- Email: [email protected]
Exaggerated initial response to warfarin following heart valve replacement
Exaggerated Initial Response to Warfarin Following Heart Valve Replacement Walter Ageno,
MD,
and Alexander G.G. Turpie,
MD
The response to initiation of oral anticoagulants at a usual dose of 5 mg of warfarin has been retrospectively evaluated in patients following heart valve replacement (HVR). Patients starting oral anticoagulants after HVR have a lower target International Normalized Ratio (INR) (1.5 to 2.6) until the pacing wires are removed after operation. The mean daily doses and INR responses after HVR and nonsurgical patients were retrospectively compared during the first 5 days of warfarin treatment. In a subset from both groups, the mean dose of warfarin was correlated with age, body weight, and albumin levels. Eighty-four HVR and 32 nonsurgical patients were studied. The mean daily warfarin dosage was 3.29 ⴞ 1.29 mg after HVR and 4.96 ⴞ1.76 mg in controls (p <0.001), and the mean INRs 2.08 ⴞ 0.60 and 1.60 ⴞ 0.54, respectively (p <0.001). Of the HVR
patients and controls, 48.8% and 21.8%, respectively, exceeded the upper level of the targeted range (p ⴝ 0.014), 86.9% and 40.6% had the dose reduced after the first 5 mg (p <0.001), and 54.7% and 28.1%, respectively, had warfarin withheld for at least 1 day (p ⴝ 0.015). Thirty-nine patients were included in the subset analysis. Patients with serum albumin levels <35 g/L required significantly less warfarin (3.84 mg/day) than patients with levels >35 g/L (5.37 mg/day; p <0.05). Thus, patients starting oral anticoagulation after HVR are significantly more sensitive to warfarin than nonsurgical patients. Patients with serum albumin levels below the normal values require less warfarin than patients with normal values during the initial phase of treatment. 䊚1999 by Excerpta Medica, Inc. (Am J Cardiol 1999;84:905–908)
he induction phase of oral anticoagulant treatment requires regular monitoring because of the large T intersubject variation in warfarin response. Manage-
METHODS
ment of patients starting oral anticoagulants after heart valve replacement (HVR) presents particular problems. To prevent postoperative morbidity and mortality resulting from cardiac arrhythmias, pacing wires are routinely placed at operation and left for approximately 5 days postoperatively.1 Most patients start oral anticoagulant treatment 24 to 48 hours after surgery, and, until the wires are removed, the levels of anticoagulation need to be maintained at low levels, with International Normalized Ratio (INR) values not exceeding 2.6. Different factors present in patients after cardiac surgery have the potential to interfere with the pharmacokinetics and pharmacodynamics of warfarin. Because of the clinical importance of this problem, we retrospectively examined our routine anticoagulation clinic records relative to the early phase of treatment to detect possible differences in the warfarin dosage and in the subsequent response to the drug between patients after HVR and a control group of nonsurgical patients. We also retrospectively examined in a subset analysis different factors potentially interfering with the induction phase of oral anticoagulation. From the Hamilton Health Sciences Corporation, Hamilton, Ontario, Canada. Manuscript received January 26, 1999; revised manuscript received March 31, 1999, and accepted May 24. Address for reprints: Alexander G.G. Turpie, MD, Hamilton Health Sciences Corporation, General Division, 237 Barton Street East, Hamilton, Ontario, Canada L8L 2X2. E-mail: turpiea@fhs. mcmaster.ca. ©1999 by Excerpta Medica, Inc. All rights reserved. The American Journal of Cardiology Vol. 84 October 15, 1999
The records of consecutive inpatients starting oral anticoagulant treatment with warfarin at the Hamilton General Hospital (Hamilton, Canada) from October 1996 to May 1997 were retrospectively reviewed. All patients considered had at least 3 daily INR values reported after treatment was started, and were treated by the same group of physicians and registered nurses with a uniform approach to oral anticoagulant therapy. The mean daily dose during the first 5 days of treatment and the mean daily prothrombin time expressed as INR were collected for each patient. All patients after HVR also received fixed low-dose subcutaneous heparin for prophylaxis of venous thromboembolism. Among nonsurgical patients, those starting on oral anticoagulants for the treatment of venous thromboembolism also received intravenous heparin adjusted to the therapeutic range of the activated partial thromboplastin time. The following data were then considered: age, average time to achieve the therapeutic range, proportion of dose adjustments, proportion of INR below or above the therapeutic range, proportion of days with warfarin therapy withheld, bleeding, need for vitamin K, and concomitant treatments potentially interfering with warfarin. In a second subset analysis, we included patients from both groups who had serum albumin levels measured on the same day that warfarin therapy was started or on the day before, and the following data were abstracted for each patient as possible factors that might affect the daily dose: age, body weight, and serum albumin levels. The daily dose of warfarin was compared between patients with albumin ⬍35 g/L and patients with albumin ⱖ35 g/L, between patients ⱕ65 and ⬎65 years, and between 0002-9149/99/$–see front matter PII S0002-9149(99)00463-4
905
TABLE I Comparison Between Patients After HVR and Nonsurgical Patients During the First Five Days of Treatment With Warfarin
Age (yrs) Target INR Loading dose (mg) Mean daily dose (mg/d) Mean INR Bleeding Need for vitamin K
HVR (n ⫽ 84)
Control (n ⫽ 32)
65.3 2.0 (1.5–2.6) 5 3.29 ⫾ 1.29 2.08 ⫾ 0.60 3 (3.6%) 7 (8.3%)
63.9 2.5 (2.0–3.0) 5 4.96 ⫾ 1.76* 1.60 ⫾ 0.54* 1 (3.1%) 1 (3.1%)
*p ⬍0.001.
patients with a body weight ⬍75 kg and ⱖ75 kg. Measured quantities were compared using the Student t test, and proportions were compared using the chisquare test. Two-tailed p values ⬍0.05 were considered significant.
RESULTS Eighty-four patients after HVR and 32 nonsurgical patients were included in the primary comparison. These patients represented the actual population during the study period who started on oral anticoagulant therapy and who were hospitalized for at least 4 days after the first dose of warfarin. Among the control group, 14 patients started anticoagulant therapy for the treatment of venous thromboembolism, 10 for prophylaxis of venous thromboembolism, 5 had atrial fibrillation, and 3 had an acute myocardial infarction. Four patients actually underwent surgical intervention, but warfarin was started from 12 to 22 days after surgery. The average age was comparable between HVR patients and control patients (Table I). During the first 5 days of treatment, including the first common dose, the average daily dosage was significantly lower in the HVR group than in the control group (p ⬍0.001) (Table I). The total proportion of dose adjustments was similar between the 2 groups, but there was a significant difference in the proportions of patients whose dose was reduced and in patients whose dose was increased (Table II). The mean INR during the time of observation was 2.08 ⫾ 0.60 in the HVR group and 1.60 ⫾ 0.54 in the control group (p ⬍0.001) (Table I). Daily mean INRs and the daily proportion of patients who underwent inadequate or excessive anticoagulation is shown in Table III. In the HVR group, 88.1% of patients achieved the therapeutic range of 1.5 to 2.6, and the average time to reach it was 36 hours from the first dose. In the control group, the therapeutic range of 2.0 to 3.0 was achieved in 40.6% of the patients (p ⬍0.001), and the average time was 53 hours (p ⬍0.001). The proportion of patients whose INR exceeded the therapeutic range, and the proportion of patients who had the treatment withheld for at least 1 day is reported in Table II. The proportion of bleeding was similar (Table I), bleeding episodes were all minor, and there was no relation between bleeding and the INR value. Finally, vitamin 906 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 84
TABLE II Differences in Dose Adjustments and INR Results Between HVR and Nonsurgical Patients HVR (n ⫽ 84) Patients whose dose was reduced from the loading dose Patients whose dose was reduced on day 1 Patients whose dose was increased from the loading dose Patients who achieved the therapeutic range Average time to achieve the therapeutic range (h) Patients with at least 1 dose adjustment Total proportion of dose adjustments Patients whose INR exceeded the therapeutic range Patients who needed a suspension of treatment
73 (86.9%)
28 (33.3%)
Control (n ⫽ 32)
p Value
13 (40.6%) ⬍0.001 7 (21.8%)
NS
8 (9.5%)
17 (53.2%) ⬍0.001
74 (88.1%)
13 (40.6%) ⬍0.001 ⬍0.001
36
53
82 (97.6%)
30 (93.8%)
NS
50 (59.5%)
17 (53.2%)
NS
41 (48.8%)
7 (21.8%)
0.014
46 (54.7%)
9 (28.1%)
0.015
Target INR for heart valve replacement patients: 1.5 to 2.6. Target INR for control patients: 2.0 to 3.0.
K was administered to a slightly, but nonsignificantly, larger number of patients after HVR (Table I). The INR values obtained after administration of vitamin K were included in the analysis. Thirty-nine patients were included in the subset analysis to determine if there was an association between warfarin dose with age, body weight, or serum albumin concentration. Eleven of these patients were treated after HVR, and had an average warfarin daily dose of 2.67 mg. Ten patients who received warfarin for prophylaxis of venous thromboembolism had an average warfarin daily dose of 5.60 mg, 10 patients for treatment of venous thromboembolism had an average of 5.20 mg, 5 patients anticoagulated for atrial fibrillation had an average 4.93 mg, and 3 starting anticoagulant therapy for acute myocardial infarction had an average dose of 6.15 mg. The trend toward a reduction in the daily dose in older patients was not significant (Table IV), and there was no difference when patients were divided into 2 groups according to body weight (Table IV). When compared according to serum albumin levels, the group of 19 patients with levels below the normal range (average 28.6 g/L) received significantly less warfarin (3.84 mg/day) than the 20 patients with normal levels (average 40.9 g/L, 5.37 mg/day; p ⬍0.05) (Table IV). The frequency of these 3 variables among the diagnostic groups were not statistically different.
DISCUSSION According to the results of this retrospective analysis, patients after HVR are more sensitive to warfarin than nonsurgical patients. After receiving a 5-mg loading dose, the mean INR was significantly higher OCTOBER 15, 1999
The potential of cardiopulmonary bypass to alter liver function and coDay Mean INR % INR ⬍1.5 % INR ⬍2.0 % INR ⬎2.6 % INR ⬎3.0 agulation has been investigated. Rittenhouse et al3 did not find structural Patients After HVR or functional abnormalities in the 1 1.39 67.8 92.8 0 0 liver after profound hypothermia and 2 2.37 22.6 41.6 30.9 23.8 60 minutes of complete circulatory 3 2.36 9.5 39.3 27.4 19.0 arrest, whereas Harker et al4 ob4 2.36 3.9 30.2 22.3 17.1 served that many of the soluble co5 2.47 5.4 30.9 30.9 12.7 agulation factors are mildly reduced Control Patients (nonsurgical) by the end of cardiopulmonary bypass, and Kalter et al5 reported a % INR ⬍2.0 % INR ⬎3.0 46% to 62% reduction in factors II, 1 1.22 96.8 0 VII, IX, and X immediately after car2 1.51 84.3 3.1 diopulmonary bypass. The concen3 1.79 68.9 6.9 tration of albumin is affected by the 4 2.15 51.8 22.2 5 2.46 31.8 18.2 amount of hemodilution during cardiopulmonary bypass,1 and some patients need infusions of albumin after the intervention. Almost 99% of warfarin is bound to plasma proteins,6 and hypoproteineTABLE IV Doses of Warfarin During the First Five Days of mia can affect its volume of distribution and half-life.7 Treatment According to Age, Body Weight, and Serum Vorum et al8 observed a reduction in the binding Albumin Levels affinity of warfarin in serum samples taken during 3 ⬎65 Years ⱕ65 Years days after surgery. Hayes et al9 observed a reduction Age in the binding capacity of warfarin in the elderly, probably due to the decrease in plasma albumin, and No. of patients 21 18 concluded that a reduced binding would allow more Average age (yrs) 72.6 50.4 Weight (kg) 76.5 78.4 free drug to be available, thus enhancing the effect of Albumin (g/L) 34.8 34.9 the drug. It is possible that decreased levels of plasma Warfarin (mg/d) 4.16 5.16 albumin and a reduction in the binding capacity of ⬍75 kg ⱖ75 kg warfarin can really affect the response to warfarin therapy after HVR, although Yacobi et al10 found no Weight correlation between the free drug concentration and No. of patients 20 19 the prothrombin time. In our subset study, albumin Average weight (kg) 64.4 91.1 levels were significantly correlated with the daily dose Age (yrs) 65.9 59.5 Albumin (g/L) 35.1 35.0 of warfarin, suggesting that, whichever is the mechaWarfarin (mg/d) 4.27 4.55 nism, serum albumin levels could be a useful marker ⬍35 g/L ⱖ35 g/L in helping to predicting the individual needs of warfarin when treatment is started. Albumin A possible correlation between dosage and age, No. of patients 19 20 which has been widely suggested during the mainteAverage albumin (g/L) 28.6 40.9 nance phase,11–14 appears controversial during the inAge (yrs) 64.7 60.2 duction phase.14 –17 Our data only show a non–statisWeight (kg) 76.8 77.9 Warfarin (mg/d) 3.84 5.37* tically significant trend toward a reduction in the daily dose in elderly patients. Concomitant anticoagulant *p ⬍0.05. drugs can be a major interacting factor. All patients after HVR received fixed low-dose subcutaneous hepin the HVR group than in the control group after 36 arin for prophylaxis, which can be a cause for the hours of treatment, with almost one third of the pa- reduced need for warfarin. Nevertheless, much higher tients exceeding the upper limit of the range. A very doses of intravenous heparin administered to the 14 high proportion of these patients had the dose reduced, patients treated for venous thromboembolism did not and more than half of them had therapy withheld to produce the same effect, thus suggesting that concomallow removal of the pacing wires. Still, the mean INR itant anticoagulants cannot be considered a sufficient on day 4 was higher than that in the control group, in explanation for such a different response to warfarin. Among other concomitant drugs, interesting obserwhich more than half of the patients had their daily dose increased. This tendency to a lower need for vations can be drawn from the use of antibiotics, many warfarin appears to be independent from the lower of which are known to affect the pharmacokinetics or target INR, and, according to other studies, is not pharmacodynamics of warfarin. Patients undergoing maintained during long-term treatment.2 Thus, it is HVR receive a prophylactic antibiotic treatment. All likely that factors responsible for such interference are our patients received cephalosporin, 82 were treated with cefazolin, and 2 received cefuroxime. Cefazolin related to the surgical procedure. TABLE III Daily Observations (day 0: first day of treatment)
VALVULAR HEART DISEASE/WARFARIN AFTER VALVE REPLACEMENT
907
is a first-generation cephalosporin, and it has been reported to cause in renally impaired patients some coagulation abnormalities and inhibition of platelet function in the absence of anticoagulant therapy.18,19 Angaran et al20 compared the effects on the prothrombin response to warfarin of 3 antibiotics— cefamandole, cefazolin, and vancomycin—in the prophylaxis of patients undergoing HVR, and found an additive effect to warfarin of cefazolin in producing hypoprothrombinemia on the third day of anticoagulant therapy, although lower than that with cefamandole, a second-generation cephalosporin. No reports on cefuroxime are currently available. Eight of 32 patients in the control group also received antibiotic drugs concomitantly with the induction phase of warfarin, and 3 had the antibiotic treatment stopped the day before warfarin was started. Six of these patients received antibiotics known to potentiate the effect of warfarin (erythromycin, ciprofloxacin, cefazolin, and trimethroprim-sulfamethoxazole),21 but their daily dose did not differ from the average; 1 patient received rifampin, which inhibits the effect of warfarin,22 but the average dose of warfarin in this patient was 3.9 mg/day. Patients were not receiving any other agent known to alter warfarin protein binding or clotting factor synthesis or degradation. Acknowledgment: We are indebted to Judy Johnson, Barbara Nowacki, and Margaret Siguenza for their invaluable collaboration in the study.
1. Kirklin JW, Barratt-Boyes BG. Postoperative care. In: Kirklin JW, BarrattBoyes BG, eds. Cardiac surgery. 2nd ed. New York: Churchill-Livingstone, 1993;195–247. 2. James AH, Britt RP, Raskino CL, Thompson SG. Factors affecting the maintenance dose of warfarin. J Clin Pathol 1992;45:704 –706. 3. Rittenhouse EA, Mohri H, Reichenbach DD, Merendino KA. Morphological alterations in vital organs after prolonged cardiac arrest at low body temperature. Ann Thorac Surg 1972;13:564 –574.
908 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 84
4. Harker LA, Malpass TW, Branson HE, Hessel EA, Slichter SJ. Mechanism of abnormal bleeding in patients undergoing cardiopulmonary bypass: acquired transient platelet dysfunction associated with selective alpha-granule release. Blood 1980;56:824 – 834. 5. Kalter RD, Saul CM, Wetstein L, Soriano C, Reiss RF. Cardiopulmonary bypass associated hemostatic abnormalities. J Thorac Cardiovasc Surg 1979;77: 427– 435. 6. Hirsh J, Dalen JE, Deykin D, Poller L, Bussey H. Oral anticoagulants. Mechanism of action, clinical effectiveness, and optimal therapeutic range. Chest 1995;108(suppl 4):231S–246S. 7. Majerus PW, Broze GJ, Miletich JP, Tollefsen DM. Anticoagulant, thrombolytic, and antiplatelet drugs. In: Hardman JG, Limbird LE, Molinoff PB, Ruddon RW, Gilman AG, eds. Goodman and Gilman’s The pharmacological basis of therapeutics. 9th ed. New York: McGraw-Hill, 1996;1341–1359. 8. Vorum H, Jorgensen HRI, Brodersen R. Variation in the binding affinity of warfarin and phenprocoumon to human serum albumin in relation to surgery. Eur J Clin Pharmacol 1993;44:157–162. 9. Hayes MJ, Langman MJS, Short AH. Changes in drug metabolism with increasing age: 1. warfarin binding and plasma proteins. Br J Clin Pharmacol 1975;2:69 –72. 10. Yacobi A, Udall JA, Levy G. Intra-subject variation of warfarin binding to protein in serum of patients with cardiovascular disease. Clin Pharmacol Ther 1976;20:300 –304. 11. O’Malley K, Stevenson CA, Ward CA, Wood AJ, Crooks J. Determinants of anticoagulant control in patients receiving warfarin. Br J Clin Pharmacol 1977; 4:309 –314. 12. Shepherd AMM, Hewick DS, Moreland TA, Stevenson IH. Age as a determinant of sensitivity to warfarin. Br J Clin Pharmacol 1977;4:315–320. 13. Fihn SD, Callahan CM, Martin DC, McDonell MB, Henikoff JG, White RH. The risk for and severity of bleeding complications in elderly patients treated with warfarin. Ann Intern Med 1996;124:970 –979. 14. Adams JN, Gautam PC. Warfarin therapy in the elderly. Br J Hosp Med 1994;51:392–393. 15. Wynne HA, Kamali F, Edwards C, Long A, Kelly P. Effects of ageing upon warfarin dose requirements: a longitudinal study. Age Ageing 1996;25:429 – 431. 16. Redwood M, Taylor C, Bain BJ, Matthews JH. The association of age with dosage requirement for warfarin. Age Ageing 1991;20:217–220. 17. Gladman JR, Dolan G. Effect of age upon the induction and maintenance of anticoagulation with warfarin. Postgraduate Med J 1995;71:153–155. 18. Khaleeli M, Giorgio AJ. Defective platelet function after cephalosporin administration. Blood 1976;48:791–794. 19. Lerner PI, Lubin A. Coagulopathy with cefazolin in uremia. N Engl J Med 1974;290:1324 –1326. 20. Angaran DM, Dias VC, Arom KV, Northrup WF, Kersten TG, Lindsay WG, Nicoloff DM. The comparative influence of prophylactic antibiotics on the prothrombin response to warfarin in the postoperative prosthetic cardiac valve patient. Ann Surg 1987;206:155–161. 21. Freedman MD, Olatidoye AG. Clinically significant drug interactions with the oral anticoagulants. Drug Safety 1994;10:381–394. 22. O’Reilly RA. Interaction of chronic daily warfarin therapy and rifampin. Ann Intern Med 1975;83:506 –508.
OCTOBER 15, 1999
MD,
and Alexander G.G. Turpie,
MD
The response to initiation of oral anticoagulants at a usual dose of 5 mg of warfarin has been retrospectively evaluated in patients following heart valve replacement (HVR). Patients starting oral anticoagulants after HVR have a lower target International Normalized Ratio (INR) (1.5 to 2.6) until the pacing wires are removed after operation. The mean daily doses and INR responses after HVR and nonsurgical patients were retrospectively compared during the first 5 days of warfarin treatment. In a subset from both groups, the mean dose of warfarin was correlated with age, body weight, and albumin levels. Eighty-four HVR and 32 nonsurgical patients were studied. The mean daily warfarin dosage was 3.29 ⴞ 1.29 mg after HVR and 4.96 ⴞ1.76 mg in controls (p <0.001), and the mean INRs 2.08 ⴞ 0.60 and 1.60 ⴞ 0.54, respectively (p <0.001). Of the HVR
patients and controls, 48.8% and 21.8%, respectively, exceeded the upper level of the targeted range (p ⴝ 0.014), 86.9% and 40.6% had the dose reduced after the first 5 mg (p <0.001), and 54.7% and 28.1%, respectively, had warfarin withheld for at least 1 day (p ⴝ 0.015). Thirty-nine patients were included in the subset analysis. Patients with serum albumin levels <35 g/L required significantly less warfarin (3.84 mg/day) than patients with levels >35 g/L (5.37 mg/day; p <0.05). Thus, patients starting oral anticoagulation after HVR are significantly more sensitive to warfarin than nonsurgical patients. Patients with serum albumin levels below the normal values require less warfarin than patients with normal values during the initial phase of treatment. 䊚1999 by Excerpta Medica, Inc. (Am J Cardiol 1999;84:905–908)
he induction phase of oral anticoagulant treatment requires regular monitoring because of the large T intersubject variation in warfarin response. Manage-
METHODS
ment of patients starting oral anticoagulants after heart valve replacement (HVR) presents particular problems. To prevent postoperative morbidity and mortality resulting from cardiac arrhythmias, pacing wires are routinely placed at operation and left for approximately 5 days postoperatively.1 Most patients start oral anticoagulant treatment 24 to 48 hours after surgery, and, until the wires are removed, the levels of anticoagulation need to be maintained at low levels, with International Normalized Ratio (INR) values not exceeding 2.6. Different factors present in patients after cardiac surgery have the potential to interfere with the pharmacokinetics and pharmacodynamics of warfarin. Because of the clinical importance of this problem, we retrospectively examined our routine anticoagulation clinic records relative to the early phase of treatment to detect possible differences in the warfarin dosage and in the subsequent response to the drug between patients after HVR and a control group of nonsurgical patients. We also retrospectively examined in a subset analysis different factors potentially interfering with the induction phase of oral anticoagulation. From the Hamilton Health Sciences Corporation, Hamilton, Ontario, Canada. Manuscript received January 26, 1999; revised manuscript received March 31, 1999, and accepted May 24. Address for reprints: Alexander G.G. Turpie, MD, Hamilton Health Sciences Corporation, General Division, 237 Barton Street East, Hamilton, Ontario, Canada L8L 2X2. E-mail: turpiea@fhs. mcmaster.ca. ©1999 by Excerpta Medica, Inc. All rights reserved. The American Journal of Cardiology Vol. 84 October 15, 1999
The records of consecutive inpatients starting oral anticoagulant treatment with warfarin at the Hamilton General Hospital (Hamilton, Canada) from October 1996 to May 1997 were retrospectively reviewed. All patients considered had at least 3 daily INR values reported after treatment was started, and were treated by the same group of physicians and registered nurses with a uniform approach to oral anticoagulant therapy. The mean daily dose during the first 5 days of treatment and the mean daily prothrombin time expressed as INR were collected for each patient. All patients after HVR also received fixed low-dose subcutaneous heparin for prophylaxis of venous thromboembolism. Among nonsurgical patients, those starting on oral anticoagulants for the treatment of venous thromboembolism also received intravenous heparin adjusted to the therapeutic range of the activated partial thromboplastin time. The following data were then considered: age, average time to achieve the therapeutic range, proportion of dose adjustments, proportion of INR below or above the therapeutic range, proportion of days with warfarin therapy withheld, bleeding, need for vitamin K, and concomitant treatments potentially interfering with warfarin. In a second subset analysis, we included patients from both groups who had serum albumin levels measured on the same day that warfarin therapy was started or on the day before, and the following data were abstracted for each patient as possible factors that might affect the daily dose: age, body weight, and serum albumin levels. The daily dose of warfarin was compared between patients with albumin ⬍35 g/L and patients with albumin ⱖ35 g/L, between patients ⱕ65 and ⬎65 years, and between 0002-9149/99/$–see front matter PII S0002-9149(99)00463-4
905
TABLE I Comparison Between Patients After HVR and Nonsurgical Patients During the First Five Days of Treatment With Warfarin
Age (yrs) Target INR Loading dose (mg) Mean daily dose (mg/d) Mean INR Bleeding Need for vitamin K
HVR (n ⫽ 84)
Control (n ⫽ 32)
65.3 2.0 (1.5–2.6) 5 3.29 ⫾ 1.29 2.08 ⫾ 0.60 3 (3.6%) 7 (8.3%)
63.9 2.5 (2.0–3.0) 5 4.96 ⫾ 1.76* 1.60 ⫾ 0.54* 1 (3.1%) 1 (3.1%)
*p ⬍0.001.
patients with a body weight ⬍75 kg and ⱖ75 kg. Measured quantities were compared using the Student t test, and proportions were compared using the chisquare test. Two-tailed p values ⬍0.05 were considered significant.
RESULTS Eighty-four patients after HVR and 32 nonsurgical patients were included in the primary comparison. These patients represented the actual population during the study period who started on oral anticoagulant therapy and who were hospitalized for at least 4 days after the first dose of warfarin. Among the control group, 14 patients started anticoagulant therapy for the treatment of venous thromboembolism, 10 for prophylaxis of venous thromboembolism, 5 had atrial fibrillation, and 3 had an acute myocardial infarction. Four patients actually underwent surgical intervention, but warfarin was started from 12 to 22 days after surgery. The average age was comparable between HVR patients and control patients (Table I). During the first 5 days of treatment, including the first common dose, the average daily dosage was significantly lower in the HVR group than in the control group (p ⬍0.001) (Table I). The total proportion of dose adjustments was similar between the 2 groups, but there was a significant difference in the proportions of patients whose dose was reduced and in patients whose dose was increased (Table II). The mean INR during the time of observation was 2.08 ⫾ 0.60 in the HVR group and 1.60 ⫾ 0.54 in the control group (p ⬍0.001) (Table I). Daily mean INRs and the daily proportion of patients who underwent inadequate or excessive anticoagulation is shown in Table III. In the HVR group, 88.1% of patients achieved the therapeutic range of 1.5 to 2.6, and the average time to reach it was 36 hours from the first dose. In the control group, the therapeutic range of 2.0 to 3.0 was achieved in 40.6% of the patients (p ⬍0.001), and the average time was 53 hours (p ⬍0.001). The proportion of patients whose INR exceeded the therapeutic range, and the proportion of patients who had the treatment withheld for at least 1 day is reported in Table II. The proportion of bleeding was similar (Table I), bleeding episodes were all minor, and there was no relation between bleeding and the INR value. Finally, vitamin 906 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 84
TABLE II Differences in Dose Adjustments and INR Results Between HVR and Nonsurgical Patients HVR (n ⫽ 84) Patients whose dose was reduced from the loading dose Patients whose dose was reduced on day 1 Patients whose dose was increased from the loading dose Patients who achieved the therapeutic range Average time to achieve the therapeutic range (h) Patients with at least 1 dose adjustment Total proportion of dose adjustments Patients whose INR exceeded the therapeutic range Patients who needed a suspension of treatment
73 (86.9%)
28 (33.3%)
Control (n ⫽ 32)
p Value
13 (40.6%) ⬍0.001 7 (21.8%)
NS
8 (9.5%)
17 (53.2%) ⬍0.001
74 (88.1%)
13 (40.6%) ⬍0.001 ⬍0.001
36
53
82 (97.6%)
30 (93.8%)
NS
50 (59.5%)
17 (53.2%)
NS
41 (48.8%)
7 (21.8%)
0.014
46 (54.7%)
9 (28.1%)
0.015
Target INR for heart valve replacement patients: 1.5 to 2.6. Target INR for control patients: 2.0 to 3.0.
K was administered to a slightly, but nonsignificantly, larger number of patients after HVR (Table I). The INR values obtained after administration of vitamin K were included in the analysis. Thirty-nine patients were included in the subset analysis to determine if there was an association between warfarin dose with age, body weight, or serum albumin concentration. Eleven of these patients were treated after HVR, and had an average warfarin daily dose of 2.67 mg. Ten patients who received warfarin for prophylaxis of venous thromboembolism had an average warfarin daily dose of 5.60 mg, 10 patients for treatment of venous thromboembolism had an average of 5.20 mg, 5 patients anticoagulated for atrial fibrillation had an average 4.93 mg, and 3 starting anticoagulant therapy for acute myocardial infarction had an average dose of 6.15 mg. The trend toward a reduction in the daily dose in older patients was not significant (Table IV), and there was no difference when patients were divided into 2 groups according to body weight (Table IV). When compared according to serum albumin levels, the group of 19 patients with levels below the normal range (average 28.6 g/L) received significantly less warfarin (3.84 mg/day) than the 20 patients with normal levels (average 40.9 g/L, 5.37 mg/day; p ⬍0.05) (Table IV). The frequency of these 3 variables among the diagnostic groups were not statistically different.
DISCUSSION According to the results of this retrospective analysis, patients after HVR are more sensitive to warfarin than nonsurgical patients. After receiving a 5-mg loading dose, the mean INR was significantly higher OCTOBER 15, 1999
The potential of cardiopulmonary bypass to alter liver function and coDay Mean INR % INR ⬍1.5 % INR ⬍2.0 % INR ⬎2.6 % INR ⬎3.0 agulation has been investigated. Rittenhouse et al3 did not find structural Patients After HVR or functional abnormalities in the 1 1.39 67.8 92.8 0 0 liver after profound hypothermia and 2 2.37 22.6 41.6 30.9 23.8 60 minutes of complete circulatory 3 2.36 9.5 39.3 27.4 19.0 arrest, whereas Harker et al4 ob4 2.36 3.9 30.2 22.3 17.1 served that many of the soluble co5 2.47 5.4 30.9 30.9 12.7 agulation factors are mildly reduced Control Patients (nonsurgical) by the end of cardiopulmonary bypass, and Kalter et al5 reported a % INR ⬍2.0 % INR ⬎3.0 46% to 62% reduction in factors II, 1 1.22 96.8 0 VII, IX, and X immediately after car2 1.51 84.3 3.1 diopulmonary bypass. The concen3 1.79 68.9 6.9 tration of albumin is affected by the 4 2.15 51.8 22.2 5 2.46 31.8 18.2 amount of hemodilution during cardiopulmonary bypass,1 and some patients need infusions of albumin after the intervention. Almost 99% of warfarin is bound to plasma proteins,6 and hypoproteineTABLE IV Doses of Warfarin During the First Five Days of mia can affect its volume of distribution and half-life.7 Treatment According to Age, Body Weight, and Serum Vorum et al8 observed a reduction in the binding Albumin Levels affinity of warfarin in serum samples taken during 3 ⬎65 Years ⱕ65 Years days after surgery. Hayes et al9 observed a reduction Age in the binding capacity of warfarin in the elderly, probably due to the decrease in plasma albumin, and No. of patients 21 18 concluded that a reduced binding would allow more Average age (yrs) 72.6 50.4 Weight (kg) 76.5 78.4 free drug to be available, thus enhancing the effect of Albumin (g/L) 34.8 34.9 the drug. It is possible that decreased levels of plasma Warfarin (mg/d) 4.16 5.16 albumin and a reduction in the binding capacity of ⬍75 kg ⱖ75 kg warfarin can really affect the response to warfarin therapy after HVR, although Yacobi et al10 found no Weight correlation between the free drug concentration and No. of patients 20 19 the prothrombin time. In our subset study, albumin Average weight (kg) 64.4 91.1 levels were significantly correlated with the daily dose Age (yrs) 65.9 59.5 Albumin (g/L) 35.1 35.0 of warfarin, suggesting that, whichever is the mechaWarfarin (mg/d) 4.27 4.55 nism, serum albumin levels could be a useful marker ⬍35 g/L ⱖ35 g/L in helping to predicting the individual needs of warfarin when treatment is started. Albumin A possible correlation between dosage and age, No. of patients 19 20 which has been widely suggested during the mainteAverage albumin (g/L) 28.6 40.9 nance phase,11–14 appears controversial during the inAge (yrs) 64.7 60.2 duction phase.14 –17 Our data only show a non–statisWeight (kg) 76.8 77.9 Warfarin (mg/d) 3.84 5.37* tically significant trend toward a reduction in the daily dose in elderly patients. Concomitant anticoagulant *p ⬍0.05. drugs can be a major interacting factor. All patients after HVR received fixed low-dose subcutaneous hepin the HVR group than in the control group after 36 arin for prophylaxis, which can be a cause for the hours of treatment, with almost one third of the pa- reduced need for warfarin. Nevertheless, much higher tients exceeding the upper limit of the range. A very doses of intravenous heparin administered to the 14 high proportion of these patients had the dose reduced, patients treated for venous thromboembolism did not and more than half of them had therapy withheld to produce the same effect, thus suggesting that concomallow removal of the pacing wires. Still, the mean INR itant anticoagulants cannot be considered a sufficient on day 4 was higher than that in the control group, in explanation for such a different response to warfarin. Among other concomitant drugs, interesting obserwhich more than half of the patients had their daily dose increased. This tendency to a lower need for vations can be drawn from the use of antibiotics, many warfarin appears to be independent from the lower of which are known to affect the pharmacokinetics or target INR, and, according to other studies, is not pharmacodynamics of warfarin. Patients undergoing maintained during long-term treatment.2 Thus, it is HVR receive a prophylactic antibiotic treatment. All likely that factors responsible for such interference are our patients received cephalosporin, 82 were treated with cefazolin, and 2 received cefuroxime. Cefazolin related to the surgical procedure. TABLE III Daily Observations (day 0: first day of treatment)
VALVULAR HEART DISEASE/WARFARIN AFTER VALVE REPLACEMENT
907
is a first-generation cephalosporin, and it has been reported to cause in renally impaired patients some coagulation abnormalities and inhibition of platelet function in the absence of anticoagulant therapy.18,19 Angaran et al20 compared the effects on the prothrombin response to warfarin of 3 antibiotics— cefamandole, cefazolin, and vancomycin—in the prophylaxis of patients undergoing HVR, and found an additive effect to warfarin of cefazolin in producing hypoprothrombinemia on the third day of anticoagulant therapy, although lower than that with cefamandole, a second-generation cephalosporin. No reports on cefuroxime are currently available. Eight of 32 patients in the control group also received antibiotic drugs concomitantly with the induction phase of warfarin, and 3 had the antibiotic treatment stopped the day before warfarin was started. Six of these patients received antibiotics known to potentiate the effect of warfarin (erythromycin, ciprofloxacin, cefazolin, and trimethroprim-sulfamethoxazole),21 but their daily dose did not differ from the average; 1 patient received rifampin, which inhibits the effect of warfarin,22 but the average dose of warfarin in this patient was 3.9 mg/day. Patients were not receiving any other agent known to alter warfarin protein binding or clotting factor synthesis or degradation. Acknowledgment: We are indebted to Judy Johnson, Barbara Nowacki, and Margaret Siguenza for their invaluable collaboration in the study.
1. Kirklin JW, Barratt-Boyes BG. Postoperative care. In: Kirklin JW, BarrattBoyes BG, eds. Cardiac surgery. 2nd ed. New York: Churchill-Livingstone, 1993;195–247. 2. James AH, Britt RP, Raskino CL, Thompson SG. Factors affecting the maintenance dose of warfarin. J Clin Pathol 1992;45:704 –706. 3. Rittenhouse EA, Mohri H, Reichenbach DD, Merendino KA. Morphological alterations in vital organs after prolonged cardiac arrest at low body temperature. Ann Thorac Surg 1972;13:564 –574.
908 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 84
4. Harker LA, Malpass TW, Branson HE, Hessel EA, Slichter SJ. Mechanism of abnormal bleeding in patients undergoing cardiopulmonary bypass: acquired transient platelet dysfunction associated with selective alpha-granule release. Blood 1980;56:824 – 834. 5. Kalter RD, Saul CM, Wetstein L, Soriano C, Reiss RF. Cardiopulmonary bypass associated hemostatic abnormalities. J Thorac Cardiovasc Surg 1979;77: 427– 435. 6. Hirsh J, Dalen JE, Deykin D, Poller L, Bussey H. Oral anticoagulants. Mechanism of action, clinical effectiveness, and optimal therapeutic range. Chest 1995;108(suppl 4):231S–246S. 7. Majerus PW, Broze GJ, Miletich JP, Tollefsen DM. Anticoagulant, thrombolytic, and antiplatelet drugs. In: Hardman JG, Limbird LE, Molinoff PB, Ruddon RW, Gilman AG, eds. Goodman and Gilman’s The pharmacological basis of therapeutics. 9th ed. New York: McGraw-Hill, 1996;1341–1359. 8. Vorum H, Jorgensen HRI, Brodersen R. Variation in the binding affinity of warfarin and phenprocoumon to human serum albumin in relation to surgery. Eur J Clin Pharmacol 1993;44:157–162. 9. Hayes MJ, Langman MJS, Short AH. Changes in drug metabolism with increasing age: 1. warfarin binding and plasma proteins. Br J Clin Pharmacol 1975;2:69 –72. 10. Yacobi A, Udall JA, Levy G. Intra-subject variation of warfarin binding to protein in serum of patients with cardiovascular disease. Clin Pharmacol Ther 1976;20:300 –304. 11. O’Malley K, Stevenson CA, Ward CA, Wood AJ, Crooks J. Determinants of anticoagulant control in patients receiving warfarin. Br J Clin Pharmacol 1977; 4:309 –314. 12. Shepherd AMM, Hewick DS, Moreland TA, Stevenson IH. Age as a determinant of sensitivity to warfarin. Br J Clin Pharmacol 1977;4:315–320. 13. Fihn SD, Callahan CM, Martin DC, McDonell MB, Henikoff JG, White RH. The risk for and severity of bleeding complications in elderly patients treated with warfarin. Ann Intern Med 1996;124:970 –979. 14. Adams JN, Gautam PC. Warfarin therapy in the elderly. Br J Hosp Med 1994;51:392–393. 15. Wynne HA, Kamali F, Edwards C, Long A, Kelly P. Effects of ageing upon warfarin dose requirements: a longitudinal study. Age Ageing 1996;25:429 – 431. 16. Redwood M, Taylor C, Bain BJ, Matthews JH. The association of age with dosage requirement for warfarin. Age Ageing 1991;20:217–220. 17. Gladman JR, Dolan G. Effect of age upon the induction and maintenance of anticoagulation with warfarin. Postgraduate Med J 1995;71:153–155. 18. Khaleeli M, Giorgio AJ. Defective platelet function after cephalosporin administration. Blood 1976;48:791–794. 19. Lerner PI, Lubin A. Coagulopathy with cefazolin in uremia. N Engl J Med 1974;290:1324 –1326. 20. Angaran DM, Dias VC, Arom KV, Northrup WF, Kersten TG, Lindsay WG, Nicoloff DM. The comparative influence of prophylactic antibiotics on the prothrombin response to warfarin in the postoperative prosthetic cardiac valve patient. Ann Surg 1987;206:155–161. 21. Freedman MD, Olatidoye AG. Clinically significant drug interactions with the oral anticoagulants. Drug Safety 1994;10:381–394. 22. O’Reilly RA. Interaction of chronic daily warfarin therapy and rifampin. Ann Intern Med 1975;83:506 –508.
OCTOBER 15, 1999