Clinical Outcomes of Aspirin Interaction with Other Non-Steroidal AntiInflammatory Drugs: A Systematic Review

- Purpose: Concomitant use of some non-Aspirin nonsteroidal anti-inflammatory drugs (NANSAIDs) reduces the extent of platelet aggregation of Aspirin (acetylsalicylic acid). This is while many observational studies and clinical trials suggest that Aspirin reduces cardiovascular (CV) risk attributed to the use of NANSAIDs. Thus, the therapeutic outcome of the interaction needs to be assessed. Methods: We searched various databases up to October 2017 for molecular interaction studies between the drugs and long-term clinical outcomes based on randomized clinical trials and epidemiological observations that reported the effect estimates of CV risks (OR, RR or HR; 95% CI) of the interacting drugs alone or in combinations. Comparisons were made between outcomes after Aspirin alone, NANSAIDs alone and Aspirin with naproxen, ibuprofen, celecoxib, meloxicam, diclofenac or rofecoxib. Results: In total, 32 eligible studies (20 molecular interactions studies and 12 observational trials) were found. Conflicting in vitro/in vivo/ex vivo platelet aggregation data were found for ibuprofen, naproxen and celecoxib. Nevertheless, for naproxen, the interaction at the aggregation level did not amount to a loss of cardioprotective effects of Aspirin. Similarly, for ibuprofen, the results overwhelmingly suggest no negative clinical CV outcomes following the combination therapy. Meloxicam and rofecoxib neither interacted with Aspirin at the level of platelet aggregation nor altered clinical outcomes. The clinical outcomes data for celecoxib and diclofenac are in conflict. Conclusion: Aspirin appears to maintain its cardioprotective effect in the presence of naproxen, ibuprofen, meloxicam and rofecoxib. The limited available data suggest that the effect of interaction at the platelet aggregation level may dissipate shortly, or the reduced platelet aggregation yielded by the interaction may be sufficient for cardioprotection; i.e., no need for near complete aggregation. In addition, cardioprotective effect of Aspirin, despite reduced platelet aggregation caused by NANSAIDs, may be through its involvement in other mechanisms such as the renin-angiotensin system and/or metabolism of arachidonic acid to biologically active compounds mediated by cytochrome P450. platelet inhibit, antiplatelet effect, and or tolmetin diclofenac or ketorolac or nabumetone indomethacin sulindac or piroxicam or meloxicam or mefenamic acid meclofenamic acid rofecoxib or celecoxib veldecoxib or paracoxib or etoricoxib or lumaricoxib or cyclooxygenase* or cyclo-oxygenase* or COX* and infarction or cerebrovascular or cardioprotect* or cardio-protect* or platelet* or platelet aggregation or platelet aggregation inhibit* or antiplatelet effect* or blood platelets and Interaction or Drug interaction or Interact* Structured Provide a structured summary including, as applicable: background; objectives; data sources; study eligibility criteria, participants, and interventions; study appraisal and synthesis methods; results; limitations; conclusions and implications of key findings; systematic review registration number.


INTRODUCTION
Acetylsalicylic acid (Aspirin) is in clinical use since mid 19 th century. In addition to being an effective analgesic, antipyretic and anti-inflammatory agent, it is used, among other indications, for its anti-platelet property to reduce all-cause mortality, cardiac death, and nonfatal myocardial infarction (MI) (1). Moreover, low-dose Aspirin, alone or in combination, is recommended for the secondary prevention of acute ischemic stroke and transient ischemic attack (2)(3)(4). In general, the anti-platelet effect of Aspirin accounted for the irreversible inhibition of platelet cyclooxygenase-1 (COX-1) enzyme. COX-1 is an enzyme that catalyzes AA to produce several prostaglandins (PG), among them thromboxane A2 (TxA2), a promoter of platelet aggregation (5,6). The inhibition of the COX-1 dependent TxA2 by Aspirin, measured by plasma thromboxane B2 (TxB2) is recommended to be near completion to significantly inhibit platelet function in vivo (7-9).
The non-Aspirin nonsteroidal anti-inflammatory drugs (NANSAIDs) are among the most commonly used medications for a variety of indications ranging from headaches to arthritis. NANSAIDs bind and inhibit the COX enzymes which lead to inhibition of prostanoids biosynthesis including PGs, prostacyclins and thromboxanes (10). Thus, the concomitant use of some NANSAIDs appear to interact with the Aspirin's anti-platelet function, thereby, although unproven, _________________________________________ Study selection and data extraction Both authors examined the titles and abstracts of studies to identify studies that potentially meet the inclusion criteria. The inclusion criteria were as follows: (i) Randomized controlled trials (RCTs) or observational studies (cohort or case-control studies) that include treatment with Aspirin alone or NANSAID alone as well as concomitant use of NANSAIDs with Aspirin. The association between the treatments and risk of CV (MI), cerebrovascular events (stroke) or all-cause mortality were assessed for studies that included odds ratios (ORs), relative risks (RRs) and/or hazard ratios (HRs) with 95% confidence interval (CI). (ii) molecular interactions trials (in vitro, in vivo or ex vivo) in human addressing the interaction at the platelet level between NANSAIDs and Aspirin.
The full texts of these potentially eligible studies were retrieved and independently assessed for eligibility. Disagreements were settled through discussion and consensus. Extracted information included study information (i.e. authors, location, publication date, type of study, number of participants, and study duration), patient characteristics (i.e. age, sex, previous CV events including stroke, Aspirin use, and NANSAIDs use), intervention and comparator (i.e. drugs and doses) and outcomes (i.e. events/total for all study population or subgroups).
The identified studies were excluded if: (i) they were reviews, questionnaire, thesis, letters, simulated studies, meeting summary, conference abstracts, editorial or commentary articles; (ii) had no eligible outcomes or did not report direct comparisons of individual NANSAIDs; (iii) used extra-oral route of administration (e.g., topical use for analgesia) or used other drugs with NANSAIDs or Aspirin.

Quality assessment
The methodological quality of the included observational studies (cohort and case-control studies) was appraised using scales adopted from the Newcastle-Ottawa quality scale (NOS) (15). Based on the study design (cohort or case-control study), each study was evaluated using the appropriate scoring system. Eight items in the included cohort and case-control studies were identified and assessed. Cohort and case-control studies with 6-9, 3-5, and 0-2 points were classified as high, fair or poor quality, respectively.

Eligible studies
Our search strategy yielded 3,563 potentially relevant articles from which 3,498 were found ineligible because they were not epidemiological studies or molecular interactions experiments. Sixtyfive articles underwent full-length article review. Twenty-five of these were excluded because they did not report the outcome of interest (MI or stroke), 5 were excluded because they did not report direct comparison of individual NANSAIDs with or without use of Aspirin, and 3 were excluded because of combination other than Aspirin with NANSAIDs or use of formulation other than oral. Twelve studies (5 cohort studies and 7 case-control studies) with 80,845 events met our eligibility criteria and were included in the analysis (12,13,(16)(17)(18)(19)(20)(21)(22)(23)(24)(25). The eligible studies scored good quality based on the calculated NOS scores (cohorts, 8-9/9 and case-controls 6-8/9) ( Table 9).
Twenty molecular interactions studies addressing the interactions between NANSAIDs and Aspirin were included. The detailed flow chart of search methodology and selection process is shown in Figure 1. Table 1 compares the outcomes of both platelet effects and clinical outcomes. Data on the selected NANSAIDs are provided in Tables 2-7. The detailed characteristics of molecular interactions experiments studies are described in Table 8. The clinical data on the interactions between Aspirin and different type of NANSAIDs are summarized in Table 9.

Platelet aggregation
The 20 eligible molecular interaction studies with the information on the interactions indicated that, in general, the anti-platelet effect of Aspirin is reduced in the presence of ibuprofen, naproxen or celecoxib (Tables 1 and 8). However, meloxicam, rofecoxib and diclofenac do not interfere with the anti-platelet effect of Aspirin.

Meloxicam
No (26) No (43) No (12, 16) NA, not available  A trend towards an increase in the rate of recurrent MI has been reported in one cohort study when subjects exposed to Aspirin and ibuprofen (HR, 1.50; CI 1.33-1.70) compared with Aspirin alone users (HR, 0.98; CI 0.94-1.03) (17). A retrospective cohort study has also concluded that patients with history of CV diseases had increased risk of mortality when exposed to Aspirin plus ibuprofen compared with users of Aspirin alone (24).

DISCUSSION
This is, to the best of our knowledge, the first systematic review that compares published Aspirin-NANSAIDs interaction at the platelet level with its long-term clinical outcomes. We have used broad inclusion criteria in many databases to capture molecular interactions experiments, RCTs and observational studies for a range of NANSAIDs and Aspirin users. However, no RCTs data were found.
We found that a NANSAID-Aspirin interaction at the platelet level does not necessarily amount to a loss of beneficial effects of Aspirin. Indeed, for naproxen, studies have consistently reported no negative clinical outcomes after addition of the drug to the Aspirin regimens (Table 3). Similarly, studies overwhelmingly suggest that Aspirin maintains it beneficial effects after addition of ibuprofen to the regimen. (Table 2).
As expected, the cardioprotective effect of Aspirin is not diminished by meloxicam and rofecoxib, two NANSAIDs that do not interact with Aspirin at the platelet level (Table 1). Interestingly, diclofenac for which its lack of effect on the antiplatelet action of Aspirin has been repeatedly reported appears to diminish the clinical benefit of the latter as reported by 2 of eligible 4 studies (Table  1).
Despite the limited number of eligible studies, meloxicam (12, 16) ( Table 4) and rofecoxib (12, 13, 17) ( Table 5) do not appear to diminish the cardioprotective effect of Aspirin. This is not unexpected since these drugs do not interact with the anti-platelet properties of Aspirin ( Table 1).
The data for celecoxib are not as conclusive as those available for naproxen and even ibuprofen since we found only 3 eligible studies. Two studies that suggest no loss of the beneficial effect of Aspirin (12, 13) contradict the other one (17). The reason for the conflicting results is unclear but it may be of relevance to mention that the latter study (17) stands out as the one that has also observed diminishing clinical benefit of Aspirin for ibuprofen, diclofenac as well. Nevertheless, in light of the conflicting data and the limited eligible studies, one cannot draw an unequivocal conclusion as to the clinical outcome of celecoxib-Aspirin interaction. Similarly, one cannot draw a definite conclusion regarding diclofenac as we found only 4 eligible studies, two in each side of the controversy. This is interesting since diclofenac does not interact with Aspirin at the platelet level (Table 1), thus, the loss of cardioprotective effect caused by the drug-drug interaction is unexpected.  The observation that not all NANSAIDs interact with Aspirin at the clinical level despite the fact that with the exception of meloxicam, rofecoxib and diclofenac, they interact with Aspirin at the platelet level (Table 1) highlights the heterogeneity of NANSAIDs (10) that is often ignored. For, example, Arfè et al. (44) who studied the risk of heart failures causes by NANSAIDs in 4 European countries noticed that only approximately one-half of the drugs used were significantly cardiotoxic. Nevertheless, they calculated the current use of any NANSAIDs, toxic or not, and concluded that the use of any NANSAID was associated with 19% increased heart failure risk.
The heterogeneity of NANSAIDs is confirmed in a crossover study (30) in which patients received 81 mg of immediate-release Aspirin followed 2 h later by ibuprofen, rofecoxib, or diclofenac for 6 days. This was followed by a washout period of 14 days, after which the same 2 medications were administered in reverse order for another 6 days. The inhibition of COX-1 was assessed by measuring serum TxB2 level, platelet aggregation induced in platelet-rich plasma and COX-2 activity by the measuring the formation of lipopolysaccharidestimulated PGE2 in whole blood. They noticed no significant interaction between Aspirin and rofecoxib or diclofenac. However, ibuprofen significantly interacted with Aspirin given before or after the NANSAID. The Aspirin-ibuprofen interaction has been confirmed by others (26-28, 31-36).
Although we have not made a comparison between molecular interactions studies and clinical trials for all NSAIDs, it is timely to reemphasize that their interaction with Aspirin is heterogeneous in nature. For example, naproxen, celecoxib, piroxicam, indomethacin, mefenamic acid, tiaprofenic acid, nimesulide, oxaprozin, flufenamic acid and dipyrone do interact, while loxoprofen, diclofenac, rofecoxib, etoricoxib, lumiracoxib, etodolac, ketorolac, meloxicam, acetaminophen, flurbiprofen, sulindac, and sodium salicylate do not (Table 8).
It has been suggested that the Aspirin-NANSAIDs interaction is due to a competition to bind to the Arginine-120 residue of the COX-1 channel which may prevent the acetylation of the serine-529 residue by Aspirin (37, 45). Nevertheless, the interference of NANSAIDs with the anti-platelet effect of Aspirin seems to have no long-term consequences as the CV protection of Aspirin remains unaffected by concomitant use of, at least, naproxen and ibuprofen. We put forward three plausible explanations for the disconnect between the results of the short-term platelet experiments and those of observational studies. (i) The interaction at the platelet level may be short-lived so that the effect dissipates shortly after its occurrence. (ii) There is no need for near complete inhibition of TxB2 inhibition to benefit from the cardioprotective properties of Aspirin so that despite a reduction in the extent of anti-platelet effect, the beneficial effect persists, or (iii) the CV effect of Aspirin may not be exclusively due to the drug's anti-platelet properties.
For all, except one eligible study, the CV risk was assessed after > 30 days exposure to the combination while typically, the effect of NANSAIDs on the anti-platelet activity of Aspirin is studied after short exposure times. Thus, the data on the therapeutic outcome of the short-term exposure to Aspirin-NANSAIDs are limited. However, the results published by Kimmel et al. (22) based on a case-control study that assessed the risk only one week before the date of onset of MI are useful in this context. They have reported that addition of NANSAIDs to Aspirin regiment does not increase the CV risk within one week post combination therapy. To this, one may add the fact that, to the best of our knowledge, there is no published report suggestive of a quick negative clinical CV outcome in individual patients who took NANSAIDs therapy while on Aspirin. Furthermore, data from a small size clinical trial, suggest that the effect of naproxen and diclofenac on the Aspirin-induced inhibition of platelet aggregation is short-lived (38). In a randomized placebo-controlled trial, Galliard-Grigioni et al. treated healthy subjects with 100 mg aspirin daily in combination with either three doses of either 1 g acetaminophen, 50 mg diclofenac, 250 mg naproxen or placebo, and assessed the platelet function. Initially, naproxen enhanced, and diclofenac reduced the anti-aggregatory action of Aspirin while acetaminophen had no effect. After 4 days of treatment, however, the platelet aggregation was equally inhibited by all Aspirin-NANSAID combinations.
In practice, a near complete inhibition of TxB2, thereby platelet aggregation is aimed to obtain cardioprotective effects of Aspirin (27). This is while the anti-platelet action of Aspirin is shown to be dose-dependent (46), i.e., low doses of the drug may not completely inhibit TxB2. Nevertheless, Aspirin has been shown to be cardioprotective after low doses (Table 9). This may suggest that to benefit from the CV properties of Aspirin, a complete inhibition of TxB2 is not needed. Thus, a reduced platelet aggregation activity of Aspirin resulted from combination therapies with NANSAIDs, unless proven through appropriately designed clinical trials, may have no significant clinical consequences.
In addition to its anti-platelet effect, Aspirin may reduce CV risks through other mechanisms. Both inflammation and some NANSAIDs appear to increase CV risks (10). Through animal studies, it has been shown that inflammatory conditions impair the balance of vasodilator/vasoconstrictor components of renin-angiotensin system (RAS) within the heart (47). The RAS is a major regulator of human physiology and has a key role in the CV homeostasis. Interestingly, NANSAIDs appear to be void of significant effects on RAS, instead, they are able to restore the imbalances that are resulted by inflammation (47). Alternatively, an altered protective/toxic balance of the cardioactive CYP450-mediated metabolites of arachidonic acid has been reported to be involved in the cardiotoxic effects of NANSAIDs (48). Whether Aspirin influences the RAS or the CYP450-mediated metabolites of arachidonic acid, remains unknown. Nevertheless, the possibility of CV protection by Aspirin through mechanisms other than its platelet effect is plausible.
The current analysis has limitations some of which are inherent to the nature of included studies. First, we have found that the published clinical evidence was sparse and has substantial limitations. To highlight this point, we were unable to assess the heterogeneity since some studies reported RR/OR while other did HR. Second, the primary outcomes of some studies that we included in our review were not CV (MI or stroke) risks as they reported the latter as secondary outcomes. Last, we were unable to perform meta-analysis as the same reference (Aspirin alone, NANSAID alone or nonusers) or outcome (OR, RR or HR) had not been used across the eligible studies.

CONCLUSION
Low-dose Aspirin is widely used to prevent MI and other CV diseases. However, there is evidence that concurrent use of some, but not all NANSAIDs, may inhibit the anti-platelet effect of Aspirin. Naproxen, meloxicam and rofecoxib do not appear to influence the cardioprotective effect of Aspirin. Similarly, a large body of evidence supports that ibuprofen coadministration with Aspirin does not antagonize the anti-platelet effect of Aspirin. Altogether, it appears that the NANSAID-Aspirin interaction at the level of platelets does not necessarily amount to a loss of beneficial effects of Aspirin. The limited available data suggest that the effect of the drug-drug interactions on the platelet aggregation may dissipate shortly. In addition, it is plausible that the reduced platelet aggregation resulted by the interaction may be sufficient for cardioprotection; i.e., no need for near complete aggregation. In addition, the cardioprotective effect of Aspirin despite reduced platelet aggregation caused by NANSAIDs may be through its involvement in other mechanisms such as the RAS and/or metabolism of arachidonic acid to biologically active compounds mediated by CYP450.

Conflict of interests:
The authors have no professional affiliation, financial interest or conflict with the subject matter or information discussed here in this manuscript to declare.
Source of Funding: King Saud University scholarship (Z. Alqahtani) and University of Alberta Self-Directed Grant (F. Jamali).
Authors' contribution: Database search, articles screening, articles review, data analysis, and manuscript preparation: Z. Alqahtani and F. Jamali. Study design and data review: F. Jamali. All authors read and approved the final manuscript.

ACKNOWLEDGMENT
We thank Janice Kung, a librarian at John W. Scott Health Sciences Library University of Alberta, for her valuable comments.     The volunteers received Aspirin (100 mg, once daily) for 6 days. Then they received either single or multiple doses of the combination of Aspirin 2 h before naproxen (500 mg, twice daily) for another 6 days. After a washout period of 14 days, the treatments were administered in reverse order.
Serum TxB2, urinary 11 dehydro-TxB2 excretion rates, platelet aggregation by aggregometry, LPSstimulated PGE2 production in whole blood Naproxen interferes with the inhibitory effect of low-dose Aspirin on platelet aggregation. Aspirin and celecoxib (alone or together) or control (saline) were added to the PRP.

Platelet aggregation (induced by AA) by aggregometry
Celecoxib interferes to a limited extent with the anti-platelet effect of low-dose Aspirin.
60s Platelets were pre-incubated with ibuprofen, naproxen, or celecoxib for 10 min. then Aspirin was added to each group.

COX-1 acetylation, TxB2 formation
A single therapeutic dose of ibuprofen or naproxen followed by Aspirin casue a potent drug-drug interaction, but not between celecoxib and Aspirin. The volunteers were randomly assigned to either Aspirin (30 mg, once daily) for 7 days, slow release diclofenac (50 mg, three times daily) or ibuprofen (800 mg, three times daily) for 1 day. Aspirin (80 mg, once daily) was given after a washout period of 14-42 days with each treatment group for 7 days.
Serum TxB2 levels Only ibuprofen interferes with the anti-platelet activity of Aspirin. The volunteers received during 4 different study periods (≥10 days washout period) either acetaminophen (1 g, three times daily), diclofenac (50 mg, three times daily), naproxen (250 mg, three times daily) or placebo plus Aspirin (100 mg, once daily) for 4 days.
PFA-100 CT Regular daily co-administration of acetaminophen, diclofenac or naproxen do not interfere with the anti-platelet activity of Aspirin. PFA-100 CT Ibuprofen, indomethacin, naproxen, and tiaprofenic acid interfere with the anti-platelet activity of Aspirin but not sulindac or celecoxib.
Healthy volunteers (aged 21-32 years, n = 10) The volunteers were randomly assigned to receive either ibuprofen (400 mg), Aspirin (325 mg) or ibuprofen (400 mg) plus one dose of Aspirin (325 mg, 2 h later). A minimum of 6 days washout period was allowed between treatments.

Platelet aggregation by aggregometry
Administration of ibuprofen before Aspirin interferes with the inhibitory effect of Aspirin on platelet aggregation. The patients were undergoing long term treatment with Aspirin (100 mg, daily), and received celecoxib (200 mg, twice daily), ibuprofen (600 mg, three times daily) or placebo for 7 days.
Serum TxB2, urinary 11 dehydro-TxB2 excretion rates, platelet aggregation by aggregometry, LPSstimulated PGE2 production in whole blood Ibuprofen interferes with antiplatelet effect of Aspirin but not celecoxib. The volunteers were randomly assigned to receive either lumiracoxib (400 mg, once daily) or placebo for 11 days. Both treatment groups received Aspirin (75mg, once daily) from day 5 to 11 (6 days

NA
Risk of bias across studies 15 Specify any assessment of risk of bias that may affect the cumulative evidence (e.g., publication bias, selective reporting within studies).

4
Additional analyses 16 Describe methods of additional analyses (e.g., sensitivity or subgroup analyses, meta-regression), if done, indicating which were pre-specified.

RESULTS
Study selection 17 Give numbers of studies screened, assessed for eligibility, and included in the review, with reasons for exclusions at each stage, ideally with a flow diagram.

4-5
Study characteristics 18 For each study, present characteristics for which data were extracted (e.g., study size, PICOS, follow-up period) and provide the citations.

18-24
Risk of bias within studies 19 Present data on risk of bias of each study and, if available, any outcome level assessment (see item 12).

22-24
Results of individual studies 20 For all outcomes considered (benefits or harms), present, for each study: (a) simple summary data for each intervention group (b) effect estimates and confidence intervals, ideally with a forest plot.

5-9
Synthesis of results 21 Present results of each meta-analysis done, including confidence intervals and measures of consistency. NA Risk of bias across studies 22 Present results of any assessment of risk of bias across studies (see Item 15).

DISCUSSION
Summary of evidence 24 Summarize the main findings including the strength of evidence for each main outcome; consider their relevance to key groups (e.g., healthcare providers, users, and policy makers).

5-9
Limitations 25 Discuss limitations at study and outcome level (e.g., risk of bias), and at review-level (e.g., incomplete retrieval of identified research, reporting bias).

12
Conclusions 26 Provide a general interpretation of the results in the context of other evidence, and implications for future research. 12-13 FUNDING Funding 27 Describe sources of funding for the systematic review and other support (e.g., supply of data); role of funders for the systematic review.