SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. SUBJECTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Objective

To investigate whether low-dose naproxen sodium (220 mg twice a day) interferes with aspirin's antiplatelet effect in healthy subjects.

Methods

We performed a crossover, open-label study in 9 healthy volunteers. They received for 6 days 3 different treatments separated by 14 days of washout: 1) naproxen 2 hours before aspirin, 2) aspirin 2 hours before naproxen, and 3) aspirin alone. The primary end point was the assessment of serum thromboxane B2 (TXB2) 24 hours after the administration of naproxen 2 hours before aspirin on day 6 of treatment. In 5 volunteers, the rate of recovery of TXB2 generation (up to 72 hours after drug discontinuation) was assessed in serum and in platelet-rich plasma stimulated with arachidonic acid (AA) or collagen.

Results

Twenty-four hours after the last dosing on day 6 in volunteers receiving aspirin alone or aspirin before naproxen, serum TXB2 was almost completely inhibited (median [range] 99.1% [97.4–99.4%] and 99.1% [98.0–99.7%], respectively). Naproxen given before aspirin caused a slightly lower inhibition of serum TXB2 (median [range] 98.0% [90.6–99.4%]) than aspirin alone (P = 0.0007) or aspirin before naproxen (P = 0.0045). All treatments produced a maximal inhibition of AA-induced platelet aggregation. At 24 hours, compared with baseline, collagen-induced platelet aggregation was still inhibited by aspirin alone (P = 0.0003), but not by aspirin given 2 hours before or after naproxen. Compared with administration of aspirin alone, the sequential administration of naproxen and aspirin caused a significant parallel upward shift of the regression lines describing the recovery of platelet TXB2.

Conclusion

Sequential administration of 220 mg naproxen twice a day and low-dose aspirin interferes with the irreversible inhibition of platelet cyclooxygenase 1 afforded by aspirin. The interaction was smaller when giving naproxen 2 hours after aspirin. The clinical consequences of these 2 schedules of administration of aspirin with naproxen remain to be studied in randomized clinical trials.

Arthritis in general and osteoarthritis in particular are increasingly becoming global problems; however, at this time, there is no known cure for osteoarthritis. Most forms of treatment therefore have dealt with the alleviation and management of chronic pain, which can affect normal functioning and quality of life of patients. Conventional medical treatment employs the use of nonsteroidal antiinflammatory drugs (NSAIDs) because they provide unmistakable and significant health benefits in the treatment of pain and inflammation (1, 2). However, traditional NSAIDs are also associated with increased risk of gastrointestinal (GI) and cardiovascular (CV) serious adverse events (i.e., upper GI bleeding and nonfatal myocardial infarction [MI]) (2–6). The safety issue of NSAIDs is considered a first priority because they are administered to elderly persons who are highly susceptible to their adverse effects, particularly due to GI and CV comorbidity. Both therapeutic and adverse effects of NSAIDs are due to the inhibition of prostanoids (7). Prostanoids play an important role in reducing the threshold to stimulation of the peripheral nociceptors, thereby increasing the excitability of spinal sensory neurons, and in stimulating thermosensitive neurons in the preoptic area of the brain (8). Further, prostanoids mediate a broad range of homeostatic functions in GI and CV systems (9, 10).

Prostanoids are produced from arachidonic acid (AA) after its release from membrane phospholipids by phospholipases (10, 11). AA is transformed by a 2-step reaction into prostaglandin H2 (PGH2) through the activity of 2 different PGH synthases, named cyclooxygenase 1 (COX-1) and COX-2, which have the same catalytic activities. In fact, COX-1 and COX-2 possess both COX and peroxidase active sites. First, AA is transformed to PGG2 by the COX active site, and then the peroxidase active site reduces a hydroperoxyl in PGG2 to a hydroxyl to form PGH2. Subsequently, PGH2 is metabolized by terminal synthases to the biologically active prostanoids (i.e., prostacyclin [PGI2], PGD2, PGF, PGE2, and thromboxane A2 [TXA2]) (10, 11). The therapeutic effects of NSAIDs (analgesic and antiinflammatory) are due to inhibition of COX- 2–dependent prostanoids (mainly PGE2) (12). Thus, selective COX-2 inhibitors (coxibs) were developed to reduce the GI toxicity of traditional NSAIDs due, at least in part, to the inhibition of cytoprotective prostanoids generated in the GI tract by COX-1 (2, 12). However, the use of coxibs unraveled the important role played by COX-2–dependent prostacyclin in the CV system (9).

Several lines of evidence suggest that the CV hazard is also associated with traditional NSAIDs through the same mechanism (4, 9). In fact, traditional NSAIDs, which are reversible inhibitors of COX-1 and COX-2, profoundly affect COX-2 in the presence of a reduction of platelet COX-1 activity (4) (i.e., <95%) that is insufficient for inhibition of platelet function (13). Thus, most traditional NSAIDs are selective for COX-2 at therapeutic doses with respect to platelet function (4). Naproxen is different among NSAIDs because it potently inhibits COX-1 and has a long half-life (1), thus affecting platelet COX-1 profoundly and persistently at therapeutic doses (14, 15). This has been proposed as one of the mechanisms by which naproxen could have a better CV safety profile than other traditional NSAIDs, such as diclofenac (4, 5). However, naproxen use in the general population is not cardioprotective for 2 major reasons. First, as a reversible inhibitor of COX-1, it is associated with marked variability in causing complete (>95%) and persistent suppression of the maximal capacity of platelets to generate TXA2 throughout the dosing interval (14, 15), which is a fundamental requisite for cardioprotection (16). Second, it can cause a coincident profound inhibition of the vasoprotective COX-2–dependent PGI2 (14, 15).

Due to the harmful or neutral effects of NSAIDs on the CV system, the coadministration of low-dose aspirin is recommended in patients with CV disease who need NSAID therapy (traditional or coxibs) to control arthritis symptoms (16). Low-dose aspirin is the only NSAID that has been shown to be cardioprotective due to its unique mechanism of action. Aspirin irreversibly acetylates Ser529 of COX-1 and COX-2, leading to irreversible enzyme inactivation (17). Because of its unusual pharmacokinetics, low doses of aspirin preferentially inhibit COX-1 in circulating platelets, thereby suppressing platelet TXA2 synthesis and attendant thrombosis. However, the coadministration of ibuprofen has been reported to interfere with the irreversible inhibition of platelet COX-1 by aspirin, leaving open the door for a potential impact on aspirin cardioprotection (18, 19).

The presence of a pharmacodynamic interaction between aspirin and naproxen on platelet COX-1 is more difficult to detect. In fact, naproxen binding to COX-1 may on the one hand prevent the irreversible acetylation of COX-1 by aspirin, while on the other hand it may have an extended direct inhibitory effect on COX-1 due to the long half-life of naproxen (∼17 hours) (1). In fact, it was shown that the extent of inhibition of both platelet COX-1 activity and AA-induced platelet aggregation up to 24 hours after dosing did not differ between long-term dosing with 500 mg naproxen twice a day administered 2 hours before or after 100 mg aspirin and dosing with aspirin alone (20). To detect potential interference of naproxen with irreversible inhibition of platelet COX-1 produced by aspirin, we studied the time-dependent recovery (up to 2 weeks) of serum TXB2 biosynthesis after coadministration of 1 single dose of naproxen (500 mg) and low-dose aspirin (100 mg) (20). A rapid recovery of COX-1 activity and function was found, which is compatible with a pharmacodynamic interaction between naproxen and aspirin (20). A limitation of that study was that the recovery of COX-1 activity after 1 single dose of aspirin alone was not evaluated.

Robust evidence is still lacking regarding the clinical consequences of the pharmacodynamic interaction between aspirin and NSAIDs. Population-based epidemiologic studies addressing this issue need to be large to address the potential effect modification with individual NSAIDs, and they are always open to some residual confounding bias (4, 21–28). Very large randomized clinical trials will be necessary, but they would be very expensive and pose major ethical dilemmas. Thus, it is urgent to develop strategies to predict and possibly minimize the drug–drug interaction.

We performed the present clinical study in healthy subjects coadministered low-dose aspirin and naproxen to address 3 questions. What is the appropriate assay to detect a pharmacodynamic interaction between aspirin and a coadministered NSAID? Can the interference of naproxen with the irreversible inhibition of platelet COX-1 by aspirin, previously detected with 500 mg naproxen twice a day (20), be reduced by lowering the daily dose of naproxen? Is the potential interaction dependent on the sequence of administration of the 2 drugs?

Naproxen is available over the counter only in a 220-mg dose. Thus, we studied the effects of its coadministration before or after aspirin versus aspirin alone on the degree of inhibition of serum TXB2 generation (a capacity index of platelet COX-1 activity in response to thrombin) and platelet aggregation induced by AA or collagen. In order to develop a suitable biochemical assay to detect the occurrence of the pharmacodynamic interaction between aspirin and naproxen, in a subgroup of individuals (after discontinuing the different treatment schedules) we compared the recovery kinetics of TXB2 generation in serum and in AA- or collagen-stimulated platelet-rich plasma by assessing the slope and y-intercept values of the least squares line using simple linear regression analysis.

SUBJECTS AND METHODS

  1. Top of page
  2. Abstract
  3. SUBJECTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Study subjects.

The study protocol was approved by the Ethics Committee of G. d'Annunzio University–Chieti. Informed consent was obtained from the 9 healthy subjects enrolled (5 men [56%]). The subjects were ages 23–37 years, within 30% of ideal body weight, and had unremarkable medical histories, physical examination findings, and routine hematologic and biochemical test results. Smokers and subjects with bleeding disorders, allergies to aspirin or any other NSAIDs, or a history of any GI, CV, or cerebrovascular disease were excluded. Subjects abstained from the use of aspirin and other NSAIDs for at least 2 weeks before enrollment.

Clinical study: design, treatments, and assessment.

We performed a crossover, open-label study to evaluate the effects of potential interactions between low-dose aspirin (100 mg/day of immediate release aspirin, not enteric coated, Aspirinetta; Bayer) and naproxen sodium (220 mg twice a day, Aleve; Roche), coadministered to 9 healthy subjects, on platelet thromboxane biosynthesis ex vivo and in vivo and on platelet aggregation induced by AA (1 mM and 2 mM) and by collagen (10 μg/ml) ex vivo. Subjects received for 5 days 3 different treatments separated by 14 days of washout, as follows: 1) naproxen sodium 220 mg twice a day (at 8:00 AM and 8:00 PM), with the first dose 2 hours before aspirin (100 mg/day at 10:00 AM); 2) 1 dose of aspirin at 8:00 AM 2 hours before the first dose of naproxen (taken at 10:00 AM and 10:00 PM); and 3) aspirin alone (at 8:00 AM). On day 6, they received only the morning doses (i.e., aspirin alone at 8:00 AM; naproxen at 8:00 AM [2 hours before aspirin]; or aspirin at 8:00 AM [2 hours before naproxen]). The dosing sequences were not randomized, but all subjects were receiving the same sequence.

The sequential administration with an interval of 2 hours between the 2 study drugs was chosen to allow aspirin or naproxen to produce their complete inhibitory effect on platelet COX-1 before the intake of the second drug. In fact, the time to peak plasma levels of naproxen is ∼2 hours (1, 15). Despite the fact that the major extent of platelet COX-1 acetylation by low-dose aspirin occurs in the presystemic circulation, further decline in platelet TXB2 generation may occur with the appearance of aspirin in blood; thus, at least 1 hour is required for plain aspirin to have a complete inhibitory effect on circulating platelet COX-1 (29, 30).

On days 1 and 6 of each sequence of treatment, the volunteers were admitted to the clinical center at SS. Annunziata Hospital for blood and urine sample collections and the administration of the first and the last study drugs. They were instructed in how to take their medications. Aspirin and naproxen were provided in 2 distinct boxes (Aspirinetta and Aleve original packages). On the sixth day of each sequence of treatment, they returned all boxes for compliance assessment by pill counts. If the returned boxes were not empty, suggesting nonadherence to the study protocol, the subjects were excluded from the study. In addition, compliance was monitored by contacting the subjects by telephone at each time of administration. After the washout period, new boxes were provided to subjects. Pill count adherence was 100%, and no enrolled subject was excluded from the study. Baseline blood samples were collected on day 1 of each treatment schedule; then, on day 6, in volunteers with sequential schedules, blood samples were collected 2, 5, 12, 24, and 48 hours after the administration of the first drug, while in volunteers taking aspirin alone, blood samples were collected 1, 24, and 48 hours after the last dose.

Blood samples were collected to assess inhibition of serum TXB2 (31) and of platelet aggregation induced by AA (1 mM and 2 mM) and collagen (10 μg/ml) in platelet-rich plasma (32). Platelet aggregation induced by AA and collagen was measured in platelet-rich plasma (32) using a Chrono-Log platelet aggregometer, whereas immunoreactive TXB2 was measured by a previously validated radioimmunoassay technique (31). For each pharmacologic schedule, urine samples were collected before treatment (overnight) and on day 6 (3 sequential collections after the first study drug was administered [i.e., from 0 to 6 hours, from 6 to 12 hours, and from 12 to 24 hours]) to assess the urinary excretion of 11-dehydro-TXB2 (TX-M), a major enzymatic metabolite of TXB2 that is an index of TXA2 biosynthesis in vivo, mainly of platelet origin (33, 34).

In a subgroup of 5 volunteers treated with naproxen sodium 2 hours before aspirin or 2 hours after aspirin, or treated with low-dose aspirin alone, blood samples were collected up to 72 hours after the administration of the first drug on day 6 of treatment to assess TXB2 levels in serum and in platelet-rich plasma in response to AA (1 mM and 2 mM) or collagen (10 μg/ml). This allowed us to calculate and compare the slope and y-intercept values of the least squares line obtained by simple linear regression analysis of time-dependent recovery of platelet TXB2 generation.

Statistical analysis.

Predrug values (measured before each treatment schedule) of serum TXB2, TX-M, and platelet aggregation (by AA or collagen) passed the normality test (by the Kolmogorov-Smirnov method); thus, data were expressed as the mean ± SD (Table 1), and statistical comparisons were made by repeated-measures analysis of variance followed by the Student-Newman-Keuls test using GraphPad Prism software. Due to heterogeneity in response to different treatment schedules on some occasions, in particular by the sequential administration of naproxen before aspirin, we analyzed the pharmacologic results using nonparametric tests. The data were expressed as the median (range) and geometric mean (95% confidence interval [95% CI]). The primary hypothesis was that 220 mg naproxen twice a day (with the first dose administered 2 hours before aspirin) would interfere with the irreversible inhibitory effect of aspirin, as assessed by the measurement of serum TXB2 (primary end point), platelet aggregation (secondary end point), and urinary excretion of TX-M (secondary end point) 24 hours after the last administration of naproxen (2 hours before aspirin) on day 6.

Table 1. Predrug measurements of serum TXB2, TX-M, and platelet aggregation in the 9 healthy volunteers*
 AspirinAspirin before naproxenNaproxen before aspirin
  • *

    Since predrug values (measured before each pharmacologic treatment) passed the normality test (by the Kolmogorov-Smirnov method), the data were expressed as the mean ± SD, and statistical comparisons were made by repeated-measures analysis of variance followed by the Student-Newman-Keuls test. TXB2 = thromboxane B2; TX-M = 11-dehydro-TXB2; AA = arachidonic acid.

  • Maximal platelet aggregation response (%).

Serum TXB2, ng/ml315 ± 99290 ± 100320 ± 130
TX-M, pg/mg of creatinine598 ± 124510 ± 150515 ± 280
Platelet aggregation   
 AA (1 mM)93 ± 792 ± 393 ± 3
 AA (2 mM)92 ± 592 ± 493 ± 3
 Collagen (10 μg/ml)95 ± 393 ± 294 ± 2

Assuming an intersubject coefficient of variation of 25% for serum TXB2 (19), 9 subjects would allow detection of a difference of 41% between the inhibitory effect of aspirin alone and the inhibitory effect of its coadministration with naproxen, with a power of 90%, on the basis of 2-tailed tests, with probability values less than the Type I error rate of 0.05. The log-transformed values of serum TXB2 and urinary TX-M concentration and those of percent of maximal aggregation were subject to nonparametric analysis (19, 35). Mixed-effects models with fixed effects of treatment or time plus random intercepts for each subject were estimated by restricted maximum likelihood in /R 2.9.0/ (http://cran.r-project.org/bin/windows/base/old/2.9.0/) by use of the function /lmer/ from package /lme4/version0.999375-31 (http://cran.uvigo.es/web/packages/Matrix/Matrix.pdf). Nonparametric estimates of the parameters and 2-tailed P values were derived by nonparametric bootstrap resampling of the residuals at each level of the model by use of 10,000 replicates (35).

The methods used were prespecified. We compared serum TXB2 levels (primary end point) and AA- and collagen-induced platelet aggregation (secondary end point) 24 hours (and other times) after dosing with predrug values for each treatment schedule. Moreover, we compared values among the 3 treatment schedules at the same time points after dosing (i.e., 24 hours and 48 hours). For TX-M, we compared the values detected at each urine collection after dosing with predrug values for each treatment schedule and among the 3 treatment schedules at the same time points.

In 5 volunteers, we assessed the relationship between the dependent variable (platelet TXB2 generation [ng/ml] in serum and platelet-rich plasma) and time after discontinuation of the different treatment schedules by linear regression analysis using Prism software. Since the variance of the outcome of TXB2 generation was not constant across the range of the predictor (time), we conducted linear regression (least squares) analysis with log10 transformation of the outcomes using Prism software. We calculated slope and y-intercept values (when x = 0), and their 95% CIs, of the least squares lines as well as the coefficient of determination (r2). Comparisons of slope and y-intercept values of linear regression lines among different treatment schedules were further assessed using Prism software. P values less than 0.05 were considered significant.

RESULTS

  1. Top of page
  2. Abstract
  3. SUBJECTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Predrug (baseline) values of serum TXB2, urinary TX-M, and maximal platelet aggregation response (percent) detected before the start of each treatment schedule did not differ significantly among the 3 different occasions (Table 1). These results show that the washout period (14 days) between treatments was long enough to bring the inhibition back to the pretreatment values.

The long-term administration of low-dose aspirin caused an almost complete inhibition of platelet COX-1 activity 1, 24, and 48 hours after dosing (at 1 hour, median 99.9% [range 99.5–100.0%], geometric mean 99.8% [95% CI 99.7–99.9%]; at 24 hours, median 99.1% [range 97.4–99.4%], geometric mean 98.9% [95% CI 98.4–99.4%]; at 48 hours, median 95.5% [range 90.9–98.5%], geometric mean 95.7% [95% CI 93.9–97.7%]) (P = 0.0001 versus baseline) (Figure 1). At the primary end point, 24 hours after the administration on day 6 of naproxen 2 hours before aspirin, serum TXB2 inhibition (median 98.0% [range 90.6–99.4%], geometric mean 95.9% [95% CI 93.4–98.8%]) was significantly lower than that detected after aspirin alone (P = 0.0007) or aspirin given 2 hours before naproxen (median 99.1% [range 98.0–99.7%], geometric mean 98.9% [95% CI 98.4–99.5%]) (P = 0.0045) (Figure 1). Forty-eight hours after the administration of naproxen 2 hours before aspirin, serum TXB2 inhibition (median 90.2% [range 76.8–98.2%], geometric mean 89.5% [95% CI 84.2–95.2%]) was significantly lower than that detected after aspirin alone (P = 0.0011), but not significantly lower than that detected after the administration of aspirin 2 hours before naproxen (median 93.9% [range 89.3–98.0%], geometric mean 94.0% [95% CI 92.1–96.0%]) assessed at the same time point (Figure 1). Compared with administration of aspirin alone, administration of aspirin 2 hours before naproxen did not cause a significant reduction of serum TXB2 both 24 hours and 48 hours after dosing (Figure 1).

thumbnail image

Figure 1. Comparison of the degree and duration of steady-state inhibition of cyclooxygenase 1 (COX-1) activity by administration for 6 days of naproxen sodium (220 mg twice a day) 2 hours before aspirin, naproxen sodium 2 hours after aspirin, or low-dose aspirin alone. Platelet COX-1 activity ex vivo (reported as the percent of inhibition [% I]), as assessed by the measurement of serum thromboxane B2 (TXB2), was evaluated in 9 healthy subjects. Data are presented as box plots, where the boxes represent the 25th to 75th percentiles, the lines within the boxes represent the median, and whiskers represent the highest and lowest values. Open symbols represent individual values. At each time point after dosing with the 3 different treatments, serum TXB2 was significantly reduced compared with predrug values (∗∗ = P = 0.0001). § = P = 0.0007 versus aspirin alone at 24 hours. † = P = 0.0045 versus aspirin before naproxen at 24 hours. # = P = 0.0011 versus aspirin alone at 48 hours. For this statistical analysis we used mixed-effects model procedures and nonparametric bootstrap resampling technique (35). WO = washout.

Download figure to PowerPoint

At each time point after dosing with 3 different treatments, platelet aggregation induced by AA (1 mM and 2 mM) was significantly reduced compared with predrug values for each pharmacologic schedule (P = 0.0001) (Figures 2A and B). Compared with administration of aspirin alone, the administration of naproxen (2 hours before or after aspirin) did not significantly affect 2 mM AA–induced platelet aggregation up to 24 hours after dosing (Figure 2A). Forty-eight hours after dosing, a statistically significant difference in platelet function response was only found between the administration of naproxen before aspirin and the administration of aspirin alone (Figure 2A). However, the administration of naproxen 2 hours after aspirin also perturbed the homogeneous suppression of 2 mM AA–induced platelet aggregation detected in all subjects receiving aspirin alone. In fact, in 2 of 9 subjects, a complete recovery of platelet aggregation was detected at 48 hours (Figure 2A). Similar results were found for 1 mM AA–induced platelet aggregation (Figure 2B).

thumbnail image

Figure 2. Inhibition of platelet function ex vivo by administration for 6 days of naproxen sodium (220 mg twice a day) 2 hours before aspirin, naproxen sodium 2 hours after aspirin, or low-dose aspirin alone. Platelet aggregation was assessed by measuring the percent of inhibition (% I). Data are presented as box plots, where the boxes represent the 25th to 75th percentiles, the lines within the boxes represent the median, and whiskers represent the highest and lowest values. Open symbols represent individual values. A, Platelet aggregation induced by 2 mM arachidonic acid (AA). At each time point after dosing with the 3 different treatments, platelet aggregation was significantly reduced compared with predrug values (∗∗ = P = 0.0001). # = P = 0.0053 versus aspirin alone at 48 hours. B, Platelet aggregation induced by 1 mM AA. At each time point after dosing with the 3 different treatments, platelet aggregation was significantly reduced compared with predrug values (∗∗ = P = 0.0001). § = P = 0.0001 versus aspirin before naproxen at 24 hours. # = P = 0.016 versus aspirin alone at 48 hours. † = P = 0.0003 versus aspirin alone at 24 hours. Φ = P = 0.04 versus aspirin alone at 48 hours. C, Platelet aggregation induced by 10 μg/ml collagen. ∗∗ = P = 0.0001 versus predrug values. Φ = P = 0.0005 versus aspirin alone at 24 hours. † = P = 0.0013 versus aspirin alone at 48 hours. # = P = 0.0045 versus predrug values. § = P = 0.01 versus predrug values. @ = P = 0.0005 versus aspirin alone at 48 hours. ∗ = P = 0.0003 versus predrug values. For this statistical analysis we used mixed-effects model procedures and nonparametric bootstrap resampling technique (35). WO = washout.

Download figure to PowerPoint

As shown in Figure 2C, low-dose aspirin alone caused a significant inhibition of collagen-induced platelet aggregation up to 48 hours after dosing, although heterogeneity of the response was detected. In contrast, collagen-induced platelet aggregation was rapidly recovered after the sequential administration of aspirin and naproxen (in both directions). Twenty-four and 48 hours after dosing with naproxen administered 2 hours before or after aspirin, collagen-induced platelet aggregation recovered to predrug values, and the degree of inhibition was significantly lower than that detected after aspirin alone at the same time points (Figure 2C).

The degree and duration of steady-state inhibition of the urinary excretion of TX-M, an index of TXA2 biosynthesis in vivo (34), was evaluated after the different treatments. As shown in Figure 3, aspirin alone profoundly reduced urinary TX-M levels. The sequential administration of naproxen and aspirin (in both directions) did not substantially affect the inhibition caused by aspirin alone. However, compared with administration of aspirin alone, we detected a slightly higher inhibition in urine samples collected 6–12 hours after dosing with aspirin given 2 hours before naproxen (P = 0.037) and 12–24 hours after dosing with naproxen given 2 hours before aspirin (P = 0.049).

thumbnail image

Figure 3. Comparison of the degree and duration of steady-state inhibition of thromboxane A2 (TXA2) biosynthesis in vivo by administration for 6 days of naproxen sodium (220 mg twice a day) 2 hours before aspirin, naproxen sodium 2 hours after aspirin, or low-dose aspirin alone. Urinary excretion of 11-dehydro-TXB2 (TX-M; an index of TXA2 biosynthesis in vivo), reported as the percent of inhibition (% I), was detected after dosing with the 3 different treatments. Data are presented as box plots, where the boxes represent the 25th to 75th percentiles, the lines within the boxes represent the median, and whiskers represent the highest and lowest values. Open symbols represent individual values. At each urine collection obtained after dosing with the 3 different treatments, urinary excretion of TX-M was significantly reduced compared with predrug values (∗∗ = P = 0.0001). § = P = 0.049 versus aspirin alone at 12–24 hours. # = P = 0.037 versus aspirin alone at 6–12 hours. For this statistical analysis we used mixed-effects model procedures and nonparametric bootstrap resampling technique (35). WO = washout.

Download figure to PowerPoint

To gain evidence that naproxen interfered with the irreversible inhibition of platelet COX-1 by aspirin, in a subgroup of 5 volunteers we compared the rate of biosynthesis of TXB2 in platelet-rich plasma stimulated with AA and collagen and in serum up to 72 hours after discontinuation of the different treatments by assessing the slope and y-intercept values of the least squares lines (using simple linear regression analysis) describing the relationship between TXB2 biosynthesis and time (Figure 4). After discontinuation of the different treatment schedules, the slopes of linear regression lines of time-dependent recovery of TXB2 biosynthesis were not significantly different. However, compared with administration of aspirin alone, the sequential administration of naproxen before or after aspirin caused a significant parallel upward shift of the regression lines (higher when naproxen was administered 2 hours before aspirin than in reverse order), as shown by increasing y-intercept values (Figure 4 and Table 2).

thumbnail image

Figure 4. Comparison of thromboxane B2 (TXB2) recovery in serum (A) or in platelet-rich plasma stimulated with 1 mM arachidonic acid (AA) (B), 2 mM AA (C), or 10 μg/ml collagen (D) up to 72 hours after discontinuation of treatment with naproxen sodium (220 mg twice a day) 2 hours before aspirin (▴), naproxen sodium 2 hours after aspirin (▪), or low-dose aspirin alone (●) in 5 healthy volunteers. The least squares lines of log10 transformation of TXB2 values were obtained using simple linear regression analysis with Prism software. All values presented were transformed back to the original scale.

Download figure to PowerPoint

Table 2. Recovery of thromboxane B2 biosynthesis in serum and platelet-rich plasma stimulated with AA or collagen up to 72 hours after discontinuation of long-term dosing with aspirin alone, aspirin 2 hours before naproxen, and naproxen 2 hours before aspirin*
 AspirinAspirin before naproxenNaproxen before aspirin
y-intercept95% CIr2Py-intercept95% CIr2Py-intercept95% CIr2P
  • *

    Y-intercept values and 95% confidence intervals (95% CIs) of the least squares lines and the coefficient of determination (r2) were calculated by linear regression analysis using Prism software. Data were log10-transformed before linear regression analysis. Shown are the antilog of the y-intercept (when x = 0) and its 95% CI. AA = arachidonic acid.

  • Describes the probability that randomly selected points would result in a regression line.

  • P = 0.013 versus aspirin before naproxen.

  • §

    P < 0.0001 versus naproxen before aspirin.

  • P = 0.028 versus aspirin before naproxen.

  • #

    P = 0.0007 versus naproxen before aspirin.

  • **

    P = 0.0025 versus aspirin before naproxen.

  • ††

    P = 0.002 versus naproxen before aspirin.

  • ‡‡

    P < 0.0001 versus aspirin before naproxen.

  • §§

    P = 0.0004 versus naproxen before aspirin.

Serum2.15§1.57–2.960.93<0.00013.02§2.35–3.900.89<0.00017.035.43–9.110.86<0.0001
Platelet-rich plasma            
 AA (1 mM)0.57§0.23–1.380.76<0.00010.73#0.41–1.310.82<0.00011.540.95–2.510.88<0.0001
 AA (2 mM)0.91§**0.57–1.460.95<0.00011.31††0.87–1.970.92<0.00013.441.78–6.630.80<0.0001
 Collagen (10 μg/ml)1.03§‡‡0.33–3.210.320.0093.73§§2.80–4.980.88<0.00017.824.63–13.20.69<0.0001

DISCUSSION

  1. Top of page
  2. Abstract
  3. SUBJECTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

NSAIDs are efficacious in reducing inflammation and pain. However, several studies have shown that their use is associated with a modest increase in CV risk (4, 6, 36). Among them, naproxen has been shown to have a better CV safety profile (5, 14, 36), presumably on the basis of its pharmacodynamic and pharmacokinetic features (1, 14, 15) (i.e., its potent and persistent inhibitory effect on platelet COX-1 activity [14, 15] due to its long half-life [1]). However, its use is associated with increased risk of GI bleeding and perforation (3); thus, it is recommended to use the lowest effective dose (3). We have recently shown that the administration of low-dose naproxen in the absence of aspirin is associated with marked variability in the inhibition of platelet COX-1 activity (15). Thus, like any other NSAID, it should be coadministered with low-dose aspirin in patients with arthritis and CV disease (16). However, the coadministration of naproxen might interfere with the irreversible inhibition of platelet COX-1 by aspirin. We performed the present study to shed some light on the determinants of a pharmacodynamic interaction between low-dose aspirin and low-dose naproxen sodium (220 mg twice a day) in healthy subjects. In particular, we aimed to characterize a biochemical model to detect the occurrence of this phenomenon.

Our results showed that the administration of low-dose naproxen 2 hours before aspirin disturbed the action of aspirin on platelet COX-1. However, 24 hours after the last dosing on day 6, serum TXB2 inhibition (the primary end point) was only slightly, even if significantly, lower than that found when naproxen was given 2 hours after aspirin or when aspirin alone was given (Figure 1). Compared with administration of aspirin alone, this effect did not translate into detectable alterations of AA-induced platelet aggregation (secondary end point) up to 24 hours after the last dosing. However, 48 hours after sequential administration of naproxen and aspirin, a heterogeneous response to AA-induced platelet aggregation was found in association with serum TXB2 inhibition <95%. These results strengthen the concept that serum TXB2 inhibition >95% is necessary to inhibit AA-induced platelet aggregation in all individuals (37, 38) (for further details, please contact the corresponding author). In contrast, when serum TXB2 inhibition is <95%, the subjects are divided into 2 groups, one with full inhibition and the other with no inhibition of AA-induced platelet aggregation (37, 38) (for further details, please contact the corresponding author).

Platelet functional consequences of the disturbance of COX-1 inhibition caused by the sequential administration of naproxen and aspirin (in both directions) were more evident when we studied platelet aggregation induced by collagen rather than by AA (secondary end points). This is explained by the finding that low concentrations of TXA2 synergize with collagen to induce full platelet aggregation (39).

In a subgroup of individuals, after discontinuing the different treatment schedules, we compared the recovery kinetics of TXB2 generation in serum and in AA- or collagen-stimulated platelet-rich plasma by assessing the slope and y-intercept values of the least squares line using simple linear regression analysis. In all these biochemical systems, compared with administration of aspirin alone, the administration of naproxen 2 hours before or after aspirin was associated with a significant parallel upward shift of the regression lines. This effect was more pronounced when naproxen was given 2 hours before aspirin. Taken together, these results may be consistent with a reduced degree of COX-1 acetylation by sequential dosing of naproxen and aspirin compared with dosing with aspirin alone. Since aspirin has a short half-life (∼20 minutes) (17, 40), the occupation of the COX-1 active site by naproxen may delay the binding of aspirin to COX-1, thus reducing its capacity to cause an irreversible inhibition of platelet TXB2 generation.

Since the interaction was less pronounced when aspirin was administered 2 hours before naproxen than in the reverse order, it would be reasonable to give naproxen at least 1 hour after plain aspirin. This allows aspirin to produce a complete inhibition of platelet COX-1, in presystemic and systemic circulation, before the administration of naproxen. For enteric-coated aspirin, which produces a delayed onset of COX-1 inhibition due to its slow releasing property (40), a longer period of time would be required before the intake of naproxen. Furthermore, it is likely that the interference of naproxen with the irreversible inhibition of platelet COX-1 by enteric-coated aspirin is more substantial. In fact, it has been shown that low doses of enteric-coated aspirin may be less bioavailable and thus prove insufficient in heavier patients with a large volume of distribution (38). Indeed, there are data showing that inhibition of platelet function with 75 mg enteric-coated aspirin daily may be incomplete in many patients with CV disease (i.e., particularly in heavier and younger patients and those with a history of MI) (38).

Further studies are needed to verify the possibility of varying the time-dependent sequence of administration based on, for example, the patient's age, other disease states, and genetic background that might affect the drug pharmacokinetic and pharmacodynamic features. Moreover, a time course would be needed in elderly patients at high risk of CV disease to accurately assess such time points after dosing with different aspirin preparations, such as plain and enteric coated.

One of the objectives of the present study was to develop a suitable biochemical assay that could be used in further studies of clinical pharmacology to detect the occurrence of the pharmacodynamic interaction between aspirin and naproxen. For this purpose we compared different biomarkers of COX-1 inhibition, such as TXB2 generation in serum and in AA- or collagen-stimulated platelet-rich plasma. We showed that the assessment of TXB2 generation in the presence of 10 μg/ml collagen was more appropriate than the other biomarkers to detect a pharmacodynamic interaction between aspirin and naproxen. Thus, the assessment of TXB2 generation in collagen-stimulated platelet-rich plasma seems more sensitive than its assessment in serum or AA-stimulated platelet-rich plasma for detecting small changes in COX-1 inhibition. COX-1 activity is regulated by the concentration of AA released close to intracellular enzyme localization and by the levels of cellular peroxides (41). In platelets, collagen, but not thrombin, has been shown to produce H2O2 (42), which is able to activate COX-1 (41). Whether this collagen-dependent signaling may play a role in our findings will be verified in further studies.

Our results suggest that the sequential administration of naproxen reduces the capacity of aspirin to cause an almost complete irreversible inactivation of COX-1. However, we showed that compared with administration of aspirin alone, the administration of aspirin 2 hours before naproxen did not significantly affect the inhibition of platelet COX-1 activity 24 hours after dosing. Similarly, AA-induced platelet aggregation was profoundly inhibited up to 24 hours after dosing with this sequential drug regimen. Also, collagen-induced platelet aggregation was significantly inhibited up to 12 hours, and the reduction of urinary excretion of TX-M by aspirin was not dampened. However, we detected a slightly higher inhibition in urine samples collected 6–12 hours after dosing. These data may suggest a small contribution of extraplatelet sources to TXA2 biosynthesis in vivo. Taken together, we suggest that the administration of low-dose aspirin 2 hours before low-dose naproxen should not translate into a significant interference with aspirin cardioprotection. Few observational studies have examined the impact of the potential interaction of concomitant use of low-dose aspirin and NSAIDs on the risk of CV events (4, 21–28). Two studies have evaluated the clinical interaction among users of naproxen and low-dose aspirin (4, 25), one showing no effect and the other reporting an estimate of risk compatible with a small (although not statistically significant) reduction in the cardioprotection afforded by low-dose aspirin.

In conclusion, sequential dosing of 220 mg naproxen twice a day and low-dose aspirin interfered with the irreversible inhibition of platelet COX-1 afforded by aspirin. The interaction was smaller when naproxen was given 2 hours after low-dose aspirin. The clinical consequences of the various schedules of the administration of naproxen with low-dose aspirin are not yet known and remain to be studied in a randomized clinical trial with preidentified CV end points.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. SUBJECTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Patrignani had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Anzellotti, Capone, Patrignani.

Acquisition of data. Anzellotti, Capone, Tacconelli, Bruno, Tontodonati, Di Francesco, Grossi, Renda, Di Gregorio.

Analysis and interpretation of data. Anzellotti, Capone, Jeyam, Tacconelli, Bruno, Merciaro, Price, Garcia Rodriguez, Patrignani.

Acknowledgements

  1. Top of page
  2. Abstract
  3. SUBJECTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

We thank the medical students of G. d'Annunzio University and the personnel of the Blood Transfusion Center of SS. Annunziata Hospital for their generous cooperation.

REFERENCES

  1. Top of page
  2. Abstract
  3. SUBJECTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES