- Email: [email protected]
Duration of increased bleeding tendency after cessation of aspirin therapy
Duration of Increased Bleeding Tendency after Cessation of Aspirin Therapy Ronan A Cahill, AFRCSI, Gerard T McGreal, FRCSI, Basil H Crowe, FIMLS, Damien A Ryan, AFRCSI, Brian J Manning, MB, AFRCSI, Mary R Cahill, MD, H Paul Redmond, MCh, FRCSI Aspirin has a significant effect on hemostasis, so it is often recommended that patients taking aspirin discontinue treatment before elective surgery. While off aspirin, these patients may be at risk of thrombosis. The optimum period of time that aspirin should be withheld is controversial. The aim of this study was to establish the duration of the antihemostatic effect of prolonged aspirin therapy. STUDY DESIGN: In a prospective study, 51 healthy volunteers were randomly assigned into 3 groups, each receiving an identical tablet for 14 days. One group received a placebo tablet; individuals in the other two groups received either 75 mg or 300 mg of aspirin once a day. Template bleeding times and specific platelet function testing (using the PFA-100; Dade Behring) were carried out on subjects before therapy and again after its completion until they returned to baseline. RESULTS: Thirty-eight volunteers complied sufficiently with the protocol to provide useful results. All bleeding times normalized within 96 hours and all platelet function tests within 144 hours after stopping aspirin. There was no demonstrable hemostatic defect in any volunteer persisting by or beyond the sixth day after treatment cessation. There was no apparent difference in duration of effect between those taking either 75 mg or 300 mg of aspirin. CONCLUSIONS: This study uses sensitive measures of platelet function to demonstrate the duration of increased bleeding tendency after withdrawal of aspirin therapy. It supports discontinuation of aspirin therapy 5 days before elective surgery (with the operation being performed on the sixth day). (J Am Coll Surg 2005;200:564–573. © 2005 by the American College of Surgeons) BACKGROUND:
dency after stopping longterm aspirin therapy has not previously been determined. Although some studies have attempted to do so, their clinical application is limited because of study design18,19 (observational studies), techniques20 (measurements of platelet function in isolation may not reflect overall haemostatic capability), or lack of a sensitive end point21 (an individual’s bleeding time may be significantly prolonged while at the same time be within the normal range for a population). Our aim was to determine the duration of increased bleeding tendency after cessation of aspirin therapy using methodology sensitive to and reflective of in vivo changes of both platelet function and global hemostatic capacity.
Aspirin has a significant effect on hemostasis;1 it prolongs bleeding time2,3 and is associated with increased intraoperative and postoperative bleeding.4-9 It is often recommended that patients on longterm aspirin therapy discontinue treatment for a period of time before elective surgical procedures,10-14 allowing their primary hemostatic mechanisms to return to normal. But while off aspirin, some patients may be at risk of thrombosis.15-17 With increasing numbers of people on aspirin, the balance between the risk of increased bleeding and that of thrombosis is of considerable importance in planning elective surgery. The actual duration of the increased bleeding tenCompeting interests declared: None.
METHODS A prospective, double blind, randomized study was performed after obtaining approval of the protocol by the local ethics committees. Fifty-one healthy volunteers were recruited, with fully informed consent, from the medical and paramedical staff of a regional hospital. All
Received October 16, 2004; Revised November 2, 2004; Accepted November 2, 2004. From the Department of Surgery, NUI (Cork), Cork University Hospital, Cork, Ireland (RA Cahill, McGreal, Ryan, Manning, Redmond) and the Department of Haematology, Mid-Western Regional Hospital, Limerick, Ireland (Crowe, MR Cahill). Correspondence address: Professor HP Redmond, Professional Unit, Department of Surgery, Cork University Hospital, Cork, Ireland.
© 2005 by the American College of Surgeons Published by Elsevier Inc.
564
ISSN 1072-7515/05/$30.00 doi:10.1016/j.jamcollsurg.2004.11.002
Vol. 200, No. 4, April 2005
Cahill et al
were asked about contraindications to commencing aspirin or history of any bleeding or wound healing problems. No volunteer was taking any regular medication and all denied recent ingestion of any drug known to have an antiplatelet effect within the previous 4 weeks. Volunteers were randomly assigned in a doubleblinded fashion into three separate groups. Members of the first group (group A) received a placebo; those in the other two groups (groups B and C) were given either of the two extremes of aspirin dosages in common clinical use, ie, 75 or 300 mg once a day, respectively. All tablets were identical in appearance. A baseline platelet count, mean platelet volume, standardized template bleeding time, and platelet function test (Platelet Function Analyser 100- PFA-100; Dade Behring) were measured. The groups were then started on a 14-day course of treatment, at the end of which the tests were performed on a daily basis until a return to baseline was demonstrated. This indicated the recovery of the individual’s hemostatic mechanisms. The value taken as the baseline score was either the pretreatment score or, in a small number of volunteers (n ⫽ 3), the repeatedly normal posttreatment scores. Volunteers were asked to abstain from alcohol and nonsteroidal antiinflammatory agents along with any other plateletmodifying medications for the 24 hours before and on each day during testing. The template bleeding time was used to test for a primary hemostatic defect (from a vascular or platelet effect) induced by aspirin. The same experienced investigator used a standardized method (the modified Ivy technique22) with a commercially available device (Simplate 1; Organon Teknika Corp) to perform the test on each volunteer. The time taken for the bleeding to stop was measured to the nearest 20 seconds. The manufacturer’s range of normal bleeding times with this method is between 2 and 10 minutes. The PFA-100 was used as a specific laboratory measure of platelet function and global hemostasis. It is reproducible and minimally manipulative.23 This device aspirates anticoagulated whole blood through a capillary toward a microscopic aperture (150 m in diameter) cut into a biologically active membrane (coated with collagen and either adenosine-5-diphosphate [ADP] or epinephrine, the latter being particularly sensitive to aspirin-induced defects). The high shear rates, along with these biochemical stimuli, mimic conditions in an actual wound, thereby activating platelets and causing a
Increased Bleeding after Ceasing Aspirin Therapy
565
Table 1. Final Composition of the Groups Studied Group
n
Mean age (ⴞ SD), y
Male:female ratio
Group A Group B Group C
9 12 17
23.7 (⫾ 0.2) 24.1 (⫾ 0.6) 24.2 (⫾ 0.3)
6:3 7:5 12:5
Group A, placebo; group B, 75 mg aspirin; group C, 300 mg aspirin.
stable plug to be slowly built at the aperture, stopping the flow of blood through it. The time required to fully occlude the aperture is recorded by a microprocessor as the closure time. A prolonged closure time using the collagen/epinephrine membrane with a normal collagen/ ADP time is associated with a specific aspirin-induced defect. The usual range of normal closure times for the collagen/epinephrine membrane is between 85 and 176 seconds, with a maximum measurement of 300 seconds and a coefficient of variability of 10%. Two separate specimens were tested for each volunteer on each day, with the mean recorded as the representative score. RESULTS Although 51 volunteers were asked to participate in the study and gave fully informed consent, 9 individuals subsequently withdrew before completing the study protocol (2 developed upper gastrointestinal symptoms while taking aspirin, and 7 were unavailable to attend at the required times). This resulted in 42 participants who took either aspirin or placebo tablets for 14 days and then underwent daily analysis of their hemostatic function after treatment cessation. But the PFA demonstrated that four of these individuals had ingested platelet-modifying drugs during the testing phase (when ingestion of these agents was precluded), so data from these individuals have also been excluded. Once the bleeding time had normalized, an additional five volunteers dropped out of the study even though their PFA tests were still prolonged. Their data on the days they attended for testing are included in the analysis. So data from 38 volunteers (13 of whom were women) are presented and analyzed. The numbers, ages, and gender of the volunteers in each group for whom data were studied are shown inTable 1; the mean of the prolongation of each individual’s tests (both bleeding times and closure times) is shown in Figure 1. Platelet number and volume measurements remained constant for each individual throughout the study period. The mean ⫾ standard deviation for the individuals’ platelet counts during the study was 231 ⫾ 14 ⫻ 103/L, while that for the platelet volumes was 9.4 ⫾ 0.2 ⫻ 103/fL.
566
Cahill et al
Increased Bleeding after Ceasing Aspirin Therapy
J Am Coll Surg
Figure 1. Mean of the prolongation (daily value– baseline value) of each individual’s tests (both bleeding and closure times) at each measured time point. A closure time of 300 was ascribed to those with closure times ⬎300 seconds (the maximum score measurable by the PFA-100), so the values shown at the first three time points are underestimates of the true value). Solid line, trend of closure times; dotted line, trends of bleeding times; Œ, individuals taking 300 mg aspirin; , those taking 75 mg aspirin.
The mean prolongation of bleeding times above the baseline value of each group on each day of testing is shown in Table 2; the daily ranges in values are shown in Figures 2, 3, and 4.Three volunteers (one in group B, two in group C) showed a hyper-response18,24 to aspirin treatment. That is, their bleeding times while on aspirin were prolonged greater than two standard deviations from the mean. Although the majority of bleeding times of those who took aspirin (n ⫽ 23) had returned to pretreatment levels by 72 hours, 6 remained prolonged until 96 hours (of these, five were in group C and two were hyper-responders). The mean time (⫾ standard deviation) of return to normal for group B was 60 (⫾ 19) hours and for group C was 70 (⫾ 22) hours. Bleeding times measured on the first posttreatment day of two volunteers in the placebo group were prolonged; that of one person taking aspirin was not significantly prolonged above the baseline value. The median and interquartile ranges for closure times using the collagen/epinephrine membrane for each group on each day are shown in Table 3 and Figures 5, 6, and 7. In
each case, as expected, the corresponding collagen/ADP measurements remained at their baseline level (data not shown). The latest significant prolongation of closure times (collagen/epinenephrine membrane) was detected at 120 hours (Fig. 8). The mean (⫾ standard deviation) time of return to baseline for these closure times was 84 (⫾ 31) hours for group B (75 mg aspirin) and 94 (⫾ 37) hours for group C (300 mg aspirin). Although there was considerable variability among individuals on aspirin as to when their hemostatic function returned back to their baseline levels (17 tests had normalized 48 hours after completion of the treatment course, and 5 tests remained prolonged after 96 hours), all tests had returned to normal by the sixth day after treatment cessation (Fig. 8). In general, the PFA-100 tests remained prolonged for 24 to 74 hours longer than their corresponding bleeding times (although the 2 tests normalized on the same day in 3 people) and the PFA100 test normalization preceded that of the bleeding time tests by 24 hours in 4 people. Although there was a
Table 2. Mean Value and Mean Prolongation Above Baseline Levels of Bleeding Times for Each Group Before and on Each Day after Completion of Tablets Group
A (n ⫽ 9) B (n ⫽ 12) C (n ⫽ 17)
Pretreatment
24 h
362 440 450
382 (22) 570 (105) 667 (211)
Data shown are mean BT (mean prolongation of BT), in seconds. BT, bleeding times; group A, placebo; group B, 75 mg aspirin; group C, 300 mg aspirin.
Posttreatment 48 h
360 446 (60) 554 (121)
72 h
96 h
380 443 460 (53)
— 320 440
Vol. 200, No. 4, April 2005
Cahill et al
Increased Bleeding after Ceasing Aspirin Therapy
567
Figure 2. Bleeding times at each measured time point for individuals taking 300 mg aspirin. ⽧, mean value; ⫺, 25th and 75th percentile; □, 5th and 95th percentiles; n, number of individuals tested at each time point. Value at 0 hours is the baseline level while testing was discontinued when each individual’s values had returned to that person’s baseline level.
trend toward a longer duration of both bleeding and closure time prolongation in the group taking the higher dose (300 mg) of aspirin, in neither case was this statistically significant. Last, there was no significant difference noted between men and women or between older and younger patients in this study (data not shown). DISCUSSION Aspirin remains in common use for the primary and secondary prevention of myocardial infarct and stroke.25 Its hemorrhagic effects are known to be mainly mediated by its effects on platelet prostaglandin synthesis because of cyclooxygenase inhibition. In addition, it is also a potent inhibitor of acetylcholinesterase (thereby reduc-
ing vascular constriction in response to trauma-induced cathecholamines).26 Aspirin ingestion may theoretically be considered to cause a bleeding risk similar to that found in patients with congenital cyclooxygenase deficiency.26,27 In addition, it may also exacerbate bleeding associated with otherwise subclinical hemostatic abnormalities (eg, von Willebrand’s disease). Increased perioperative blood loss has been observed in patients on aspirin undergoing many different procedures,4,28 including cardiothoracic,5,6 neurosurgical,7 orthopaedic,9 gynecologic, and general surgical10 operations. That the increased blood loss in some studies does not reach statistical significance does not imply that it is not clinically relevant. Overall,
Figure 3. Bleeding times at each measured time point for individuals taking 75 mg aspirin, ⽧, mean value; ⫺, 25th and 75th percentile; □, 5th and 95th percentiles; n, number of individuals tested at each time point. Value at 0 hours is the baseline level while testing was discontinued when each individual’s values had returned to that person’s baseline level.
568
Cahill et al
Increased Bleeding after Ceasing Aspirin Therapy
J Am Coll Surg
Figure 4. Bleeding times at each measured time point for individuals taking placebo tablets. ⽧, mean value; ⫺, 25th and 75th percentile; □, 5th and 95th percentiles; n, number of individuals tested at each time point. Value at 0 hours is the baseline level while testing was discontinued when each individual’s values had returned to that person’s baseline level.
it has been concluded that aspirin’s effect on hemostasis is “mild, variable but significant.”11 So, many surgeons advise that aspirin therapy be withdrawn before elective surgical intervention. This study provides evidence to support the withdrawal of aspirin before elective surgery and aims to guide clinicians who do not operate electively on patients taking this drug. Although no definite data exist on the exact risk of thrombotic events in individuals who have discontinued their treatment for a short period of time, it may increase the thrombotic risk to withhold therapy for longer than is necessary.29 Indeed, such are the beneficial effects of aspirin that increasingly, aspirin therapy is continued throughout the perioperative period of cardiovascular procedures despite its associated hemorrhagic effects.30,31 The optimal period of time for which aspirin should be withheld before elective surgical intervention is uncertain,32 but is often suggested to be several days.11,33,34 In accordance with published recommendations,35 it had been our practice to advise discontinuation of therapy 10 days before planned operations. Because the antiplatelet effects of aspirin are mediated
through the irreversible inhibition of cyclooxygenase36 and the lifespan of the platelet is approximately 10 days,37 this time period should allow an entirely normal platelet population to be reestablished in the circulation. Our data demonstrate that sensitive measures of platelet function and primary hemostasis can detect a measurable defect for up to 5 days after longterm aspirin therapy is ceased. There may be an increased intraoperative bleeding risk during this period. Although some individuals discontinuing aspirin may have “normalized” their hemostatic function before this point, it would seem most prudent to use the maximum period of time taken for all the tests used in this study to return to their baseline level as the duration on which to base guidelines. Based on this, our study suggests that withdrawal of aspirin 5 days before planned surgery, with the operation taking place on the sixth day, is optimum. Although this has not been demonstrated before, it is consistent with a clear understanding of platelet kinetics, because the influence of the following factors should allow recovery of adequate clotting mechanisms before renewal of the entire platelet population.
Table 3. Median Closure Times for Each Group on Each Day (n ⫽ numbers tested each day) Group
Pretreatment
24 h
48 h
A (n ⫽ 9) B (n ⫽ 12) C (n ⫽ 17)
146 (n ⫽ 9) 140 (n ⫽ 12) 137 (n ⫽ 17)
135 (n ⫽ 9) 230 (n ⫽ 11) ⬎300 (n ⫽ 17)
139 (n ⫽ 9) 258 (n ⫽ 11) ⬎300 (n ⫽ 16)
Post-treatment, sec 72 h 96 h
162 (n ⫽ 3) 204 (n ⫽ 9) 226 (n ⫽ 16)
164 (n ⫽ 1) 150 (n ⫽ 9) 199 (n ⫽ 14)
120 h
144 h
(n ⫽ 0) 179 (n ⫽ 2) 255 (n ⫽ 5)
(n ⫽ 0) 179 (n ⫽ 1) 165 (n ⫽ 1)
Data represent the median value for the individuals tested on each time point, so values before and after testing do not equate. Testing was discontinued when all individuals’ values had returned to baseline levels (⫾ 10%). Group A, placebo; group B, 75 mg aspirin; group C, 300 mg aspirin.
Vol. 200, No. 4, April 2005
Cahill et al
Increased Bleeding after Ceasing Aspirin Therapy
569
Figure 5. Closure times (PFA-100 tests performed using the collagen/epinephrine membranes) at each measured time point for individuals taking 300 mg aspirin. Maximum measurable closure time ⫽ 300 seconds; ⽧, median value; ⫺, 25th and 75th percentiles; □, 5th and 95th percentiles; n, number of individuals tested at each time point. Value at 0 hours is the baseline level while testing was discontinued when each individual’s values had returned to that person’s baseline level ⫾ 10%.
Even while a patient is taking aspirin, some platelet function persists (circa 20%). In addition, although aspirin also affects platelets that have not yet been released by megakaryocytes, these nucleated cells can regenerate cyclooxygenase and recover function about 12 hours af-
ter treatment cessation.38 Also, platelet turnover may be increased in patients with clinically obvious vascular disease.39 Newly formed platelets may contribute more to overall hemostatic effectiveness than senescent ones,37 while thromboxane A2 produced by newly formed, un-
Figure 6. Closure times (PFA-100 tests performed using the collagen/epinephrine membranes) at each measured time point for individuals taking 75 mg aspirin. Maximum measurable closure time ⫽ 300 seconds; ⽧, median value; ⫺, 25th and 75th percentiles; □, 5th and 95th percentiles; n, number of individuals tested at each time point. Value at 0 hours is the baseline level while testing was discontinued when each individual’s values had returned to that person’s baseline level ⫾ 10%.
570
Cahill et al
Increased Bleeding after Ceasing Aspirin Therapy
J Am Coll Surg
Figure 7. Closure times (PFA-100 tests performed using the collagen/epinephrine membranes) at each measured time point for individuals taking the placebo tablet. Maximum measurable closure time ⫽ 300 seconds; ⽧, median value; ⫺, 25th and 75th percentiles; □, 5th and 95th percentiles; n, number of individuals tested at each time point. Value at 0 hours is the baseline level while testing was discontinued when each individual’s values had returned to that person’s baseline level ⫾ 10%.
inhibited platelets can stimulate aggregation of other older platelets unable to produce their own thromboxane A2.40 In addition to thromboxane A2, platelet aggregation in a wound is stimulated by other factors (eg, flow rates and shear forces) and substances present (eg, collagen). Such factors are not measured in specific platelet aggregrometry tests. An entirely normal platelet population is not necessary to have adequate clotting mechanisms (approximately one-half of the normal number of platelets is sufficient for hemostasis during
most surgical procedures).41 In addition, aspirin also inhibits endothelial cell production of prostacyclin, an effect that, although reversible, counters its antiplatelet effects. So even though a proportion of platelets may still show an aspirin-induced defect (demonstrable by platelet release assays), overall platelet function should be adequate to allow normal clotting in significantly less time than 10 days. The subjects in this study differ from actual patients on aspirin in that they are young and healthy. Ethical
Figure 8. Numbers of tests remaining abnormal each day, demonstrating that all bleeding times had normalized by 96 hours and all closure times had normalized within 144 hours.
Vol. 200, No. 4, April 2005
Cahill et al
considerations prevented this study from being performed using actual patients because the group taking the placebo would have to interrupt their therapy for the purposes of the study. Interestingly, related studies have suggested that platelet function in patients with clinically apparent cardiovascular disease39 should recover more quickly from aspirin’s effects than younger, healthier patients, so our recommendations on when to withhold therapy should be applicable to actual patients taking aspirin. The relatively high rate of noncompliance with the study protocol most likely reflects the arduous nature of the requested regimen over a considerable period of time (up to 21 days in total for some of those involved), and because all participants were volunteers, anyone could withdraw at any time. Nine volunteers withdrew from the study after commencing their tablets but before completing their 2-week course; 5 more attended for posttreatment hemostatic testing only for a limited duration (only providing complete sets of bleeding time measurements while their closure time data were incomplete). Four of the individuals who did attend as requested were excluded from analysis because they seemed to have taken aspirin-like medications during their testing period (which obviously skewed their results). The choice of normalization of tests as the study end point for participating individuals rather then a set number of days accounts for the dwindling number of tests measured at the later time points, and reflects both our initial uncertainty about how long aspirin’s effects may persist (the point of the study) and our desire to limit the number of (relatively unpleasant) tests preformed for ethical reasons. The dose of aspirin usually used clinically in Europe ranges between 75 and 300 mg once a day, so these were the doses used in this study. Although this study was not designed or powered to examine for differences in the persistence of the effects of different doses of aspirin, as there was no apparent difference found in the duration of increased bleeding tendency in the two groups used in this study, it seems logical to extrapolate that the same advice should also apply to those taking an intermediate dose. Because hemostatic recovery from aspirin’s effects depends primarily on sufficient replenishment of a functioning platelet population, similar recommendations may be considered for those taking greater daily dosages. Although people with a higher baseline platelet count theoretically may be expected to recover from aspirin’s
Increased Bleeding after Ceasing Aspirin Therapy
571
effects differently than do people with a low platelet count, we found no correlation between recovery from aspirin and either platelet volume or number. But the range of these values in our volunteers was narrow. Although many patients who come off their aspirin before surgery will have been taking this drug for longer than the 2-week duration of our volunteers, the fact that aspirin’s antihemostatic effects are mediated by irreversible platelet function inhibition (so are limited by platelet life-cycle dynamics) makes the duration of pretreatment used in this study adequate to allow the full antihemostatic effects of aspirin therapy to develop.42 Indeed, previous studies have demonstrated that maximal inhibition of platelet response to aspirin occurs after much shorter treatment periods.43-45 Additionally, rather than more prolonged usage leading to a cumulative effect, recent evidence suggests that the efficacy of aspirin’s antiplatelet effects may actually diminish with time.46 Bleeding times and PFA-100 tests can only approximate platelet activity in a surgical wound, but they are better tests than aggregrometry, which relies on a separated sample. The standardized template bleeding time (TBT) attempts to measure the overall effectiveness of the hemostatic mechanism. Although it is known that bleeding times are prolonged for up to 4 days after single doses of aspirin,47 here we have demonstrated that this is also the time taken for a return to normal after a prolonged course of aspirin in the therapeutic range. The TBT has its limitations.48-50 It may, in fact, reflect aspirin’s effects on the vascular endothelium rather than the platelet reaction to injury.26 Because aspirin’s antiplatelet effect is considered to be the more significant contribution to the overall increase in bleeding tendency and may also be its most persistent effect, a platelet function test (a more specific and reproducible means of determining recovery from aspirin’s prohemorrhagic effects) was also included in this study. The PFA-100 assesses primary hemostasis in whole blood by attempting to recreate conditions found in injured blood vessels. Although now well established as a useful measure, its use as an instrument to monitor the effects of drugs such as aspirin is novel. Although platelet aggregation tests are useful in the study of specific aspects of platelet activity, they do not reflect overall platelet capabilities. In addition, unlike aggregometry, the PFA-100 is quick, easy to perform, and is reproducible.23,51 The delay in the return to baseline of this platelet function test compared with the TBTs in the majority
572
Cahill et al
Increased Bleeding after Ceasing Aspirin Therapy
of volunteers may correlate with the faster recovery of the vasculature compared with the platelets from the effects of aspirin. While the TBT plays no role as a routine single preoperative screening test, the PFA-100 may have a clinically useful application in detecting the presence of an increased bleeding tendency in patients admitted for elective surgery who may have inadvertently not discontinued their aspirin for the recommended period of time (by comparing the results obtained by using the collagen/epinephrine versus the collagen/ADP membranes). It may also allow consideration of agents to reverse aspirin effects (eg, desmopressin acetate,52 aprotinin,53 or platelet transfusion54) in patients who require emergency surgery. In summary, this study demonstrates that the duration of increased bleeding tendency after withdrawal of aspirin therapy persists for 5 days. This period of time is consistent with a close understanding of the mechanisms of aspirin’s action and of platelet kinetics. We believe that aspirin therapy should be discontinued 5 days before elective surgery and the operation planned for the sixth day after treatment cessation. This should allow physicians and surgeons to minimize the risks of stopping aspirin without increasing the risk of excessive perioperative bleeding. Author Contributions Study conception and design: RA Cahill, McGreal, MR Cahill, Redmond Acquisition of data: RA Cahill, Crowe, Ryan, Manning Analysis and interpretation of data: RA Cahill, Crowe, MR Cahill Drafting of manuscript: RA Cahill, McGreal, MR Cahill, Redmond Critical revision: RA Cahill, MR Cahill, Redmond Statistical expertise: RA Cahill, MR Cahill Obtaining funding: RA Cahill, Redmond Supervision: MR Cahill, Redmond REFERENCES 1. George JN, Shattil SJ. Acquired disorders of platelet function. In: Hoffman R, Benz EJ, Shattil SJ, et al, eds. Haematology. Basic principles and practice. PA: Churchhill Livingstone; 1991: 2172–2186. 2. WarkenhemTE, Kelten JG. Acquired platelet disorders. In: Bloom AL, Forbes CD, Thomas DP, Tuddenham EG, eds. Haemostasis and thrombosis. Singapore: Churchhill Livingstone;1994:765–815.
J Am Coll Surg
3. Bick RL, Adams T, Schmalhorst WR. Bleeding times, platelet adhesion and aspirin. Am J Clin Pathol 1976;65:69–72. 4. Anti-Platelet Trialists’ Collaboration. Collaborative overview of randomised trials of anti-platelet therapy. Reduction in venous thrombosis and pulmonary embolism by anti-platelet prophylaxis among surgical and medical patients. BMJ 1994;308:235– 246. 5. Ferraris VA, Ferraris SP, Joseph O, et al. Aspirin and postoperative bleeding after coronary bypass grafting. Ann Surg 2002; 235:820–827. 6. Kallis P, Tooze JA, Talbot S, et al. Pre-operative aspirin decreases platelet aggregation and increases post-operative blood loss—a prospective, randomised, placebo-controlled, double-blind clinical trial in 100 patients with chronic, stable angina. Eur J Cardiothorac Surg 1994;8:404–409. 7. Palmer JD, Sparrow OC, Iannotti F. Post-operative haematomata: a 5-year survey and identification of avoidable risk factors. Neurosurgery 1994;35:1061–1064. 8. Amrein PC, Ellman L, Harris WH. Aspirin-induced prolongation of the bleeding time and perioperative blood loss. JAMA 1981;245:1825–1828. 9. Kitchen L, Erichson RB, Sideropoulos H. Effect of druginduced platelet dysfunction on surgical bleeding. Am J Surg 1982;143:215–217. 10. Mak S, Amoroso P. Stop those antiplatelet agents before surgery! BJU International 2003;91:593–594. 11. George JN, Shattil SJ. The clinical importance of acquired abnormalities of platelet function. N Engl J Med 1991;324:27– 39. 12. Giangrande PLF, Littlewood TJ. Medical problems in surgical patients. Haematological problems. In: Morris PJ, Malt RA, eds. Oxford textbook of surgery. Oxford: Oxford University Press; 1994:212–221. 13. Harkess JW. Arthroplasty of the hip. In: Crenshaw AH, ed. Campbell’s operative orthopaedics. St Louis: CV Mosby; 1992: 441–626. 14. Rarick MU, Feinstein DI. Haematological problems and complications. In: Apuzzo MLJ, ed. Brain surgery: Complication avoidance and management. Edinburgh: Churchhill Livingstone; 1993:145–154. 15. Matsuzaki K, Matsui K, Haraguchi N, et al. Ischemic heart attacks following cessation of aspirin before coronary bypass surgery: A report of two cases. Ann Thorac Cardiovasc Surg 1999;5:121–122. 16. Kovich O, Otley CC. Thrombotic complications related to discontinuation of warfarin and aspirin perioperatively for cutaneous operations. J Am Acad Dermatol 2003;48:233–237. 17. Mitchell SM, Sethia KK. Hazards of aspirin withdrawal before transurethral prostatectomy BJU International 1999;84:550. 18. Fiore LD, Brophy MT, Lopez A, et al. The bleeding time in response to aspirin: identifying the hyper-responder. Am J Clin Path 1990;94:292–296. 19. Benzon HT, Brunner EA, Vaisrub N. Bleeding time and nerve blocks after aspirin. Reg Anesth 1983;9:86–89. 20. Trentalange MJ, Walts LF. A comparison of thromboelastogram and template bleeding time in the evaluation of platelet function after aspirin. J Clin Anaesth 1991;3:377–381. 21. Sonksen JR, Kong KL, Holder R. Magnitude and time course of impaired primary haemostasis after stopping chronic low and medium dose aspirin in healthy volunteers. Br J Anaesth 1999; 82:360–365. 22. Miltich JP. Bleeding time. In: Beutler E, Lichtmann MA, Coller
Vol. 200, No. 4, April 2005
23. 24. 25.
26.
27. 28. 29.
30. 31. 32. 33. 34. 35.
36. 37.
38.
Increased Bleeding after Ceasing Aspirin Therapy
Cahill et al
BS, Kipps TJ, eds. Williams Haematology. USA:McGraw-Hill; 1995:L111–112. Kundu SK, Heilmann EJ, Sio R, et al. Description of an in vitro platelet function analyzer-PFA-100. Semin Thromb Hemost 1995;21[Suppl 2]:106–112. Lanas AI, Arroyo MT, Esteva F, et al. Aspirin related gastrointestinal bleeders have an exaggerated bleeding time response due to aspirin use. Gut 1996;39:654–660. Anti-Platelet Trialists’ Collaboration. Collaborative overview of randomised trials of anti-platelet therapy-1: Prevention of death, myocardial infarct and stroke by prolonged anti-platelet therapy in various categories of patients. BMJ 1994;308:81– 106. Bick RL. Hereditary and acquired vascular bleeding disorders. In: Bick RL, Bennett MD, Brynes RK, et al, eds. Haematology— clinical and laboratory practice. St Louis: Mosby; 1993:1325– 1336. Paereti FI, Mannucci PM, D’Angelo A. Congenital deficiency of thromboxane and prostacyclin. Lancet 1980;1:898–901. Scher KS. Unplanned operation for rebleeding. Am Surg 1996; 62:52–55. Samama CM, Bastien O, Forestier F, et al. Antiplatelet agents in the perioperative period: expert recommendations of the French Society of Anesthesiology and Intensive care (SFAR) 2001— summary statement. Can J Anaesth 2002;49:S25–35. Topol EJ. Aspirin with bypass surgery—from taboo to new standard of care. N Engl J Med 2002;347:1359–1360. Nellipovitz DT, Bryson GL, Nichol G. The effect of perioperative aspirin therapy in peripheral vascular surgery. A decision analysis. Anesth Analg 2001;93:573–580. Devereaux PJ, Ghali WA, Gibson NE, et al. Physician’s recommendations for patients who undergo noncardiac surgery. Clin Invest Med 2000;23:116–123. Russell MW, Jobes D. What should we do with aspirin, NSAIDs and glycoprotein receptor inhibitors? Int Anesthiol Clin 2002; 40:63–76. Bullingham A, Strunin L. Prevention of postoperative venous thromboembolism. Br J Anaesth 1995;75:622–630. Nielsen JD, Holm-Nielsen A, Jespersen J, et al. The effect of low-dose acetylsalicyclic acid on bleeding after transurethral prostatectomy—a prospective, randomised double blind, placebo-controlled study. Scand J Urol Nephrol 2000;34:194– 198. Bick RL. Antithrombotic therapy. In: Bick RL, Bennett MD, Brynes RK, et al, eds. Haematology—clinical and laboratory practice. St Louis: Mosby; 1993:1603–1624. George JN, Dale GL. Platelet kinetics. In: Bloom AL, Forbes CD, Thomas DF, Tuddenham EG, eds. Haemostasis and thrombosis. Singapore: Churchhill Livingstone; 1994:1202– 1205. Van Pampus EC, Huijgens PC, Zevenbergen A, et al. Influence
39. 40. 41. 42. 43.
44.
45.
46. 47. 48. 49. 50.
51. 52. 53.
54.
573
of aspirin on human megakaryocyte prostaglandin synthesis. Eur J Haematol 1993;50:264–268. Winocour PD. Platelet turnover in advanced diabetes. Eur J Clin Invest 1994;24 [Suppl 1]:34–37. Packham MA. Role of platelets in thrombosis and haemostasis. Can J Physiol Pharmacol 1994;72:278–284. Counts RB, Haisch C, Simon TL, et al. Haemostasis in massively transfused patients. Ann Surg 1979;190:91–99. Jakubowski JA, Stampfer MJ, Vaillancourt R, Deykin D. Cumulative antiplatelet effect of low-dose enteric coated aspirin. Br J Haematol 1985;60:635–642. May JA, Heptinstall S, Cole AT, Hawkey CJ. Platelet responses to several agonists and combinations of agonists in whole blood: a placebo controlled comparison of the effects of a once daily dose of plain aspirin 300 mg, plain aspirin 75 mg and enteric coated aspirin 300 mg in man. Thromb Res 1997;88:183–192. Kuster LJ, Frolich JC. Platelet aggregation and thromboxane release induced by arachidonic acid, collagen, ADP and plateletactivating factor following low dose acetylsalicyclic acid in man. Prostaglandins 1986;32:415–423. Kallmann R, Nieuwenhuis HK, de Groot PG, et al. Effects of low doses of aspirin on bleeding time, thromboxane production and 6-keto-PGF1 alpha excretion in healthy subjects. Thromb Res 1987;45:355–361. Pulcinelli FM, Pignatelli P, Celestini A, et al. Inhibition of platelet aggregation by aspirin progressively decreases in long-term treated patients. J Am Coll Cardiol 2004;43:979–984. Mielke CHJ. Aspirin prolongation of the template bleeding time: influence of venostasis and direction of incision. Blood 1982;60:1139–1143. De Rossi SS, Glick M. Bleeding time: an unreliable predictor of clinical haemostasis. J Oral Maxillofac Surg 1996;54:1119– 1120. Channing Rodgers RP, Levin J. A critical reappraisal of the bleeding time. Semin Thromb Haemost 1990;16:1–20. DeCaterina R, Lanza M, Manca G, et al. Bleeding time and bleeding: An analysis of the relationship of the bleeding time test with parameters of surgical bleeding. Blood 1994;84:3363– 3370. Nicholson NS, Panzer-Knodle SG, Haas NF, et al. Assessment of platelet function assays. Am H J 1988;135:S170–178. Porte RJ, Leebeck FW. Pharmacological strategies to decrease transfusion requirements in patients undergoing surgery. Drugs 2002;62:2193–2211. Schonberger JP, Bredee JJ, van Oeursen W, et al. Preoperative therapy of low dose aspirin in internal artery bypass operations with or without low dose aprotinin. J Thorac Cardiovasc Surg 1993;106:262–267. Hirsch J, Fuster V, Salzman E. Dose anti-platelet agents: the side-effects, and anti-thrombotic effectiveness. Chest 1986;89 [Suppl 2]:4S–10S.
dency after stopping longterm aspirin therapy has not previously been determined. Although some studies have attempted to do so, their clinical application is limited because of study design18,19 (observational studies), techniques20 (measurements of platelet function in isolation may not reflect overall haemostatic capability), or lack of a sensitive end point21 (an individual’s bleeding time may be significantly prolonged while at the same time be within the normal range for a population). Our aim was to determine the duration of increased bleeding tendency after cessation of aspirin therapy using methodology sensitive to and reflective of in vivo changes of both platelet function and global hemostatic capacity.
Aspirin has a significant effect on hemostasis;1 it prolongs bleeding time2,3 and is associated with increased intraoperative and postoperative bleeding.4-9 It is often recommended that patients on longterm aspirin therapy discontinue treatment for a period of time before elective surgical procedures,10-14 allowing their primary hemostatic mechanisms to return to normal. But while off aspirin, some patients may be at risk of thrombosis.15-17 With increasing numbers of people on aspirin, the balance between the risk of increased bleeding and that of thrombosis is of considerable importance in planning elective surgery. The actual duration of the increased bleeding tenCompeting interests declared: None.
METHODS A prospective, double blind, randomized study was performed after obtaining approval of the protocol by the local ethics committees. Fifty-one healthy volunteers were recruited, with fully informed consent, from the medical and paramedical staff of a regional hospital. All
Received October 16, 2004; Revised November 2, 2004; Accepted November 2, 2004. From the Department of Surgery, NUI (Cork), Cork University Hospital, Cork, Ireland (RA Cahill, McGreal, Ryan, Manning, Redmond) and the Department of Haematology, Mid-Western Regional Hospital, Limerick, Ireland (Crowe, MR Cahill). Correspondence address: Professor HP Redmond, Professional Unit, Department of Surgery, Cork University Hospital, Cork, Ireland.
© 2005 by the American College of Surgeons Published by Elsevier Inc.
564
ISSN 1072-7515/05/$30.00 doi:10.1016/j.jamcollsurg.2004.11.002
Vol. 200, No. 4, April 2005
Cahill et al
were asked about contraindications to commencing aspirin or history of any bleeding or wound healing problems. No volunteer was taking any regular medication and all denied recent ingestion of any drug known to have an antiplatelet effect within the previous 4 weeks. Volunteers were randomly assigned in a doubleblinded fashion into three separate groups. Members of the first group (group A) received a placebo; those in the other two groups (groups B and C) were given either of the two extremes of aspirin dosages in common clinical use, ie, 75 or 300 mg once a day, respectively. All tablets were identical in appearance. A baseline platelet count, mean platelet volume, standardized template bleeding time, and platelet function test (Platelet Function Analyser 100- PFA-100; Dade Behring) were measured. The groups were then started on a 14-day course of treatment, at the end of which the tests were performed on a daily basis until a return to baseline was demonstrated. This indicated the recovery of the individual’s hemostatic mechanisms. The value taken as the baseline score was either the pretreatment score or, in a small number of volunteers (n ⫽ 3), the repeatedly normal posttreatment scores. Volunteers were asked to abstain from alcohol and nonsteroidal antiinflammatory agents along with any other plateletmodifying medications for the 24 hours before and on each day during testing. The template bleeding time was used to test for a primary hemostatic defect (from a vascular or platelet effect) induced by aspirin. The same experienced investigator used a standardized method (the modified Ivy technique22) with a commercially available device (Simplate 1; Organon Teknika Corp) to perform the test on each volunteer. The time taken for the bleeding to stop was measured to the nearest 20 seconds. The manufacturer’s range of normal bleeding times with this method is between 2 and 10 minutes. The PFA-100 was used as a specific laboratory measure of platelet function and global hemostasis. It is reproducible and minimally manipulative.23 This device aspirates anticoagulated whole blood through a capillary toward a microscopic aperture (150 m in diameter) cut into a biologically active membrane (coated with collagen and either adenosine-5-diphosphate [ADP] or epinephrine, the latter being particularly sensitive to aspirin-induced defects). The high shear rates, along with these biochemical stimuli, mimic conditions in an actual wound, thereby activating platelets and causing a
Increased Bleeding after Ceasing Aspirin Therapy
565
Table 1. Final Composition of the Groups Studied Group
n
Mean age (ⴞ SD), y
Male:female ratio
Group A Group B Group C
9 12 17
23.7 (⫾ 0.2) 24.1 (⫾ 0.6) 24.2 (⫾ 0.3)
6:3 7:5 12:5
Group A, placebo; group B, 75 mg aspirin; group C, 300 mg aspirin.
stable plug to be slowly built at the aperture, stopping the flow of blood through it. The time required to fully occlude the aperture is recorded by a microprocessor as the closure time. A prolonged closure time using the collagen/epinephrine membrane with a normal collagen/ ADP time is associated with a specific aspirin-induced defect. The usual range of normal closure times for the collagen/epinephrine membrane is between 85 and 176 seconds, with a maximum measurement of 300 seconds and a coefficient of variability of 10%. Two separate specimens were tested for each volunteer on each day, with the mean recorded as the representative score. RESULTS Although 51 volunteers were asked to participate in the study and gave fully informed consent, 9 individuals subsequently withdrew before completing the study protocol (2 developed upper gastrointestinal symptoms while taking aspirin, and 7 were unavailable to attend at the required times). This resulted in 42 participants who took either aspirin or placebo tablets for 14 days and then underwent daily analysis of their hemostatic function after treatment cessation. But the PFA demonstrated that four of these individuals had ingested platelet-modifying drugs during the testing phase (when ingestion of these agents was precluded), so data from these individuals have also been excluded. Once the bleeding time had normalized, an additional five volunteers dropped out of the study even though their PFA tests were still prolonged. Their data on the days they attended for testing are included in the analysis. So data from 38 volunteers (13 of whom were women) are presented and analyzed. The numbers, ages, and gender of the volunteers in each group for whom data were studied are shown inTable 1; the mean of the prolongation of each individual’s tests (both bleeding times and closure times) is shown in Figure 1. Platelet number and volume measurements remained constant for each individual throughout the study period. The mean ⫾ standard deviation for the individuals’ platelet counts during the study was 231 ⫾ 14 ⫻ 103/L, while that for the platelet volumes was 9.4 ⫾ 0.2 ⫻ 103/fL.
566
Cahill et al
Increased Bleeding after Ceasing Aspirin Therapy
J Am Coll Surg
Figure 1. Mean of the prolongation (daily value– baseline value) of each individual’s tests (both bleeding and closure times) at each measured time point. A closure time of 300 was ascribed to those with closure times ⬎300 seconds (the maximum score measurable by the PFA-100), so the values shown at the first three time points are underestimates of the true value). Solid line, trend of closure times; dotted line, trends of bleeding times; Œ, individuals taking 300 mg aspirin; , those taking 75 mg aspirin.
The mean prolongation of bleeding times above the baseline value of each group on each day of testing is shown in Table 2; the daily ranges in values are shown in Figures 2, 3, and 4.Three volunteers (one in group B, two in group C) showed a hyper-response18,24 to aspirin treatment. That is, their bleeding times while on aspirin were prolonged greater than two standard deviations from the mean. Although the majority of bleeding times of those who took aspirin (n ⫽ 23) had returned to pretreatment levels by 72 hours, 6 remained prolonged until 96 hours (of these, five were in group C and two were hyper-responders). The mean time (⫾ standard deviation) of return to normal for group B was 60 (⫾ 19) hours and for group C was 70 (⫾ 22) hours. Bleeding times measured on the first posttreatment day of two volunteers in the placebo group were prolonged; that of one person taking aspirin was not significantly prolonged above the baseline value. The median and interquartile ranges for closure times using the collagen/epinephrine membrane for each group on each day are shown in Table 3 and Figures 5, 6, and 7. In
each case, as expected, the corresponding collagen/ADP measurements remained at their baseline level (data not shown). The latest significant prolongation of closure times (collagen/epinenephrine membrane) was detected at 120 hours (Fig. 8). The mean (⫾ standard deviation) time of return to baseline for these closure times was 84 (⫾ 31) hours for group B (75 mg aspirin) and 94 (⫾ 37) hours for group C (300 mg aspirin). Although there was considerable variability among individuals on aspirin as to when their hemostatic function returned back to their baseline levels (17 tests had normalized 48 hours after completion of the treatment course, and 5 tests remained prolonged after 96 hours), all tests had returned to normal by the sixth day after treatment cessation (Fig. 8). In general, the PFA-100 tests remained prolonged for 24 to 74 hours longer than their corresponding bleeding times (although the 2 tests normalized on the same day in 3 people) and the PFA100 test normalization preceded that of the bleeding time tests by 24 hours in 4 people. Although there was a
Table 2. Mean Value and Mean Prolongation Above Baseline Levels of Bleeding Times for Each Group Before and on Each Day after Completion of Tablets Group
A (n ⫽ 9) B (n ⫽ 12) C (n ⫽ 17)
Pretreatment
24 h
362 440 450
382 (22) 570 (105) 667 (211)
Data shown are mean BT (mean prolongation of BT), in seconds. BT, bleeding times; group A, placebo; group B, 75 mg aspirin; group C, 300 mg aspirin.
Posttreatment 48 h
360 446 (60) 554 (121)
72 h
96 h
380 443 460 (53)
— 320 440
Vol. 200, No. 4, April 2005
Cahill et al
Increased Bleeding after Ceasing Aspirin Therapy
567
Figure 2. Bleeding times at each measured time point for individuals taking 300 mg aspirin. ⽧, mean value; ⫺, 25th and 75th percentile; □, 5th and 95th percentiles; n, number of individuals tested at each time point. Value at 0 hours is the baseline level while testing was discontinued when each individual’s values had returned to that person’s baseline level.
trend toward a longer duration of both bleeding and closure time prolongation in the group taking the higher dose (300 mg) of aspirin, in neither case was this statistically significant. Last, there was no significant difference noted between men and women or between older and younger patients in this study (data not shown). DISCUSSION Aspirin remains in common use for the primary and secondary prevention of myocardial infarct and stroke.25 Its hemorrhagic effects are known to be mainly mediated by its effects on platelet prostaglandin synthesis because of cyclooxygenase inhibition. In addition, it is also a potent inhibitor of acetylcholinesterase (thereby reduc-
ing vascular constriction in response to trauma-induced cathecholamines).26 Aspirin ingestion may theoretically be considered to cause a bleeding risk similar to that found in patients with congenital cyclooxygenase deficiency.26,27 In addition, it may also exacerbate bleeding associated with otherwise subclinical hemostatic abnormalities (eg, von Willebrand’s disease). Increased perioperative blood loss has been observed in patients on aspirin undergoing many different procedures,4,28 including cardiothoracic,5,6 neurosurgical,7 orthopaedic,9 gynecologic, and general surgical10 operations. That the increased blood loss in some studies does not reach statistical significance does not imply that it is not clinically relevant. Overall,
Figure 3. Bleeding times at each measured time point for individuals taking 75 mg aspirin, ⽧, mean value; ⫺, 25th and 75th percentile; □, 5th and 95th percentiles; n, number of individuals tested at each time point. Value at 0 hours is the baseline level while testing was discontinued when each individual’s values had returned to that person’s baseline level.
568
Cahill et al
Increased Bleeding after Ceasing Aspirin Therapy
J Am Coll Surg
Figure 4. Bleeding times at each measured time point for individuals taking placebo tablets. ⽧, mean value; ⫺, 25th and 75th percentile; □, 5th and 95th percentiles; n, number of individuals tested at each time point. Value at 0 hours is the baseline level while testing was discontinued when each individual’s values had returned to that person’s baseline level.
it has been concluded that aspirin’s effect on hemostasis is “mild, variable but significant.”11 So, many surgeons advise that aspirin therapy be withdrawn before elective surgical intervention. This study provides evidence to support the withdrawal of aspirin before elective surgery and aims to guide clinicians who do not operate electively on patients taking this drug. Although no definite data exist on the exact risk of thrombotic events in individuals who have discontinued their treatment for a short period of time, it may increase the thrombotic risk to withhold therapy for longer than is necessary.29 Indeed, such are the beneficial effects of aspirin that increasingly, aspirin therapy is continued throughout the perioperative period of cardiovascular procedures despite its associated hemorrhagic effects.30,31 The optimal period of time for which aspirin should be withheld before elective surgical intervention is uncertain,32 but is often suggested to be several days.11,33,34 In accordance with published recommendations,35 it had been our practice to advise discontinuation of therapy 10 days before planned operations. Because the antiplatelet effects of aspirin are mediated
through the irreversible inhibition of cyclooxygenase36 and the lifespan of the platelet is approximately 10 days,37 this time period should allow an entirely normal platelet population to be reestablished in the circulation. Our data demonstrate that sensitive measures of platelet function and primary hemostasis can detect a measurable defect for up to 5 days after longterm aspirin therapy is ceased. There may be an increased intraoperative bleeding risk during this period. Although some individuals discontinuing aspirin may have “normalized” their hemostatic function before this point, it would seem most prudent to use the maximum period of time taken for all the tests used in this study to return to their baseline level as the duration on which to base guidelines. Based on this, our study suggests that withdrawal of aspirin 5 days before planned surgery, with the operation taking place on the sixth day, is optimum. Although this has not been demonstrated before, it is consistent with a clear understanding of platelet kinetics, because the influence of the following factors should allow recovery of adequate clotting mechanisms before renewal of the entire platelet population.
Table 3. Median Closure Times for Each Group on Each Day (n ⫽ numbers tested each day) Group
Pretreatment
24 h
48 h
A (n ⫽ 9) B (n ⫽ 12) C (n ⫽ 17)
146 (n ⫽ 9) 140 (n ⫽ 12) 137 (n ⫽ 17)
135 (n ⫽ 9) 230 (n ⫽ 11) ⬎300 (n ⫽ 17)
139 (n ⫽ 9) 258 (n ⫽ 11) ⬎300 (n ⫽ 16)
Post-treatment, sec 72 h 96 h
162 (n ⫽ 3) 204 (n ⫽ 9) 226 (n ⫽ 16)
164 (n ⫽ 1) 150 (n ⫽ 9) 199 (n ⫽ 14)
120 h
144 h
(n ⫽ 0) 179 (n ⫽ 2) 255 (n ⫽ 5)
(n ⫽ 0) 179 (n ⫽ 1) 165 (n ⫽ 1)
Data represent the median value for the individuals tested on each time point, so values before and after testing do not equate. Testing was discontinued when all individuals’ values had returned to baseline levels (⫾ 10%). Group A, placebo; group B, 75 mg aspirin; group C, 300 mg aspirin.
Vol. 200, No. 4, April 2005
Cahill et al
Increased Bleeding after Ceasing Aspirin Therapy
569
Figure 5. Closure times (PFA-100 tests performed using the collagen/epinephrine membranes) at each measured time point for individuals taking 300 mg aspirin. Maximum measurable closure time ⫽ 300 seconds; ⽧, median value; ⫺, 25th and 75th percentiles; □, 5th and 95th percentiles; n, number of individuals tested at each time point. Value at 0 hours is the baseline level while testing was discontinued when each individual’s values had returned to that person’s baseline level ⫾ 10%.
Even while a patient is taking aspirin, some platelet function persists (circa 20%). In addition, although aspirin also affects platelets that have not yet been released by megakaryocytes, these nucleated cells can regenerate cyclooxygenase and recover function about 12 hours af-
ter treatment cessation.38 Also, platelet turnover may be increased in patients with clinically obvious vascular disease.39 Newly formed platelets may contribute more to overall hemostatic effectiveness than senescent ones,37 while thromboxane A2 produced by newly formed, un-
Figure 6. Closure times (PFA-100 tests performed using the collagen/epinephrine membranes) at each measured time point for individuals taking 75 mg aspirin. Maximum measurable closure time ⫽ 300 seconds; ⽧, median value; ⫺, 25th and 75th percentiles; □, 5th and 95th percentiles; n, number of individuals tested at each time point. Value at 0 hours is the baseline level while testing was discontinued when each individual’s values had returned to that person’s baseline level ⫾ 10%.
570
Cahill et al
Increased Bleeding after Ceasing Aspirin Therapy
J Am Coll Surg
Figure 7. Closure times (PFA-100 tests performed using the collagen/epinephrine membranes) at each measured time point for individuals taking the placebo tablet. Maximum measurable closure time ⫽ 300 seconds; ⽧, median value; ⫺, 25th and 75th percentiles; □, 5th and 95th percentiles; n, number of individuals tested at each time point. Value at 0 hours is the baseline level while testing was discontinued when each individual’s values had returned to that person’s baseline level ⫾ 10%.
inhibited platelets can stimulate aggregation of other older platelets unable to produce their own thromboxane A2.40 In addition to thromboxane A2, platelet aggregation in a wound is stimulated by other factors (eg, flow rates and shear forces) and substances present (eg, collagen). Such factors are not measured in specific platelet aggregrometry tests. An entirely normal platelet population is not necessary to have adequate clotting mechanisms (approximately one-half of the normal number of platelets is sufficient for hemostasis during
most surgical procedures).41 In addition, aspirin also inhibits endothelial cell production of prostacyclin, an effect that, although reversible, counters its antiplatelet effects. So even though a proportion of platelets may still show an aspirin-induced defect (demonstrable by platelet release assays), overall platelet function should be adequate to allow normal clotting in significantly less time than 10 days. The subjects in this study differ from actual patients on aspirin in that they are young and healthy. Ethical
Figure 8. Numbers of tests remaining abnormal each day, demonstrating that all bleeding times had normalized by 96 hours and all closure times had normalized within 144 hours.
Vol. 200, No. 4, April 2005
Cahill et al
considerations prevented this study from being performed using actual patients because the group taking the placebo would have to interrupt their therapy for the purposes of the study. Interestingly, related studies have suggested that platelet function in patients with clinically apparent cardiovascular disease39 should recover more quickly from aspirin’s effects than younger, healthier patients, so our recommendations on when to withhold therapy should be applicable to actual patients taking aspirin. The relatively high rate of noncompliance with the study protocol most likely reflects the arduous nature of the requested regimen over a considerable period of time (up to 21 days in total for some of those involved), and because all participants were volunteers, anyone could withdraw at any time. Nine volunteers withdrew from the study after commencing their tablets but before completing their 2-week course; 5 more attended for posttreatment hemostatic testing only for a limited duration (only providing complete sets of bleeding time measurements while their closure time data were incomplete). Four of the individuals who did attend as requested were excluded from analysis because they seemed to have taken aspirin-like medications during their testing period (which obviously skewed their results). The choice of normalization of tests as the study end point for participating individuals rather then a set number of days accounts for the dwindling number of tests measured at the later time points, and reflects both our initial uncertainty about how long aspirin’s effects may persist (the point of the study) and our desire to limit the number of (relatively unpleasant) tests preformed for ethical reasons. The dose of aspirin usually used clinically in Europe ranges between 75 and 300 mg once a day, so these were the doses used in this study. Although this study was not designed or powered to examine for differences in the persistence of the effects of different doses of aspirin, as there was no apparent difference found in the duration of increased bleeding tendency in the two groups used in this study, it seems logical to extrapolate that the same advice should also apply to those taking an intermediate dose. Because hemostatic recovery from aspirin’s effects depends primarily on sufficient replenishment of a functioning platelet population, similar recommendations may be considered for those taking greater daily dosages. Although people with a higher baseline platelet count theoretically may be expected to recover from aspirin’s
Increased Bleeding after Ceasing Aspirin Therapy
571
effects differently than do people with a low platelet count, we found no correlation between recovery from aspirin and either platelet volume or number. But the range of these values in our volunteers was narrow. Although many patients who come off their aspirin before surgery will have been taking this drug for longer than the 2-week duration of our volunteers, the fact that aspirin’s antihemostatic effects are mediated by irreversible platelet function inhibition (so are limited by platelet life-cycle dynamics) makes the duration of pretreatment used in this study adequate to allow the full antihemostatic effects of aspirin therapy to develop.42 Indeed, previous studies have demonstrated that maximal inhibition of platelet response to aspirin occurs after much shorter treatment periods.43-45 Additionally, rather than more prolonged usage leading to a cumulative effect, recent evidence suggests that the efficacy of aspirin’s antiplatelet effects may actually diminish with time.46 Bleeding times and PFA-100 tests can only approximate platelet activity in a surgical wound, but they are better tests than aggregrometry, which relies on a separated sample. The standardized template bleeding time (TBT) attempts to measure the overall effectiveness of the hemostatic mechanism. Although it is known that bleeding times are prolonged for up to 4 days after single doses of aspirin,47 here we have demonstrated that this is also the time taken for a return to normal after a prolonged course of aspirin in the therapeutic range. The TBT has its limitations.48-50 It may, in fact, reflect aspirin’s effects on the vascular endothelium rather than the platelet reaction to injury.26 Because aspirin’s antiplatelet effect is considered to be the more significant contribution to the overall increase in bleeding tendency and may also be its most persistent effect, a platelet function test (a more specific and reproducible means of determining recovery from aspirin’s prohemorrhagic effects) was also included in this study. The PFA-100 assesses primary hemostasis in whole blood by attempting to recreate conditions found in injured blood vessels. Although now well established as a useful measure, its use as an instrument to monitor the effects of drugs such as aspirin is novel. Although platelet aggregation tests are useful in the study of specific aspects of platelet activity, they do not reflect overall platelet capabilities. In addition, unlike aggregometry, the PFA-100 is quick, easy to perform, and is reproducible.23,51 The delay in the return to baseline of this platelet function test compared with the TBTs in the majority
572
Cahill et al
Increased Bleeding after Ceasing Aspirin Therapy
of volunteers may correlate with the faster recovery of the vasculature compared with the platelets from the effects of aspirin. While the TBT plays no role as a routine single preoperative screening test, the PFA-100 may have a clinically useful application in detecting the presence of an increased bleeding tendency in patients admitted for elective surgery who may have inadvertently not discontinued their aspirin for the recommended period of time (by comparing the results obtained by using the collagen/epinephrine versus the collagen/ADP membranes). It may also allow consideration of agents to reverse aspirin effects (eg, desmopressin acetate,52 aprotinin,53 or platelet transfusion54) in patients who require emergency surgery. In summary, this study demonstrates that the duration of increased bleeding tendency after withdrawal of aspirin therapy persists for 5 days. This period of time is consistent with a close understanding of the mechanisms of aspirin’s action and of platelet kinetics. We believe that aspirin therapy should be discontinued 5 days before elective surgery and the operation planned for the sixth day after treatment cessation. This should allow physicians and surgeons to minimize the risks of stopping aspirin without increasing the risk of excessive perioperative bleeding. Author Contributions Study conception and design: RA Cahill, McGreal, MR Cahill, Redmond Acquisition of data: RA Cahill, Crowe, Ryan, Manning Analysis and interpretation of data: RA Cahill, Crowe, MR Cahill Drafting of manuscript: RA Cahill, McGreal, MR Cahill, Redmond Critical revision: RA Cahill, MR Cahill, Redmond Statistical expertise: RA Cahill, MR Cahill Obtaining funding: RA Cahill, Redmond Supervision: MR Cahill, Redmond REFERENCES 1. George JN, Shattil SJ. Acquired disorders of platelet function. In: Hoffman R, Benz EJ, Shattil SJ, et al, eds. Haematology. Basic principles and practice. PA: Churchhill Livingstone; 1991: 2172–2186. 2. WarkenhemTE, Kelten JG. Acquired platelet disorders. In: Bloom AL, Forbes CD, Thomas DP, Tuddenham EG, eds. Haemostasis and thrombosis. Singapore: Churchhill Livingstone;1994:765–815.
J Am Coll Surg
3. Bick RL, Adams T, Schmalhorst WR. Bleeding times, platelet adhesion and aspirin. Am J Clin Pathol 1976;65:69–72. 4. Anti-Platelet Trialists’ Collaboration. Collaborative overview of randomised trials of anti-platelet therapy. Reduction in venous thrombosis and pulmonary embolism by anti-platelet prophylaxis among surgical and medical patients. BMJ 1994;308:235– 246. 5. Ferraris VA, Ferraris SP, Joseph O, et al. Aspirin and postoperative bleeding after coronary bypass grafting. Ann Surg 2002; 235:820–827. 6. Kallis P, Tooze JA, Talbot S, et al. Pre-operative aspirin decreases platelet aggregation and increases post-operative blood loss—a prospective, randomised, placebo-controlled, double-blind clinical trial in 100 patients with chronic, stable angina. Eur J Cardiothorac Surg 1994;8:404–409. 7. Palmer JD, Sparrow OC, Iannotti F. Post-operative haematomata: a 5-year survey and identification of avoidable risk factors. Neurosurgery 1994;35:1061–1064. 8. Amrein PC, Ellman L, Harris WH. Aspirin-induced prolongation of the bleeding time and perioperative blood loss. JAMA 1981;245:1825–1828. 9. Kitchen L, Erichson RB, Sideropoulos H. Effect of druginduced platelet dysfunction on surgical bleeding. Am J Surg 1982;143:215–217. 10. Mak S, Amoroso P. Stop those antiplatelet agents before surgery! BJU International 2003;91:593–594. 11. George JN, Shattil SJ. The clinical importance of acquired abnormalities of platelet function. N Engl J Med 1991;324:27– 39. 12. Giangrande PLF, Littlewood TJ. Medical problems in surgical patients. Haematological problems. In: Morris PJ, Malt RA, eds. Oxford textbook of surgery. Oxford: Oxford University Press; 1994:212–221. 13. Harkess JW. Arthroplasty of the hip. In: Crenshaw AH, ed. Campbell’s operative orthopaedics. St Louis: CV Mosby; 1992: 441–626. 14. Rarick MU, Feinstein DI. Haematological problems and complications. In: Apuzzo MLJ, ed. Brain surgery: Complication avoidance and management. Edinburgh: Churchhill Livingstone; 1993:145–154. 15. Matsuzaki K, Matsui K, Haraguchi N, et al. Ischemic heart attacks following cessation of aspirin before coronary bypass surgery: A report of two cases. Ann Thorac Cardiovasc Surg 1999;5:121–122. 16. Kovich O, Otley CC. Thrombotic complications related to discontinuation of warfarin and aspirin perioperatively for cutaneous operations. J Am Acad Dermatol 2003;48:233–237. 17. Mitchell SM, Sethia KK. Hazards of aspirin withdrawal before transurethral prostatectomy BJU International 1999;84:550. 18. Fiore LD, Brophy MT, Lopez A, et al. The bleeding time in response to aspirin: identifying the hyper-responder. Am J Clin Path 1990;94:292–296. 19. Benzon HT, Brunner EA, Vaisrub N. Bleeding time and nerve blocks after aspirin. Reg Anesth 1983;9:86–89. 20. Trentalange MJ, Walts LF. A comparison of thromboelastogram and template bleeding time in the evaluation of platelet function after aspirin. J Clin Anaesth 1991;3:377–381. 21. Sonksen JR, Kong KL, Holder R. Magnitude and time course of impaired primary haemostasis after stopping chronic low and medium dose aspirin in healthy volunteers. Br J Anaesth 1999; 82:360–365. 22. Miltich JP. Bleeding time. In: Beutler E, Lichtmann MA, Coller
Vol. 200, No. 4, April 2005
23. 24. 25.
26.
27. 28. 29.
30. 31. 32. 33. 34. 35.
36. 37.
38.
Increased Bleeding after Ceasing Aspirin Therapy
Cahill et al
BS, Kipps TJ, eds. Williams Haematology. USA:McGraw-Hill; 1995:L111–112. Kundu SK, Heilmann EJ, Sio R, et al. Description of an in vitro platelet function analyzer-PFA-100. Semin Thromb Hemost 1995;21[Suppl 2]:106–112. Lanas AI, Arroyo MT, Esteva F, et al. Aspirin related gastrointestinal bleeders have an exaggerated bleeding time response due to aspirin use. Gut 1996;39:654–660. Anti-Platelet Trialists’ Collaboration. Collaborative overview of randomised trials of anti-platelet therapy-1: Prevention of death, myocardial infarct and stroke by prolonged anti-platelet therapy in various categories of patients. BMJ 1994;308:81– 106. Bick RL. Hereditary and acquired vascular bleeding disorders. In: Bick RL, Bennett MD, Brynes RK, et al, eds. Haematology— clinical and laboratory practice. St Louis: Mosby; 1993:1325– 1336. Paereti FI, Mannucci PM, D’Angelo A. Congenital deficiency of thromboxane and prostacyclin. Lancet 1980;1:898–901. Scher KS. Unplanned operation for rebleeding. Am Surg 1996; 62:52–55. Samama CM, Bastien O, Forestier F, et al. Antiplatelet agents in the perioperative period: expert recommendations of the French Society of Anesthesiology and Intensive care (SFAR) 2001— summary statement. Can J Anaesth 2002;49:S25–35. Topol EJ. Aspirin with bypass surgery—from taboo to new standard of care. N Engl J Med 2002;347:1359–1360. Nellipovitz DT, Bryson GL, Nichol G. The effect of perioperative aspirin therapy in peripheral vascular surgery. A decision analysis. Anesth Analg 2001;93:573–580. Devereaux PJ, Ghali WA, Gibson NE, et al. Physician’s recommendations for patients who undergo noncardiac surgery. Clin Invest Med 2000;23:116–123. Russell MW, Jobes D. What should we do with aspirin, NSAIDs and glycoprotein receptor inhibitors? Int Anesthiol Clin 2002; 40:63–76. Bullingham A, Strunin L. Prevention of postoperative venous thromboembolism. Br J Anaesth 1995;75:622–630. Nielsen JD, Holm-Nielsen A, Jespersen J, et al. The effect of low-dose acetylsalicyclic acid on bleeding after transurethral prostatectomy—a prospective, randomised double blind, placebo-controlled study. Scand J Urol Nephrol 2000;34:194– 198. Bick RL. Antithrombotic therapy. In: Bick RL, Bennett MD, Brynes RK, et al, eds. Haematology—clinical and laboratory practice. St Louis: Mosby; 1993:1603–1624. George JN, Dale GL. Platelet kinetics. In: Bloom AL, Forbes CD, Thomas DF, Tuddenham EG, eds. Haemostasis and thrombosis. Singapore: Churchhill Livingstone; 1994:1202– 1205. Van Pampus EC, Huijgens PC, Zevenbergen A, et al. Influence
39. 40. 41. 42. 43.
44.
45.
46. 47. 48. 49. 50.
51. 52. 53.
54.
573
of aspirin on human megakaryocyte prostaglandin synthesis. Eur J Haematol 1993;50:264–268. Winocour PD. Platelet turnover in advanced diabetes. Eur J Clin Invest 1994;24 [Suppl 1]:34–37. Packham MA. Role of platelets in thrombosis and haemostasis. Can J Physiol Pharmacol 1994;72:278–284. Counts RB, Haisch C, Simon TL, et al. Haemostasis in massively transfused patients. Ann Surg 1979;190:91–99. Jakubowski JA, Stampfer MJ, Vaillancourt R, Deykin D. Cumulative antiplatelet effect of low-dose enteric coated aspirin. Br J Haematol 1985;60:635–642. May JA, Heptinstall S, Cole AT, Hawkey CJ. Platelet responses to several agonists and combinations of agonists in whole blood: a placebo controlled comparison of the effects of a once daily dose of plain aspirin 300 mg, plain aspirin 75 mg and enteric coated aspirin 300 mg in man. Thromb Res 1997;88:183–192. Kuster LJ, Frolich JC. Platelet aggregation and thromboxane release induced by arachidonic acid, collagen, ADP and plateletactivating factor following low dose acetylsalicyclic acid in man. Prostaglandins 1986;32:415–423. Kallmann R, Nieuwenhuis HK, de Groot PG, et al. Effects of low doses of aspirin on bleeding time, thromboxane production and 6-keto-PGF1 alpha excretion in healthy subjects. Thromb Res 1987;45:355–361. Pulcinelli FM, Pignatelli P, Celestini A, et al. Inhibition of platelet aggregation by aspirin progressively decreases in long-term treated patients. J Am Coll Cardiol 2004;43:979–984. Mielke CHJ. Aspirin prolongation of the template bleeding time: influence of venostasis and direction of incision. Blood 1982;60:1139–1143. De Rossi SS, Glick M. Bleeding time: an unreliable predictor of clinical haemostasis. J Oral Maxillofac Surg 1996;54:1119– 1120. Channing Rodgers RP, Levin J. A critical reappraisal of the bleeding time. Semin Thromb Haemost 1990;16:1–20. DeCaterina R, Lanza M, Manca G, et al. Bleeding time and bleeding: An analysis of the relationship of the bleeding time test with parameters of surgical bleeding. Blood 1994;84:3363– 3370. Nicholson NS, Panzer-Knodle SG, Haas NF, et al. Assessment of platelet function assays. Am H J 1988;135:S170–178. Porte RJ, Leebeck FW. Pharmacological strategies to decrease transfusion requirements in patients undergoing surgery. Drugs 2002;62:2193–2211. Schonberger JP, Bredee JJ, van Oeursen W, et al. Preoperative therapy of low dose aspirin in internal artery bypass operations with or without low dose aprotinin. J Thorac Cardiovasc Surg 1993;106:262–267. Hirsch J, Fuster V, Salzman E. Dose anti-platelet agents: the side-effects, and anti-thrombotic effectiveness. Chest 1986;89 [Suppl 2]:4S–10S.