Controlled Trials of Antihypertensive Drugs in Pregnancy Christopher w.G. Redman, FRCP • It Is taken for granted that severe hypertension In pregnancy should be treated, although the principle has not been formally tested by properly controlled trials. There is less certainty about treating mild to moderate hypertension (140/90 to 1691109 mm Hg). The risk of chronic hypertension in pregnancy depends on that of superimposed preeclampsia, which must be prevented by control of the blood pressure if antihypertensive treatment Is to be beneficial. There Is no a priori reason why lowering the blood pressure should have this effect. Most of the trials of treatment have been too small to provide conclusive answers. Usually treatment has been started too late to give a realistic expectation of Influencing the evolution of superimposed preeclampsia. However, the largest trial of the early use of methyldopa In women with mild chronic hypertension, showed clearly that treatment does not prevent the superimposition of preeclampsia. !3-Adrenergic blocking agents, If used from the second start of the trimester, are associated with a major risk of severe growth retardation and are therefore contraindicated. Methyldopa has the best safety record, which includes long-term follow-up to assess the development of children exposed to methyldopa in utero. The ineffectiveness of antihypertensive drugs In preventing or ameliorating preeclampsia needs to be contrasted with the consistent evidence for the effectiveness of anti platelet therapy. This Is consistent with the increasing evidence that preeclampsia is not primarily, or even necessarily, a hypertensive disease. © 1991 by the National Kidney Foundation, Inc. INDEX WORDS: Antihypertensive treatment; controlled trials; anti platelet therapy; hypertension In pregnancy.

THE THRESHOLDS FOR ANTIHYPERTENSIVE TREATMENT

I

T IS TAKEN for granted that severe hypertension in pregnancy should be treated . The principle has never been formally tested , but is based on circumstantial reasoning. It is consistent with the management of nonpregnant hypertensive individuals,l has some validity derived from animal experiments,2 and takes account of the well-known fact that when women die of preeclampsia/eclampsia, it is most commonly of cerebral hemorrhage , 3 thought to be caused by severe hypertension. Severe hypertension usually means blood pressures at or above 170/110 mm Hg . There has been no trial of the treatment of severe hypertension in pregnancy and none is likely to be performed now because, despite the lack of firm proof of efficacy, it would be considered to be unethical. There are more uncertainties about treating mild to moderate hypertension. In general medical practice, an initial enthusiasm has moderated as the results of large controlled trials have become available. In obstetric practice, the issue is unresolved. Although mild to moderate hypertension (140/90 to 1691109 mm Hg) is associated with an increased perinatal morbidity, association need not imply causation. But, most important, the pathology that the treatment is intended to prevent-the "damage" caused by modest hypertension-is undefined .

THE RATIONALE FOR TREATMENT OF MILD TO MODERATE HYPERTENSION IN PREGNANCY

The problem is complicated by the heterogeneity of causes for the hypertension (preeclampsia, chronic hypertension, or combinations of both conditions) and confusion about whether antecedent or chronic hypertension is, of itself, a disability in pregnancy. The conventional definition for chronic hypertension is that the affected woman presents before the 20th week of pregnancy with a diastolic pressure of 90 mm Hg or more. In our practice, approximately 2 % of women are in this category.4 As expected, they are a little bit older, heavier, and less likely to be parous than women who develop preeclampsia (Table 1). They suffer from more proteinuria and a higher perinatal mortality, the latter disability being confined to those who are chronically hypertensive and in their first pregnancy (Table 1). Of all the groups examined, multigravid women with chronic hypertension had the lowest perinatal mortality (Table 1). This underlines the fact that raised blood pressure is not of From the Nuffield Department of Obstetrics and Gynaecology. John Radcliffe Hospital, University of Oxfo rd, Oxford, England. Address reprint requests to Christopher WG. Redman , FRCp, Department of Obstetrics and Gynaecology, John Radcliffe Hospital, Oxford 0X3 9DU, United Kingdom. © 1991 by the National Kidney Foundation , Inc. 0272-6386/9111702-0011$3.00/0

American Journal of Kidney Diseases, Vol XVII. No 2 (February). 1991: pp 149-153

149

150

CHRISTOPHER w.G. REDMAN Table 1. Characteristics of Women With and Without Hypertension in Pregnancy Normal

14,109 No. of subjects Age (yr)* 27.5±5.0 WeighUheight2 (g/cm 2) * 24±4 Primigravidae 40.6% Proteinuria 0.5% Birthweight (kg)* 3.34±0.52 Maturity at delivery (d)* 279 ± 13 Perinatal deaths Primigravidae 51/5,731 (0.89%) Multigravidae 75/8,378 (0.90%)

Chronic Hypertension Preeclampsia

485 28.5±5.2

1,617 26.6±4.9

27±6 41.4% 2.4% 3.26±0.63

24±4 60.7% 5.5% 3.17±0.65

273± 17

274± 16

61201

11/981

(2.99%)

(1.12%)

21284

8/636

(0.70%)

(1.26%)

NOTE. All women presented for care in the John Radcliffe Hospital, Oxford, before 20 weeks of gestation in 1981 to 1984. The definitions classifying the groups are given by Redman and Jeffries. 36 *Means ± SO.

itself dangerous. The association with primiparity points to the involvement of a preeclamptic process. Comparable observations have been made by others 5 and it is commonplace clinical experience that hypertensive women who do not develop superimposed preeclampsia have normal outcomes, despite relatively severe hypertension. In short, it is not the raised blood pressure that is a problem in pregnancy, but the associated predisposition to preeclampsia, the likelihood of which is increased three to seven times in affected women. 6·8 With respect to the treatment of chronic hypertension in pregnancy, the crucial question is: Does lowering the blood pressure reduce the likelihood of developing preeclampsia to that experienced by normotensive women? In other words, is the hypertension itself a causative factor? With respect to the treatment of hypertension arising early in preeclampsia, the question is whether treatment can halt or retard worsening of the disease. In this case, the hypothesis is that the raised blood pressure is contributing to the progression of the disorder. Little is known about the pathogenesis of preeclampsia, but what evidence there is provides no encouragement for the concept that the hypertension is anything but a consequence of other more primary processes. Preeclampsia is not a hemodynamic, but a trophoblastic disease. The affected woman must be pregnant, but she need not carry a

fetus.· because she may get the syndrome with a hydatidiform mole. 9 Removal of the placenta cures the condition, whereas treatment of the blood pressure never does. The problem appears to be poor placentation, which leads to deficient arterialization in the maternoplacental circulation; the spiral arteries retain their musculoelastic wall, which normally is replaced by thin fibrinoid thus dilating the vessels. 10.11 Placentation is established between weeks 6 and 18. The appearance of preeclampsia later in pregnancy appears to be the sequel to these abnormalities, perhaps as a result of increasing placental ischemia, which would be consistent with the methods of inducing preeclamptic syndromes in experimental animals. 12 . J3 At the time ofthe overt illness, further lesions in the spiral arteries appear. They are obstructive aggregates of fibrin, aggregated platelets, and fatfilled macrophages or lipophages, the process being called acute atherosis. These are endothelial lesions, not comparable to any sort of arterial injury known to be related to hypertension, either acute or long-standing. They also occur in normotensive pregnancies complicated by severe intrauterine growth retardation l4 ; thus, it is not possible to suppose they directly result from hypertension per se. They are much more likely to be related to a generalized endothelial injury affecting the maternal vessels. 15 However, they will aggravate any tendency to placental ischemia. As far as it is known, the processes affecting the placental spiral arteries are what treatment must influence. Since the majority of cases of preeclampsia occur in previously normotensive women, it is not possible to impute that hypertension is the cause of deficient placentation. Acute atherosis may be the result of hypertensive injury, but the question must be answered why it does not complicate hypertension in any other context. The trials of antihypertensive treatment in pregnancy, all of which concern women suffering only from mild to moderate hypertension, need to be interpreted with these considerations in mind. END POINTS

What are the appropriate end points for trials of the treatment of moderate hypertension in pregnancy? From what has been briefly summarized, it should be apparent that the most important is the development of preeclampsia. In the case of chronic hypertension, treated early, the purpose is

151

HYPERTENSION IN PREGNANCY

to prevent superimposition of preeclampsia. Pregnancy-induced hypertension includes at least two presentations: nonproteinuric preeclampsia-in which case the intention of treatment is to halt progression; and reversal of the pregnancy-induced hypotension of midgestation revealing previously inapparent chronic hypertension-in which case treatment will be beneficial if it halts later superimposition of preeclampsia. In the latter instance, the hypothesis, never clearly stated, is that even over a short period of time in the last weeks of pregnancy, a mildly raised blood pressure causes preeclampsia. Thus, whether the indication for treatment is chronic hypertension, pregnancyinduced hypertension, hypertension unqualified, or preeclampsia, the relevant end points need to measure the development of preeclampsia and its severity; the underlying hypotheses being that treatment either prevents superimposition of preeclampsia on antecedent hypertension, or retards its progression in its earlier stages. The next question is how to measure the occurrence and severity of preeclampsia. The problem is that because the disease is not understood, there is no specific test; it can be recognized only as a syndrome comprising a cluster of various nonspecific signs. The complete syndrome has two components-maternal and fetal 16-the fetus suffering from the consequences of either poor placental function, namely poor growth and degrees of respiratory impairment culminating in asphyxia and intrauterine death; or obstetric intervention, which may, for example, cause prematurity because ending the pregnancy remains the only definitive treatment; or operative delivery. Eclampsia and maternal death are maternal end points, but so rare as not to be discernible in the small trials that have been done, of which most have been confined to a single center. One end point, the severity of hypertension, is of limited interest. Avoidance of severe hypertension is an advantage. But in most trials it has been a reason for starting treatment in the women assigned to the control group, that is, the efficacy of treatment has been assumed if the problem is severe. To claim that antihypertensive treatment confers benefit by ameliorating moderate hypertension is only valid if this degree of hypertension is shown to cause problems, which in fact depends on the outcomes of trials in which end points other than hypertension are used to measure benefit. In short, it may be assumed that antihy-

pertensive drugs lower the blood pressure; it cannot be assumed that this is of itself beneficial if the hypertension is not severe. Other indicators of preeclampsia include the development of new proteinuria, new hyperuricemia, or fetal complications ascribable to the disorder. The last includes perinatal death, or growth retardation. Unreliable indicators include the clinical response to the disease as it is perceived, for example admission to hospital for bed rest, or induction of labor, or operative delivery. The second two have "knock on" effects on the perinatal outcome, causing prematurity and other neonatal complications. These end points are unreliable, because the use of antihypertensive agents can change the clinical perception ofthe severity of the disease and alter the intervention pattern. If this normally includes unnecessary intervention, then all that the drug treatment is achieving is to protect the patient from her doctors. THE DRUGS THAT HAVE BEEN USED

Diuretics were the first, but they are not considered here. Methyldopa, and latterly ,8-adrenergic blocking agents, have been the principle agents tested. Trials of the management of severe hypertension are rare, but test the efficacy of the standard obstetric agent-parenteral hydralazineagainst calcium channel blockers or labetalol. 17 ,18 THE RESULTS OF TRIALS TO DATE

Results of trials to date are summarized in Table 2. It is important to notice that no trial has been large enough to have the power to give unequivocal results. Second, the time at which treatment is started has varied from 16 weeks to the peripartum period, and the indications for treatment have similarly varied, to include every diagnostic category. Thus, the trials have not been equivalent, although in a summary like this they are treated as such. With respect to maternal end points, the only consistent feature is that lower blood pressures were achieved in the treated groups. 19-21 Of itself this is not beneficial, it merely demonstrates that the drugs used had the intended effect. One important result has been that the early treatment of chronic hypertension does not prevent superimposed preeclampsia. 19 Treatment with atenolol started in the third trimester reduced the incidence of proteinuria in one trial,22 but this Was not as so-

152

CHRISTOPHER w.G. REDMAN Table 2.

Controlled Trials of Antihypertensive Treatment in Pregnancy

Reference

No. of Subjects

Entry Time (wk)

Diagnosis

Active Treatment (Control Group)

Effects of Treatment>

27 19

100 207

NR 21-22

HT CHT

Methyldopa ± diuretic (none) Methyldopa (none)

28 2:?

25 86

NR 34

CHT HT

Methyldopa (place bot) Atenolol (place bot)

24 29 20 30

29 152 100 52

16 34 33 33

CHT PIH PE HT

Atenolol (place bot) Labetolol (placebot) Labetolol (bed rest) Metoprolol (placebot)

21

155

28

PIH

23 31 25 32 18 17

183 100 60 72 60 31

29-30 30-31 NR 31

HT HT HT HT PHT PPHT

Oxprenolol ± hydralazine (placebot) Oxprenolol (methyldopa) Oxprenolol (methyldopa) Atenolol (methyldopa) Labetolol (methyldopa) Labetolol (hydralazine) Nifedipine (placebot)

None Fewer perinatal and midtrimester losses None Less proteinuria, fewer hospital admissions, more neonatal bradycardia Intrauterine growth retardation None Intrauterine growth retardation Fetal bradycardia, lower umbilical venous lactate Fewer caesarean sections, less prolonged neonatal care Better intrauterine growth None Lower placental weigt'lts None None More urine output in first 24 h after delivery

Abbreviations: NR, not recorded; PIH, pregnancy-induced hypertension; CHT, chronic hypertension; HT, hypertension unqualified; PPHT, postpartum hypertension; PHT, peripartum hypertension; PE, preeclampsia. 'Excluding lower blood pressures and statistically nonsignificant effects. tDouble-blind allocation of treatment.

ciated with the expected improvement in fetal end points that would have been observed if preeclampsia had been prevented. The observation remains unexplained and has not been replicated in other studies. With respect to fetal or neonatal end points, a better perinatal outcome has been reported only once, 19 and in this case the excess fetal losses in the control group, which included several midpregnancy losses, could not be ascribed to hypertensive pathology. With one exception,23 there has been a tendency for ,8-blockers to be associated with smaller fetuses 20 .24 or placentas. 25 The effect is so striking in one recently published study,24 that it must now be stated that ,8-blocking agents are contraindicated in the long-term treatment of hypertension in pregnancy. Some of the benefits claimed, for example less hospital admissions, 20 or fewer cesarean sections,21 concern extrinsic end points, and the possibility that the clinical staff were merely reacting to differences in blood pressure, rather than to genuine alterations in disease severity, cannot be excluded. The conclusion that treatment is conferring no significant benefit is confirmed in a recent extensive series of metaanalyses of these trials. 26

CONCLUSIONS

The trials performed thus far do not establish a case for treating mild to moderate hypertension in pregnancy; treatment should be reserved for women with severe hypertension, that is blood pressures of 170/110 mm Hg or more. ,8Adrenergic blocking agents retard fetal growth if given over long periods of time, and therefore should not be used. Methyldopa remains the drug of choice, particularly as it is the only agent for which there is a long-term follow-up of children exposed in utero, and this is reassuring. 33 The ineffectiveness of antihypertensive drugs in preventing or ameliorating preeclampsia needs to be contrasted with the consistent evidence for the effectiveness of antiplatelet therapy. 34 Perhaps it is time for clinicians to recognize that preeclampsia is not primarily or even necessarily a hypertensive disease. 35 REFERENCES 1. Breckenridge A, Dollery CT, Parry EHO: Prognosis of treated hypertension. Q J Med 39:411-429, 1970 2. Byrom FB: Pathogenesis of hypertensive encephalopathy and its relation to the malignant phase of hypertension: Experi-

HYPERTENSION IN PREGNANCY mental evidence from the hypertensive rat. Lancet 2:201-211, 1954 3. Department of Health and Social Security: Report on Confidential Enquiries Into Maternal Deaths in England and Wales 1982-84. London, England, Her Majesty's Stationery Office, 1989, pp 10-19 4. Redman CWG: Hypertension in pregnancy, in de Swiet M (ed): Medical Disorders of Pregnancy. London, England, Blackwell Scientific, 1989, pp 249-305 5. Chamberlain G, Philipp E, Howlett B, et al: British Births 1970, vol 2. Obstetric Care. London, Engand, Heinemann, 1978, pp 39-53 6. Harley IMG: Pregnancy in the chronic hypertensive woman. Proc R Soc Med 39:835-838, 1966 7. Chesley LC, Annitto IE, larvis DG: A study of the interaction of pregnancy and hypertensive disease. Am 1 Obstet GynecoI53:851-863, 1947 8. Butler NR, Bonham DG: Toxemia in pregnancy, in Perinatal Mortality. Edinburgh, England, Livingstone, 1963, pp 87-100 9. Chun D, Braga C, Chow C, et al: Clinical observations on some aspects of hydatidiform moles. 1 Obstet Gynaecol Br Commonw 71:180-184, 1964 10. Brosens lA, Robertson WB, Dixon HG: The role of the spiral arteries in the pathogenesis of pre-eclampsia, in Wynn RM (ed): Obstetrics and Gynecology Annual. Norwalk, CT, Appleton Century-Crofts, 1972, pp 177-191 11. Robertson WB, Brosens I, Dixon HG: The pathological response of the vessels of the placental bed to hypertensive pregnancy. 1 Pathol 93:581-592, 1967 12. Abitbol MM: Hemodynamic studies in experimental toxemia of the dog. Obstet Gynecol 50:293-298, 1977 13. Cavanagh D, Rao PS, Tsai CC, et al: Experimental toxemia in the pregnant primate. Am 1 Obstet Gynecol 128:75-85, 1977 14. Sheppard BL, Bonnar 1: An ultrastructural study of utero-placental spiral arteries in hypertensive and normotensive pregnancy and fetal growth retardation. Br 1 Obstet Gynaecol 88:695-795, 1981 15. Roberts 1M, Taylor RN, Musci Tl, et al: Preeclampsia: An endothelial cell disorder. Am 1 Obstet Gynecol 161:12001204, 1989 16. Redman CWG: Hypertension in pregnancy, in Thrnbull AC, Chamberlain GVP (eds): Textbook of Obstetrics. London, England, Churchill Livingstone, 1989, pp 515-541 17. Barton lR, Hiett AK, Conover WB: The use of nifedipine during the postpartum period in patients with severe preeclampsia. Am J Obstet GynecoI162:788-792, 1990 18. Mabie WC, Gonzalez AR, Sibai BM, et al: A comparative trial of labetolol and hydralazine in the acute management of severe hypertension complicating pregnancy. Obstet Gynecol 70:328-333, 1987 19. Redman CWG, Beilin U, Bonnar 1, et al: Fetal outcome in trial of antihypertensive treatment in pregnancy. Lancet 2:753-756, 1976 20. Sibai BM, Gonzalez AR, Mabie WC, et al: A comparison of labetalol plus hospitalization versus hospitalization alone

153 in the management of preeclampsia remote from term. Obstet Gynecol 70:323-327, 1987 21. Plouin P-F, Breart G, Llado 1, et al: A randomized comparison of early with conservative use of antihypertensive drugs in the management of pregnancy-induced hypertension. Br 1 Obstet Gynaecol 97:134-141, 1990 22. Rubin PC, Butters L, Clark DM, et al: Placebo-controlled trial of atenolol treatment of pregnancy-associated hypertension. Lancet 1:431-434, 1983 23. Gallery EDM, Ross ME, Gyory AZ: Antihypertensive treatment in pregnancy: Analysis of different responses to oxprenolol and methyldopa. Br Med 1 291:563-566, 1985 24. Butters L, Kennedy S, Rubin PC: Atenolol in the management of essential hypertension during pregnancy. Br Med 1301:587-589, 1990 25. Thorley KJ: Randomised trial of atenolol and methyl dopa in pregnancy related hypertension. Clin Exp Hypertens B3: 168, 1984 26. Collins R, Wallenburg HCS: Pharmacological prevention and treatment of hypertensive disorders in pregnancy, in Chalmers I, Enkin M, Keirse MINC (eds): Effective Care in Pregnancy and Childbirth, vol 1. Oxford, England, Oxford University Press, 1989, pp 512-533 27. Leather HM, Humphreys DM, Baker P, et al: A controlled trial of hypotensive agents in hypertension in pregnancy. Lancet 2:488-490, 1968 28. Weitz C, Khouzami V, Maxwell K, et al: Treatment of hypertension in pregnancy with methyldopa: A randomised double blind study. Int 1 Gynaecol Obstet 25:35-40, 1987 29. Pickles Cl, Symonds EM, Broughton Pipkin F: The fetal outcome in a randomized double-blind controlled trial of labetalol versus placebo in pregnancy-induced hypertension; Br 1 Obstet Gynaecol 96:38-43, 1989 30. Wichman K, Ryden G, Karlberg BE: A placebo controlled trial of metoprolol in the treatment of hypertension in pregnancy. Scand 1 Clin Lab Invest 44:90-95, 1984 (suppl) 31. Fidler 1, Smith V, Fayers P, De Swiet M: Randomised controlled comparative study of methyl dopa and oxprenolol for the treatment of hypertension in pregnancy. Br Med 1 286:1927-1930, 1983 32. Redman CWG: A controlled trial of the treatment of hypertension in pregnancy: Labetalol compared with methyl dopa, in Riley A, Symonds EM (eds): Investigation of Labeta101 in the Management of Hypertension in Pregnancy. International Congress Series 591, Excerpta Medica, 1982, pp 10 1110 33. Cockburn 1, Moar VA, Ounsted M, et al: Final report of study on hypertension during pregnancy: The effects of specific treatment on the growth and development of the children. Lancet 1:647-649, 1986 34. Cunningham FG, Gant NF: Prevention of preeclampsia-A reality. N Engl J Med 321:606-607, 1989 35. Redman CWG: Platelets and the genesis of preeclampsia. N Engl 1 Med 323:478-480, 1990 36. Redman CWG, leffries M: A revised definition of preeclampsia. Lancet 1:809-812, 1988