Zisman, Eliyahu; Erport, Angelika; Kohanovsky, Eugene; Ballagulah, Marc; Cassel, Aliza; Quitt, Miriam; Pizov, Reuven

Introduction

Aspirin is one of the most widely used prophylactic and injury-limiting drugs in atherosclerotic arterial disease. Aspirin acts as an antiplatelet agent via its inhibition of the release of thromboxane, but it also inhibits a release of tissue prostacyclin and prostaglandin in a dose-dependent manner.^1–8 It has previously been shown that the lower the aspirin daily dose (up to 100 mg), the more platelet thromboxane synthesis relative to prostaglandin synthesis is decreased.^1,2 Concerns about perioperative aspirin-induced haemorrhagic complications have prompted recommendations to discontinue routine use of the drug 7–10 days before elective surgery.⁹ This policy is based on data from retrospective studies,^9,10 case–control studies,¹¹ prospective surgical studies,¹² and cardiac surgical studies.^13,14 There has been some recent evidence that platelet function recovers earlier.^14–16 Platelet aggregation with arachidonic acid in patients undergoing cardiac surgery was found to be significantly more impaired among those who ingested aspirin 2 days or less preoperatively compared with those who stopped aspirin 3–7 days before surgery.¹⁴ Bleeding time test results in volunteers revealed that most of the haemostatic defects had disappeared 48 h after aspirin cessation.¹⁵ Coleman and Alberts¹⁶ measured recovery of platelet function after cessation of aspirin administration in healthy volunteers using a point-of-care assay device. They found that platelet function fully recovered 5 days after the last dose of aspirin. Unfortunately, the point-of-care device was not compared with the standard test of platelet function in vitro, such as aggregometry.¹⁷ To the best of our knowledge, there has been no previous study on accurate day-by-day evaluation of platelet function recovery, either in volunteers or in patients.

Additionally, it became apparent in recent years that the duration of interruption of aspirin and other antiplatelet therapy before surgery could be life-threatening, particularly in patients after angioplasty with different intracoronary stents.¹⁸ The main conclusion of this and other studies was that the cessation of any antiplatelet therapy in high-risk patients should be as short as possible, if at all, to prevent postoperative complications.^18,19

Based on the accumulated knowledge that the effect of aspirin on platelets is shorter than 7 days, our goal was to clarify the process of platelet function recovery after termination of aspirin treatment in terms of precise day of recovery and to find out whether there is a difference in platelet function restoration between healthy volunteers and patients undergoing elective surgery. We used platelet aggregometry (PAG) with different agonists as the reference in-vitro test for the assessment of platelet function to prove our hypothesis.

Patients and methods

Patients

The study sample consisted of 14 healthy, untreated, male volunteers aged 20–60 years and 58 male and female (all postmenopausal) patients receiving regular treatment with 100 mg aspirin who were scheduled for elective orthopaedic or urological surgery at a major tertiary medical centre. Patients with a history of excessive bleeding or bruising, peptic disease, bronchial asthma or alcohol intake were excluded, as were patients in whom laboratory tests revealed an abnormality in haemoglobin level or blood platelet count, prothrombin time (PT), or partial thromboblastin time (PTT), and patients taking coagulation-modifying medications or food supplements such as garlic, Ginkgo biloba, ginger, cayenne, or bilberry.

The study was approved by the Institutional Review Board (protocol number 920020427 from 06/20/2002 and protocol number 920040317 from 05/19/2004) and written informed consent was obtained from all participants.

Protocol

Healthy volunteers

The volunteers received aspirin in a dose of 100 mg per day for 10 days and were followed for an additional 7 days after drug cessation. PAG was performed at onset of treatment (baseline), on day 10 of aspirin intake, and on days 1, 2, 3, 4 and 7 after cessation of aspirin treatment. Blood tests, including complete blood count (CBC), PT, PTT and fibrinogen were performed before onset of treatment, on day 10 of aspirin intake, and 7 days after aspirin cessation.

Patients

Patients were approached to participate in the study in the hospital's surgery clinics, at the time when their surgical procedure was scheduled. It is the policy of our orthopaedic and urological departments to stop routine aspirin therapy 10 days before elective surgery. For the present study, PAG was performed three times: during the preoperative visit, when the patients were still taking aspirin, and at one interim point: the third, fourth or sixth day after cessation of aspirin intake, and on the day of surgery, when they had been free of aspirin for 10 days. The interpoint days were chosen based on the data from volunteers. The interim day for PAG was allocated among the patients at random, and each patient was assigned his/her scheduled day in a sealed envelope at enrolment. Patients also underwent blood tests for CBC, PT, PTT and fibrinogen at the preoperative visit and on the day of operation.

Platelet aggregation test

Platelet-rich plasma (PRP) was prepared by centrifugation at 300g for 5 min at 20°C. Platelet-poor plasma (PPP) was prepared by centrifugation at 2000g for 15 min at 10°C. Platelet aggregation tests were performed with a platelet aggregometer (PACK-4; Helena Laboratories, Beaumont, Texas, USA). The platelet count of the PRP was adjusted to 150–200 × 10⁹ l⁻¹ by dilution of the PRP with PPP. The aggregometer was calibrated using normal PRP for 0% light transmission and PPP for 100% light transmission. The following aggregation agents were used: arachidonic acid at a concentration of 500 μg ml⁻¹; epinephrine (EPI) at a concentration of 3 μg ml⁻¹; and adenosine diphosphate (ADP) at a concentration of 4 μg ml⁻¹.^14,15 An aggregation of at least 65% was considered normal according to conventional standard normal values for the PAG test.^1,17

Urine thromboxane assessment

The urine thromboxane level was measured twice: once during aspirin intake and once at least 7 days after cessation of aspirin. Specifically, we measured the stable metabolite of thromboxane in urine, 11-dehydro-thromboxane B₂ (11dhTx B₂), using the AspirinWorks ELISA test kit (Corgenix Inc., Broomfield, Colorado, USA). The concentration of 11dhTx B₂ was expressed in picograms per millilitre.

Statistical analysis

Analysis of variance (ANOVA) with repeated measures was used to compare PAG values over time in the patient and volunteer groups. Student's t-test or Fisher's exact test was used to compare demographic and comorbidity variables of the two groups as appropriate. The χ² test was performed to compare comorbidity among patients tested by PAG on three different interim days. The relationship between changes (absolute and relative) in PAG values and urine 11dhTx B₂ levels was evaluated using Spearman's correlation coefficient. The values for both volunteers and patients at the same day after aspirin cessation were compared using the Mann–Whitney test, including urine 11dhTx B₂ concentration. Data are presented as mean ± SD. A P value less than 0.05 was considered significant.

Results

Of the 14 volunteers enrolled in the study, two were excluded from the final analysis because of findings of an abnormal coagulation profile at baseline before aspirin treatment was initiated. Four of the 58 patients were excluded from the analysis because of a failure to undergo the second PAG test or cancellation of surgery.

The characteristics and comorbidities of the healthy volunteers and surgical patients are shown in Table 1.

Table 1

In the 12 volunteers included in the study, findings for all coagulation parameters were within the normal range before and after the onset of treatment (Table 2). During aspirin treatment, platelet aggregation in response to arachidonic acid and EPI was significantly lower than normal. Aspirin had a slightly lesser effect on the response to EPI, though the difference from normal was still significant. Levels normalized on day 3 after discontinuation of the drug (Table 3, Fig. 1). Platelet aggregation in response to ADP did not change significantly throughout the study. There were no volunteers who developed aspirin-related side effects of abdominal discomfort or pain, bleeding or bruising, dyspnoea or pulmonary wheezing. The 11dhTx B₂ level in urine measured 4146 ± 1939 pg ml⁻¹ before aspirin intake and decreased to 1858 ± 1172 pg ml⁻¹ during treatment (P < 0.05; Table 4).

Table 2

Table 3

Fig. 1

Table 4

In the patients, the coagulation parameters showed a normal profile with lower haemoglobin concentrations than those in volunteers (Table 2). Platelet aggregation in response to arachidonic acid and EPI was significantly lower than normal during aspirin treatment, though the difference from normal was more prominent in response to arachidonic acid than to EPI (Table 3, Fig. 1). Platelet aggregation recovered gradually after the drug intake was stopped (Table 5, Fig. 1). Of the 17 patients tested on interim day 3 after aspirin cessation, six had significantly low levels of platelet aggregation. Of the 20 patients tested on interim day 4, findings were normal in all but one, who still showed significantly impaired aggregation (Fig. 2). Among the 17 patients tested on day 6, all had a normal platelet response. By day 10, values were normal in 52 patients of 54. Of those two patients who presented an abnormal pattern of platelet function recovery, one showed a slow rate of recovery of aggregation with arachidonic acid: from 34.5% on full aspirin to 25% 3 days after aspirin cessation and 52% on the day of surgery (no aspirin). This slow pattern of platelet recovery measured by aggregation was supported by a minimal change in the urine level of 11dhTx B₂ from 947 pg ml⁻¹ on full aspirin to 1124 pg ml⁻¹ at the day of surgery. In another patient, aggregation in response to arachidonic acid failed to progress, with values of 16.4% during aspirin treatment, 18.6% on day 4 after aspirin cessation, and 13.2% on day 10 after aspirin cessation. Her urine 11dhTx B₂ level presented the expected pattern, rising from 801 pg ml⁻¹ during aspirin treatment to 2121 pg ml⁻¹ at 10 days after treatment cessation.

Table 5

Fig. 2

The value of aggregometry with arachidonic acid was very similar for the volunteers and the patients during aspirin intake and without aspirin and significantly lower at day 4 after aspirin cessation in the patients (Fig. 1). The values at day 3 were not significantly different (P = 0.126) owing to the wide variability in both the volunteer and patient groups.

In response to ADP, platelet aggregation decreased somewhat, although it remained far above the lower limit of normal (Table 5).

In an attempt to isolate the effect of aspirin from other medications being taken by the patient group (Table 6), we stratified the PAG results by the specific comorbidities and by the different medications. There were no differences in platelet aggregation in response to arachidonic acid or EPI between patients with or without ischaemic heart disease, noninsulin-dependent diabetes mellitus, hypertension, hyperlipidaemia, or chronic renal failure. Furthermore, there was no noticeable interaction between any of the medications being taken at the time of the study (nitrates, β-blockers, calcium-channel blockers, statins, antiglycaemic drugs, diuretics) and PAG. The only exception was angiotensin converting enzyme (ACE) inhibitor: values for PAG with EPI were lower during aspirin therapy in patients who were taking ACE inhibitors than in patients who were not (44.4 ± 11.8 vs. 55.4 ± 15.3%, P = 0.005). This difference, however, disappeared at 10 days after aspirin cessation (81.2 ± 21.6 vs. 87.1 ± 10.8%, P = 0.7). No such difference was noted for PAG with arachidonic acid between patients taking ACE inhibitors or not, either during aspirin therapy (16.5 ± 8.5 vs. 15.6 ± 7.2%, P = 0.21) or after aspirin cessation (90.2 ± 9.3 vs. 89.7 ± 4.4%, P = 0.8).

Table 6

Additionally, comparison of the patients by interim day on which PAG was performed (3, 4 or 6 after aspirin cessation) revealed no differences in demographic data, preoperative haematological results, comorbidities or other medications used (Table 6).

The urine level of 11dhTx B₂ in the patients was 1669 ± 965 pg ml⁻¹ during aspirin treatment and rose to 5927 ± 5991 pg ml⁻¹ after cessation of aspirin (P < 0.05; Table 4). This finding was highly significant despite the fact that two patients had opposite results: a higher thromboxane level during treatment with a full dose of aspirin than without aspirin. Both patients showed the expected PAG pattern of low aggregation during aspirin therapy and normal aggregation without aspirin.

There was no correlation of the changes in urine 11dhTx B₂ level with platelet aggregation in either the volunteer or the patient group, and no correlation of thromboxane level with any of the patients' comorbidities or medications.

Of the total of seven patients who exhibited a slow return to normal levels of platelet aggregation, six were tested on interim day 3 and one was tested on interim day 4 (Fig. 2); all had similar 11dhTx B₂ levels to those in rest of the patients.

Discussion

The present study shows that, following the cessation of regular intake of aspirin, platelet function, measured in vitro, usually returns to normal within 3 days in healthy individuals and within 4 days in patients. Platelet recovery in some patients, however, can be delayed. These results, first, do not support the widespread practice of discontinuing regular aspirin intake 7–10 days before elective surgery^9–15 and, second, demonstrate the difference in platelet recovery between volunteers and patients.

We used a prospective, longitudinal, two-phase study design. Measurements were performed with aggregometry, the most commonly used method for the quantitative assessment of platelet function in vitro.¹⁴ Different agonists were used (arachidonic acid, EPI, ADP) in order to evaluate different pathways of platelet activation. Arachidonic acid is specific to aspirin because it serves as a substrate for platelet cyclooxygenase (COX). Aspirin inhibits COX, thereby preventing discharge of the potent pro-aggregant substance, thromboxane A₂, from platelets while, in low doses, sparing the synthesis of prostacyclin. By contrast, the platelet receptor for EPI is complex and affected by different substances by different mechanisms in different directions.²⁰ The difference between the receptors may explain the more significant effect of aspirin on platelet aggregation in response to arachidonic acid than to EPI in the present study (Table 3). Alberts et al.²¹ reported that, in patients treated with aspirin for stroke, platelet aggregation in response to EPI remained normal. In our study, aggregation with ADP was not significantly affected by aspirin, whereas Ferrell et al.²² noted a distortion of the second phase of aggregation when ADP was used as the agonist.

We assessed aspirin at a dose of 100 mg per day for several reasons: it was the dose most frequently used in our patients; it is not sufficient to cause gastrointestinal complications such as bleeding;²³ it is sufficient to neutralize more than 90% of thromboxane formation by platelets;²² its effect on platelet function is no different from the effect of aspirin at lower (75 mg) or higher (up to 300 mg) doses;¹⁵ this effect on platelet function is significantly more prominent than the effect of aspirin at a dose of 40 mg;²⁴ and low doses of aspirin, as mentioned, preferentially inhibit platelet thromboxane A₂ synthesis but not the production of antithrombotic prostacyclin by endothelial cells.^1,2,5

Sonksen et al.,¹⁵ in a study of volunteers, reported that, after cessation of aspirin treatment, bleeding time recovered already after 48–72 h. Moreover, in a later study, Coleman and Alberts¹⁶ observed that, in healthy volunteers, a significant amount of platelet function recovers already 3 days after aspirin cessation and activity returns to normal after 5 days using a point-of-care assay device but not standard PAG.

The recommendation to wait 7–10 days until platelet function recovers is based on the assumption that aspirin inhibition of platelet COX is irreversible and neutralizes platelets for their whole lifespan.⁵ However, in the present study, according to the aggregometry measurements with both agonists, platelet function recovered earlier: after 3 days in volunteers and after 4 days in most patients. To explain this finding, we suggest that the effect of aspirin on megakaryocytes in bone marrow is not permanent, in contrast to platelets that are permanently impaired by the action of aspirin and are later removed from the circulation. The megakaryocytes in affected bone marrow can regenerate COX within 12 h after cessation of aspirin, constantly regenerating about 35 × 10⁹ l⁻¹ new normal platelets that are added to the circulation daily.²⁵ In this manner, up to 15% of aspirin-affected platelets are replaced daily by intact platelets. If the platelets are not re-exposed to aspirin, the generation of new platelets leads to the gradual recuperation of overall platelet function.

The difference in recovery between the volunteers and patients might be explained by age (the volunteers were generally younger with more rapid metabolic and regeneration processes), general health status (the patients were sicker), and medications (taken only by the patients on a regular basis).

Various medications are known to affect platelet function. ACE inhibitors were found to lower levels of plasminogen activator inhibitor type 1 and decrease platelet aggregability²⁶ and to inhibit platelet activation in congestive heart failure.²⁷ Others noted that ACE inhibitors augment ecto-ATP diphosphohydrolase activity and enhance endothelial antiplatelet function in human umbilical vein endothelial cells,²⁸ thereby impairing platelet aggregation, which could possibly augment the effect of aspirin. The effects of statins on platelet function and their interaction with aspirin are still controversial.^29–32 According to our data, statin intake did not significantly alter the effect of aspirin on platelet function or the recovery of platelet function after aspirin cessation.

Comorbidities, too, may influence platelet function. Hypertension has a prothrombotic effect associated with a decrease in fibrinolytic activity, which is expressed as an increase in levels of plasma plasminogen activator inhibitor type 1 and greater platelet aggregability.²⁶ Also, in diabetic patients, high glucose levels enhance platelet activation and aggregation and are associated with poor platelet responsiveness to aspirin in vitro.^33–35

Faraday et al.³⁶ reported that the phenotypic variability in platelet response to aspirin is genetically determined, though the specific genes involved are not known. Apparently, the relationship between genotype, aspirin-response phenotype, and clinical consequences is very complicated. We suggest that, in addition to clinical factors such as age, comorbidities, and medications, the slower recovery of platelet aggregation in some of the patients in our study (Fig. 2), compared with the volunteers and to other patients, might be attributable to individual genotype.

Our study has several limitations. First, we recruited only male volunteers in order to avoid the influence of female cyclic hormonal changes on the coagulation system in general and on platelets in particular.^37–40 Second, we could not apply exactly the same protocol in the patients as we used in the volunteers, because we thought it unethical to draw blood from the patients on a daily basis. Third, we did not evaluate the clinical consequences of aspirin cessation, such as bleeding or cardiovascular events. We are aware, however, that patients treated preoperatively with aspirin have an increased tendency to bleeding after cardiac surgery and a lower incidence of postoperative myocardial infarction.^41,42 Fourth, our sample size is limited and could not provide sufficient data for constructing a demographic and comorbidity model of platelet function recovery after aspirin.

Despite these drawbacks, we think that this study has proven that the aspirin effect on the platelet function vanished in the majority of patients in 4 days and in all patients 6 days after cessation of its intake. However, a minority of the patients showed a slower process of platelet recovery from aspirin influence up to the sixth day of aspirin cessation. This knowledge, on one hand, could be important for patients undergoing surgical procedures when serious perioperative bleeding could put a patient at risk. On the other hand, the precise timing of aspirin cessation is important because of the possible detrimental effect of interruption of aspirin therapy that can lead to ischaemic cerebral events.^18,42 Mangano¹⁹ showed that early postoperative reinstitution of aspirin decreases cardiovascular complications. The duration of interruption of aspirin and other antiplatelet therapy before surgery could be life-threatening or life-saving, particularly in patients after angioplasty with different intracoronary stents.¹⁸ The main conclusion of this review¹⁸ was that the cessation of any antiplatelet therapy in patients with coronary stents should be as short as possible, if at all, to prevent stent thrombosis.

In conclusion, the present study shows that, in the majority of the patients treated with aspirin, platelet function recovers in 4 days after aspirin cessation compared with 3 days in healthy volunteers. Despite the fact that platelet function recovers after 4 days in the majority of patients, some patients might nevertheless have delayed recovery. We, therefore, suggest that the duration of aspirin cessation before elective surgical procedures with a high risk of potential bleeding should be 6 days, and that platelets should be monitored by aggregometry if this is not feasible.

Acknowledgements

We acknowledge Dr Ada Tamir for the statistical analysis; Drs Miriam David and Alisa Kessel, Lady Davis Carmel Medical Center, for their laboratory assistance; Dr Simon Gelman, Chairman Emeritus, Harvard University, Boston, Massachusetts, for the creative assistance during study design and preparation of the article.

The present study was partly supported by a grant from the Dankner Family Foundation, 2006.

References