- Email: [email protected]
The long-term risk of warfarin sodium therapy and the incidence of thromboembolism in children after prosthetic cardiac valve replacement
J
THORAC CARDIOVASC SURG
1987;93:551-4
The long-term risk of warfarin sodium therapy and the incidence of thromboembolism in children after prosthetic cardiac valve replacement Thirty children less than 18 years of age underwent cardiac valve replacement with a prosthetic valve
between 1967and 1984and havebeenfoUowed up for a meanof 6 years (range 1 to 17 years). Their mean age at the time of operation was 13 years (range 6 to 17 years). AU patients were begun on a regimenof warfarin before bospital discharge. One major and four minor bleeding episodes occurred in 211 patient-years of warfarin therapy, an incidence of 2.3 per 100 patient-years.Three of those five episodes occurred in patients wbo werereceiving excessively anticoagulation or wbo were participating in pbysical activities inappropriate for a patient on warfarin therapy. Thus, the majority of the bleeding episodes were preventable. There were five thromboembolic events in 211 patient-years, an incidence of 2.3 per 100 patient-years. Three of those five patients had intentionally stopped their warfarin therapy. The majority of thromboembolic episodes, like the bleeding episodes, were preventable. Eight teenage patients were noncompliant with the warfarin therapy. More than one third of that group experienced a thromboembolic event, an incidence of 5.5 per 100 patient-years(55 patient-years). Twenty-two patients adhered to the warfarinregimen and only two (9 %) of them had a thromboembolic event, an incidence of 1.3 per 100 patient-years (156 patient-years). Warfarin therapy presented no greater risk of serious bleeding to this pediatric age group than it does to an adult age group. The incidence of thromboembolism among these patients was less than that which is generaUy reported for adult patients. Discontinuation of or noncompliance witb warfarin therapy substantiaUy increased the risk of thromboembolism. Continuous warfarin tberapy is recommended for every child after prosthetic valve replacement.
Scott Stewart, M.D., Diane Cianciotta, R.N., Chloe Alexson, M.D., and James Manning, M.D., Rochester, N. Y.
h e operative risk of cardiac valve replacement in children has declined significantly in the past decade and nowapproachesthat of the adult in several reported series.':' However, the ongoing threat of thromboembolismand the frequentlyexpressed concern regarding the riskof warfarin sodium (Coumadin) therapy continue to stimulate discussion and controversy as to the most efficacious way to provideanticoagulation in children. A perceived increased risk of warfarin therapy in the child and its failure to provide complete protection against From the Division of Cardiothoracic Surgery and the Department of Pediatrics, University of Rochester Medical Center, Rochester, N. Y. Received for publication March 10, 1986. Accepted for publication May 20, 1986. Address for reprints: Scott Stewart, M.D., Department of Surgery, University of Rochester Medical Center, 601 Elmwood Ave., Rochester, N. Y. 14642.
thromboembolism has led some to advocate the alternative use of aspirin and dipyridamole.': 5 Others have used no anticoagulation in smaller series of patients.s 3, 6 The late follow-up experience of every pediatric patient undergoing prosthetic cardiac valve replacement at the University of Rochester Medical Center was reviewed to determine how effective a deterrent warfarin was to thromboembolism and the incidence of bleeding associated with its use. Metbods and patients The hospitaland office chart and the operative record of every patient less than 18 years of age who had undergonea cardiac valvereplacement at the University of Rochester Medical Center were reviewed. The patient's cardiologist and primary physician were contacted by both letter and telephone for up-to-date follow-up information. In addition, direct telephone contact was made with the patient's parents or, when 551
5 5 2 Stewart et al.
appropriate, the patient himself. Information was specifically sought concerning (1) the current clinical status of the patient, (2) history relevant to unusual bleeding episodes or thromboembolism, and (3) compliance with warfarin therapy. Thirty patients were identified who had been discharged from the hospital after prosthetic valve replacement and had at least 1 year of follow-up. This study includes all patients operated on between 1967 and 1984. There were 16 male and 14 female patients. Their mean age was 13 years with a range of 6 to 17 years. Twenty-eight patients had a single valve replaced and two had double valve replacement. Three patients, each having had a mitral valve replacement, required a second mitral valve replacement. One patient who had double valve replacement required a second operation to replace both of those valves and the tricuspid valve as well. Thus, 30 patients ultimately had 38 valve replacements. Twelve patients had aortic valve replacement; 14, mitral valve replacement; one, pulmonary valve replacement; two, aortic and mitral valve replacement; and one, tricuspid valve replacement. Nineteen patients had congenital valve disease, 10 had rheumatic valve disease, and one had a valve injury. Ten of 12 patients receiving aortic valve replacement had congenital valve disease and seven of 14 having mitral valve replacement had congenital disease. Since this study includes all patients undergoing valve replacement over a 17 year period by three different surgeons, a wide variety of prostheses have been used. The Starr-Edwards and, more recently, the Bjork-Shiley prostheses have been the predominant prostheses used for mitral valve replacement. The Starr-Edwards and, again, more recently, the Lillehei-Kaster prostheses have been the predominant prostheses for aortic valve replacement. These 30 patients have been followed up for a mean of 6 years. The range of the follow-up period is 1 to 17 years. Four patients have been lost to current follow-up, but accurate information was available for each of them for a minimum of 1 year after operation. One died of aortic dissection 1 year after operation; one died at the time of operation to replace another valve 3 years after the initial operation, and two did not return for followup after 3 and 11 years of consecutive visits. The total number of follow-up is 211 patient-years.
Results Twenty-three patients were in New York Heart Association Class I at the time of their most recent
The Journal of Thoracic and Cardiovascular Surgery
follow-up contact. Six patients were in Class II and one was in Class III. There were three late deaths. The potential complications of valve replacement and warfarin therapy, thromboembolism, and bleeding were reviewed in terms of their incidence and severity. In addition, bleeding events were examined to determine if they might have been prevented either by a more appropriate prothrombin time or modification of patient activity. The relationship of thromboembolism to warfarin compliance was also reviewed. Bleeding. Bleeding complications were classified as fatal, serious, or minor as suggested by Edmunds.' Twenty-five patients had either no problems or only minimal bleeding problems such as epistaxis or easy bruising. Four had minor bleeding episodes and one had a serious bleeding episode. Three had bleeding associated with trauma but none required hospitalization or transfusion. One was injured while riding a minibike when his prosthrombin time was inordinately high. One had bleeding about a joint after a football injury, and the third had a single episode of hematuria after a prolonged period of jogging. One has had heavy menses in the presence of an appropriate prothrombin time. Finally, one had hematuria with typhoid fever at a time when he was inappropriately receiving both heparin and warfarin anticoagulation. He required transfusion. Two teenage females had intentionally overdosed on warfarin but did not have any deleterious consequences. There have been no instances of intracerebral or gastrointestinal hemorrhage. Thus, five patients have had bleeding episodes that they otherwise might not have had if they had not been on warfarin therapy. No episode per se necessitated hospitalization and only one necessitated transfusion. Three of the episodes occurred in patients who either were participating in activities that they had been advised against or who were receiving excessive anticoagulation. Thus, only two patients have had bleeding episodes that were probably not preventable. There were a total of five bleeding episodes in 211 patient-years of warfarin therapy or an incidence of 2.3 per 100 patientyears. Thromboembolism. Twenty-five patients have not had a thromboembolic event. Three patients have had a single cerebral embolus. One had hemiparesis, one had a temporal lobe embolus with a brief period of amnesia, and the third had an episode of expressiveaphasia. Each patient has had a complete recovery without residual neurologic deficit. Two of those patients were appropriately anticoagulated while the third patient had intentionally stopped taking warfarin. A fourth patient had
Volume 93 Number 4 April 1987
stopped taking warfarin and had a myocardial infarction that was likely the result of an embolic event. A fifth patient also intentionally stopped taking warfarin and, subsequently, the mitral valve prosthesis thrombosed. The valve was replaced at an emergency operation and she survived. Thus, three of the five thromboembolic episodes were preventable. There were five thromboembolic episodes in 211 patient-years or an incidence of 2.3 per 100 patient-years. The incidence was 1.0 per 100 patientyears for patients with a prosthesis in the aortic position and 3.9 for the mitral position. Warfarin compliance. Eight patients were identified who were noncompliant with their warfarin regimen. More than one third (3/8) of them experienced a thromboembolic event whereas only 9% (2/22) of those complying with their medication schedule experienced a similar event. The incidence of thromboembolism in the noncompliant group was 5.5 per 100 patient-years (55 patient-years) whereas it was only 1.3 per 100 patientyears (156 patient-years) among those who took their warfarin on a regular basis.
Discussion The best method to reduce the risk of thromboembolism and yet minimize the hazard of serious bleeding in the child who has received a prosthetic cardiac valve remains controversial. Three medication management programs for these children have been discussed in the literature; (1) warfarin.v->" (2) aspirin with or without dipyridamole.!" and (3) no anticoagulation.s '" There have been three major reports of children who have been managed without anticoagulation after prosthetic valve replacement. 2,3.6 The length of follow-up was 53 and 80 patient-years in two of those reports and could not be ascertained in the third. One series used the St. Jude Medical valve exclusively and reported a thromboembolic incidence of 2.6 per 100 patient-years during a mean follow-up of about 2 years." The second report had an 80 patient-year follow-up and the Starr-Edwards prosthesis was used predominantly.' The thromboembolism incidence was 2.9 per 100 patient-years in the mitral position and 4.9 in the aortic position. The third report was limited to aortic valve replacement and does not provide data that would permit calculating the number of patient-years followUp. 2 There were only 14 patients in this series and none had thromboemboli, but it is not possible to determine how many of them had tissue valves and how many had prosthetic valves. These limited data suggest that children, like adults, who have prosthetic cardiac valves in
Prosthetic valve replacement in children 5 5 3
place are at a greater risk of thromboembolism if they are left unprotected against thrombus formation than if they are receiving warfarin. The combination of aspirin and dipyridamole has been used as a compromise to provide protection against thrombus formation and still avoid the potential bleeding complications of warfarin therapy. Although this combination, or aspirin alone, has been used sporadically for a number of years there are only two recent reports that have specifically evaluated its effectiveness in children. r.s Both reports are preliminary since their patient-year follow-up is limited to only 14 and 39 years, respectively. Weinstein, Mavroudis, and Ebert' reported no incidence of thromboembolism among 12 patients. A late follow-up of this series will be very important. In the second series, 10 patients had been followed up for 16 patient-years and the incidence of thromboembolism was 12 per 100 patient-years. The brief length of follow-up in these two reports precludes any meaningful conclusion from their data. However, even if the data from both reports were combined the incidence of thromboembolism would still be an alarming 6 per 100 patient-years, considerably higher than what is generally reported with warfarin therapy. This suggests that the combination of aspirin and dipyridamole is not an effective method of preventing thromboembolism in the child after prosthetic valve replacement. There are only two other series of children taking warfarin in which the follow-up exceeds 200 patientyears. Gardner and associates' reported a thromboembolism incidence of 0.8 per 100 patient years in 203 patient-years. Schachner and colleagues" described a series of children who had a 224 patient-year follow-up and a thromboembolism incidence of 2 per 100 patientyears. Other series that have considerably less follow-up have reported an incidence of thromboembolism between 0 and 2 per 100 patient-years.v"!' Thus, the data indicate that warfarin therapy in the child is at least as effective in preventing thromboembolism as it is in the adult. 7 Those who have not prescribed warfarin for children have repeatedly alluded to an inherent increased risk of serious bleeding among the active patients in this age group. The data do not support this concern. The incidence of serious bleeding has ranged between 0 and 4 per 100 petient-years except in one series in which the bleeding incidence was an inordinately high 22 per 100 patient-years.v" 8, 10. II This is in the same low range that is reported for adults.' Few reports have sought out the incidence of thromboembolism among those patients who are not just
554
Stewart et al.
thought to be taking warfarin but who are actually compliant with the medication schedule. When this point was examined in this series the incidence of thromboembolism among those patients actually taking warfarin was an extremely low 1.3 per 100 patient years. Thus, warfarin is very effective in preventing thromboembolism in the child. This information has not been taken into considerabion by those who suggest that the protection that warfarin therapy provides does not outweigh the alleged risks of bleeding. We conclude from these data that warfarin therapy provides the best protection against thromboembolism in the child after prosthetic cardiac valve replacement. The risk of serious bleeding from warfarin is extremely small and does not outweight the risk of thromboembolism when warfarin is omitted. Compliance with medication schedules in the teenager is less than that in the younger or older age groups and therefore they deserve close supervision. REFERENCES 1. Weinstein GS, Mavroudis C, Ebert PA. Preliminary experience with aspirin for anticoagulation in children with prosthetic cardiac valves. Ann Thorac Surg 1982;33:549-53. 2. Smith JM, Cooley DA, Ott DA, Ferreira W, Reul GJ. Aortic valve replacement in preteenage children. Ann Thorac Surg 1980;29:512-8.
The Journal of Thoracic and Cardiovascular Surgery
3. Stansel HC, Nudel DB, Berman MA, Talner NS. Prosthetic valve replacement in children. Arch Surg 1975;110:1397-1400. 4. Gardner TJ, Roland J-MA, Neill CA, Donahoo JS. Valve replacement in children. J THoRAc CARDIOVASC SURG 1982;83:178"85. 5. Bradley LM, Midgley FM, Watson DC, Getson PR, Scott LP. Anticoagulation therapy in children with mechanical prosthetic cardiac valves. Am J Cardiol 1985;56:533-5. 6. Pass HI, Sade RM, Crawford FA, Hohn AR. Cardiac valve prostheses in children without anticoagulation. J THORAC CARDIOVASC SURG 1984;87:832-5. 7. Edmunds LH. Thromboembolic complications of current cardiac valvular prostheses. Ann Thorac Surg 1982;34:96105. 8. Klint R, Hernandez A, Weldon C, Hartmann AF, Goldring D. Replacement of cardiac valves in children. J Pediatr 1972;80:980-7. 9. Wada J, Yokoyama M, Hashimoto A, et al. Long-term follow-up of artificial valves in patients under 15 years old. Ann Thorac Surg 1980;29:519-21. 10. Mathews RA, Park SC, Neches WH, et al. Valve replacement in children and adolescents. J THORAC CAR. D10VASC SURG 1977;73:872-6. II. Chen S, Laks H, Fagan L, et al. Valve replacement in children. Circulation 1977;56(Pt 2):11117-21. 12. Schachner A, Salomon J, Levinsky L, Blieden LC, Levy MJ. Prosthetic valve replacement in infants and children. J Cardiovasc Surg 1984;25:537-44.
THORAC CARDIOVASC SURG
1987;93:551-4
The long-term risk of warfarin sodium therapy and the incidence of thromboembolism in children after prosthetic cardiac valve replacement Thirty children less than 18 years of age underwent cardiac valve replacement with a prosthetic valve
between 1967and 1984and havebeenfoUowed up for a meanof 6 years (range 1 to 17 years). Their mean age at the time of operation was 13 years (range 6 to 17 years). AU patients were begun on a regimenof warfarin before bospital discharge. One major and four minor bleeding episodes occurred in 211 patient-years of warfarin therapy, an incidence of 2.3 per 100 patient-years.Three of those five episodes occurred in patients wbo werereceiving excessively anticoagulation or wbo were participating in pbysical activities inappropriate for a patient on warfarin therapy. Thus, the majority of the bleeding episodes were preventable. There were five thromboembolic events in 211 patient-years, an incidence of 2.3 per 100 patient-years. Three of those five patients had intentionally stopped their warfarin therapy. The majority of thromboembolic episodes, like the bleeding episodes, were preventable. Eight teenage patients were noncompliant with the warfarin therapy. More than one third of that group experienced a thromboembolic event, an incidence of 5.5 per 100 patient-years(55 patient-years). Twenty-two patients adhered to the warfarinregimen and only two (9 %) of them had a thromboembolic event, an incidence of 1.3 per 100 patient-years (156 patient-years). Warfarin therapy presented no greater risk of serious bleeding to this pediatric age group than it does to an adult age group. The incidence of thromboembolism among these patients was less than that which is generaUy reported for adult patients. Discontinuation of or noncompliance witb warfarin therapy substantiaUy increased the risk of thromboembolism. Continuous warfarin tberapy is recommended for every child after prosthetic valve replacement.
Scott Stewart, M.D., Diane Cianciotta, R.N., Chloe Alexson, M.D., and James Manning, M.D., Rochester, N. Y.
h e operative risk of cardiac valve replacement in children has declined significantly in the past decade and nowapproachesthat of the adult in several reported series.':' However, the ongoing threat of thromboembolismand the frequentlyexpressed concern regarding the riskof warfarin sodium (Coumadin) therapy continue to stimulate discussion and controversy as to the most efficacious way to provideanticoagulation in children. A perceived increased risk of warfarin therapy in the child and its failure to provide complete protection against From the Division of Cardiothoracic Surgery and the Department of Pediatrics, University of Rochester Medical Center, Rochester, N. Y. Received for publication March 10, 1986. Accepted for publication May 20, 1986. Address for reprints: Scott Stewart, M.D., Department of Surgery, University of Rochester Medical Center, 601 Elmwood Ave., Rochester, N. Y. 14642.
thromboembolism has led some to advocate the alternative use of aspirin and dipyridamole.': 5 Others have used no anticoagulation in smaller series of patients.s 3, 6 The late follow-up experience of every pediatric patient undergoing prosthetic cardiac valve replacement at the University of Rochester Medical Center was reviewed to determine how effective a deterrent warfarin was to thromboembolism and the incidence of bleeding associated with its use. Metbods and patients The hospitaland office chart and the operative record of every patient less than 18 years of age who had undergonea cardiac valvereplacement at the University of Rochester Medical Center were reviewed. The patient's cardiologist and primary physician were contacted by both letter and telephone for up-to-date follow-up information. In addition, direct telephone contact was made with the patient's parents or, when 551
5 5 2 Stewart et al.
appropriate, the patient himself. Information was specifically sought concerning (1) the current clinical status of the patient, (2) history relevant to unusual bleeding episodes or thromboembolism, and (3) compliance with warfarin therapy. Thirty patients were identified who had been discharged from the hospital after prosthetic valve replacement and had at least 1 year of follow-up. This study includes all patients operated on between 1967 and 1984. There were 16 male and 14 female patients. Their mean age was 13 years with a range of 6 to 17 years. Twenty-eight patients had a single valve replaced and two had double valve replacement. Three patients, each having had a mitral valve replacement, required a second mitral valve replacement. One patient who had double valve replacement required a second operation to replace both of those valves and the tricuspid valve as well. Thus, 30 patients ultimately had 38 valve replacements. Twelve patients had aortic valve replacement; 14, mitral valve replacement; one, pulmonary valve replacement; two, aortic and mitral valve replacement; and one, tricuspid valve replacement. Nineteen patients had congenital valve disease, 10 had rheumatic valve disease, and one had a valve injury. Ten of 12 patients receiving aortic valve replacement had congenital valve disease and seven of 14 having mitral valve replacement had congenital disease. Since this study includes all patients undergoing valve replacement over a 17 year period by three different surgeons, a wide variety of prostheses have been used. The Starr-Edwards and, more recently, the Bjork-Shiley prostheses have been the predominant prostheses used for mitral valve replacement. The Starr-Edwards and, again, more recently, the Lillehei-Kaster prostheses have been the predominant prostheses for aortic valve replacement. These 30 patients have been followed up for a mean of 6 years. The range of the follow-up period is 1 to 17 years. Four patients have been lost to current follow-up, but accurate information was available for each of them for a minimum of 1 year after operation. One died of aortic dissection 1 year after operation; one died at the time of operation to replace another valve 3 years after the initial operation, and two did not return for followup after 3 and 11 years of consecutive visits. The total number of follow-up is 211 patient-years.
Results Twenty-three patients were in New York Heart Association Class I at the time of their most recent
The Journal of Thoracic and Cardiovascular Surgery
follow-up contact. Six patients were in Class II and one was in Class III. There were three late deaths. The potential complications of valve replacement and warfarin therapy, thromboembolism, and bleeding were reviewed in terms of their incidence and severity. In addition, bleeding events were examined to determine if they might have been prevented either by a more appropriate prothrombin time or modification of patient activity. The relationship of thromboembolism to warfarin compliance was also reviewed. Bleeding. Bleeding complications were classified as fatal, serious, or minor as suggested by Edmunds.' Twenty-five patients had either no problems or only minimal bleeding problems such as epistaxis or easy bruising. Four had minor bleeding episodes and one had a serious bleeding episode. Three had bleeding associated with trauma but none required hospitalization or transfusion. One was injured while riding a minibike when his prosthrombin time was inordinately high. One had bleeding about a joint after a football injury, and the third had a single episode of hematuria after a prolonged period of jogging. One has had heavy menses in the presence of an appropriate prothrombin time. Finally, one had hematuria with typhoid fever at a time when he was inappropriately receiving both heparin and warfarin anticoagulation. He required transfusion. Two teenage females had intentionally overdosed on warfarin but did not have any deleterious consequences. There have been no instances of intracerebral or gastrointestinal hemorrhage. Thus, five patients have had bleeding episodes that they otherwise might not have had if they had not been on warfarin therapy. No episode per se necessitated hospitalization and only one necessitated transfusion. Three of the episodes occurred in patients who either were participating in activities that they had been advised against or who were receiving excessive anticoagulation. Thus, only two patients have had bleeding episodes that were probably not preventable. There were a total of five bleeding episodes in 211 patient-years of warfarin therapy or an incidence of 2.3 per 100 patientyears. Thromboembolism. Twenty-five patients have not had a thromboembolic event. Three patients have had a single cerebral embolus. One had hemiparesis, one had a temporal lobe embolus with a brief period of amnesia, and the third had an episode of expressiveaphasia. Each patient has had a complete recovery without residual neurologic deficit. Two of those patients were appropriately anticoagulated while the third patient had intentionally stopped taking warfarin. A fourth patient had
Volume 93 Number 4 April 1987
stopped taking warfarin and had a myocardial infarction that was likely the result of an embolic event. A fifth patient also intentionally stopped taking warfarin and, subsequently, the mitral valve prosthesis thrombosed. The valve was replaced at an emergency operation and she survived. Thus, three of the five thromboembolic episodes were preventable. There were five thromboembolic episodes in 211 patient-years or an incidence of 2.3 per 100 patient-years. The incidence was 1.0 per 100 patientyears for patients with a prosthesis in the aortic position and 3.9 for the mitral position. Warfarin compliance. Eight patients were identified who were noncompliant with their warfarin regimen. More than one third (3/8) of them experienced a thromboembolic event whereas only 9% (2/22) of those complying with their medication schedule experienced a similar event. The incidence of thromboembolism in the noncompliant group was 5.5 per 100 patient-years (55 patient-years) whereas it was only 1.3 per 100 patientyears (156 patient-years) among those who took their warfarin on a regular basis.
Discussion The best method to reduce the risk of thromboembolism and yet minimize the hazard of serious bleeding in the child who has received a prosthetic cardiac valve remains controversial. Three medication management programs for these children have been discussed in the literature; (1) warfarin.v->" (2) aspirin with or without dipyridamole.!" and (3) no anticoagulation.s '" There have been three major reports of children who have been managed without anticoagulation after prosthetic valve replacement. 2,3.6 The length of follow-up was 53 and 80 patient-years in two of those reports and could not be ascertained in the third. One series used the St. Jude Medical valve exclusively and reported a thromboembolic incidence of 2.6 per 100 patient-years during a mean follow-up of about 2 years." The second report had an 80 patient-year follow-up and the Starr-Edwards prosthesis was used predominantly.' The thromboembolism incidence was 2.9 per 100 patient-years in the mitral position and 4.9 in the aortic position. The third report was limited to aortic valve replacement and does not provide data that would permit calculating the number of patient-years followUp. 2 There were only 14 patients in this series and none had thromboemboli, but it is not possible to determine how many of them had tissue valves and how many had prosthetic valves. These limited data suggest that children, like adults, who have prosthetic cardiac valves in
Prosthetic valve replacement in children 5 5 3
place are at a greater risk of thromboembolism if they are left unprotected against thrombus formation than if they are receiving warfarin. The combination of aspirin and dipyridamole has been used as a compromise to provide protection against thrombus formation and still avoid the potential bleeding complications of warfarin therapy. Although this combination, or aspirin alone, has been used sporadically for a number of years there are only two recent reports that have specifically evaluated its effectiveness in children. r.s Both reports are preliminary since their patient-year follow-up is limited to only 14 and 39 years, respectively. Weinstein, Mavroudis, and Ebert' reported no incidence of thromboembolism among 12 patients. A late follow-up of this series will be very important. In the second series, 10 patients had been followed up for 16 patient-years and the incidence of thromboembolism was 12 per 100 patient-years. The brief length of follow-up in these two reports precludes any meaningful conclusion from their data. However, even if the data from both reports were combined the incidence of thromboembolism would still be an alarming 6 per 100 patient-years, considerably higher than what is generally reported with warfarin therapy. This suggests that the combination of aspirin and dipyridamole is not an effective method of preventing thromboembolism in the child after prosthetic valve replacement. There are only two other series of children taking warfarin in which the follow-up exceeds 200 patientyears. Gardner and associates' reported a thromboembolism incidence of 0.8 per 100 patient years in 203 patient-years. Schachner and colleagues" described a series of children who had a 224 patient-year follow-up and a thromboembolism incidence of 2 per 100 patientyears. Other series that have considerably less follow-up have reported an incidence of thromboembolism between 0 and 2 per 100 patient-years.v"!' Thus, the data indicate that warfarin therapy in the child is at least as effective in preventing thromboembolism as it is in the adult. 7 Those who have not prescribed warfarin for children have repeatedly alluded to an inherent increased risk of serious bleeding among the active patients in this age group. The data do not support this concern. The incidence of serious bleeding has ranged between 0 and 4 per 100 petient-years except in one series in which the bleeding incidence was an inordinately high 22 per 100 patient-years.v" 8, 10. II This is in the same low range that is reported for adults.' Few reports have sought out the incidence of thromboembolism among those patients who are not just
554
Stewart et al.
thought to be taking warfarin but who are actually compliant with the medication schedule. When this point was examined in this series the incidence of thromboembolism among those patients actually taking warfarin was an extremely low 1.3 per 100 patient years. Thus, warfarin is very effective in preventing thromboembolism in the child. This information has not been taken into considerabion by those who suggest that the protection that warfarin therapy provides does not outweigh the alleged risks of bleeding. We conclude from these data that warfarin therapy provides the best protection against thromboembolism in the child after prosthetic cardiac valve replacement. The risk of serious bleeding from warfarin is extremely small and does not outweight the risk of thromboembolism when warfarin is omitted. Compliance with medication schedules in the teenager is less than that in the younger or older age groups and therefore they deserve close supervision. REFERENCES 1. Weinstein GS, Mavroudis C, Ebert PA. Preliminary experience with aspirin for anticoagulation in children with prosthetic cardiac valves. Ann Thorac Surg 1982;33:549-53. 2. Smith JM, Cooley DA, Ott DA, Ferreira W, Reul GJ. Aortic valve replacement in preteenage children. Ann Thorac Surg 1980;29:512-8.
The Journal of Thoracic and Cardiovascular Surgery
3. Stansel HC, Nudel DB, Berman MA, Talner NS. Prosthetic valve replacement in children. Arch Surg 1975;110:1397-1400. 4. Gardner TJ, Roland J-MA, Neill CA, Donahoo JS. Valve replacement in children. J THoRAc CARDIOVASC SURG 1982;83:178"85. 5. Bradley LM, Midgley FM, Watson DC, Getson PR, Scott LP. Anticoagulation therapy in children with mechanical prosthetic cardiac valves. Am J Cardiol 1985;56:533-5. 6. Pass HI, Sade RM, Crawford FA, Hohn AR. Cardiac valve prostheses in children without anticoagulation. J THORAC CARDIOVASC SURG 1984;87:832-5. 7. Edmunds LH. Thromboembolic complications of current cardiac valvular prostheses. Ann Thorac Surg 1982;34:96105. 8. Klint R, Hernandez A, Weldon C, Hartmann AF, Goldring D. Replacement of cardiac valves in children. J Pediatr 1972;80:980-7. 9. Wada J, Yokoyama M, Hashimoto A, et al. Long-term follow-up of artificial valves in patients under 15 years old. Ann Thorac Surg 1980;29:519-21. 10. Mathews RA, Park SC, Neches WH, et al. Valve replacement in children and adolescents. J THORAC CAR. D10VASC SURG 1977;73:872-6. II. Chen S, Laks H, Fagan L, et al. Valve replacement in children. Circulation 1977;56(Pt 2):11117-21. 12. Schachner A, Salomon J, Levinsky L, Blieden LC, Levy MJ. Prosthetic valve replacement in infants and children. J Cardiovasc Surg 1984;25:537-44.