Main

There is a considerable body of evidence that a stimulus during a critical or sensitive period can program later structure, physiology, or behavior of an organism (1). For example, we have shown that the diet fed to preterm infants in the neonatal period can program later neurodevelopmental status (2–4), and epidemiologic studies have suggested that poor growth performance in utero programs higher later blood pressure (5), although the mechanisms for these observations are not fully understood.

Three studies have reported lower blood pressure in infants delivered by cesarean section rather than vaginally (6–8): a 1956 study reported that it remained significantly lower for the first 3 wk of life (6); in a study by Kraus (7), it was lower at age 20 min and 5 d; and in a recent study, it was lower at 3–4 d (8). No study has investigated whether mode of delivery influences blood pressure later in life.

The aim of this study was to investigate whether mode of delivery influenced childhood blood pressure. We have tested this hypothesis using data already collected from a cohort of preterm infants whose blood pressure was measured at 7.5 to 8 y (9), and we also conducted a small pilot study to look for evidence that mode of delivery influences blood pressure in 7- to 9-y-old children born at term. Subsequently, we examined data from a cohort of children born in Tasmania, Australia, and recruited into the Tasmanian Infant Health Survey, which was established initially to examine risk factors for sudden infant death syndrome (10).

METHODS

In five centers in the United Kingdom (Cambridge, Ipswich, King's Lynn, Norwich, and Sheffield), 926 preterm infants born between 1982 and 1985, weighing <1850 g, were enrolled into trials of infant feeding within 48 h of birth. Extensive social, antenatal, and neonatal data were recorded prospectively by trained research nurses. Mode of delivery was recorded from maternal obstetric records and categorized as cesarean section, spontaneous vaginal, or forceps delivery, and presentation as cephalic or breech. In some cases (mostly when the infant was delivered elsewhere and transferred as a sick neonate to one of our recruiting centers), we were unable to determine reliably whether the cesarean section was elective, so all cesarean-delivered children are combined. Size for gestation was calculated as birth weight ratio (birth weight divided by mean birth weight for sex and gestation). Altogether 833 of 926 subjects survived into childhood. These studies were approved by the ethics committee in all centers. Observers were not aware that the blood pressure data would subsequently be used to test this hypothesis.

The UK cohort of singleton term infants was identified retrospectively from birth records at the Rosie Maternity Hospital in Cambridge, and their parents were approached by letter. We planned to recruit equal numbers of children born by breech, forceps, elective cesarean delivery (no labor), and spontaneous cephalic vaginal delivery. Inasmuch as breech deliveries were least common, we selected all children born that way, and by the closest date of birth for children born by other modes. The study was approved by the Cambridge ethics committee. The observer (K.K.) was aware of the hypothesis to be tested and identified potential subjects, so was not blinded to mode of delivery.

In the United Kingdom, children were seen at 7.5–8 y (preterm study; G.L., C.L.-P., and R.M.) and between 7 and 9 y (term study; K.K.). Weight and height were measured using Soehnle digital scales (to the nearest 100 g) and a Harpenden stadiometer (to the next succeeding millimeter). In the majority of cases (74% of preterm and 100% of term infants), an automated device Accutor (Datascope Corp, Paramus, NJ) was used to measure blood pressure, but during a period when two of these were not available to us, standard sphygmomanometers were used in some centers in the preterm study; diastolic pressure was then measured at Korotkov phase V. Blood pressure was measured three times after the child had rested for at least 5 min, and the mean of the last two values was used in analyses. The arm was resting on a surface level with the heart, and cuff size was selected on the basis of mid-upper arm circumference (11).

Children were recruited into the Tasmanian Infant Health Survey at birth on the basis of a risk score for sudden infant death syndrome (10). Information about obstetric, neonatal, social, and demographic factors, as well as detailed anthropometry at birth and in infancy, was collected by trained research nurses. Children were followed up at 7–8 y for a study focusing on the association between birth weight and childhood blood pressure. This was approved by the University of Tasmania Ethics Committee. The researchers were not aware that the blood pressure data would subsequently be used to test this hypothesis. Blood pressure was measured using a Dinamap Adult/Pediatric Vital Signs Monitor (Critikon, Arlington, TX) by four trained research nurses. Blood pressure was measured in the same way as in the UK studies, but the mean of three systolic and diastolic blood pressure values was calculated and used in analyses. Height and weight were measured in a standardized fashion using a stadiometer and bathroom scales calibrated daily with known weights, respectively. For this study we selected singleton children born after ≥37 completed wk of gestation. There were insufficient children born preterm to permit separate analyses.

Intergroup differences in blood pressure were assessed using t test and ANOVA;t test and χ2 were used to compare demographic variables. Regression models were used to adjust for intergroup demographic differences and (in the preterm cohort) the type of instrument used to measure blood pressure. Results for the preterm cohort were also tabulated separately for the two types of instrument.

RESULTS

Children Born Preterm

Blood pressure was measured in 756 of 833 (91%) surviving children from the preterm cohort. Mean age, weight, and height at the time of measurement were 7.64 y, 23.0 kg, and 121.8 cm, respectively: values for the four delivery groups differed very little. Birth weight was below the 10th centile for gestational age in 339 (44%) and above the 90th centile in 23 (3%) children. Those born by cesarean delivery had significantly higher mean gestational age and lower size for gestation than those born vaginally; mean birth weight was very similar (Table 1). Children born by cesarean section were least likely to have their blood pressure measured using the automated device. This was because of intercenter differences in cesarean section rates.

Table 1 Demographic data and mean (SD) systolic and diastolic blood pressure in relation to mode of delivery for preterm infants Comparison of cesarean vs vaginal delivery. *p = 0.003 by t test; †p < 0.0005 by t test; ‡p < 0.0001 by t test; §p < 0.05 by χ2. Comparison of cesarean vs spontaneous vaginal vs forceps vs breech. ¶p < 0.01 by ANOVA.
Full size table

Mean systolic and diastolic pressures were 100.4 ± 9.7 (SD) and 60.6 ± 7.3 mm Hg, respectively. Using the sphygmomanometer, mean systolic pressures were lower than with the automated device (Table 1). Such differences have been seen in other studies (12). Mean blood pressure for each delivery group is shown in Table 1. There was a significant trend (by ANOVA) to higher blood pressure in those born by breech versus forceps versus spontaneous vaginal delivery versus cesarean delivery (Fig. 1). No such trend was seen for diastolic pressure. Children born by cesarean section had significantly lower systolic blood pressure than those born vaginally (99.3 ± 10.0 versus 104.4 ± 9.4 mm Hg, 95% CI for difference, −0.69 to −3.46;p = 0.003). The influence of mode of delivery on later systolic blood pressure was less clear in the smaller group measured by sphygmomanometer.

Figure 1
figure 1

Systolic blood pressure (in mm Hg) of children in the preterm cohort by mode of delivery. Values are means ± SEM.

Full size image

In preliminary regression analyses, cesarean delivery was found to be significantly positively related to maternal age, having either toxemia or antepartum hemorrhage, or having a baby that was small for gestational age, but not to social class, evidence of birth asphyxia, or subsequent duration of ventilatory support. Regression models were constructed with type of delivery, gestation, birth weight ratio, sex, current weight, height, age, and the type of instrument used (automated versus conventional sphygmomanometer), as well as the factors associated with cesarean delivery (above) as independent variables and with blood pressure as the dependent variable. Even after such adjustment, systolic pressure was 2.4 mm Hg lower in children born by cesarean section rather than by vaginal delivery (95% CI, −3.9 to −0.9, p < 0.002).

There was no evidence that being born small for gestation or to a mother with toxemia influenced childhood blood pressure in this cohort, nor was there evidence of an interaction between gestational age and the influence of mode of delivery on later systolic pressure. For example, in children born with gestational ages <30 wk (n = 221), 30–32 wk (n = 319), and >32 wk of gestation (n = 216), mean systolic pressures were lower by 0.9, 2.8, and 2.1 mm Hg, respectively, in those born by cesarean section rather than vaginally.

Children Born at Term

UK Cohort.

Blood pressure was measured in 166 children, 44% of the 374 approached; 80 (48%) of these were boys. Mean values for age, weight, and height of children were 7.97 y, 26.6 kg, and 126.9 cm, respectively: they differed very little among the delivery groups. Altogether, 52 infants were delivered spontaneously vaginally, 48 by elective cesarean section, 21 by breech, and 45 by forceps delivery. Children delivered by cesarean section were more likely to be male and had significantly lower mean gestational age (Table 2).

Table 2 Demographic data and mean (SD) systolic and diastolic blood pressure in relation to mode of delivery for UK term cohort Comparison of cesarean vs vaginal delivery. *p < 0.05 by χ2; †p < 0.0001 by t test.
Full size table

Mean systolic and diastolic pressures were 100.5 ± 8.1 and 54.9 ± 6.7 mm Hg, respectively. Table 2 shows mean systolic and diastolic blood pressure for the different delivery groups. No significant differences were found even after adjusting for sex, gestation, current age, weight, and height, and there was no evidence of lower blood pressure in children delivered by cesarean section.

Among children delivered vaginally, those born by spontaneous cephalic delivery had lower systolic (100.2 ± 7.3 versus 101.5 ± 6.9 mm Hg; 95% CI, −3.8 to +1.4) and diastolic pressures (53.3 ± 7.0 versus 55.7 ± 5.9 mm Hg; 95% CI, −4.7 to −0.05;p = 0.046) than those born by breech or by forceps delivery.

Tasmanian Cohort.

There were 651 children recruited into the Tasmanian Infant Health Survey in 1989–1990, and followed up in 1996 to 1997, who fulfilled the selection criteria for this study. Mode of delivery was known for 650 of them, and 73% (474 of 650) were boys. This was because male sex is a risk factor for sudden infant death syndrome and was one of the scoring factors (below) used to select subjects for the study (10). Mean age, weight, and height were 8.1 y, 28.3 kg, and 128.3 cm, respectively; values were very similar in the three delivery groups. Mean systolic and diastolic pressures were 107.3 ± 8.9 and 63.8 ± 6.4 mm Hg, respectively. There were 562 children born by vaginal delivery, 48 born by emergency cesarean section, and 40 born by elective cesarean section. Characteristics and mean systolic and diastolic blood pressures for these delivery groups are shown in Table 3. No differences in blood pressure were seen among the delivery groups, either before (p = 0.80) or after adjustment for the scoring factors (sex, maternal age, season of birth, mothers' intention to breast-feed, duration of second-stage of labor, and birth weight) used to select subjects for entry to the Tasmanian Infant Health Survey (10) (p = 0.83) or additionally for gestation (p = 0.84), current age (p = 0.70), current height (p = 0.69), or current weight (p = 0.80).

Table 3 Demographic data and mean (SD) systolic and diastolic blood pressure in relation to mode of delivery for Tasmanian term cohort
Full size table

DISCUSSION

In a cohort of 756 children born with birth weight <1850 g, systolic blood pressure was significantly lower at 7.5 to 8 y in those born by cesarean delivery (6–8). There was a gradient in systolic pressure, with the lowest values found in subjects born by cesarean delivery and the highest in those born by breech delivery (Fig. 1).

In light of our findings in the preterm cohort, we undertook a pilot study of 166 children born at term in the United Kingdom and found no evidence of lower blood pressure in those born by cesarean section. The marginally higher diastolic pressure in children born by breech or by forceps delivery, rather than by spontaneous cephalic delivery, should be interpreted with caution given the small sample size and lack of association across the four categories of mode of delivery. We also analyzed data from a cohort of 650 singleton children born after ≥37 wk of gestation in Tasmania, Australia, and again found no evidence that blood pressure was lower in those born by elective cesarean section.

There is support from animal studies for the concept that mode of delivery can have a long-term programming influence on function in adult life. Young adult rats born by uncomplicated cesarean section have been shown to have alterations in steady-state levels of dopamine in their CNS (13) and greater responses to amphetamine or repeated stress in terms of dopamine release and behavior (14), in comparison with vaginally born controls.

During vaginal delivery, the infant is exposed to major uterine contractions, and this results in substantially increased circulating concentrations of catecholamines (15–21), chromogranin A (15), neuropeptide Y (21), angiotensin II (22), and cortisol (23). Mode of delivery significantly influences these, with consistently lower values in newborn infants delivered by elective cesarean section rather than vaginally. For example, Falconer and Lake (16) showed that, in term infants, umbilical arterial norepinephrine values increased linearly among infants born by cesarean section, spontaneously vaginal delivery, forceps delivery, or breech delivery, a similar gradient to that found for blood pressure by mode of delivery in the preterm infants we studied (Fig. 1). There is evidence that such substances could program later vascular function at a critical period before full-term, and there is evidence from animal studies that the programming influence of an insult depends on stage of gestation. For example, giving dexamethasone to pregnant sheep programmed later raised blood pressure in offspring exposed in utero at 22–29 d of gestation, but not in those exposed at 59–66 d (24). Furthermore, rats exposed in utero to dexamethasone in the last third of gestation had a 25% increase in hepatic expression of glucocorticoid receptor mRNA and a 60% increase in phosphoenolpyruvate carboxykinase mRNA in adult life, whereas exposure in the first two thirds of gestation had no such influence (25).

The difference in our findings between the term and preterm cohorts could be because the long-term programming influence of mode of delivery on later blood pressure was restricted to or greater during a critical or sensitive period of development preterm. Although exploratory analyses failed to demonstrate that the association found in the preterm cohort was gestation dependent, it is possible that the mechanism for the observed influence of delivery mode on later blood pressure only operates before full-term.

In the preterm study, 44% had birth weights below the 10th centile, compared with only one child in the UK term cohort and 19% in the Tasmanian term cohort. This raised the possibility that the effect was seen mainly in children born small for gestational age. However, the mean difference in systolic pressure between cesarean section and vaginally delivered preterm children was similar in those born small for gestational age and those born with more appropriate birth weights (2.2 versus 1.95 mm Hg).

The fact that we used two different instruments in the preterm cohort could have influenced our findings, because blood pressure was higher with the automated devices and these were used less often to measure blood pressure of children born by cesarean section or spontaneous vaginal delivery. However, after adjustment in regression models for the type of instrument used, an intergroup blood pressure difference remained. Furthermore, data are shown separately for the two types of instrument in Table 2, and in the larger group measured with the automated device, the same trend was seen.

The difference in findings between preterm and term cohorts may have been because of inadequate sample size in the latter (although data from the two term cohorts were consistent), or differences in selection bias or unidentified confounding in the term and preterm studies. However, our data raise the hypothesis that there is a sensitive period before full-term for programming of later blood pressure by factors associated with mode of delivery. This needs to be tested in another cohort.

We previously reported, from analyses of data from the same preterm cohort, that we were unable to find the expected association between lower birth weight or size for gestation and higher later blood pressure in children born before 34 wk of gestation (9). Indeed, we found an insignificant trend to higher blood pressure with increasing size for gestation. Infants born small for gestational age were more likely to be born by cesarean section, and we have found here that this type of delivery is associated with lower childhood blood pressure. We considered the possibility that this could be the reason why we failed to find the expected blood pressure–size for gestation association. However, our findings (9) were not changed by inclusion of mode of delivery in the regression models described. Our findings relating to maternal smoking in pregnancy and later blood pressure in the offspring (26) were similarly unchanged after inclusion of mode of delivery in regression models.

The study of children born preterm may give important insights into the development and programming of vasomotor tone, and despite the potential for confounding in this group (27), we suggest that this opportunity should not be ignored.