Eur J Cardio-thorac
Surg (1991) 5:300-305
Coumarin anticoagulation during pregnancy in patients with mechanical valve prostheses M. Cotrufo, T. S. L. de Luca, R. Calabro, G. Mastrogiovati, Institute of Medical and Surgical Cardiology,
and D. Lama
First Medical School, University
of Naples, V. Monaldi Hospital,
Naples, Italy
Abstract. Between January 1987 and December 1989,20 female patients with one mechanical valve prosthesis (MVP) for at least 1 year postoperatively were studied while on coumarin therapy for the full length of pregnancy. In each case, caesarean section was scheduled for the 38th week. Patients were selected according to the following criteria: (1) prothrombin ratios remaining within the therapeutic range for more than 85% of their total estimations in the previous 12 months with mean daily doses of warfarin less than 5 mg; (2) stable cardiac status; (3) no previous obstetric diseases and (4) full acceptance of the risks involved in the protocol. The patients were in NYHA functional class I or II. Their ages ranged from 23 to 31 years (mean 26 f 3). Ten patients had a mitral prosthesis and 10 an aortic prosthesis. Among the 20 mechanical valve prostheses, 10 were Sorin, 6 Starr-Edwards, 2 Bjsrk-Shiley, and 2 Lillehei-Kaster. Eighteen patients were in sinus rythm, 1 in chronic atrial fibrillation, and 1 had a permanent endocardial pacemaker. Nineteen were delivered by caesarean section: warfarin was withdrawn 48 h before surgery and resumed 24 h thereafter. One patient had premature spontaneous delivery at 36 weeks. The mean prothrombin ratio measured weekly in the 20 patients was 2.06 f 0.45 INR, using a mean daily warfarin dose of 4.1 mg +, 1.63. The mean value of the prothrombin ratio during caesarean section for the 19 patients was 1.23 f 0.38 INR. In the 20 live births, the mean birth weight was 2.9 kg f 0.40. There were no observed thromboembolic or haemorragic complications, no spontaneous abortions, stillbirths or deaths. All infants underwent cardiac, neurological, X-ray, and echocardiographic examination. No evidence of warfarin embryopathy was found. These findings strongly suggest that (1) the use of strictly controlled coumarin therapy with medium to low doses of warfarin, and (2) delivery by caesarean section may afford pregnant patients with mechanical valve prostheses satisfactory anticoagulation without embryopathy or haemorrhagic complications. [Eur J Cardio-thorac Surg (1991) 5:300-3051 Key words: Pregnancy - Heart valve prosthesis - Warfarin - Embryopathy
The thromboembolic risk after implantation of mechanical valve prostheses remains significantly high, even with the newer generation valves. Long-term anticoagulation is therefore necessary. The most effective therapy is obtained with oral coumarin derivatives. Pregnancy is associated with a hypercoagulability state which further emphasizes the need for anticoagulation. However, problems arise because oral anticoagulants cross the placenta. Indeed, foetal coagulability is depressed more than maternal, mainly due to the immaturity of the foetal liver. Anticoagulants continued throughout pregnancy produce a greater risk of teratogenie lesions in the first trimester, foetal haemorrhagic complications in later months, maternal haemorrhage
Read at the 4th Annual Meeting of the European Association Cardio-thoracic Surgery, Naples, September 17-19, 1990
for
during delivery, and foetal malformations or death with an overall incidence of 30%. Discontinuing anticoagulation during pregnancy, even temporarily, has given rise to disastrous thrombotic complications. Alternative anticoagulation protocols using heparin or antiplatelet agents do not solve the problem. Bioprostheses implant in women of childbearing age may facilitate pregnancy and avoid the need for anticoagulation. They may however require early reoperation due to the rapid deterioration of the biological tissue induced by young age and pregnancy. On this basis, management of pregnant patients with mechanical valve prostheses remains controversial. It is mandatory to inform patients about the risks and limitations they will face with pregnancy and eventually to recommend therapeutic abortion. Nevertheless, many women, especially those with no children, are highly motivated to continue pregnancy even at the risk of their
301
own lives. In these cases, we have a great responsibility for their proper management. In our Institution, mechanical valve prostheses have been implanted in 3240 patients since 1976. Up until now, anticoagulation was routinely induced using sodium warfarin. The mean value for prothrombin activity which we consider within the therapeutic range is 30% t_ 5%, corresponding to a mean value for the prothrombin ratio of 2.07 + 0.30 INR. Recently, dypiridamole has been added to warfarin in cases with high thromboembolic risk (Starr-Edwards ball valves, giant left atrium, previous thromboembolic episodes). This study followed 20 patients with mechanical valve prostheses who became pregnant while on warfarin therapy alone. Literature reports by Quick [33] Kraus, Perlow, and Singer [23] and Hirsh, Cade and Gallus [18] sought to predict the risk factors for foetal mortality in dogs and rabbits exposed to coumarin derivatives during pregnancy. They found that large doses of the drug, administration of coumarin until term, and foetal exposure to the trauma of spontaneous delivery all significantly increased the risk of foetal death. To our knowledge, no clinical trials have been done to investigate these experimental data. The purpose of our study was to show that the adverse effects of warfarin during pregnancy could be reduced in selected patients. Starting with our clinical observations as well as those in the aforementioned reports, we hypothesized adequate anticoagulation could be achieved in selected patients with lower doses of warfarin. The basic points of the study include (a) avoiding excessive doses of the drug, (b) avoiding the trauma of spontaneous delivery, and (c) allowing anticoagulation to be discontinued just prior to caesarean section. Material and methods All patients of childbearing age discharged from our Institution after implantation of one or more valve prostheses were informed of the risks of pregnancy during anticoagulation and were asked to refer to us in case of pregnancy. Between January 1987 and December 1989,29 women were observed in the Out-patient clinic. Twenty women were selected for an anticoagulation protocol which included (1) sodium warfarin therapy throughout pregnancy, (2) programmed caesarian section during the 38th week, (3) discontinuance of anticoagulation 2 days before surgery and 1 day after, (4) weekly estimations of the prothrombin ratios done at our Institution, (5) monthly cardiological and obstetrical examinations in the Cardiology Clinic and in the High Risk Pregnancy Clinic affiliated to our Institution, (6) echo studies of the foetus in the 2nd, 5th and 8th month. The selection criteria included: normal coagulation system, prothrombin ratios within the therapeutic range for more than 85% of total estimations in the previous 12 months with mean daily doses of warfarin equal to or lower than 5 mg, stable cardiac status, no previous obstetric disease and a full knowledge and acceptance of the risks included in the protocol. During the same period of time, 9 women were not admitted to the trial and were urged to discontinue pregnancy. The exclusion was due to high daily doses of warfarin in 5 cases during the 12 months preceding pregnancy, unstable cardiac state in 2 cases and obstetric complications in 1 case. Four further pregnant women were treated in 1990 with the protocol described. and the results confirmed those of the present report.
The age of the 20 selected women ranged from 23 to 31 years (mean 26 years f 3); 14 were primigravid and 6 multigravid; 10 had an aortic valve prosthesis (9 tilting disk Sorin prosthesis and 1 valved Sorin conduit); 10 had a mitral valve prosthesis (6 Starr-Edwards, 2 Bjsrk-Shiley, 2 Lillehei-Kaster). All patients were in NYHA functional class I or II; 18 patients were in sinus rhythm. 1 in chronic atria1 fibrillation and 1 had a permanent endocardial pacemaker; 8 patients were taking digitalis and diuretics during pregnancy. All patients became pregnant without planning the date of conception and while on warfarin therapy. The mean time interval between implantation of the valve prosthesis and pregnancy was 3.5 yearsk2.3. None of the patients were hospitalized when pregnancy was confirmed and no special limitations were imposed on their routine activities. All patients remained under our direct surveillance during pregnancy, delivery and puerperium. The prothrombin activity was determined every week during pregnancy and puerperium and every day during the week of hospitalization for caesarian section using rabbit brain calcium thromboplastin. The prothrombin ratio was determined in blood samples from the infant soon after birth and 12 h later after vitamin K had been administered. The neonates underwent immediate examination by the clinical geneticist, pediatrician, cardiologist and neurologist and have been followed by echo and X-ray studies during the following 12-24 months.
Results Twenty-nine pregnant women with one mechanical valve prosthesis in place for at least 1 year were observed from January 1987 until December 1989. Nine patients were advised to discontinue their high risk pregnancies and 20 women were followed throughout pregnancy and puerperium. All were anticoagulated with daily doses of sodium warfarin throughout pregnancy with a mean daily dose of 4.1 mgt_ 1.63. The prothrombin ratio was determined 12 h after the oral ingestion of the drug and the mean value for the prothrombin ratios of the weekly measurements in the 20 patients was 2.06kO.45 INR. Nineteen patients delivered by caesarian section during the 38th week: warfarin was interrupted 48 h before surgery and restarted 24 h later. In no case was vitamin K administered to the mother before surgery. The mean value of the prothrombin ratio during caesarian section in 19 patients was 1.23 If:0.38 INR. One patient had a premature spontaneous vaginal delivery during the 36th week. No significant complications were observed during pregnancy and puerperium. One patient showed right heart failure during the 3rd month which was succesfully treated with digitalis and diuretics. No thromboembolic or haemorrhagic complications, abortions, stillbirths or deaths were observed. All 20 babies were born alive: the mean birth weight was 2.9 kgk0.40. The prothrombin ratio was determined in blood samples from 6 neonates soon after birth and before vitamin K was administered and the mean value was 2.90 INR. Eight male and 12 female neonates were examined soon after birth and 3.6 and 12 months later by clinical geneticist and no sign of warfarin embryopathy was found. The clinical follow-up by X-ray and echo studies demonstrated a small muscular VSD, a fetal growth defect and obesity in one case each but no skeletal abnor-
302
malities, mental or neurological retardation. All mothers were allowed to breast-feed their infants and no haemorrhagic complications were observed [32]. Discussion
The most effective prevention of thromboembolic complications after implantation of mechanical valve prostheses is achieved using coumarin derivatives [I I]. A hypercoagulable state exists during pregnancy and the postpartum period [17, 25, 35, 371 when concentrations of clotting factors, platelet turnover and viscosity are increased and fibrinolysis is diminished [31]. These changes increase the risk of prosthetic valve thrombosis and the need for anticoagulation; discontinuing the drug, even temporarily, has given rise to disastrous complications and even death [29]. Several reports indicate that vitamin K antagonists are the most effective anticoagulants during pregnancy [21] but they are of low molecular weight and may cross the placenta, as Quick demonstrated in experimental animals [33]. The foetus becomes overanticoagulated because coumarin crosses the placenta but vitamin K dependent coagulation factors in the mother’s blood do not. In addition, the immature foetal liver is more sensitive to the anti vitamin K effect of coumarin [2, 311. Thus the foetal prothrombin time may be considerably prolonged even when that of the mother is normal [l]. The use of coumarin derivatives during pregnancy carries serious risks to the foetus such as coumarin embryopathy and death, especially if the drugs are administered during the first 3 months [20, 34, 391. The warfarin syndrome is characterized by chondrodysplasia with nasal hypoplasia, optic atrophy and dwarfism [37]. It has been suggested that the teratogenic effects may develop when the embryo is exposed to these agents between the 6th and 9th week of gestation [15]. The use of coumarin derivatives at the time of delivery has been related to an increased incidence of stillbirths and neonatal deaths primarily as a result of haemorrhage [27, 281. Neurological complications, eye abnormalities, placentae haemorrhage and spontaneous abortions which have occurred after the administration of coumarin later on in pregnancy are almost certainly attributable to haemorrhage [31]. It is uncertain whether the “warfarin syndrome” is produced by the chemical structure of the drug or by the effect of anticoagulation on the developing cartilagenous tissue. It was reported that both teratogenic and adverse growth effects induced by warfarin exposure during pregnancy may show up subsequent to birth [19]. No significant difference was found between pregnant patients exposed to acenocoumarin and those exposed to sodium warfarin [20]. Little information has been reported on the use of phenindione [30]. In an extensive study of over 418 pregnancies compiled from various sources, Hall, Pauli and Wilson [15] revealed severe complications due to coumarin derivatives in 30%. Several alternative anticoagulation protocols have been used in pregnancy to avoid the risks of coumarin
derivatives. Because heparin does not cross the placenta, it has been administered subcutaneously [34] or intravenously for full anticoagulation. However, both routes of administration carried a substantial risk [22]. Moreover, heparin can cause osteoporosis, thrombocytopenia and alopecia [8, lo]. The long term use of heparin in pregnancy is inconvenient as it must be given parenterally; monitoring is difficult and several cases of valve thrombosis during heparin anticoagulation have been reported [31]. Even though heparin is not teratogenic, there are case reports showing more than 20% foetal complications and a higher rate of maternal complications including haemorrhage, abortion and prematurity [14]. Dipyridamole and acetylsalicylic acid are efficient blockers of white thrombus formation when associated with coumarin therapy [16,40] but they are not reliable when used without oral anticoagulants in terms of preventing thrombosis of the prosthesis or embolic complications [4, 20, 34,371. Moreover it has been suggested that the addition of dipyridamole to warfarin may increase the rate of foetal wastage [7, 361. The use of bioprostheses in patients of childbearing age initially seemed to be a solution to the problem [6,12, 26,341. The fact remains that anticoagulation is necessary in a significant number of patients due to previous embolism, poor haemodynamic state, left mega-atrium with or without thrombosis, atria1 fibrillation, and a high thromboembolic risk secondary to malfunction of the bioprosthesis. It has been clearly demostrated that young age and pregnancy induce rapid deterioration of the bioprostheses requiring early reoperation [24]. Almost three decades of experimental and clinical research on the management of pregnant patients with artificial valve prostheses remain inconclusive: Oakley in 1968 [3] suggested routine sterilization of young women with valve prostheses and the same author in 1987 [31] confirmed that women with artificial valves may even be advised against pregnancy because of the serious risks to both mother and child. Our experience with artificial valve prostheses beginning in 1976 and based on a patient population of 3240 convinced us that the most effective and safe long-term anticoagulation is obtained using coumarin derivatives. The therapeutic range for the prothrombin ratio is obtained by giving daily doses of sodium warfarin, the dosage varying widely between 1 mg and 15 mg from patient to patient based on the patient’s own control ratio. The same observations have been made in women of childbearing age whereby they could be classified in two groups according to the daily dose of sodium warfarin which was medium to low in the first group and medium to high in the second group. Pregnancy does not seem to modify the warfarin dosage. Considering that warfarin has a direct effect on the foetus which reacts independently from the mother, the correlation between drug dose and foetal effects seems to be fundamental and the absence of warfarin embryopathy observed with low doses of anticoagulants is probably due to the fact that haemorrhagic phenomena in the foetus are prevented. These observations are not unique. Quick [33] and Kraus et al. [23] in pregnant dogs and rabbits exposed to
303
coumarin derivatives demonstrated large doses of the drug given up to term to be the greatest single risk factor for foetal haemorrhage and death. Hirsh et al. confirmed this result in rabbits and further suggested that an additional risk factor was the trauma of spontaneous delivery [I 81. Since then, a few clinical reports have mentioned foetal complications during coumarin therapy to be more likely to occurr when oral anticoagulation is excessive and when there is a previous history of miscarriages [5, 13, 18, 201. Sareli et al. [36] reported a low maternal risk for thromboembolism and haemorrhage with the newer generation mechanical heart valves of lower thrombogenic potential and with lower levels of anticoagulation. Nevertheless the rate of foetal wastage and neonatal morbidity associated with warfarin anticoagulation remains high. Oakley’s recently reported experience [31] suggests that a successful outcome can be expected in about two thirds of cases when coumarin derivatives are used throughout pregnancy. The risk of a newborn child having a disabling coumarin embryopathy seems to be less than 5%. The present study shows that the incidence of complications of coumarin therapy used throughout pregnancy may be reduced to almost zero in pregnancies where women are on a coumarin regimen with medium-low daily doses of the drug. The main haemorrhage risk is spontaneous delivery on full anticoagulation and is avoided through caesarean section during a short warfarin discontinuance. The same therapeutic attitude is acceptable in patients with chronic or intermittent atria1 fibrillation.
Conclusions
The combination of mechanical valve prostheses, pregnancy and anticoagulation represents a high risk which is very difficult to manage. Strongly motivated women in a stable cardiac state and without previous obstetric disease may have a good chance to endure pregnancy if the thromboembolic risk is reduced by continuous warfarin anticoagulation and the haemorrhagic risk is prevented by delivering through caesarian section while briefly interrupting warfarin therapy. This policy seems to reduce significantly the incidence of warfarin embryopathy allowing pregnancy to term in those women who maintain a stable prothrombin ratio within the therapeutic range when ingesting low to medium daily doses of the drug.
References Becker RM (1983) Intracardiac surgery in pregnant women. Ann Thorac Surg 36: 453-458 Ben lsmail M, Fekih M, Tartak M, Chelli M (1979) Protheses valvulaires cardiaques et grossesse. Arch Ma1 Coeur 72: 192-199 Bennett GG, Oakley CM (1968) Pregnancy in a patient with a mitral valve prosthesis. Lancet I: 616-618 Biale Y. Cantos A, Lewen Thal H, Guerron M (1980) The course of pregnancy: patients with artificial heart valve treated with dipyridamole. lnt J Obstet Gynecol 81: 1288132
5. Bloomfield DK (1970) Fetal deaths and malformations associated with the use of coumarin derivatives in pregnancy. A critical review. Am J Obstet Gynecol 107: 883-886 6. Chen WWC, Chan CS, Lee PK, Wang RYC, Wong VCW (1982) Pregnancy in patients with prosthetic heart valves: an experience with 45 pregnancies. Q J Med 51: 358-365 7. Chesebro JH, Adams PC, Fuster V (1986) Antithrombotic therapy in patients with valvular heart disease and prosthetic heart valves. JACC 8:41B-56B 8. Chong BH, Pitney WR, Castaldi PA (1982) Heparin induced thrombocytopenia. Lancet II: 1246- 1248 9. De Swiet M (1976) Pregnancy and maternal heart disease. Br J Hosp Med 15:353-359 10. De Swiet M. Dorrington Ward P. Fidler J (1983) Prolonged heparin therapy in pregnancy causes bone demineralisation. Br J Obstet Gynecol 10: 1129-1134 11. Edmunds LH Jr (1982) Thromboembolic complications of current cardiac valvular prostheses. Ann Thorac Surg 34: 96- 106 12. Fernandez E, Montoy L, Recaesens E (1976) Protesis valvulares y embarazo. Rev Clin Esp 140: 537-541 13. Fillmore SJ, MC DeVitt E (1970) Effects of coumarin compounds on the fetus. Ann Intern Med 73: 731-734 14. Gervin AS (1975) Complications of heparin therapy. Surg Gynecol Obstet 140:789-794 15. Hall JG, Pauli RM. Wilson KM (1980) Maternal and fetal sequelae of anticoagulation during pregnancy. Am J Med 68: 122-140 16. Harker LA, Siichter SJ (1970) Studies of platelet and fibrinogen kinetics in patients with prosthetic heart valves. N Engl J Med 283: 1302-1304 17. Hedstrand H. Cullhed I (1968) Pregnancy in patients with prosthetic heart valves (Starr-Edwards). Stand J Thorac Cardiov Surg 2: 1966198 18. Hirsh J, Cade JF, Gallus AS (1970) Fetal effects of coumadin administered during pregnancy. Blood 36: 623 -627 19. Holzgreve W, Carey JC, Hall BD (1976) Warfarin induced fetal abnormalities. Lancet II:9144915 20. lbarra-Perez C. Azevalo Toledo N, Alvarez de La Cadena 0, Noriega Guerra L (1976) The course of pregnancy in patients with artificial heart valves. Am J Med 61: 5044512 21 lsmail BM, Abid F, Trabelsi S. Tartak M, Ferih M (1986) Cardiac valve prostheses, anticoagulation and pregnancy. Br Heart J 55: 101- 105 22 lturbe Alessio I, Del Carmen Fonseca M, Mutchinik 0, Santos MA, Zajarias A, Salazar E (1986) Risks of anticoagulant therapy in pregnant women with artificial heart valves. N Engl J Med 315: 1390-1393 23 Kraus AP, Perlow A, Singer K (1949) Danger of dicoumarol treatment in pregnancy. JAMA 139: 7588761 24. Lapiedra OJ, Bernal JM, Ninot S. Gonzalez I, Pastor E, Miralles PJ (1986) Open heart surgery for thrombosis of prosthetic mitral valve during pregnancy. J Cardiovasc Surg 27: 217-220 25. Laros RK Jr, Alger LS (1979) Thromboembolism and pregnancy. Clin Obstet Gynecol 22: 871-878 26. Larreca JL. Nuanez L. Reque JA, Gil Agnado M. Matarros P, Minguez JA (1983)Pregnancy and mechanical valve prostheses: a high risk situation for the mother and the fetus. Ann Thorac Surg 36: 459-463 27. Luth DJ, Noller KL, Spike11 JA Jr, Danielson GK, Fish CR (1978)Pregnancy and its complications following cardiac valve prostheses. Ann J Obstet Gynecol 131:460-468 28. Mahairas GH, Veingold AB (1963) Fetal hazard with anticoagulant therapy. Am J Obstet Gynecol 85: 234-237 29. Oakley C, Doherty P (1976) Pregnancy in patients after valve replacement. Br Heart J 38: 1140-l 146 30. Oakley CM, Hawkins DF (1983) Pregnancy in patients with prosthetic heart valves. Br Med J 287: 358-362 31. Oakley C (1987) Valve prostheses and pregnancy. Br Heart J 58:303-305 32. Orme MLE, Lewis PJ, De Swiet M (1977) May mothers given warfarin breast-feed their infants? Br Med J i: 1564-1565
304 33. Quick AJ (1946) Experimentally induced changes in the prothrombin level of the blood. Prothrombin concentration of newborn pups of a mother given dicoumarol before parturition. J Biol Chem 164: 371-374 34. Salazar E, Zajarias A, Gutierez N, Iturbe I(l984) The problem of cardiac valve prostheses, anticoagulants and pregnancy. Circulation 70 [Suppl I]: 169-177 35. Saka DM, Marx GF (1976) Management of a partourient with cardiac valve prosthesis. Anaesth Analg 55:214-216 36. Sareli PS, England MJ, Berk MR, Marcus RH, Epstein M, Driscoll J, Meyer T, McIntyre J, Van Gelderen C (1989) Maternal and fetal sequelae of anticoagulation during pregnancy in patients with mechanical heart valve prostheses. Am J Cardiol 63: 1462-1465 37. Schaffer AT (1985) The hypercoagulable states. Ann Intern Med 102:814-828 38. Shauld WL, Hall JG (1977) Multiple congenital anomalies associated with oral anticoagulants. Am J Obstet 127: 191-201
39. Stevenson RE, Burton OM, Ferlanto GS, Taylor HA (1980) Hazards of oral anticoagulants during pregnancy. JAMA 243:1549-1551 40. Sullivan JM, Harken DE, Gorlin R (1968) Pharmacological control of thromboembolic complications of cardiac valve replacement. N Engl J Med 279: 576-579
Maurizio Cotrufo, MD Institute of Medical and Surgical Cardiology Monaldi Hospital Via L. Bianchi I-80131 Naples Italy
305
Discussion Dr. M. J. Antunes (Coimbra, Portugal). The authors are to be congratulated for the excellent results obtained with their group of 20 patients with mechanical cardiac valve prostheses receiving anticoagulation during pregnancy. These results are in accordance with those published last year in the American Journal of Cardiology by the Johannesburg group in a series of 50 similar patients, except for the absence of foetal wastage which occurred in 16 of our cases (32 percent) (I feel at ease to discuss these patients, as I was directly or indirectly involved in their surgery and postoperative follow-up). I am not certain why there should be such a difference, as high foetal wastage has also been reported by most other authors. This is presumably not related only to the low mean daily dose of Warfarin (4.1 mg in the authors’ series as compared to 5.1 mg in our series). Like Dr. Cotrufo and coworkers. we had also observed a very low rate of maternal complications and of congenital malformations. I am, therefore. in complete agreement with their conclusions which indicate that pregnancy is not contraindicated in patients with mechanical prosthetic heart valves. This is based upon three assumptions: firstly, abortion and other forms of foetal wastage are not a cause for major concern as they occur naturally quite frequently in noncardiac women and are usually well accepted by the patient; secondly, the incidence of thromboembolic complications can be maintained at low levels if adequate control of anticoagulation can be achieved throughout pregnancy; thirdly, the incidence of congenital anomalies in live-born babies is low (below 5%) and those which occur are of a relatively minor severity. I would also concur with the authors in that warfarin should probably be continued throughout the pregnancy and that caesarean section should be performed preventively after temporary and brief discontinuation of anticoagulation. Nevertheless, nobody is encouraging young women with mechanial prostheses to get pregnant; only that those who have reasonably good reasons to do so should not be discouraged. This brings us to the reverse side of the coin - which prosthesis to use in the woman of childbearing age who requires valve replacement. Bioprostheses have repeatedly been shown to have accelerated biodegradation in young patients and during pregnancy, presumably because of the increased calcium metabolism. Furthermore. thromboembolic complications and valve-related mortality appear to occur more frequently than initially anticipated. Finally, the added risks of an almost certainly needed reoperation make the use of these valves an unacceptable alternative except under special circumstances. The evidence uncovered by the studies of the Naples group, of ours and of others permits a greater reassurance to these patients. Therefore, I strongly disagree with the authors’ conclusion that pregnancy should “be allowed to term only to those women who maintain a stable prothrombin ratio within therapeutic range with the ingestion of low-medium daily doses of the drug.” Not only does their study and its results not allow such conclusion but also there is no other consistent scientific evidence that could permit to advocate termination of pregnancy on sound legal and ethical grounds. I have one or two questions to Dr. Cotrufo. Are you advising lowering the doses of warfarin after the patient became pregnant? I certainly would not be party to such a recommendation. The second question is, what is the mean length of pregnancy before patients came to your clinic? Could it be that some cases of abortion were not detected or not included in your study, thus justifying the absence of foetal wastage? I wish to thank the authors for allowing me to review their manuscript and congratulate them again on the tine results obtained. I also wish to thank the Association for granting me the privilege of the floor.
Dr. K. Shyamprasad (Vellore, h&z). The anticoagulation policy of these patients must entail a compromise between a risk and benefit. The maternal risk is due to both the physiological fluid overload and anticoagulation. The risk is increased when the patient is in poor functional status and in atria1 fibrillation. The onset of labor further increases the load on the heart. For these reasons pregnancy should be dissuaded in most patients with mechanical valves. The maternal complications are less with coumarinic anticoagulation. The adequacy of anticoagulation achieved with heparin is difftcult to monitor and there are at least six reported cases of thrombosed tilting disc valves in this situation with a fatal outcome. Warfarin, of course, is easy to administer. The risk to the foetus is largely due to anticoagulation alone and is considerably reduced by using heparin which does not cross the placental barrier. Doctor Cotrufo has shown that in 20 of these patients foetal complications were absent even with coumarinic anticoagulation. In our series of 47 pregnancies, there was 1 stillbirth with no evidence of warfarin embryopathy. These numbers are small for conclusions and earlier reports of chondrodysplasia cannot be ignored. It is reasonable to presume that low dose coumarinic anticoagulation reduces the risk of foetal abnormalities. What is not known is the incidence of early foetal wastage or failure to conceive. The 20 patients in Dr. Cotrufo’x study were selected from a group of patients who were already confirmed as pregnant, but what we do not know is how many of them failed to conceive because of being on coumarinic anticoagulation. The reported incidence varies between 30% and 40% in most reports. When I reviewed our statistics, out of a total of 688 eligible women, only 48 patients conceived. Heparin in this situation may help patients in conceiving if they are desperate for a child. Maternal safety has been emphasized and therefore I disagree with the authors on the management in the last trimester. Our policy has been to switch over to heparin during the last 4 weeks. There are good reasons for this. One is the very high frequency of premature labor ~ you had one in your series - which can jeopardize what has been achieved until then. A vaginal delivery can impose a great risk to the child. And the other is the unpredictable anteparturn haemorrhage which is very much more difficult to manage if the patient is on coumadin anticoagulation. Hence, these problems can be altered if anticoagulation is switched over to heparin at least during the last 4 weeks when the mother can be hospitalized for effective anticoagulation.
Dr. M. Cotrufo: I thank the discussants for their interesting remarks. First of all I need to state that it was never demonstrated that the chemical structure of coumadin derivatives induces teratogenic effects. It means that all the negative effects on the foetus could be associated with haemorrhagic phenomena. To Dr. Antunes I reply that we do not advise the patients who become pregnant to lower the coumadin doses because it could be very dangerous. We suggest a therapeutical interruption of pregnancy to those women who need high doses of coumadin to keep the prothrombin ratio in the therapeutical range. All the patients of childbearing age discharged from our Institute after implantation of one or more mechanical valve prostheses and in coumadin anticoagulation are informed regarding the risks of an eventual pregnancy and are requested to refer to us shortly after pregnancy has been demonstrated. To Dr. Shyamprased I am pleased to reply that all our patients conceived during coumadin therapy. We think that heparin anticoagulation during the last month of pregnancy increases the risks of maternal and foetal complications and the most recent literature reports confirm our data.
Surg (1991) 5:300-305
Coumarin anticoagulation during pregnancy in patients with mechanical valve prostheses M. Cotrufo, T. S. L. de Luca, R. Calabro, G. Mastrogiovati, Institute of Medical and Surgical Cardiology,
and D. Lama
First Medical School, University
of Naples, V. Monaldi Hospital,
Naples, Italy
Abstract. Between January 1987 and December 1989,20 female patients with one mechanical valve prosthesis (MVP) for at least 1 year postoperatively were studied while on coumarin therapy for the full length of pregnancy. In each case, caesarean section was scheduled for the 38th week. Patients were selected according to the following criteria: (1) prothrombin ratios remaining within the therapeutic range for more than 85% of their total estimations in the previous 12 months with mean daily doses of warfarin less than 5 mg; (2) stable cardiac status; (3) no previous obstetric diseases and (4) full acceptance of the risks involved in the protocol. The patients were in NYHA functional class I or II. Their ages ranged from 23 to 31 years (mean 26 f 3). Ten patients had a mitral prosthesis and 10 an aortic prosthesis. Among the 20 mechanical valve prostheses, 10 were Sorin, 6 Starr-Edwards, 2 Bjsrk-Shiley, and 2 Lillehei-Kaster. Eighteen patients were in sinus rythm, 1 in chronic atrial fibrillation, and 1 had a permanent endocardial pacemaker. Nineteen were delivered by caesarean section: warfarin was withdrawn 48 h before surgery and resumed 24 h thereafter. One patient had premature spontaneous delivery at 36 weeks. The mean prothrombin ratio measured weekly in the 20 patients was 2.06 f 0.45 INR, using a mean daily warfarin dose of 4.1 mg +, 1.63. The mean value of the prothrombin ratio during caesarean section for the 19 patients was 1.23 f 0.38 INR. In the 20 live births, the mean birth weight was 2.9 kg f 0.40. There were no observed thromboembolic or haemorragic complications, no spontaneous abortions, stillbirths or deaths. All infants underwent cardiac, neurological, X-ray, and echocardiographic examination. No evidence of warfarin embryopathy was found. These findings strongly suggest that (1) the use of strictly controlled coumarin therapy with medium to low doses of warfarin, and (2) delivery by caesarean section may afford pregnant patients with mechanical valve prostheses satisfactory anticoagulation without embryopathy or haemorrhagic complications. [Eur J Cardio-thorac Surg (1991) 5:300-3051 Key words: Pregnancy - Heart valve prosthesis - Warfarin - Embryopathy
The thromboembolic risk after implantation of mechanical valve prostheses remains significantly high, even with the newer generation valves. Long-term anticoagulation is therefore necessary. The most effective therapy is obtained with oral coumarin derivatives. Pregnancy is associated with a hypercoagulability state which further emphasizes the need for anticoagulation. However, problems arise because oral anticoagulants cross the placenta. Indeed, foetal coagulability is depressed more than maternal, mainly due to the immaturity of the foetal liver. Anticoagulants continued throughout pregnancy produce a greater risk of teratogenie lesions in the first trimester, foetal haemorrhagic complications in later months, maternal haemorrhage
Read at the 4th Annual Meeting of the European Association Cardio-thoracic Surgery, Naples, September 17-19, 1990
for
during delivery, and foetal malformations or death with an overall incidence of 30%. Discontinuing anticoagulation during pregnancy, even temporarily, has given rise to disastrous thrombotic complications. Alternative anticoagulation protocols using heparin or antiplatelet agents do not solve the problem. Bioprostheses implant in women of childbearing age may facilitate pregnancy and avoid the need for anticoagulation. They may however require early reoperation due to the rapid deterioration of the biological tissue induced by young age and pregnancy. On this basis, management of pregnant patients with mechanical valve prostheses remains controversial. It is mandatory to inform patients about the risks and limitations they will face with pregnancy and eventually to recommend therapeutic abortion. Nevertheless, many women, especially those with no children, are highly motivated to continue pregnancy even at the risk of their
301
own lives. In these cases, we have a great responsibility for their proper management. In our Institution, mechanical valve prostheses have been implanted in 3240 patients since 1976. Up until now, anticoagulation was routinely induced using sodium warfarin. The mean value for prothrombin activity which we consider within the therapeutic range is 30% t_ 5%, corresponding to a mean value for the prothrombin ratio of 2.07 + 0.30 INR. Recently, dypiridamole has been added to warfarin in cases with high thromboembolic risk (Starr-Edwards ball valves, giant left atrium, previous thromboembolic episodes). This study followed 20 patients with mechanical valve prostheses who became pregnant while on warfarin therapy alone. Literature reports by Quick [33] Kraus, Perlow, and Singer [23] and Hirsh, Cade and Gallus [18] sought to predict the risk factors for foetal mortality in dogs and rabbits exposed to coumarin derivatives during pregnancy. They found that large doses of the drug, administration of coumarin until term, and foetal exposure to the trauma of spontaneous delivery all significantly increased the risk of foetal death. To our knowledge, no clinical trials have been done to investigate these experimental data. The purpose of our study was to show that the adverse effects of warfarin during pregnancy could be reduced in selected patients. Starting with our clinical observations as well as those in the aforementioned reports, we hypothesized adequate anticoagulation could be achieved in selected patients with lower doses of warfarin. The basic points of the study include (a) avoiding excessive doses of the drug, (b) avoiding the trauma of spontaneous delivery, and (c) allowing anticoagulation to be discontinued just prior to caesarean section. Material and methods All patients of childbearing age discharged from our Institution after implantation of one or more valve prostheses were informed of the risks of pregnancy during anticoagulation and were asked to refer to us in case of pregnancy. Between January 1987 and December 1989,29 women were observed in the Out-patient clinic. Twenty women were selected for an anticoagulation protocol which included (1) sodium warfarin therapy throughout pregnancy, (2) programmed caesarian section during the 38th week, (3) discontinuance of anticoagulation 2 days before surgery and 1 day after, (4) weekly estimations of the prothrombin ratios done at our Institution, (5) monthly cardiological and obstetrical examinations in the Cardiology Clinic and in the High Risk Pregnancy Clinic affiliated to our Institution, (6) echo studies of the foetus in the 2nd, 5th and 8th month. The selection criteria included: normal coagulation system, prothrombin ratios within the therapeutic range for more than 85% of total estimations in the previous 12 months with mean daily doses of warfarin equal to or lower than 5 mg, stable cardiac status, no previous obstetric disease and a full knowledge and acceptance of the risks included in the protocol. During the same period of time, 9 women were not admitted to the trial and were urged to discontinue pregnancy. The exclusion was due to high daily doses of warfarin in 5 cases during the 12 months preceding pregnancy, unstable cardiac state in 2 cases and obstetric complications in 1 case. Four further pregnant women were treated in 1990 with the protocol described. and the results confirmed those of the present report.
The age of the 20 selected women ranged from 23 to 31 years (mean 26 years f 3); 14 were primigravid and 6 multigravid; 10 had an aortic valve prosthesis (9 tilting disk Sorin prosthesis and 1 valved Sorin conduit); 10 had a mitral valve prosthesis (6 Starr-Edwards, 2 Bjsrk-Shiley, 2 Lillehei-Kaster). All patients were in NYHA functional class I or II; 18 patients were in sinus rhythm. 1 in chronic atria1 fibrillation and 1 had a permanent endocardial pacemaker; 8 patients were taking digitalis and diuretics during pregnancy. All patients became pregnant without planning the date of conception and while on warfarin therapy. The mean time interval between implantation of the valve prosthesis and pregnancy was 3.5 yearsk2.3. None of the patients were hospitalized when pregnancy was confirmed and no special limitations were imposed on their routine activities. All patients remained under our direct surveillance during pregnancy, delivery and puerperium. The prothrombin activity was determined every week during pregnancy and puerperium and every day during the week of hospitalization for caesarian section using rabbit brain calcium thromboplastin. The prothrombin ratio was determined in blood samples from the infant soon after birth and 12 h later after vitamin K had been administered. The neonates underwent immediate examination by the clinical geneticist, pediatrician, cardiologist and neurologist and have been followed by echo and X-ray studies during the following 12-24 months.
Results Twenty-nine pregnant women with one mechanical valve prosthesis in place for at least 1 year were observed from January 1987 until December 1989. Nine patients were advised to discontinue their high risk pregnancies and 20 women were followed throughout pregnancy and puerperium. All were anticoagulated with daily doses of sodium warfarin throughout pregnancy with a mean daily dose of 4.1 mgt_ 1.63. The prothrombin ratio was determined 12 h after the oral ingestion of the drug and the mean value for the prothrombin ratios of the weekly measurements in the 20 patients was 2.06kO.45 INR. Nineteen patients delivered by caesarian section during the 38th week: warfarin was interrupted 48 h before surgery and restarted 24 h later. In no case was vitamin K administered to the mother before surgery. The mean value of the prothrombin ratio during caesarian section in 19 patients was 1.23 If:0.38 INR. One patient had a premature spontaneous vaginal delivery during the 36th week. No significant complications were observed during pregnancy and puerperium. One patient showed right heart failure during the 3rd month which was succesfully treated with digitalis and diuretics. No thromboembolic or haemorrhagic complications, abortions, stillbirths or deaths were observed. All 20 babies were born alive: the mean birth weight was 2.9 kgk0.40. The prothrombin ratio was determined in blood samples from 6 neonates soon after birth and before vitamin K was administered and the mean value was 2.90 INR. Eight male and 12 female neonates were examined soon after birth and 3.6 and 12 months later by clinical geneticist and no sign of warfarin embryopathy was found. The clinical follow-up by X-ray and echo studies demonstrated a small muscular VSD, a fetal growth defect and obesity in one case each but no skeletal abnor-
302
malities, mental or neurological retardation. All mothers were allowed to breast-feed their infants and no haemorrhagic complications were observed [32]. Discussion
The most effective prevention of thromboembolic complications after implantation of mechanical valve prostheses is achieved using coumarin derivatives [I I]. A hypercoagulable state exists during pregnancy and the postpartum period [17, 25, 35, 371 when concentrations of clotting factors, platelet turnover and viscosity are increased and fibrinolysis is diminished [31]. These changes increase the risk of prosthetic valve thrombosis and the need for anticoagulation; discontinuing the drug, even temporarily, has given rise to disastrous complications and even death [29]. Several reports indicate that vitamin K antagonists are the most effective anticoagulants during pregnancy [21] but they are of low molecular weight and may cross the placenta, as Quick demonstrated in experimental animals [33]. The foetus becomes overanticoagulated because coumarin crosses the placenta but vitamin K dependent coagulation factors in the mother’s blood do not. In addition, the immature foetal liver is more sensitive to the anti vitamin K effect of coumarin [2, 311. Thus the foetal prothrombin time may be considerably prolonged even when that of the mother is normal [l]. The use of coumarin derivatives during pregnancy carries serious risks to the foetus such as coumarin embryopathy and death, especially if the drugs are administered during the first 3 months [20, 34, 391. The warfarin syndrome is characterized by chondrodysplasia with nasal hypoplasia, optic atrophy and dwarfism [37]. It has been suggested that the teratogenic effects may develop when the embryo is exposed to these agents between the 6th and 9th week of gestation [15]. The use of coumarin derivatives at the time of delivery has been related to an increased incidence of stillbirths and neonatal deaths primarily as a result of haemorrhage [27, 281. Neurological complications, eye abnormalities, placentae haemorrhage and spontaneous abortions which have occurred after the administration of coumarin later on in pregnancy are almost certainly attributable to haemorrhage [31]. It is uncertain whether the “warfarin syndrome” is produced by the chemical structure of the drug or by the effect of anticoagulation on the developing cartilagenous tissue. It was reported that both teratogenic and adverse growth effects induced by warfarin exposure during pregnancy may show up subsequent to birth [19]. No significant difference was found between pregnant patients exposed to acenocoumarin and those exposed to sodium warfarin [20]. Little information has been reported on the use of phenindione [30]. In an extensive study of over 418 pregnancies compiled from various sources, Hall, Pauli and Wilson [15] revealed severe complications due to coumarin derivatives in 30%. Several alternative anticoagulation protocols have been used in pregnancy to avoid the risks of coumarin
derivatives. Because heparin does not cross the placenta, it has been administered subcutaneously [34] or intravenously for full anticoagulation. However, both routes of administration carried a substantial risk [22]. Moreover, heparin can cause osteoporosis, thrombocytopenia and alopecia [8, lo]. The long term use of heparin in pregnancy is inconvenient as it must be given parenterally; monitoring is difficult and several cases of valve thrombosis during heparin anticoagulation have been reported [31]. Even though heparin is not teratogenic, there are case reports showing more than 20% foetal complications and a higher rate of maternal complications including haemorrhage, abortion and prematurity [14]. Dipyridamole and acetylsalicylic acid are efficient blockers of white thrombus formation when associated with coumarin therapy [16,40] but they are not reliable when used without oral anticoagulants in terms of preventing thrombosis of the prosthesis or embolic complications [4, 20, 34,371. Moreover it has been suggested that the addition of dipyridamole to warfarin may increase the rate of foetal wastage [7, 361. The use of bioprostheses in patients of childbearing age initially seemed to be a solution to the problem [6,12, 26,341. The fact remains that anticoagulation is necessary in a significant number of patients due to previous embolism, poor haemodynamic state, left mega-atrium with or without thrombosis, atria1 fibrillation, and a high thromboembolic risk secondary to malfunction of the bioprosthesis. It has been clearly demostrated that young age and pregnancy induce rapid deterioration of the bioprostheses requiring early reoperation [24]. Almost three decades of experimental and clinical research on the management of pregnant patients with artificial valve prostheses remain inconclusive: Oakley in 1968 [3] suggested routine sterilization of young women with valve prostheses and the same author in 1987 [31] confirmed that women with artificial valves may even be advised against pregnancy because of the serious risks to both mother and child. Our experience with artificial valve prostheses beginning in 1976 and based on a patient population of 3240 convinced us that the most effective and safe long-term anticoagulation is obtained using coumarin derivatives. The therapeutic range for the prothrombin ratio is obtained by giving daily doses of sodium warfarin, the dosage varying widely between 1 mg and 15 mg from patient to patient based on the patient’s own control ratio. The same observations have been made in women of childbearing age whereby they could be classified in two groups according to the daily dose of sodium warfarin which was medium to low in the first group and medium to high in the second group. Pregnancy does not seem to modify the warfarin dosage. Considering that warfarin has a direct effect on the foetus which reacts independently from the mother, the correlation between drug dose and foetal effects seems to be fundamental and the absence of warfarin embryopathy observed with low doses of anticoagulants is probably due to the fact that haemorrhagic phenomena in the foetus are prevented. These observations are not unique. Quick [33] and Kraus et al. [23] in pregnant dogs and rabbits exposed to
303
coumarin derivatives demonstrated large doses of the drug given up to term to be the greatest single risk factor for foetal haemorrhage and death. Hirsh et al. confirmed this result in rabbits and further suggested that an additional risk factor was the trauma of spontaneous delivery [I 81. Since then, a few clinical reports have mentioned foetal complications during coumarin therapy to be more likely to occurr when oral anticoagulation is excessive and when there is a previous history of miscarriages [5, 13, 18, 201. Sareli et al. [36] reported a low maternal risk for thromboembolism and haemorrhage with the newer generation mechanical heart valves of lower thrombogenic potential and with lower levels of anticoagulation. Nevertheless the rate of foetal wastage and neonatal morbidity associated with warfarin anticoagulation remains high. Oakley’s recently reported experience [31] suggests that a successful outcome can be expected in about two thirds of cases when coumarin derivatives are used throughout pregnancy. The risk of a newborn child having a disabling coumarin embryopathy seems to be less than 5%. The present study shows that the incidence of complications of coumarin therapy used throughout pregnancy may be reduced to almost zero in pregnancies where women are on a coumarin regimen with medium-low daily doses of the drug. The main haemorrhage risk is spontaneous delivery on full anticoagulation and is avoided through caesarean section during a short warfarin discontinuance. The same therapeutic attitude is acceptable in patients with chronic or intermittent atria1 fibrillation.
Conclusions
The combination of mechanical valve prostheses, pregnancy and anticoagulation represents a high risk which is very difficult to manage. Strongly motivated women in a stable cardiac state and without previous obstetric disease may have a good chance to endure pregnancy if the thromboembolic risk is reduced by continuous warfarin anticoagulation and the haemorrhagic risk is prevented by delivering through caesarian section while briefly interrupting warfarin therapy. This policy seems to reduce significantly the incidence of warfarin embryopathy allowing pregnancy to term in those women who maintain a stable prothrombin ratio within the therapeutic range when ingesting low to medium daily doses of the drug.
References Becker RM (1983) Intracardiac surgery in pregnant women. Ann Thorac Surg 36: 453-458 Ben lsmail M, Fekih M, Tartak M, Chelli M (1979) Protheses valvulaires cardiaques et grossesse. Arch Ma1 Coeur 72: 192-199 Bennett GG, Oakley CM (1968) Pregnancy in a patient with a mitral valve prosthesis. Lancet I: 616-618 Biale Y. Cantos A, Lewen Thal H, Guerron M (1980) The course of pregnancy: patients with artificial heart valve treated with dipyridamole. lnt J Obstet Gynecol 81: 1288132
5. Bloomfield DK (1970) Fetal deaths and malformations associated with the use of coumarin derivatives in pregnancy. A critical review. Am J Obstet Gynecol 107: 883-886 6. Chen WWC, Chan CS, Lee PK, Wang RYC, Wong VCW (1982) Pregnancy in patients with prosthetic heart valves: an experience with 45 pregnancies. Q J Med 51: 358-365 7. Chesebro JH, Adams PC, Fuster V (1986) Antithrombotic therapy in patients with valvular heart disease and prosthetic heart valves. JACC 8:41B-56B 8. Chong BH, Pitney WR, Castaldi PA (1982) Heparin induced thrombocytopenia. Lancet II: 1246- 1248 9. De Swiet M (1976) Pregnancy and maternal heart disease. Br J Hosp Med 15:353-359 10. De Swiet M. Dorrington Ward P. Fidler J (1983) Prolonged heparin therapy in pregnancy causes bone demineralisation. Br J Obstet Gynecol 10: 1129-1134 11. Edmunds LH Jr (1982) Thromboembolic complications of current cardiac valvular prostheses. Ann Thorac Surg 34: 96- 106 12. Fernandez E, Montoy L, Recaesens E (1976) Protesis valvulares y embarazo. Rev Clin Esp 140: 537-541 13. Fillmore SJ, MC DeVitt E (1970) Effects of coumarin compounds on the fetus. Ann Intern Med 73: 731-734 14. Gervin AS (1975) Complications of heparin therapy. Surg Gynecol Obstet 140:789-794 15. Hall JG, Pauli RM. Wilson KM (1980) Maternal and fetal sequelae of anticoagulation during pregnancy. Am J Med 68: 122-140 16. Harker LA, Siichter SJ (1970) Studies of platelet and fibrinogen kinetics in patients with prosthetic heart valves. N Engl J Med 283: 1302-1304 17. Hedstrand H. Cullhed I (1968) Pregnancy in patients with prosthetic heart valves (Starr-Edwards). Stand J Thorac Cardiov Surg 2: 1966198 18. Hirsh J, Cade JF, Gallus AS (1970) Fetal effects of coumadin administered during pregnancy. Blood 36: 623 -627 19. Holzgreve W, Carey JC, Hall BD (1976) Warfarin induced fetal abnormalities. Lancet II:9144915 20. lbarra-Perez C. Azevalo Toledo N, Alvarez de La Cadena 0, Noriega Guerra L (1976) The course of pregnancy in patients with artificial heart valves. Am J Med 61: 5044512 21 lsmail BM, Abid F, Trabelsi S. Tartak M, Ferih M (1986) Cardiac valve prostheses, anticoagulation and pregnancy. Br Heart J 55: 101- 105 22 lturbe Alessio I, Del Carmen Fonseca M, Mutchinik 0, Santos MA, Zajarias A, Salazar E (1986) Risks of anticoagulant therapy in pregnant women with artificial heart valves. N Engl J Med 315: 1390-1393 23 Kraus AP, Perlow A, Singer K (1949) Danger of dicoumarol treatment in pregnancy. JAMA 139: 7588761 24. Lapiedra OJ, Bernal JM, Ninot S. Gonzalez I, Pastor E, Miralles PJ (1986) Open heart surgery for thrombosis of prosthetic mitral valve during pregnancy. J Cardiovasc Surg 27: 217-220 25. Laros RK Jr, Alger LS (1979) Thromboembolism and pregnancy. Clin Obstet Gynecol 22: 871-878 26. Larreca JL. Nuanez L. Reque JA, Gil Agnado M. Matarros P, Minguez JA (1983)Pregnancy and mechanical valve prostheses: a high risk situation for the mother and the fetus. Ann Thorac Surg 36: 459-463 27. Luth DJ, Noller KL, Spike11 JA Jr, Danielson GK, Fish CR (1978)Pregnancy and its complications following cardiac valve prostheses. Ann J Obstet Gynecol 131:460-468 28. Mahairas GH, Veingold AB (1963) Fetal hazard with anticoagulant therapy. Am J Obstet Gynecol 85: 234-237 29. Oakley C, Doherty P (1976) Pregnancy in patients after valve replacement. Br Heart J 38: 1140-l 146 30. Oakley CM, Hawkins DF (1983) Pregnancy in patients with prosthetic heart valves. Br Med J 287: 358-362 31. Oakley C (1987) Valve prostheses and pregnancy. Br Heart J 58:303-305 32. Orme MLE, Lewis PJ, De Swiet M (1977) May mothers given warfarin breast-feed their infants? Br Med J i: 1564-1565
304 33. Quick AJ (1946) Experimentally induced changes in the prothrombin level of the blood. Prothrombin concentration of newborn pups of a mother given dicoumarol before parturition. J Biol Chem 164: 371-374 34. Salazar E, Zajarias A, Gutierez N, Iturbe I(l984) The problem of cardiac valve prostheses, anticoagulants and pregnancy. Circulation 70 [Suppl I]: 169-177 35. Saka DM, Marx GF (1976) Management of a partourient with cardiac valve prosthesis. Anaesth Analg 55:214-216 36. Sareli PS, England MJ, Berk MR, Marcus RH, Epstein M, Driscoll J, Meyer T, McIntyre J, Van Gelderen C (1989) Maternal and fetal sequelae of anticoagulation during pregnancy in patients with mechanical heart valve prostheses. Am J Cardiol 63: 1462-1465 37. Schaffer AT (1985) The hypercoagulable states. Ann Intern Med 102:814-828 38. Shauld WL, Hall JG (1977) Multiple congenital anomalies associated with oral anticoagulants. Am J Obstet 127: 191-201
39. Stevenson RE, Burton OM, Ferlanto GS, Taylor HA (1980) Hazards of oral anticoagulants during pregnancy. JAMA 243:1549-1551 40. Sullivan JM, Harken DE, Gorlin R (1968) Pharmacological control of thromboembolic complications of cardiac valve replacement. N Engl J Med 279: 576-579
Maurizio Cotrufo, MD Institute of Medical and Surgical Cardiology Monaldi Hospital Via L. Bianchi I-80131 Naples Italy
305
Discussion Dr. M. J. Antunes (Coimbra, Portugal). The authors are to be congratulated for the excellent results obtained with their group of 20 patients with mechanical cardiac valve prostheses receiving anticoagulation during pregnancy. These results are in accordance with those published last year in the American Journal of Cardiology by the Johannesburg group in a series of 50 similar patients, except for the absence of foetal wastage which occurred in 16 of our cases (32 percent) (I feel at ease to discuss these patients, as I was directly or indirectly involved in their surgery and postoperative follow-up). I am not certain why there should be such a difference, as high foetal wastage has also been reported by most other authors. This is presumably not related only to the low mean daily dose of Warfarin (4.1 mg in the authors’ series as compared to 5.1 mg in our series). Like Dr. Cotrufo and coworkers. we had also observed a very low rate of maternal complications and of congenital malformations. I am, therefore. in complete agreement with their conclusions which indicate that pregnancy is not contraindicated in patients with mechanical prosthetic heart valves. This is based upon three assumptions: firstly, abortion and other forms of foetal wastage are not a cause for major concern as they occur naturally quite frequently in noncardiac women and are usually well accepted by the patient; secondly, the incidence of thromboembolic complications can be maintained at low levels if adequate control of anticoagulation can be achieved throughout pregnancy; thirdly, the incidence of congenital anomalies in live-born babies is low (below 5%) and those which occur are of a relatively minor severity. I would also concur with the authors in that warfarin should probably be continued throughout the pregnancy and that caesarean section should be performed preventively after temporary and brief discontinuation of anticoagulation. Nevertheless, nobody is encouraging young women with mechanial prostheses to get pregnant; only that those who have reasonably good reasons to do so should not be discouraged. This brings us to the reverse side of the coin - which prosthesis to use in the woman of childbearing age who requires valve replacement. Bioprostheses have repeatedly been shown to have accelerated biodegradation in young patients and during pregnancy, presumably because of the increased calcium metabolism. Furthermore. thromboembolic complications and valve-related mortality appear to occur more frequently than initially anticipated. Finally, the added risks of an almost certainly needed reoperation make the use of these valves an unacceptable alternative except under special circumstances. The evidence uncovered by the studies of the Naples group, of ours and of others permits a greater reassurance to these patients. Therefore, I strongly disagree with the authors’ conclusion that pregnancy should “be allowed to term only to those women who maintain a stable prothrombin ratio within therapeutic range with the ingestion of low-medium daily doses of the drug.” Not only does their study and its results not allow such conclusion but also there is no other consistent scientific evidence that could permit to advocate termination of pregnancy on sound legal and ethical grounds. I have one or two questions to Dr. Cotrufo. Are you advising lowering the doses of warfarin after the patient became pregnant? I certainly would not be party to such a recommendation. The second question is, what is the mean length of pregnancy before patients came to your clinic? Could it be that some cases of abortion were not detected or not included in your study, thus justifying the absence of foetal wastage? I wish to thank the authors for allowing me to review their manuscript and congratulate them again on the tine results obtained. I also wish to thank the Association for granting me the privilege of the floor.
Dr. K. Shyamprasad (Vellore, h&z). The anticoagulation policy of these patients must entail a compromise between a risk and benefit. The maternal risk is due to both the physiological fluid overload and anticoagulation. The risk is increased when the patient is in poor functional status and in atria1 fibrillation. The onset of labor further increases the load on the heart. For these reasons pregnancy should be dissuaded in most patients with mechanical valves. The maternal complications are less with coumarinic anticoagulation. The adequacy of anticoagulation achieved with heparin is difftcult to monitor and there are at least six reported cases of thrombosed tilting disc valves in this situation with a fatal outcome. Warfarin, of course, is easy to administer. The risk to the foetus is largely due to anticoagulation alone and is considerably reduced by using heparin which does not cross the placental barrier. Doctor Cotrufo has shown that in 20 of these patients foetal complications were absent even with coumarinic anticoagulation. In our series of 47 pregnancies, there was 1 stillbirth with no evidence of warfarin embryopathy. These numbers are small for conclusions and earlier reports of chondrodysplasia cannot be ignored. It is reasonable to presume that low dose coumarinic anticoagulation reduces the risk of foetal abnormalities. What is not known is the incidence of early foetal wastage or failure to conceive. The 20 patients in Dr. Cotrufo’x study were selected from a group of patients who were already confirmed as pregnant, but what we do not know is how many of them failed to conceive because of being on coumarinic anticoagulation. The reported incidence varies between 30% and 40% in most reports. When I reviewed our statistics, out of a total of 688 eligible women, only 48 patients conceived. Heparin in this situation may help patients in conceiving if they are desperate for a child. Maternal safety has been emphasized and therefore I disagree with the authors on the management in the last trimester. Our policy has been to switch over to heparin during the last 4 weeks. There are good reasons for this. One is the very high frequency of premature labor ~ you had one in your series - which can jeopardize what has been achieved until then. A vaginal delivery can impose a great risk to the child. And the other is the unpredictable anteparturn haemorrhage which is very much more difficult to manage if the patient is on coumadin anticoagulation. Hence, these problems can be altered if anticoagulation is switched over to heparin at least during the last 4 weeks when the mother can be hospitalized for effective anticoagulation.
Dr. M. Cotrufo: I thank the discussants for their interesting remarks. First of all I need to state that it was never demonstrated that the chemical structure of coumadin derivatives induces teratogenic effects. It means that all the negative effects on the foetus could be associated with haemorrhagic phenomena. To Dr. Antunes I reply that we do not advise the patients who become pregnant to lower the coumadin doses because it could be very dangerous. We suggest a therapeutical interruption of pregnancy to those women who need high doses of coumadin to keep the prothrombin ratio in the therapeutical range. All the patients of childbearing age discharged from our Institute after implantation of one or more mechanical valve prostheses and in coumadin anticoagulation are informed regarding the risks of an eventual pregnancy and are requested to refer to us shortly after pregnancy has been demonstrated. To Dr. Shyamprased I am pleased to reply that all our patients conceived during coumadin therapy. We think that heparin anticoagulation during the last month of pregnancy increases the risks of maternal and foetal complications and the most recent literature reports confirm our data.
