Antihypertensive treatment during pregnancy and functional development at primary school age in a historical cohort study


MMHJ van Gelder, Department of Epidemiology, Biostatistics, and HTA (HP 133), Radboud University Nijmegen Medical Centre, PO Box 9101, 6500 HB Nijmegen, the Netherlands. Email


Please cite this paper as: Pasker-de Jong P, Zielhuis G, van Gelder M, Pellegrino A, Gabreëls F, Eskes T. Antihypertensive treatment during pregnancy and functional development at primary school age in a historical cohort study. BJOG 2010;117:1080–1087.

Objective  To determine the functional development of children born after treatment of mild-to-moderate gestational hypertension with labetalol versus methyldopa, and no antihypertensive treatment.

Design  Historical cohort study.

Setting  Twelve Dutch hospital departments of obstetrics.

Population  Live-born children born in these hospitals and prenatally exposed to labetalol, methyldopa, or bed rest because of mild-to-moderate gestational hypertension.

Methods  Central nervous system development was measured with standard tests at 4–10 years of age. Linear regression techniques and Pearson’s chi-square tests were used to compare the groups with regard to the outcome measures.

Main outcome measures  Intelligence quotient (IQ), concentration, motor development, and behaviour at primary school age.

Results  A total of 202 children were included in the analyses. More children exposed to labetalol had attention deficit hyperactivity disorder (ADHD) than those exposed to methyldopa (OR 2.3; 95% CI 0.7–7.3), or those born to women who had been admitted for bed rest (OR 4.1; 95% CI 1.2–13.9). Sleeping problems seemed to be reported more frequently after prenatal methyldopa exposure than after exposure to labetalol (OR 3.2; 95% CI 0.6–16.7) or bed rest (OR 4.5; 95% CI 0.9–23.2), although the differences were not statistically significant. Test scores on other aspects of functional development did not differ between the three groups.

Conclusions  In this hypothesis-generating study, labetalol exposure in utero seemed to increase the risk of ADHD among children of primary school age, whereas prenatal methyldopa exposure might influence sleep. Further studies with appropriate sample sizes are warranted to determine the long-term effects of antihypertensive medications.


Approximately 7% of pregnant women experience a significant rise in blood pressure during pregnancy [pregnancy induced hypertension (PIH) or pregnancy aggravated hypertension (PAH)].1 Although medical treatment is indicated for women whose diastolic blood pressures increase to 110 mmHg or more,2,3 the question has been raised whether treatment is necessary in mild-to-moderate cases of gestational hypertension (diastolic blood pressure 90–110 mmHg), especially in the absence of proteinuria.4 In addition, the rapidity of the increase in blood pressure may be more important than the actual level of the blood pressure in PIH/PAH. Antihypertensive treatment for mild-to-moderate hypertension is beneficial for the mother by reducing the risk of developing severe hypertension, but pregnancy outcome has been found to be similar or even worse: i.e. more small-for-gestational-age infants, when comparing antihypertensive drug(s) with placebo or no antihypertensive drug treatment.5,6 If treatment of mild-to-moderate gestational hypertension does not improve pregnancy outcome, but only improves the maternal condition, knowledge of the long-term effects of prenatal exposure is necessary for a rational treatment of PIH/PAH in which the risks for the infant are balanced against the benefits for the mother. However, information on the long-term effects of antihypertensive drug treatment during pregnancy in humans is scarce.

In the period 1983–1987, Dutch obstetricians gradually switched from methyldopa to labetalol as the first choice of treatment, because methyldopa was found to have central side effects in the women that were absent with the combined peripheral α- and β-blocking agent labetalol. Currently, both drug treatments are still being used in the treatment of mild-to-moderate gestational hypertension.7 In a follow-up study of a large clinical trial on the effectiveness of methyldopa versus no antihypertensive treatment, no differences were found in the functional development of children at 4 and 7.5 years of age.8,9 However, it has not been established whether labetalol is safe with regard to fetal central nervous system development. Ten out of 14 children exposed to labetalol examined at the age of 6 months seemed to develop normally,10 but no comparison was made with unexposed children. No other follow-up studies on the developmental effects of prenatal exposure to these antihypertensive medications have been published to date.

During the switch from methyldopa to labetalol, the choice for either one of these treatments seemed to be predominantly a reflection of the change in preference of the attending obstetrician. The situation in the 4-year period therefore had many characteristics of a natural experiment. This gave us the opportunity to conduct a hypothesis-generating historical cohort study on the functional development of children born between 1983 and 1987 to mothers treated with labetalol or methyldopa for gestational hypertension to learn about the functional development of children. Although methyldopa is considered safe with regard to developmental outcome,8,9 a group of children born to mothers with gestational hypertension who were not given any antihypertensive drug treatment (i.e. who were admitted for bed rest) was added as a second reference group.


The departments of obstetrics of all nine teaching hospitals and ten general hospitals in the Netherlands were invited to participate in this study. We only included hospitals for which data on the highest measured diastolic blood pressure during pregnancy could be retrieved through the hospital registry or records of the Dutch Obstetrics Registry. This information was not available for four general hospitals. Twelve of the eligible departments finally participated in the study: seven teaching hospitals and five general hospitals.

The hospital records of women whose diastolic blood pressure exceeded 90 mmHg at least once during pregnancy, and who delivered in one of the participating hospitals between 1 January 1983 and 31 December 1987 were reviewed. Only women aged between 20 and 35 years, diagnosed with PIH or PAH, with singleton pregnancies, who delivered a live-born child after at least 32 weeks of gestation, and who did not have diabetes or epilepsy in the index pregnancy, were eligible. We only included women treated with labetalol or methyldopa for PIH or PAH in the study. Women who used both medications were excluded. A sample taken from the hospital records of women admitted for bed rest because of PIH or PAH was included as a second reference group. The medical ethical committee approved the study protocol. In addition, a written informed consent was obtained from every participant.

The women were sent a letter with an invitation to participate in this study from the department where they had delivered. A reminder was sent to the non-responders, and addresses were traced (if possible) when the letter was returned as undeliverable. In 1991–1992, the children whose parents agreed to participate were tested at the child’s home by one of six (university-)trained child-psychology test assistants, who were unaware of the exposure status of the children. Six developmental aspects were assessed: namely the child’s concentration (Bourdon Vos test),11 intelligence quotient [tested with the shortened Revision Amsterdam Child Intelligence Test (RAKIT)],12 memory (a subtest of RAKIT was used for visual memory, the ten-words test was used for auditive memory, and verbal word-pair associations were used for associative memory),12,13 gross motor development (subtest of Hamm Marburger Körper Koordinationstest für Kinder),14 and fine motor development (prick test).13

Information about the pregnancy and development of the child was obtained from hospital records and the parents using a standardised questionnaire. We collected data on child behaviour from the parents using the Dutch version of the Child Behaviour Checklist, and from the child’s teacher using the Dutch version of the Teacher Report Form.15 We arranged the questions on the Teacher Report Form about the habits and characteristics of the children compatible with Attention Deficit Hyperactive Disorder (ADHD) in three subcategories: ‘lack of concentration’, ‘hyperactivity’ and ‘distractedness’. Children were classified as having ADHD if they conformed to at least three items per subcategory, in accordance with the American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders III (DSM-III) definition of ADHD.16 The duration of symptoms was not included in the criteria for ADHD.

In the statistical analyses, we compared the outcomes in the labetalol group with those in the methyldopa and bed-rest groups. Test scores were adjusted for age at testing where appropriate, as children in the labetalol group were younger on average than children in the methyldopa and bed-rest groups. Potential confounding factors (maternal education, parity, alcohol use during pregnancy, proteinuria, start week of treatment, delivery complications, birthweight, and otitis media, vision problems and concussion in childhood) were discarded if they were equally distributed over the exposure groups. Potential confounding factors were discarded from the models when their removal did not change the effect estimate for antihypertensive drug treatment by more than 10% (manual backward selection). Continuous outcomes were compared using multivariable linear regression analysis. We used Pearson’s chi-square tests to study the associations between the three exposures and binary outcome data. There were insufficient events to calculate adjusted odds ratios. All statistical analyses were performed using spss v12.0.1. for Windows (SPSS Inc., Chicago, IL, USA).


In total, 4000 hospital records were reviewed. After applying the inclusion and exclusion criteria, and drawing a random sample of women treated with bed rest, 355 mother–child pairs were included: 99 treated with labetalol, 101 treated with methyldopa, and 155 treated with bed rest. A total of 275 out of the 355 women included in the study responded to the invitation. Of these, 203 were eligible and agreed to participate (follow-up rate 57%). However, checking by telephone revealed that about half of the 80 non-responders did not live at the address the letter had been sent to. One child exposed to methyldopa was not tested for practical reasons. Therefore, 202 subjects were left for analysis. Some characteristics of the participants, refusers and non-responders (including the untraceables) are shown in Table 1.

Table 1.   Characteristics of participants, non-participants, and non-responders in a Dutch historical cohort study on functional development after prenatal exposure to antihypertensive treatment
CharacteristicParticipants (= 202)Refusers (n = 72)Non-responders (= 80)
Maternal age (years)27.90.326.90.526.80.4
Group (%)
Labetalol (n = 99)29.1 22.2 30.8 
Methyldopa (n = 100)30.0 25.0 26.9 
Bed rest (n = 155)40.9 52.8 42.3 
Birthweight (g)27076128522462993114
Gestational age (weeks)37.90.338.50.338.20.3
Age of the child on 1 January 19937.

The maternal and pregnancy characteristics of the participants are presented in Table 2. The bed-rest group was different from both medication groups. These women had later onset of illness and were more often diagnosed with PIH, and they did not receive any other drugs for pregnancy hypertension, whereas almost 20% in the labetalol and methyldopa groups received phenobarbital. Furthermore, the bed-rest group consisted of less severe cases of PIH/PAH. In addition, the infants in the bed-rest group were less likely to be born very preterm compared with infants in the medication groups. However, they were more often small for gestational age (using cut-off points for Dutch infants as reported by Visser et al.17). The three groups were comparable on the modes of delivery and most aspects of the occurrence of pregnancy complications, including breech presentation, fetal distress, premature rupture of the membranes, umbilical cord entanglement, and meconium-stained amniotic fluid. In the methyldopa group, treatment started earlier in pregnancy than in the labetalol group, and more often they received other medical treatments for hypertension, whereas their maximum diastolic blood pressure was lower. In addition, the proportion of primiparae in the methyldopa group was lower than in the bed-rest group.

Table 2.   General characteristics of the pregnancies of participants in a Dutch historical cohort study on functional development after prenatal exposure to antihypertensive treatment
CharacteristicLabetalol (= 58)Methyldopa (= 61)Bed rest (N = 83)
  1. DBP, diastolic blood pressure.

  2. *Educational level: low, ≤10 years of education; intermediate, trade school > 10 years; high, higher education > 10 years.

  3. **Birthweight lower than 2.3rd percentile for gestational age using the Dutch reference curves for birthweight by gestational age.17 Data were missing for 24.1% of labetalol-exposed infants, for 16.4% of methyldopa-exposed infants, and for 7.2% of bed-rest infants.

  4. ***Includes breech presentation, fetal distress, premature rupture of the membranes, umbilical cord entanglement, and meconium-stained amniotic fluid.

Maternal age in years28.
Maternal education (%)*
Low38.6 23.6 38.7 
Intermediate40.4 54.6 41.3 
High21.0 21.8 20.0 
Primiparae (%)68.4 55.7 76.5 
Birthweight (g)261390927288712761844
Gestational age (%)
≤33 weeks15.5 16.4 3.6 
34–37 weeks24.1 24.6 20.5 
≥38 weeks60.3 57.4 74.7 
Small for gestational age (%)**8.6 4.9 20.5 
Complications at delivery (%)***40.7 39.0 39.2 
Mode of delivery (%)      
Spontaneous vaginal birth27.6 39.3 36.1 
Labour inducement27.6 21.3 32.5 
Caesarean section36.2 36.1 22.9 
Assisted vaginal birth3.4 3.3 6.0 
Other5.2 0.0 2.4 
Maximal DBP (mmHg)11212.710512.11048.0
In week28.49.829.09.635.53.3
Start treatment in week26.410.419.712.534.24.8

Table 3 shows the crude and adjusted developmental outcomes at primary school age for the three groups. Correction for maternal education, parity, and birthweight of the child altered the association between treatment for PIH/PAH and the developmental outcome of the children with more than 10%, and these were included as confounding factors in the multivariable analyses. The other potential confounding factors did not fulfil the criteria to be included in the multivariable analyses. The scores on the concentration test were higher (i.e. worse) in the labetalol group than in the methyldopa group, but the difference diminished after correction for confounding factors, and was no longer statistically significant (P = 0.17). Children in the labetalol and bed-rest groups on average seemed to have better gross motor development than the children in the methyldopa group (P = 0.10). More children who were prenatally exposed to labetalol had ADHD compared with children exposed to methyldopa (OR 2.3; 95% CI 0.7–7.3) and children exposed to bed rest (OR 4.1; 95% CI 1.2–13.9). The parents of the children in the methyldopa group seemed to report more frequently that their children had sleeping problems than parents of children in the labetalol (OR 3.2; 95% CI 0.6–16.7) and bed-rest groups (OR 4.5; 95% CI 0.9–23.2), athough these differences were not statistically significant. The groups did not differ materially in other aspects of functional development of the central nervous system.

Table 3.   Crude and adjusted developmental outcome measures of 4–10-year-old children in a Dutch historical cohort study on functional development after prenatal exposure to antihypertensive treatment (labetalol versus methyldopa and bed rest) conducted in 1991–1992
 Labetalol (n = 58)Methyldopa (n = 61)Bed rest (n = 83)
Crude (SD)AdjustedCrude (SD)AdjustedDifference (95% CI)*Crude (SD)AdjustedDifference (95% CI)*
Concentration score110.3 (19.6)111.0**102.7 (17.2)104.1**6.9 (–2.6, 16.4)107.6 (13.6)109.5**1.5 (–7.7, 10.7)
IQ (RAKIT)***106.6 (15.5)105.1**107.7 (15.4)103.2**1.9 (–4.1, 8.0)106.0 (15.2)103.6**1.5 (–4.1, 7.1)
Gross motor development15.2 (6.2)14.3****14.1 (8.3)12.4****1.9 (–0.4, 4.3)15.6 (8.5)14.4****–0.1 (–2.0, 1.8)
Fine motor development60.3 (16.1)59.4****61.7 (17.1)59.0****0.4 (–4.9, 5.7)61.4 (19.5)58.8****0.6 (–3.7, 4.8)
Memory task
Visual memory17.1 (5.5)17.9**16.7 (4.7)17.0**0.9 (–1.3, 3.1)16.2 (5.4)16.9**1.0 (–1.0, 3.1)
Associative memory11.1 (4.4)10.9****11.2 (3.6)10.3****0.6 (–0.9, 2.1)11.2 (3.5)10.3****0.6 (–0.6, 1.8)
Auditive memory*****9 (7–10)8.4****10 (9–10)8.2****0.2 (–0.5, 0.8)10 (9–10)8.6****–0.2 (–0.7, 0.4)
 Labetalol (n = 58)Methyldopa (n = 61)Bed rest (n = 83)
n (%)n (%)OR (95% CI)******n (%)OR (95% CI)******
  1. ADHD, Attention deficit hyperactivity disorder.

  2. For confounding factor selection, backward selection was used: all a priori selected potential confounding factors were excluded from the models if their removal did not change the effect estimate by more than 10%.

  3. *Labetalol versus methyldopa and labetalol versus bed rest, respectively; with 95% confidence intervals.

  4. **Adjusted for maternal education (intermediate education), parity (first born), and small for gestational age (no).

  5. ***Intelligence quotient, assessed with the shortened Revision Amsterdam Child Intelligence Test.

  6. ****Adjusted for maternal education (intermediate education), parity (first born), small for gestational age (no), and age at testing (6.5 years).

  7. *****Skewed distribution: crude median and interquartile range.

  8. ******Odds ratio with 95% confidence interval for prenatal labetalol exposure.

ADHD (%)10 (17.2)5 (8.2)2.3 (0.7–7.3)4 (4.8)4.1 (1.2–13.9)
Sleeping disorders (%)2 (3.4)6 (10.3)0.3 (0.1–1.6)2 (2.5)1.4 (0.2–10.2)


The results of this hypothesis-generating study suggest that prenatal exposure to labetalol for gestational hypertension may increase the risk of AHDH at primary school age, compared with treatment with methyldopa or bed rest. Although the difference was not statistically significant, labetalol-exposed children also seemed to score worse on the concentration test than methyldopa-exposed children. Furthermore, children exposed to methyldopa during pregnancy more often seemed to have sleeping problems and decreased gross motor development than children of mothers treated with labetalol or bed rest. However, several assumptions were made, and the study design resulted in some methodological drawbacks that limit the clinical interpretation of our findings, including insufficient events of ADHD and sleeping disorders to calculate adjusted odds ratios.

As the follow-up rate was only 57%, the question arises whether selection bias might have occurred. The pregnancy characteristics of the participants, refusers and non-responders were similar for the three treatment groups, so that differential participation over these potential confounding factors is unlikely. However, as a strikingly high proportion of infants in the bed-rest group was born small for gestational age, selective participation in this group cannot be excluded. In addition, it is very likely that participation depended on the developmental outcome of the child. We know of one woman who did not participate because her child was too mentally retarded to be examined. If this selective participation occurred with equal frequency in all three groups, it will not have influenced our outcome estimates. Selective participation associated with the treatment of gestational hypertension, which is unlikely to have occurred, would have biased our results.

The assumption that the choice of medical treatment by the obstetricians reflected natural randomisation was confirmed by the actual prescription pattern in each of the twelve participating hospitals (data not shown). However, there were differences in population characteristics and treatment regimes between the medication groups. Maximal diastolic blood pressures in the labetalol group were higher, whereas methyldopa treatment started earlier and was more often supplemented with additional treatments. The difference in maximal diastolic blood pressure between groups may therefore reflect the efficacy of early treatment. Maximal diastolic blood pressure may not be a good indicator of the severity of the disease once treatment with antihypertensive medication has been started. The bed-rest group was clearly different from the two medication groups in gestational age, birthweight, severity of hypertension, prevalence of PIH and PAH, and other treatments for PIH/PAH. As women in the bed-rest group were likely to have had milder disease, and therefore may not be comparable with women in the labetalol or methyldopa groups, the results found in this study must be interpreted with caution.

To date, no studies on the long-term effects of prenatal exposure to labetalol in humans have been published. Nevertheless, several ways of action of labetalol can be postulated to explain the effects on the child’s development. Prenatal exposure to labetalol has been suggested to increase the risk of being born small for gestational age,18 which, in turn, is an important marker for neurological dysfunction, including lower intelligence and behavioural problems, including ADHD.19,20 It may also affect blood flow through the placenta, as labetalol has been found to increase placental vascular resistance,21,22 whereas exposure to methyldopa decreases the resistance.23 However, this was not confirmed in other studies.24,25 Moreover, antihypertensive drug treatment, and especially beta-blocking agents, have been associated with an increased risk for neonatal bradycardia,26 which might affect neurodevelopmental outcome. All beta-blocking agents cross the human placenta.

We regarded methyldopa as a medical drug that does not affect the functional development of the exposed child. Ounsted et al.8 concluded from a follow-up study that methyldopa might counteract any small negative effect that pregnancy hypertension may have on the development of the child. In the present study, however, the methyldopa group seemed to have worse outcome distributions with regard to sleeping problems, and possibly gross motor development, compared with a group of children born in the same period to mothers with hypertension treated with labetalol or bed rest. Therefore, the results of this study indicate that the use of methyldopa during pregnancy may affect the functional development of the child at primary school age. In a case report, Shimohira et al.27 also linked sleeping problems and delayed motor development with prenatal exposure to methyldopa. As methyldopa can be found in the fetus in concentrations equal to or higher than those in the mother,28 an effect on sleep may be mediated through the noradrenergic receptors in the fetal brain.27,29 Sleeping problems are also a side effect of methyldopa in the mother. Furthermore, sleeping disturbances after prenatal exposure to clonidine, another centrally active antihypertensive, have been reported in a cohort study,30 which made us enquire about sleeping problems in the present study. It should be noted that we only obtained data on sleeping problems of the children from a standardised parent questionnaire. Although the prevalence of reported sleeping problems in our study is comparable with that reported in the literature,31 the reliability of these parent reports may be questioned. It is most likely, however, that this potential misclassification was non-differential among the three study groups, and therefore would have led to an underestimation of the effect of methyldopa on the occurrence of sleeping problems. The odds ratio estimates of this association would have been overestimated only if mothers in the methyldopa group were more likely to incorrectly report sleeping problems than mothers in the other two groups.


This hypothesis-generating study suggests that labetalol and methyldopa, when used in the treatment of mild-to-moderate gestational hypertension, may influence functional development in children of primary school age. We found an increased risk for developing ADHD at primary school age after prenatal exposure to labetalol. The effects of labetalol on functional development are not very surprising, as effects of exposure during pregnancy have been reported both on birthweight and on placental blood flow. Furthermore, the use of methyldopa during pregnancy may be associated with sleeping problems in children between 4 and 10 years of age. The study population was large enough to detect any differences considered to be clinically relevant in concentration scores, intelligence quotient scores and in other continuous outcome measures, if they existed. However, correcting for confounding factors decreased our study power. Although parts of our results did not attain statistical significance, and our study design resulted in some methodological drawbacks, the consistency of the findings suggests that the conclusions are valid. However, it should be kept in mind that our results may not apply to the the use of labetalol and methyldopa in the acute setting, and in the management of severe pre-eclampsia. Prospective research with comparable groups should be conducted to study the long-term effects of antihypertensive medications used during pregnancy.

Disclosure of interests

None declared.

Contribution to authorship

PCMPdJ study design, data analysis, interpretation of results, and writing of the manuscript. GAZ study design, interpretation of results, and critical review of the manuscript. MMHJvG data analysis, interpretation of results, and critical review of the manuscript. AP coordination of data collection and critical review of the manuscript. FJMG provided clinical direction and a critical review of the manuscript. TKABE provided clinical direction and a critical review of the manuscript.

Details of ethics approval

As stated in the Methods section, the local institution has approved this study: Institutional Review Board Project #9107-5035, obtained on 30 September 1991.


This work was supported by a grant from the Dutch Praevention Fund (grant no. 28-2137).


We would like to thank the departments of obstetrics of the participating hospitals. We would also like to thank Claire Assman-Hulsmans MSc for her valuable contributions to this paper, and Prof. Fred Lotgering for his helpful comments on the manuscript.

Editor’s Commentary

The US Food and Drug Administration categorises methyldopa as Pregnancy Category B (Animal studies have failed to demonstrate a risk to the fetus and there are no adequate and well-controlled studies in pregnant women OR Animal studies have shown an adverse effect, but adequate and well-controlled studies in pregnant women have failed to demonstrate a risk to the fetus in any trimester) and labetalol as Pregnancy Category C (Animal studies have shown an adverse effect on the fetus and there are no adequate and well-controlled studies in humans, but potential benefits may warrant use of the drug in pregnant women despite potential risks). Very few drugs ever attain an A category rating (Adequate and well-controlled studies have failed to demonstrate a risk to the fetus in the first trimester of pregnancy and there is no evidence of risk in later trimesters) because of the dearth of data on the long-term effects of many of the drugs used in pregnancy. Antihypertensive agents are frequently prescribed during pregnancy and Pasker-de Jong et al. report a hypothesis-generating retrospective historical cohort study that begins to address our long-term data deficit.

These investigators studied 202 Dutch children (aged 4–10 years) whose mothers were treated with labetalol, methyldopa or bed rest for gestational hypertension. The aim was to evaluate the effects of labetalol and methyldopa on the functional development and behavioural activity of the children compared with bed rest. The authors concluded that in utero labetalol exposure may increase the risk of attention deficit and hyperactivity disorder (ADHD) among children at primary school age, whereas prenatal methyldopa exposure may influence their sleep patterns, and they recommended that larger studies are warranted to determine the long-term effects of antihypertensive medications.

  • 1 These findings call for better-designed studies on the effects of these, and other, antihypertensive drugs used in pregnancy. The benefit of antihypertensive therapy for mild gestational hypertension is usually limited to the mother, with few if any advantages to the child. Sometimes antihypertensive drugs may be associated with fetal harm, for example beta blockers have been associated with decreased fetal growth. However, in women with higher blood pressure (>150/100 mmHg) these drugs help to control hypertension and prevent complications. It therefore behoves us to interpret these data with caution and avoid overinterpretation, which might discourage the use of these drugs, to the mother’s disadvantage. The data in this paper are retrospective and collected over 20 years ago.
  • 2 Gestational age was not controlled for and was reported as not being significantly different between the three groups. The numbers are however small with the potential for a type II error. Furthermore, a large proportion of the children in the labetalol and methyldopa groups were born between 34 and 37 weeks of gestation, whereas in the bed rest group the mean gestational age was higher (38 ± 2 weeks). This raises the possibility of bias because moderate prematurity (32–36 weeks of gestation) is associated with neurobehavioural outcomes such as ADHD (van Baar et al., Pediatrics 2009;124:251–7; Aarnoudse-Moens et al., Pediatrics 2009;124:717–28. Epub 2009 Jul 27).
  • 3 The duration of therapy for each of the interventions was not stated but a significant difference in duration of therapy could also be a potential bias. The urgency of delivery in these women could have influenced the outcomes, and it would be important to know how many pregnancies ended in emergency delivery for superimposed pre-eclampsia, and whether there was a difference in the primary outcome measures based on the development of this complication. Other questions include how many of the children who developed ADHD were born preterm versus term in a compromised state, and whether there was a preponderance of ADHD in babies delivered as emergencies compared with those who underwent a planned delivery?
  • 4 The authors reported a high rate (40%) of ‘complications’ at delivery but provided no details on how many children in each group had conditions such as meconium aspiration, infection or metabolic acidosis, which could have contributed to their ultimate outcome. As with many studies of this type, the response rate was low (57%), and there were some missing data (low birthweight; 24% in labetalol group, 16% in methyldopa group and 7% of bed-rest group). Finally, these data reflect relatively long-term use of orally administered drugs in women with mainly chronic hypertension and should not be extrapolated to apply to the acute management of severe pre-eclampsia.

In the USA, at least, intravenous labetalol is a mainstay of acute severe hypertension management in pre-eclampsia, and avoidance of appropriate antihypertensive therapy in a severely pre-eclamptic woman because of a theoretical concern about ADHD based on these data would be premature. Sadly, we have so few data on either the long-term or short-term effects of other agents such as hydralazine, ketanserin, prazocin and nifedipine that they cannot be regarded as being any safer.

The ‘fetal origins of adult disease’ hypothesis mandates that the long-term effects on the fetus of any treatment used in pregnancy should be properly established. Despite the obvious importance of long-term follow up of drug exposures in pregnancy, such studies are few and far between. This is clearly an area that should interest funding agencies and I hope that this paper will stimulate more research into this critical area of obstetric care.

Michael Belfort
Department of Obstetrics and Gynecology,
University of Utah School of Medicine
Salt Lake City, Utah, USA