Anesthesiology 2005; 103:712–7 © 2005 American Society of Anesthesiologists, Inc. Lippincott Williams & Wilkins, Inc. Dose-dependent Inhibition of Platelet Function by Acetaminophen in Healthy Volunteers Edward Munsterhjelm, M.D.,* Nina M. Munsterhjelm, B.Pharm.Sc.,† Tomi T. Niemi, M.D., Ph.D.,‡ Olavi Ylikorkala, M.D., Ph.D.,§ Pertti J. Neuvonen, M.D., Ph.D.,࿣ Per H. Rosenberg, M.D., Ph.D.# Background: Acetaminophen (paracetamol) is widely used Normal platelet function is dependent on the production for postoperative analgesia. Its mechanism of action is inhibi- of proaggregatory thromboxane A2 (TxA2) through tion of prostaglandin synthesis in the central nervous system, COX-1, and acetaminophen has been shown to inhibit and acetaminophen is traditionally not considered to influence 5 platelet function. The authors studied the dose-dependent inhi- platelet function both in vitro and in high intravenous 6 bition of platelet function by acetaminophen in healthy volun- doses in vivo. However, oral administration of conven- Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/103/4/712/360846/0000542-200510000-00009.pdf by guest on 29 September 2021 teers. tional doses (approximately 1 g) of acetaminophen does Methods: Thirteen healthy male volunteers (aged 19–26 yr) not alter platelet function.7,8 were given placebo or 15, 22.5, or 30 mg/kg acetaminophen Acetaminophen is widely used for postoperative anal- intravenously in a double-blind, crossover study. Ten and 90 9 10,11 min after infusion, platelet function was assessed by photomet- gesia, although the optimal dose is debatable. In ric aggregometry and by measuring release of thromboxane B2, pediatric patients, no analgesic ceiling effect was de- analgesia by cold pressor test, and plasma acetaminophen con- tected when acetaminophen was administered rectally centrations by high-performance liquid chromatography. in doses up to 60 mg/kg.12 However, high doses of ␮ Results: When triggered with 500 M arachidonic acid, me- acetaminophen may alter platelet function through pe- dian platelet aggregation (area under the curve) was 25.7, 22.8, -or 3.6 ؋ 103 area units (P < 0.001) 10 min after placebo or ripheral COX-1 inhibition. Because proper platelet func ,4.1 15, 22.5, or 30 mg/kg acetaminophen, respectively. An increas- tion is essential for adequate intraoperative and postop- ing concentration of arachidonic acid attenuated the antiaggre- erative hemostasis, we studied the dose-dependent gatory effect. After 90 min, platelet function was recovering. effect of acetaminophen on platelet function in healthy Release of thromboxane B was also dose-dependently inhib- 2 volunteers. We also measured the analgesic effect of ited by acetaminophen. Although plasma concentration of acet- aminophen increased linearly with the dose, no analgesic effect acetaminophen with the cold pressor test as a painful was detected in the cold pressor test. stimulus. Conclusions: Acetaminophen, which is a weak inhibitor of platelet cyclooxygenase 1, has a dose-dependent antiaggrega- tory effect. This property may become clinically significant Materials and Methods in patients with intrinsic or drug-induced impairment of hemostasis. The protocol was approved by the Ethics Committee for Studies in Healthy Subjects and Primary Care in the CYCLOOXYGENASE (COX), the key enzyme in prosta- Hospital District of Helsinki and Uusimaa (Helsinki, Fin- glandin formation, is an important pharmacologic tar- land) and by the National Agency for Medicines in Fin- get.1 The antithrombotic effect of acetylsalicylic acid is land. Fifteen healthy, nonsmoking men aged between 19 caused by irreversible inhibition of COX-1, constitutively and 26 yr volunteered in this double-blinded, random- expressed in platelets, whereas the analgesic effect of ized, placebo-controlled, crossover study. Written in- nonsteroidal antiinflammatory drugs (NSAIDs) is medi- formed consent was obtained from each subject before ated through inhibition of COX-2, induced during inflam- the study. Normal plasma alanine transaminase and as- mation. The main mechanism of action of acetamino- partate aminotransferase activities were a prerequisite phen is inhibition of prostaglandin synthesis in the for participation. Two volunteers withdrew their con- central nervous system,2 the recently characterized sent before completing the study. The use of acetylsali- COX-3 being a possible target.3 However, acetamino- cylic acid was forbidden for 10 days and that of other phen has also peripheral COX-1 inhibiting properties.4 drugs affecting platelet function was forbidden for 1 week before each experiment. * Resident, Department of Anesthesiology and Intensive Care Medicine, Hel- Experimental Procedures sinki University Hospital, Helsinki, Finland. † Researcher, ‡ Assistant Professor, Every volunteer participated in four experiments with # Professor, Department of Anesthesiology and Intensive Care Medicine, § Pro- fessor, Department of Obstetrics and Gynecology, ࿣ Professor, Department of at least a 1-week interval. After3hoffasting, 15, 22.5, or Clinical Pharmacology, University of Helsinki, Helsinki, Finland. 30 mg/kg acetaminophen (Perfalgan®; Bristol-Myers Received from the Department of Anesthesiology and Intensive Care Medi- Squibb, New York, NY) or placebo (0.9% NaCl; Braun, cine, University of Helsinki, Helsinki, Finland. Submitted for publication February 15, 2005. Accepted for publication June 17, 2005. Supported by departmental Kronberg, Germany) was given as a 10-min intravenous funding and a grant from the Medical Society of Finland, Helsinki, Finland. infusion through a 20-gauge cannula (Venflon; Becton Presented in part as a poster at the XXth Congress of the International Society on Thrombosis Hemostasis, Sydney, Australia, August 6–12, 2005. Dickinson, Franklin Lakes, NJ) in a dorsal vein of the Address reprint requests to Dr. Munsterhjelm: P.O. Box 340 (P-floor), FIN- hand. The infusions were blinded and administered in 00029 HUS, Finland. Address electronic mail to: edward.munsterhjelm@hus.fi. Individual article reprints may be purchased through the Journal Web site, random order. The code was not broken until all exper- www.anesthesiology.org. iments had been performed. Anesthesiology, V 103, No 4, Oct 2005 712 ACETAMINOPHEN AND PLATELETS IN HEALTHY VOLUNTEERS 713 Table 1. Plasma Concentration of Acetaminophen (mg/l) Dose of Acetaminophen 0 mg/kg 15 mg/kg 22.5 mg/kg 30 mg/kg Time after drug administration 10 min Ͻ 0.1 18.8 (17.0–20.6) 30.2 (27.2–33.3) 39.1 (34.4–43.8) 90 min Ͻ 0.1 10.5 (9.5–11.5) 16.3 (14.6–17.9) 22.2 (20.3–24.2) Data are reported as mean values (95% confidence intervals). Each volunteer (n ϭ 13) was given placebo and 15, 22.5, and 30 mg/kg acetaminophen. Venous blood samples (approximately 35 ml/sample) stored at Ϫ20°C. Acetaminophen concentration was de- were drawn from an antecubital vein before infusion and termined using high-performance liquid chromatogra- Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/103/4/712/360846/0000542-200510000-00009.pdf by guest on 29 September 2021 10 and 90 min after cessation of infusion. The first phy.16 The limit of quantification was 0.1 mg/l, and the sample after infusion was drawn from the contralateral day-to-day coefficients of variation were 7.4% at 16.8 mg/l hand. A 20-gauge needle (PrecisionGlide; Becton Dick- and 4.2% at 36.8 mg/l (n ϭ 6). inson) was used, and samples for aggregation studies were collected into polypropylene tubes (Vacuette; Statistical Analysis Greiner bio-one, Kremsmu¨nster, Austria) containing The sample size needed was estimated in advance as 3.2% buffered citrate, giving a volume ratio of 1:10. described in the statistical literature.17 The study was Samples for acetaminophen concentration measurement designed to discover a difference in platelet aggregation were collected into plastic EDTA tubes (5.9 mg K2EDTA, between each acetaminophen group and placebo VenoSafe; Terumo Europe, Haasrode, Belgium). greater than 1 SD, with a power of 80% (␣ error ϭ 5%, Cold Pressor Test. Immediately after each blood sam- Bonferroni correction was applied). The sample size pling, a cold pressor test was performed. The volunteer needed was n ϭ 11. A difference smaller than 1 SD was immersed his arm, halfway to the elbow, into an ice considered of minor clinical significance. Data distribution bath, and the times elapsed until first sensation of pain was tested with the Kolmogorov-Smirnov test, and non- and sensation of strong pain were recorded. There was parametric statistics were used for nonnormally distributed at least a 50-min interval between the cold pressor test data. Nonnormally distributed data are presented as medi- and the next blood sampling. ans and 25th/75th percentiles; normally distributed data are presented as mean values and 95% confidence intervals. Laboratory Tests The difference between all groups was analyzed with the Platelet Aggregation. Platelet aggregation was mea- Friedman test (repeated-measures analysis of variance on sured with a four-channel photometric aggregometer ranks), and when a significant difference was encountered, (Packs-4; Helena Laboratories, Beaumont, TX) based on 13 each acetaminophen group was further compared with the method of Born. Platelet-rich plasma and platelet- placebo using the Wilcoxon matched pairs signed rank poor plasma were prepared by centrifuging as described sum test and applying the Bonferroni correction. Statistical 14 ␮ previously. Aggregation was induced in 270 l platelet- testing was with SigmaStat for Windows Version 2.03 (SPSS ␮ rich plasma by adding 30 l of one of the following Inc., Chicago, IL). Confidence intervals were calculated triggers: adenosine diphosphate (ADP) to a final concen- using the appropriate t distribution. tration of 1.5 or 3 ␮M; arachidonic acid to a final con- centration of 500, 750, or 1,000 ␮M; or epinephrine to a final concentration of 5 ␮M. Reagents were purchased Results from Sigma-Aldrich (St. Louis, MO) and Calbiochem (San Diego, CA). Based on previous experience, these con- Thirteen volunteers completed the study according to centrations are known to cause platelet aggregation.