23-25 years of age. Blood samples were collected, an anthropometric assessment was performed, and information was obtained regarding occupation, schooling, life habits and chronic diseases. Blood pressure (BP) was classifi ed as: 1) BBP: systolic BP (SBP) ≥ 130 mm Hg and < 140 mm Hg and/or diastolic BP (DBP) ≥ 85 mm Hg and < 90 mm Hg; and 2) HT: SBP ≥ 140 mm Hg and/or DBP ≥ 90 mm Hg. A polytomic logistic regression model was used.

Essential hypertension is an important risk factor for morbidity and mortality from coronary heart disease, stroke, and renal disease. 26Although clinical manifestations of hypertension (HT) do not generally emerge until middle age, the pathophysiologic precursors of adult HT are thought to originate very early in life even during fetal development.
Fetal growth disorders are associated with chronic noncommunicable diseases in adulthood, among them HT, according to the fetal origins of adult disease hypothesis. 6However, it is not clear whether early life factors such as weight and length at birth are also associated with borderline blood pressure (BBP).

INTRODUCTION
young adulthood still needs elucidation.BBP, a new risk category for Latin American and European countries, includes a population at high risk of developing HT and in which lifestyle modifi cations are needed.This might be useful not only in assessing future risk of hypertension but also in prevention and intervention algorithms. 11is study aimed to describe the prevalence of BBP and HT among young adults and to assess the association between size at birth and BBP/HT.

METHODS
Study carried out with data from the fi rst Ribeirão Preto Birth Cohort Study conducted in the city of Ribeirão Preto, southeastern Brazil, that started in 1978-1979.Data were obtained at birth and at young adult age (23 to 25 years).There were recruited 9,067 liveborn infants delivered at the city's eight maternity hospitals from June 1 st , 1978 to May 31, 1979 (98% of all live births).Infants whose families did not reside in the city (2,094) and twins (146) were excluded from the study, remaining 6,827 live births. 9Information about mother's age, schooling and smoking was obtained through interviews with the mothers after delivery and review of medical records at the maternity hospitals.Birth weight and length were measured within 30 minutes of birth by trained personnel with appropriate devices that were donated by the research team to all hospitals.The infants were weighed naked on scales that were calibrated on a weekly basis with 10-g precision (Filizola, São Paulo, Brazil).Two trained staff members measured length at birth with the infants lying in the supine position on a neonatometer with a fi xed vertical headpiece and a smooth sliding vertical footpiece.Measurements were taken to the nearest 0.5 cm.
From the original cohort, 30% of subjects were selected to participate in the adult follow-up (between 2002  and 2004).This sample size was enough to test the study hypothesis.Of 6,827 singleton liveborns, 343 subjects were found to be deceased and 819 could not be traced, leaving 5,665 singletons.Subjects were classifi ed according to the income of the head of the family (geoeconomic classifi cation) to ensure a representative sample of socioeconomic groups.One in every three subjects from the same geographic area was invited to undergo medical examination.The fi rst of every three names was selected from a list sorted by birth date in each geographic area, and if unavailable the next name down was selected.Losses to follow-up (n = 705) occurred because of refusal to participate, imprisonment, death after 20 years of age, or failure to show up for the interview.Losses were replaced using the same sampling method, resulting in 2,063 young adults (Figure ). 5 Data were obtained by structured questionnaires, physical examination, and blood collection.Anthropometric measurements (weight, height, and waist circumference) were taken according to standardized techniques with subjects barefoot and wearing light clothing.The instruments used were a precision scale periodically calibrated, anthropometers for standing and sitting measurements, a non-extensible metric tape and a caliper.Blood pressure was measured three times using a digital sphygmomanometer with a cuff of the same size which was adjusted to the arm circumference.The measurements were taken by the same examiner at 15-min intervals, with the subject resting in the sitting position with the left arm at the height of his/her heart.The mean of the last two measurements was then calculated.Three of 2,063 subjects were excluded because of technical diffi culties in recording BP (extreme obesity), totaling 2,060 young adults.A sample size of 1,291 was needed to test the study hypothesis assuming the event had a 10% prevalence with a 5% probability of type I error and 1% precision.Details regarding the methodology are available elsewhere. 5,9ood pressure was the primary outcome measure and was classifi ed into three groups: 1) normal blood pressure (NT): systolic BP (SBP) < 130 mm Hg and/or diastolic BP (DBP) < 85 mm Hg; 2) borderline blood pressure (BBP): SBP 130 mm Hg to 139 mm Hg and/or DBP 85 mm Hg to 89 mm Hg; 3) hypertension (HT): SBP ≥ 140 mm Hg and/or DBP ≥ 90 mm Hg according to Argentinean, Brazilian, and European guidelines. 12,23,25ciodemographic characteristics were evaluated at birth: mother's age (years), maternal schooling (years), occupation of the head of the family according to the International Standard Classifi cation of Occupation 21 and maternal smoking (no smoking; 1-10; > 10 cigarettes).Gender, gestational age, birth weight (BW), length, ponderal index (PI, kg/m 3 ), and adverse perinatal outcomes (low birth weight [LBW, < 2,500 g],  Adulthood sociodemographic characteristics evaluated were: occupation of the head of the family; smoking habit (no smoking; 1-10; > 10 cigarettes); physical activity (active, suffi ciently active, and inactive according to the International Physical Activity Questionnaire scoring protocol); a alcohol consumption (g/day; high > 31 g/day; low ≤ 31 g/day). 4e recommendation of the World Health Organization (WHO) of a daily intake of < 5.0 g of salt (2,000 mg of sodium) was adopted to defi ne the cut-off of high sodium intake in mg/day. 28esity in adulthood was defi ned as BMI ≥ 30 kg/m 2 .
Abdominal obesity was based on the waist circumference cut-offs proposed for the metabolic syndrome classifi cation by the International Diabetes Federation. 2 Fasting blood glucose was determined by the GOD/ PAP human diagnostic colorimetric enzymatic method (Chronolab AG, Zug, Switzerland) and was considered high when ≥ 100 mg/dL. 2 Total cholesterol, HDL cholesterol, and triglycerides were determined by an enzymatic colorimetric method using the Dade Behring XPand device (Dade Behring, Liederbach, Germany) and reagents of Dade Behring Dimension clinical chemistry.Insulin resistance was estimated by a so-called surrogate marker, 18 the triglyceride-to-high density lipoprotein (TG/HDL) ratio, which has been proposed as one of the most accurate surrogate markers; optimal cut-off was 2.90 or less.
All variables are presented as proportions (n, %) and the three levels of blood pressure were compared by the chi-square test.
A univariate risk analysis between BBP/HT and covariates was performed.The variables with a p < 0.05 were included in a polytomic logistic regression analysis to evaluate the risk of both BBP and HT; SBP < 130 mm Hg and/or DBP < 85 mm Hg were the reference values.
To be consistent with the literature, we opted to keep birth weight as a covariate in the regression models.
Three models were constructed: the fi rst one (1) was adjusted for sociodemographic and clinical conditions at birth; the second one (2) include sociodemographic and clinical conditions and biochemical markers during young adulthood; and the third one (3) was adjusted for variables included in previous models.

RESULTS
The study sample at birth and at 23-25 years of age was comparable with the original population regarding birth variables: birth weight (p = 0.618) and length (p = 0.507), SGA (p = 0.513), and maternal age at delivery (p = 0.065).There were slight differences between those followed up and those who did not participate.There was a predominance of women (p = 0.004), preterms (p = 0.037), belonging to families with more qualifi ed occupations (p < 0.001), mothers with fi ve years or more of schooling (p < 0.001), married mothers (p < 0.001), and who did not smoke at the time of delivery (p < 0.001) (data not shown).
Males In the model 1 (birth) of the polytomic logistic regression analysis, the magnitude of the association of male gender with BBP/HT and birth length with BBP persisted after adjustment.In the model 2 (adulthood), the risk for BBP increased in males.The positive associations between alcohol consumption, sodium intake ≥ 2,000 mg/day, smoking, and triglyceride/HDL cholesterol ratio were no longer signifi cant, whereas additional adjustment removed part of the effect of BMI ≥ 30 kg/m 2 and central obesity on the risk of BBP and HT.Similarly, the positive association between high fasting blood glucose and HT was reduced but remained signifi cantly associated, while for BBP it was reduced and was no longer signifi cant after adjustment.
The model 3 shows the combined effect of birth plus adulthood variables.For both BBP and HT, the associations were mostly consistent with those observed in previous models for male gender, BMI ≥ 30 kg/m 2 and central obesity.Birth length ≥ 50 cm remained associated with BBP, while high fasting blood glucose was only associated with HT.BW was not associated with BBP or HT.The replacement of BW and birth length by PI was found to have an inverse association with BBP (adjusted RR [adjRR] 1.91, 95%CI 1.14;3.20)and HT (adjRR 2.29, 95%CI 1.17; 4.4) (Table 3).
There was no evidence of collinearity in the models (Table 3).Interaction terms were tested, but none was signifi cant at the 0.05 level.

DISCUSSION
One of a few studies conducted in Brazil and Latin American middle-income countries has investigated the suggested association between weight and size at birth and BBP/HT at 23-25 years of age in a population-based sample.Among indicators of size at birth, only birth length was independently associated with BBP at young adulthood.When BBP was replaced by prehypertension (SBP 120 mm Hg to 139 mm Hg or DBP 80 mm Hg to 89 mm Hg according to the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure [JNC 7]) 10 in the model 3, birth length showed a signifi cant positive association with BBP but to a lesser extent (RR 1.079, 95%CI 1.002;1.162)while BW was not associated with either prehypertension or HT.
Consistent with the Barker hypothesis, low birth weight is related to higher blood pressure later in life in prospective cohort studies.However, other studies have not been able to support this association. 13,20e prevalence of HT between 18 and 39 years of age was 7.2% in The National Health and Nutrition Examination Survey (NHANES) III, 15 a proportion that is close to that seen in our study (9.5%).
The higher prevalence of HT among men compared to women is noteworthy but at a lesser rate compared with similar studies. 22,8This may be attributable to a protective effect of estrogen, 19 higher smoking rates among men (most of the female subjects were non-smokers), higher male height (1.76; standard deviation [SD] 6.48 m for men vs. 1.62;SD 6.45 m for women, p < 0.001), elevated sodium intake ≥ 2,000 mg/day (70 vs. 52%, p < 0.001) and abnormal triglyceride/HDL cholesterol ratio (27% vs. 11%, p < 0.001) in males compared to females (data not shown).
There was a lower risk of central obesity (RR 0.78 [95%CI 0.69;0.89],p < 0.001) and biochemical markers among women compared to men (data not shown).
Increases in body fat may have different effects in women than men, and that a greater degree of adiposity must be achieved in women to obtain a signifi cant rise in blood pressure and an increase in a lipid risk profi le comparable with that of men.
Birth length was positively associated with BBP at the age of 23-25 years.The association between birth length and blood pressure is controversial.Studies from high-income countries, where intrauterine growth restriction is rare, have either failed to report an association or found inverse associations. 3,17On the other hand, studies from low-and middle-income settings have reported positive associations in accordance with the present fi ndings and a previous Brazilian study. 20his is in agreement with a recent systematic review which showed that the association between birth size and later blood pressure varies according to the level of economic development of the country of birth. 1 The biological mechanisms by which birth length may infl uence blood pressure are not yet clear.High blood pressure seen in individuals with higher birth length could be associated to greater height in adulthood.
This was confi rmed in a subsequent analysis in which height above the mean of this population (169.0 cm, SD 9.3 cm) was statistically associated with an increase in the risk of either BBP (OR 6.01, 95%CI 4.37;8.27,p < 0.001) or HT (OR 6.84, 95%CI 4.63;10.14,p < 0.001) compared with those at or below 169.0 cm (data not shown).
Further studies are needed to confi rm this association and to clarify potential biological mechanisms.
Subjects with BBP and hypertension showed as young adults an increased prevalence of cardiovascular risk factors for the metabolic syndrome, such as measures of generalized and visceral obesity, adverse glucose homeostasis, and dyslipidemia (Table 2).
Near fi fty percent of these young adults lead a sedentary life, which could contribute to higher adiposity and impaired cardiovascular health later in life, in accordance with similar studies. 13r study agrees with Bergvall et al 7 fi ndings that the association of size at birth and blood pressure appears not to be confounded by socioeconomic or family effects.
Subjects with hypertension during young adulthood showed high alcohol consumption, sodium intake ≥ 2,000 mg/day, and components of the metabolic syndrome (Tables 2-3).These results are consistent with the tracking concept (persistence) that risk factors remain relatively constant over time.Persistence of high blood pressure over time has been demonstrated in both pediatric and adult populations.This is an important public health issue because of the morbidity and mortality related to HT. Monitoring of BP in these individuals is recommended.
Some of the mechanisms underlying the associations in this study are unclear.Nevertheless, several recognized mechanisms could be postulated such as obesity, hyperinsulinemia/insulin resistance, high sodium intake, and infl ammatory markers. 11is is the fi rst study that included BBP, a new category of blood pressure classifi cation in Latin American countries.The sample followed up from birth to young adulthood showed better sociodemographic conditions than the original population.
The results of the present study expand a limited body of evidence supporting the hypothesis that BBP increases the risk of progression to full-blown hypertension in young adults.
To the extent that unmeasured risk factors present at baseline may be positively correlated with BBP, there is a chance that BBP was overestimated.Previous observational studies 10,25 have reported that the average BP in the second visit may be lower than in the initial visit due to regression to the mean or the familiarization of the participants to the clinic setting (white coat effect).Consequently, the prevalence of BBP/HT may be overestimated.
Our control variables have other shortcomings mainly because this study started collecting data more than 30 years ago when risk factors and levels of risk factors were less well established.Misreporting of alcohol consumption may have occurred.
The validity of associations between birth weight and health in later life might be due in part to inappropriate statistical adjustment for variables on the causal pathway which creates an artifactual statistical effect, known as the reversal paradox. 16e data obtained for this large community-based birth cohort support the hypothesis that birth length is positively and independently related to blood pressure in early adulthood but do not support an inverse association between birth weight and blood pressure.Adult risk factors explained most of the increase in the levels of BBP and HT.
Moreover, the simultaneous inclusion of current BMI with weight and length at birth in the regression model (indicating an effect of postnatal growth) did not substantially change the effect on the risk of BBP.
Young adults with either BBP or HT showed excess adiposity.These results, when analyzed in the context of the upward secular trends in adiposity and blood pressure in Brazilian youth, underscore the importance of controlling excess adiposity early in life in the general population.
Investigation of causative and synergistic interactions of high blood pressure and other cardiovascular risk factors and of precise physiologic mechanisms associated with early onset of high blood pressure is warranted.

Table 2 .
Sociodemographic, clinical and biochemical characteristics in young adulthood according to blood pressure levels.Ribeirão Preto Birth Cohort Study, Ribeirão Preto, SoutheasternBrazil, 1978Brazil,  -1979.a   .a 2 Values are numbers (percentage) * p-values are for the chi-squared test BBP: bordeline blood pressure; HT: hypertension; BMI: body mass index; HDL: high density lipoprotein Physical activity: Classifi cation based on the International Physical Activity Questionnaire scoring protocol; Abdominal obesity: Classifi cation based on the International Diabetes Federation (waist circumference altered if  90 cm for male and  80 cm for female)2; High fasting glycemia = Glycemia ≥ 100 mg/dl2

Table 3 .
Risk of borderline blood pressure and hypertension in young adults, adjusted by socio-demographic characteristics, clinical conditions and biochemical markers by polytomic logistic regression analysis in three models.Ribeirão Preto Birth Cohort Study, Ribeirão Preto, SoutheasternBrazil, 1978-1979.
2 Normotension as reference value.All models were adjusted for gestational age BBP: bordeline blood pressure; HT: hypertension; Physical activity: Classifi cation based on the International Physical Activity Questionnaire scoring protocol; Alcohol consumption = low:  31 g/day, high: > 31 g/day; Abdominal obesity: Classifi cation based on the International Diabetes Federation (waist circumference altered if  90 cm for male and  80 cm for female)2; High fasting glycemia = Glycemia ≥ 100 mg/dl2