Introduction

Hypertension is a major cardiovascular risk factor in adults,1, 2, 3, 4 and the risk of developing high blood pressure (BP) during adolescence can be predicted based on BP in childhood.5 Children with high BP are at an increased risk of sustained hypertension as well as metabolic syndrome in adulthood.6 Meanwhile, self-measurement of BP at home (HBP), which has been recognized as a useful tool for the accurate diagnosis and treatment of hypertension in adults,7, 8 would be superior to the conventional BP (CBP) measurements in children as well, regarding the long-term reproducibility and the prognostic significance of the presence of subclinical end-organ damage.9, 10 Nevertheless, current data on HBP in children are limited.

Pregnancy-induced hypertension (PIH), diagnosed by increased BP with or without proteinuria (preeclampsia or gestational hypertension (GH)), is a common complication in pregnancy.11, 12 PIH occurs in ~7–12% of all singleton pregnancies worldwide.13 Even in children and adolescents, the offspring from preeclampsia or GH mothers reportedly have higher ambulatory BP or CBP.14, 15, 16, 17, 18 To the best of our knowledge, however, no studies have investigated whether maternal hypertension during pregnancy affects self-measured HBP in mothers and their offspring. We aimed to investigate the association between maternal hypertensive disorders during pregnancy and both HBP and CBP in young children and mothers after 7 years of the delivery.

Methods

Participants

The Tohoku Study of Child Development, conducted in the northeastern region of Japan, is a prospective birth cohort study investigating the health and development of children. Details of the Tohoku Study of Child Development have been described previously.19, 20 From January 2001 to September 2005, we enrolled 1566 pregnant women. Mothers were eligible if they had no severe diseases such as thyroid dysfunction, hepatitis, immune deficiency, malignant tumors or mental diseases; if they had not undergone in vitro fertilization; and if they spoke Japanese as their native language. A total of 1348 babies fulfilled the inclusion criteria as shown in Figure 1: the absence of congenital anomalies or severe diseases and a singleton birth at term (36–42 weeks of gestation) with a birth weight of >2400 g. Of those, 953 children at 7 years of age and their mothers were surveyed in a quiet examination room by 31 March 2013. We excluded 138 children from the current analysis because either their BP was unavailable (n=71), their HBP in the morning was measured for <3 days (n=37), the mothers were taking antihypertensive medication during the survey (n=3), or the required characteristics or outcomes at the 7-year survey for analysis were not fully available (n=27). We further excluded mothers who experienced preeclampsia because of the limited number (n=2). Therefore, the study included 813 mother-offspring pairs (Figure 1). There were no significant differences in gestational age, birth weight and height, maternal age, prevalence of sex or parity between the 813 analyzed children and the 138 who were not included (P>0.12). We obtained written informed consent from the mothers, and the Ethics Committee of Tohoku University Graduate School of Medicine approved the study.19, 20

Figure 1
figure 1

Flowchart of the study population.

Full size image

Information on mothers during pregnancy

Characteristics of mothers during pregnancy were obtained from medical records during the perinatal period. A urinary protein dip-stick test was performed at every 2- to 4-week visit to the hospital. CBP during pregnancy was also measured at every visit after the participant rested for a few minutes in a sitting position. The median number of visits to the hospital during pregnancy was 12 (interquartile range, 11–13). Information on smoking and drinking habits during pregnancy, parity and education was collected by self-administered questionnaires at enrollment or 4 days after the delivery.

Hypertension during pregnancy was defined as CBP ⩾140 mm Hg systolic and/or BP ⩾90 mm Hg diastolic. Proteinuria was diagnosed if the protein reading on dip-stick testing was 1+ or more.21 Preeclampsia was defined as de novo hypertension on at least two occasions after 20 weeks of gestation accompanied by proteinuria.11, 12 GH was defined as de novo hypertension on at least two occasions after 20 weeks of gestation, but without proteinuria.11, 12 We found no participants with superimposed preeclampsia, which was defined as fulfilling preeclampsia criteria along with hypertension or proteinuria on at least one occasion before 20 weeks of gestation.11, 12 We therefore defined PIH mothers as those with preeclampsia and GH.

BP measurement and other information on the 7-year survey

On the 7-year survey, the CBP and HBP in mothers and children were measured using an OMRON HEM-7080IC (Omron Healthcare, Kyoto, Japan), a device based on the cuff-oscillometric method that generates a digital display of both systolic and diastolic BP values. The device used in the present study is equivalent to the OMRON HEM-705IT and has been validated.20, 22 We downloaded the BP readings directly to a computer from the device without any measurement bias.20 Small, standard and large cuffs had inflatable bladders of 16 × 9 cm, 22 × 12 cm and 30 × 15 cm, respectively, and were used in mothers and offspring depending on upper-arm circumference (under 22 cm, 22–32 cm and over 32 cm, respectively).

The CBP was measured by trained survey staff; the measurements were performed once on mothers and children in a seated position after resting for at least 2 min. The study participants were also instructed and trained how to measure their HBP, and they were asked to take their own measurements for 2 weeks. Their HBP was measured once in the morning before breakfast, in the sitting position, within 1 h after awaking and after 2 min or more of rest. HBP in each child was measured with the mother’s assistance. Although many participants measured their HBP twice or more per occasion, we used the first value from each measurement in our analysis to exclude individual selection bias, and the mean of all first measurements were used as HBP. The initial HBP was defined as first HBP measurement from the first day for each participant. Maternal conventional hypertension was defined as CBP ⩾140 mm Hg for systolic and/or BP ⩾90 mm Hg for diastolic, and maternal home hypertension as HBP ⩾135/⩾85 mm Hg.

The heights and weights of the children at 7 years of age were measured using a KS-502Gp automatic analyzer (Kansai Seiki, Kusatsu, Japan), whereas the participants wore light indoor clothes and no shoes. We calculated the body mass index (BMI) using the following formula: weight (kg)/height (m)2.

Statistical analyses

We used SAS software (version 9.3, SAS Institute Cary, NC, USA) for database management and statistical analysis. All data are presented as the mean±s.d. unless otherwise stated. Statistical significance was established at P<0.05.

The variables for the characteristics of the study participants and the children’s outcomes at the 7-year survey were compared using Fisher’s exact test or Student’s t-test appropriately. For the comparison of CBP and HBP between the two groups, we used analysis of covariance adjusted by the following factors according to the previous studies:15, 20, 23 for the children, sex, birth weight, age at 7-year survey, BMI at 7-year survey, parental education status, maternal age at delivery, parity, maternal smoking during pregnancy, maternal BMI before pregnancy, cesarean section and gestational week; for the mothers, age and BMI at 7-year survey, education (⩾13 or <13 years), parity, family history of hypertension, and current drinking and smoking status. BP variability, represented by s.d. and coefficient of variation, was also compared between the two groups.

Along with the whole-population analysis, we conducted matched case–control analysis to equalize baseline characteristics in the participants. Each mother diagnosed with PIH was paired with two controls matched for maternal age at delivery (an age range of ±6 years with respect to each case), maternal BMI before pregnancy (±6 kg m−2), parity and residence—factors thought to have great influence on maternal PIH; the remaining controls were discarded.

Results

Of the 813 children at age 7 (mean 84.7±1.7 months) and their mothers included in this study, 28 mothers (3.4%) were diagnosed with GH. On the 7-year survey, 42 (5.2%) and 46 (5.7%) mothers had conventional and home hypertension, respectively. Table 1 shows the characteristics of the GH, non-GH and matched non-GH groups of mothers and children. Maternal BMI before pregnancy and on the 7-year survey in the GH group were 23.8±4.7 and 23.8±4.9 kg m−2, respectively, which were significantly higher than those in the non-GH group, 21.0±2.8 and 21.4±3.0 kg m−2, respectively (P=0.004/0.01). The rate of family history of hypertension was also higher in mothers with GH than those with non-GH (67.9% vs. 39.0%, P=0.003). No significant differences were observed in other variables (P>0.12). After matching with maternal age, maternal BMI, parity and residence, we found no significant differences in their characteristics between 28 matched cases and 56 controls in the case-matching analysis (P⩾0.07) except for drinking during pregnancy (P=0.03).

Table 1 Characteristics of mothers and children by GH categories
Full size table

In our population, the 95th percentile of children’s conventional and home BP were 109/72 and 106/65 mm Hg, respectively. After adjusting for the confounding factors, there were no significant differences (P⩾0.38) in children’s BP between the non-GH group (93.1±9.2/57.4±9.1 mm Hg for CBP and 94.1±6.9/56.1±5.7 mm Hg for HBP) and the GH group (90.0±6.4/56.7±7.0 mm Hg for CBP and 93.5±5.7/55.9±4.1 mm Hg for HBP) except for systolic CBP (P=0.04). The mean (95% confidence intervals) differences between the two groups were 3.6 (0.2–7.0)/0.8 (−2.6–4.2) mmHg for CBP and 1.1 (−1.3–3.5)/0.6 (−1.5–2.7) mmHg for HBP (Figure 2). Maternal HBP as well as CBP in the GH group (123.7±20.3/79.8±12.8 mm Hg for CBP and 120.3±12.7/76.8±10.3 mm Hg for HBP) was significantly higher than in the non-GH group (109.5±12.6/70.5±9.6 mm Hg for CBP and 110.4±10.9/68.6±8.3 mm Hg for HBP); the mean (95% confidence intervals) differences were 10.6 (6.3–15.0)/6.9 (3.5–10.3) mmHg for CBP and 6.8 (3.2–10.5)/6.1 (3.3–9.0) mmHg for HBP after applying multivariable adjustment for possible confounders (P<0.0002) (Figure 3). The results were essentially similar when the initial HBP instead of the average HBP was used (Table 2). No significant differences in the s.d. and coefficient of variation of systolic/diastolic HBP were observed between the two groups of children (P>0.22) and mothers (P>0.07).

Figure 2
figure 2

Children’s CBP and HBP in the non-GH group (open bar; n=785) and the GH group (solid bar; n=28). Panel a, systolic blood pressure; panel b, diastolic blood pressure. CBP and HBP were analyzed using analysis of variance adjusted for sex, birth weight, age at 7-year survey, children’s body mass index at 7-year survey, parental education status, maternal age at delivery, parity, maternal smoking during pregnancy, maternal body mass index before pregnancy, cesarean section and gestational week. Error bars denote s.d. P-values express the differences between GH and non-GH after adjusting for sex, birth weight, age at 7-year survey, children’s body mass index at 7-year survey, parental education status, maternal age at delivery, parity, maternal smoking during pregnancy, maternal body mass index before pregnancy, cesarean section and gestational week. GH, gestational hypertension; CBP, conventional blood pressure; HBP, home blood pressure.

Full size image
Figure 3
figure 3

Maternal CBP and HBP in the non-GH group (open bar; n=785) and the GH group (solid bar; n=28). Panel a, systolic blood pressure; panel b, diastolic blood pressure. CBP and HBP were analyzed using analysis of variance adjusted for maternal age and body mass index at 7-year survey, education status, parity and current drinking and smoking status. Error bars denote s.d. P-values express the differences between GH and non-GH after adjusting for maternal age and body mass index at 7-year survey, educational status, parity, family history of hypertension, and current drinking and smoking status. GH, gestational hypertension; CBP, conventional blood pressure; HBP, home blood pressure.

Full size image
Table 2 Comparison of initial home blood pressure at 7-year survey by GH categories
Full size table

Figure 4 shows the 1:2 matched case–control analysis. Although there were no significant differences between the matched non-GH and GH groups in terms of their characteristics (Table 1), maternal HBP in the GH group was significantly higher than in the matched non-GH group even after adjustment (P<0.002). However, children’s HBP was statistically at the same level between the two groups (Figure 4; P>0.41). In children, only systolic CBP was significantly higher in the non-GH group than in the GH group (3.9 (95% confidence intervals: 0.8–6.9) mmHg, P=0.01).

Figure 4
figure 4

Maternal and children’s HBP in the matched non-GH group (open bar; n=56) and the GH group (solid bar; n=28). Panel a, systolic HBP; panel b, diastolic HBP. Error bars denote s.d. P-values express the differences between GH and non-GH after adjusting for the following variables: for children, sex, birth weight, age at 7-year survey, body mass index at 7-year survey, parental education status, maternal smoking during pregnancy, cesarean section and gestational week; for mothers, education status and current drinking and smoking status. GH, gestational hypertension; HBP, home blood pressure.

Full size image

Results were confirmed when we included two preeclampsia mothers and their children into the study population as well as the evaluated BP levels in 30 PIH mothers and their children (Table 3).

Table 3 Blood pressure levels at 7-year survey in relation to pregnancy-induced hypertension
Full size table

Discussion

This is the first study investigating the association between GH and self-measured HBP as well as CBP captured 7 years after delivery. In this birth cohort population, maternal GH did not affect BP levels or BP variability in offspring. However, experience with GH affected the mother’s own high HBP levels even when multivariable-adjustment models were applied. We obtained confirmatory results when the initial HBP value was used. The use of self-measured HBP has been reported to have a better reproducibility and prognostic value than CBP in both adults8, 24 and children.9, 10 Our findings suggest that measurement not only by CBP but also by self-measured HBP provides more useful prognostic information for women who experienced GH at an earlier or middle age.

Women with GH or preeclampsia have an increased risk of cardiovascular disease,25, 26 cerebrovascular disease23, 25 and hypertension23 in later life. Our results were consistent with previous studies. Mori et al.27reported that BP in women with preeclampsia significantly decreased a month after delivery, whereas BP remained higher than in normotensive mothers. Furthermore, nighttime systolic BP in the ambulatory setting was significantly (P=0.030) elevated in Swedish women with preeclampsia compared with those without preeclampsia even 11 years after delivery, although this was not the case for the other BP data (P⩾0.073).28 According to the guidelines of the International Society for the Study of Hypertension in Pregnancy, PIH is defined as hypertension with or without proteinuria that resolves within 12 weeks postpartum.12 Although we did not collect BP data beyond 12 weeks after delivery, HBP and CBP in the GH women were significantly higher than those in non-GH women even 7 years after giving birth. Our findings support the importance for women who developed hypertension during pregnancy to manage their own health even in middle age or earlier.

The Avon Longitudinal Study of Parents and Children is a large, population-based pregnancy cohort.29, 30 A total of 13 617 mothers with a live singleton birth consented to have their obstetric data abstracted from medical records. The mean (s.d.) age of the eligible 2888 offspring was 208.5 (11.2) months old, and the proportion of preterm delivery before 37 weeks was 4.4%.15 CBP in the Avon Longitudinal Study of Parents and Children was significantly higher in 17-year-old offspring of mothers with GH (120.5±11.3/66.0±7.2 mm Hg) or preeclampsia (120.2±10.1/66.6±7.0 mm Hg) compared with non-PIH (117.6±10.4/64.5±6.8 mm Hg) even when adjusting for potential confounding factors in multivariable linear regression analysis.15

Inconsistent with earlier studies,15, 31 including the Avon Longitudinal Study of Parents and Children, there were no significant associations between maternal GH and HBP in children at 7 years old in the present study. These findings were confirmatory when the 1:2 matched-case analyses were performed. We obtained contradictory findings for systolic CBP (93.9 mm Hg in matched non-GH group vs. 90.0 mm Hg in GH group), although this opposite result might be due to the lower number of CBP measurements and inaccuracy of BP information compared with home measurements.9 In the present study, the rate of children with a preterm birth was 1.6%. This prevalence is lower than the Avon Longitudinal Study of Parents and Children, in which 4.4% of children were preterm, whereas the inclusion criteria of birth weight in our study was >2400 g. Preeclampsia at term, accompanied by offspring that appear to be unaffected by the condition, may represent a mixture of conditions, ranging from mild preeclampsia with moderate placental involvement to hypertensive conditions without placental dysfunction.32 Offspring born to preeclamptic mothers at term have fetal growth similar to that of babies born to normotensive mothers.33 Our study includes fewer preterm births or low-birth-weight infants compared with previous studies, suggesting that maternal GH might have little influence on BP in 7-year-old children who were born at term and had normal birth weight.

The present study must be interpreted with the context of its limitations. First, because our study population consisted of residents from specific Japanese districts, our findings might not be generalizable to other populations. Second, we could not collect mothers’ BP before pregnancy and after 12 weeks postpartum. According to the guidelines, the definition of transient hypertension of pregnancy is that maternal CBP returns to normal by 12 weeks postpartum.11, 12 Some mothers with chronic hypertension might have inadvertently been included in our present analysis. However, we excluded mothers with severe complications at the entry period and those who were under antihypertensive drug medication at the 7-year survey. We may have included mothers with white-coat hypertension in the GH group because we did not measure maternal HBP during pregnancy. Third, we could not assess GH and preeclampsia separately because of the limited number of preeclamptic women (n=2), and there were only 30 women with PIH in our population. We might have overlooked small differences in children’s BP between GH and non-GH because it may reflect a beta error due to the comparably small number of women with GH. Further studies with large sample sizes will be required to confirm the associations of maternal GH and preeclampsia with HBP in children and whether the differences in population characteristics (for example, birth weight, rate of preterm) affect these associations. Finally, our findings might not be applicable to preterm and low-birth-weight children born to PIH mothers who exhibit preterm birth and intrauterine growth restriction.34, 35 Because our study population consists primarily of term children, our results could be applied to term children born to PIH mothers.

In conclusion, maternal GH did not affect HBP in offspring with normal birth weight and at term but strongly affected maternal HBP even 7 years after giving birth. Preventive strategies focusing on PIH may lead to lowering PIH women’s morbidity and mortality of cardiovascular disease in the future. Our findings support the importance for women who experienced hypertensive disorders during pregnancy to manage their own health even in middle age or earlier and indicate that growth environment has an important influence on HBP in children.