Use of Non-Steroidal Antiinflammatory Drugs and Type-Specific Risk of Acute Coronary Syndrome

Article Outline

• Abstract
• Methods
• Results
• Discussion
• Acknowledgment
• Supplementary Data
• References
• Copyright

The exact mechanism by which nonsteroidal anti-inflammatory drugs (NSAIDs) may increase coronary risk is not completely understood. The aim of this study was to quantify the risk for each type of acute coronary syndrome (ACS) associated using NSAIDs and the role played by dose, duration, and patient characteristics. A prospective case-control study was performed, interviewing 2,954 patients hospitalized for ACS at 32 Spanish hospitals and a similar number of age-matched controls using a structured questionnaire collecting information on the use of NSAIDs, risk factors, and cardiovascular history. Odds ratios (ORs) for any type and each ACS type were calculated adjusted for gender, body mass index, other risk factors, and concomitant medications by conditional logistic regression. The adjusted OR of ACS associated with the current use of NSAIDs was 1.16 (95% confidence interval [CI] 0.95 to 1.42). The risk was increased in patients consuming high doses (OR 1.64, 95% CI 1.06 to 2.53) and those with previous ischemic heart disease (OR 1.84, 95% CI 1.13 to 3.00). The hazard was driven mostly by the increase in the risk for non-ST-segment elevation ACS (OR 1.20, 95% CI 0.99 to 1.47), whereas NSAIDs did not increase the risk for ST-segment elevation myocardial infarction (OR 1.00, 95% CI 0.80 to 1.26). In conclusion, the use of NSAIDs was associated with a small, nonsignificant overall coronary risk that was more apparent for non-ST-segment elevation ACS. This risk was stronger when NSAIDs were used at high doses or in patients with previous ischemic heart disease.

There is evidence suggesting that there is a cardiovascular hazard associated with the administration of nonsteroidal anti-inflammatory drugs (NSAIDs) selective for cyclooxygenase (COX)–2 and some traditional NSAIDs¹, ², ³, ⁴, ⁵, ⁶ through a common mechanism involving the inhibition of COX-2-dependent prostacyclin.⁷ Moreover, it is believed that increased incidence of thrombotic events associated with profound inhibition of COX-2-dependent prostacyclin can be mitigated, if not obliterated, by the complete suppression of platelet COX-1 activity.⁶, ⁷ However, a number of questions related to the cardiovascular risk of NSAIDs, such as whether the increased risk is common to all types of acute coronary syndromes (ACS), the role played by dose, treatment duration, and the importance of patients' clinical characteristics, have not been fully answered.

Back to Article Outline

Methods 

This was a prospective, multicenter, hospital-based, case-control study with the aim of quantifying the risk for each type of ACS associated with traditional NSAIDs as a group and with diclofenac, ibuprofen, and naproxen individually, as well as ascertaining a potential dose and duration response in risk. Investigators from a set of Spanish hospitals integrated in the Red de Investigación Cardiovascular (RECAVA) del Instituto de Salud Carlos III research network⁸ or the Working Group on Ischemic Heart Disease of the Spanish Society of Cardiology were requested to enroll a minimum of 100 consecutive patients with ACS and corresponding controls during the recruitment period. Patients were defined as individuals admitted to 1 of the participant centers for clinical events diagnosed as ACS at discharge. Inclusion criteria were a definite diagnosis of ACS, age 40 to 85 years, and capability and willingness to personally answer the questionnaire. Controls were recruited from individuals visiting the patients (relatives or not) of similar age (±5 years) who fulfilled the same eligibility criteria as patients. Patients were not enrolled if (1) reliable information could not be obtained because of their clinical conditions or social circumstances, (2) they refused to participate, or (3) they had diagnoses of malignant neoplasms in the year before the interview.

ACS were defined as clinical events characterized by episodes of chest pain or other symptoms consistent with myocardial ischemia that also met the following criteria: (1) required emergent hospitalization (<24 hours after symptom onset); (2) received 1 of the following definitive diagnoses: ACS, unstable angina, or acute myocardial infarction; (3) presented with ≥1 of the following signs of myocardial ischemia: increase and decrease of any myocardial necrosis marker or dynamic changes in electrocardiographic repolarization; and (4) was not triggered by extreme pathophysiologic conditions, such as tachycardia >120 beats/min, severe hypotension or hypertension (systolic blood pressure <80 or >200 mm Hg), or severe anemia (hemoglobin <9 g/dl). ACS were classified as unstable angina when no elevation of myocardial necrosis markers occurred and as myocardial infarction when there were elevations of myocardial necrosis markers within the first 24 hours of admission. These were subsequently classified into ST-segment elevation myocardial infarction, when presenting electrocardiography showed persistent (lasting ≥30 minutes and not resolved after nitroglycerin administration) ST-segment elevation ≥0.1 mV in ≥2 contiguous leads, or left bundle branch block, and non-ST-segment elevation myocardial infarction when presenting electrocardiography did not show ST-segment elevation or left bundle branch block.

After obtaining written informed consent, patients were interviewed by a trained staff member (physician, nurse, or medical student) unaware of the study's main hypothesis using a structured questionnaire that included information on sociodemographic features, co-morbidities, symptoms and signs that led to hospitalization, information related to the ACS episode, and drugs of interest consumed in the past year. Co-morbidity reported by interviewees had to be confirmed by medical staff members. These included the traditional risk factors, previous cardiovascular disease (ischemic heart disease [IHD] angina, myocardial infarction, or percutaneous or surgical coronary revascularization), heart failure, peripheral arterial disease or intermittent claudication, and cerebrovascular disease (stroke or transient ischemic attack), and osteoarticular disease (osteoarthritis, rheumatoid arthritis, or other inflammatory arthritis).

Patients and their respective controls were interviewed by the same individual to avoid differential information bias. A patient was considered a current user when intake occurred <7 days before index date (date of ACS episode onset for patients and date of interview for controls). The date of onset was defined as the day on which the first symptom suggestive of ACS occurred (patients). Past users were individuals who had their most recent use 8 to 365 days before the index date. Nonusers were individuals who did not take the drugs during the year before the index date. Current users were subdivided into single and multiple users. Single users were those who used only 1 specific NSAID in the month before the index date, and multiple users were those who used either 2 NSAIDs concomitantly or ≥2 NSAIDs at different times during the month before the index date. We categorized current single users into low- to medium-dose or high-dose users. Specific cut-off values for high NSAID dose were as follows: aceclofenac 100 mg, celecoxib 200 mg, diclofenac 100 mg, ibuprofen 1,200 mg, indomethacin 75 mg, ketoprofen 150 mg, ketorolac 10 mg, meloxicam 7.5 mg, naproxen 750 mg, and piroxicam 10 mg. Doses less than or equal to the cut-off values were grouped under low to medium doses, and doses greater than the cut-off value were grouped under high doses. We divided current single users into 2 categories according to the duration of the treatment. Treatments <1 month were classified as of short-term duration, and treatments >30 days were classified as of mid- to long-term duration. To help patients and controls recall the name of the drugs, interviews were always performed with the support of a booklet that included pictures of all anti-inflammatory and cardiovascular drugs available on the Spanish market at that time.

The sample size was calculated according to a previous study performed in Spain, which showed that 9% of controls were using NSAIDs the week previous to an interview.⁹ Similarly, the percentages of use for ibuprofen, diclofenac, and naproxen were estimated at 3%, 2.5%, and 1%, respectively. The percentage of use of any of the 3 main NSAIDs was estimated in 2% of the control population. For an α value of 0.05 and a β value of 0.80, and assuming that NSAIDs could increase the risk for ACS by a factor of 1.5, sample sizes of 3,000 patients and 3,000 controls were estimated to be necessary for the study. The odds ratios (ORs) and 95% confidence intervals (CIs) of all types of ACS as well as for each specific ACS type associated with using NSAIDs and other factors were estimated using conditional logistic regression (Stata/IC version 10.0; StataCorp LP, College Station, Texas). All estimates were adjusted by gender, traditional risk factors, body mass index, previous IHD, and the use of statins, antiplatelet drugs, and antihypertensive drugs. The effects of daily dose and duration were investigated in current single users. Unconditional logistic regression was performed in stratified analyses (previous IHD) or subtype ACS analyses because of reduced power of the conditional model and no virtual changes in the estimates compared to the conditional model. In addition to the variables used in the conditional regression models, age was entered in the unconditional regression models.

Back to Article Outline

Results 

From February to November 2007, 5,908 study subjects were interviewed at 32 Spanish hospitals (see online Appendix). Patients and controls were well matched for age (mean age 65.2 ± 12.4 years in patients, 64.4 ± 12.2 years in controls), but there was a significantly higher proportion of women in the control group (50% vs 26%, p <0.001). The proportion of risk factors was consistently higher in patients than in controls (current smoking, 30% vs 19%, p <0.0001; diabetes, 30% vs 13%, p <0.001; hypertension, 60% vs 42%, p <0.001; dyslipidemia, 50% vs 31%, p <0.001; and body mass index ≥30 kg/m², 25% vs 21%, p <0.001), as was the prevalence of previous cardiovascular disease (IHD, 36% vs 10%, p <0.001; myocardial infarction, 21% vs 5%, p <0.001; angina, 26% vs 6%, p <0.001; percutaneous coronary revascularization, 19% vs 4%, p <0.001; cerebrovascular disease, 6% vs 4.4%, p = 0.038; peripheral arterial disease, 10% vs 5.3%, p <0.001; heart failure, 6.2% vs 3.5%, p <0.001; and aortic aneurysm, 1.0% vs 0.4%, p = 0.011). Gout (5.3% vs 3.3%, p = 0.003) but not rheumatoid arthritis (2.3% vs 3.2%, p = 0.22) was also more prevalent in patients.

There was a significantly higher proportion of current NSAID use in women compared to men (21% vs 11%, p <0.001). After adjustment for gender, traditional risk factors, and previous cardiovascular history, the OR of overall ACS associated with current use of NSAIDs was 1.16 (95% CI 0.95 to 1.42). There was no increased risk for ACS associated with occasional NSAID use, consumption <30 days, or use of low to medium daily doses (Table 1). We found the use of high doses of NSAIDs to be associated with a significant increase in the risk for ACS (OR 1.64, 95% CI 1.06 to 2.53). When individual NSAIDs were considered, differences in ACS risks between the most frequently used agents were found. No increased risk was found with ibuprofen (OR 0.88, 95% CI 0.68 to 1.14) and naproxen (OR 1.30, 95% CI 0.60 to 2.82), while estimates for diclofenac (OR 1.56, 95% CI 0.99 to 2.45) and aceclofenac (OR 3.50, 95% CI 1.45 to 8.47) were compatible with an increased risk. Celecoxib was the only coxib with some use in our population (10 exposed patients and 5 controls) and had an OR of 4.96 (95% CI 1.09 to 22.47).

Table 1. Odds ratios and 95% confidence intervals of any type of acute coronary syndrome according to recency, duration, daily dose, and nonsteroidal anti-inflammatory drug intake pattern^⁎

Variable	Patients (n = 2,954)	Controls (n = 2,954)	Adjusted OR⁎	95% CI
Recency
Nonuse	1,917	1,831	1	—
Current use	420	464	1.16	0.95–1.42
Single	370	422	1.13	0.92–1.39
Multiple	50	42	1.53	0.88–2.66
Past use	617	659	1.01	0.84–1.21
Duration
Nonuse	1,917	1,831	1	—
Use ≤30 days	55	65	1.01	0.62–1.64
Use >30 days	305	333	1.19	0.95–1.50
Daily dose
Nonuse	1,917	1,831	1	—
Low to medium dose	288	347	1.03	0.82–1.29
High dose	79	72	1.64	1.06–2.53
Intake pattern
Nonuse	1,917	1,831	1	—
Occasional use	214	249	1.05	0.77–1.41
Regular use	155	170	1.19	0.92–1.55

When we analyzed the effect according to the type of ACS, we found that most of the observed ACS risk with NSAIDs was due to an increased risk for non-ST-segment elevation ACS, whereas no association was found with ST-segment elevation myocardial infarction (Table 2). Additionally, patients free of IHD presented no increased risk for ACS when using NSAIDs, while a history of IHD greatly predisposed users of NSAIDs to increased ACS risk, again mostly concentrated in an increase in non-ST-segment elevation ACS (Table 3).

Table 2. Odds ratios and 95% confidence intervals of different acute coronary syndromes according to recency, duration, and daily dose of nonsteroidal anti-inflammatory drug exposure compared to nonuse

	Controls (n = 2,954)	Unstable Angina			Non-ST-segment Elevation Myocardial Infarction			ST-Segment Elevation Myocardial Infarction
		Patients (n = 819)	Adjusted OR	95% CI	Patients (n = 1,006)	Adjusted OR	95% CI	Patients (n = 1,120)	Adjusted OR⁎	95% CI
Recency
Nonuse	1,831	542	1	—	650	1	—	719	1	—
Current use	464	115	1.22	0.93–1.60	155	1.23	0.97–1.56	149	1.00	0.80–1.26
Duration
Nonuse	1,831	542	1	—	650	1	—	719	1	—
Use ≤30 days	65	8	0.66	0.29–1.52	24	1.32	0.77–2.27	23	1.02	0.60–1.75
Use >30 days	333	100	1.43	1.06–1.92	108	1.21	0.92–1.58	96	0.92	0.70–1.20
Daily dose
Nonuse	1,831	542	1	—	650	1	—	719	1	—
Low to medium dose	347	84	1.17	0.86–1.60	103	1.09	0.83–1.44	101	0.89	0.68–1.16
High dose	72	24	1.64	0.93–2.88	29	1.51	0.91–2.50	25	1.23	0.74–2.06

Table 3. Odds ratios and 95% confidence intervals of different acute coronary syndrome types by history of ischemic heart disease according to recency of nonsteroidal anti-inflammatory drug use compared to nonuse

	Patients Without IHD			Patients With IHD
	Patients/Controls	Adjusted OR⁎	95% CI	Patients/Controls	Adjusted OR⁎	95% CI
Unstable angina
n	365/2,655			452/288
Recency
Nonuse	226/1,602	1	—	314/224	1	—
Current use	52/399	1.13	0.80–1.59	56/20	1.94	1.09–3.45
Non-ST-segment elevation myocardial infarction
n	609/2,655			391/288
Recency
Nonuse	381/1,602	1	—	265/224	1	—
Current use	84/399	1.07	0.80–1.42	49/20	1.85	1.03–3.32
ST-segment elevation myocardial infarction
n	889/2,655			227/288
Recency
Nonuse	562/1,602	1	—	154/224	1	—
Current use	108/399	0.96	0.74–1.25	18/20	1.06	0.52–2.16

Back to Article Outline

Discussion 

This large case-control study suggests that the use of NSAIDs is not associated with a major significant risk for ACS overall in the general population. However, if taken in high doses and during prolonged time periods, an increase in the risk for non-ST-segment elevation ACS is present, with a greater association in patients with previous IHD.

NSAIDs are associated with an increased risk for gastrointestinal toxicity,⁹ renal functional impairment,¹⁰ increase in blood pressure,¹¹ heart failure,¹² and myocardial infarction.¹, ², ³, ⁴, ⁵, ⁶ The widespread and growing use of these drugs in the population means that any increase in cardiovascular risk, however small in relative terms, will have a large impact in the population. On average, observational studies published to date suggest a 10% increased risk for myocardial infarction in users of traditional NSAIDs.⁴ Our finding of a nonsignificant 16% increase in ACS risk for current traditional NSAID use is in line with previous evidence, but in contrast, it indicates for the first time that this risk may be concentrated in non-ST-segment elevation ACS, whereas the risk for ST-segment myocardial infarction seems to be little influenced by NSAID use. There is presently no clear explanation for this finding. The inhibition of prostacyclin synthesis produced by NSAIDs in endothelial cells⁷ may facilitate the development of small platelet thrombi in the vascular surface but might not be sufficient to trigger a larger thrombus formation resulting in occlusion of a large epicardial coronary artery, usually the cause of ST-segment elevation myocardial infarctions, in most patients. This speculation should be confirmed in experimental animal models and future clinical studies. However, it should be pointed out that the long-term risk for non-ST-segment elevation and ST-segment elevation myocardial infarctions is similar.

One of the key questions on the cardiovascular safety of traditional NSAIDs is whether this risk is uniform across the population or if there are patient-specific characteristics that predict higher coronary risk, as suggested previously.¹³ Our study confirms that patients with antecedent IHD cluster most of the risk. Thus, although patients without IHD seem to have no harm on NSAIDs, those patients with IHD taking NSAIDs presented a near doubling of coronary risk compared to those who did not use these drugs. In concordance with previous studies,⁶ our results suggest that high doses and longer treatment durations are the treatment patterns that significantly increase the coronary risk, particularly in patients previously at high baseline risk.

Results for the 3 most widely used NSAIDs (ibuprofen, diclofenac, and naproxen) in our study population confirm estimates of risk observed in previous studies.⁴ Ibuprofen was associated with no increased risk for ACS, whereas diclofenac was associated with an increased risk. The 2 agents are traditional NSAIDs lacking an effect on platelet function⁷ but differ in their degrees of inhibition of COX-2 (an index of drug potency and exposure) at therapeutic doses, with ibuprofen presenting COX-2 inhibition of 75% and diclofenac inhibition close to 99%.⁶ It is worth noting that aceclofenac, which requires conversion to diclofenac to affect COX-2, presented a similar increased risk for ACS as diclofenac. Naproxen has been postulated as the safest traditional NSAID,¹ and it has been speculated that it could have an aspirin-like effect¹⁴ because of its capacity to inhibit platelet COX-1 throughout dosing intervals, when administered at high doses, such as 1,000 mg/day.¹⁵ Therefore, some investigators have recommended naproxen as the first choice among traditional NSAIDs for patients with high cardiovascular risk.¹⁶ However, we were not able to confirm the reduction in coronary risk found by others.¹, ¹⁷, ¹⁸ This is consistent with the marked interindividual heterogeneity in the persistent and complete inhibition of platelet COX-1 activity detected in clinical studies.¹⁹

Some limitations must be considered when interpreting our results. Despite the large number of subjects interviewed in this hospital-based study, the absolute number of patients taking individual NSAIDs except ibuprofen was small, so the study had reasonable statistical power to test the primary hypothesis but was underpowered for secondary analyses, such as the risk assessment associated with individual NSAIDs, particularly coxibs, as well as for subgroup analyses. As with all field-based case-control studies, selection of controls is always a concern to be taken into account. The dose-dependent risk of NSAIDs may be confounded by the use of higher NSAID doses by more chronically ill patients with greater baseline risk for ACS.

Back to Article Outline

Acknowledgment 

This study was developed with the cooperation of RECAVA del Instituto de Salud Carlos III, Ministerio de Ciencia e Innovación, the Working Group on Ischemic Heart Disease of the Spanish Society of Cardiology, Centro Español de Investigación Farmacoepidemiológica, and Pfizer S.A. (Madrid, Spain), which funded the study through an unrestricted grant but had no participation in data gathering or storage, statistical analyses, or report drafting, writing, or editing.

Back to Article Outline

Supplementary Data 

Back to Article Outline

References 

1. Kearney PM, Baigent C, Godwin J, Halls H, Emberson JR, Patrono C. Do selective cyclo-oxygenase-2 inhibitors and traditional non-steroidal anti-inflammatory drugs increase the risk of atherothrombosis? (Meta-analysis of randomised trials). BMJ. 2006;332:1302–1308
   • View In Article
   • CrossRef
2. García Rodríguez LA, Varas-Lorenzo C, Maguire A, González-Pérez A. Nonsteroidal antiinflammatory drugs and the risk of myocardial infarction in the general population. Circulation. 2004;109:3000–3006
   • View In Article
   • CrossRef
3. Chan AT, Manson JE, Albert CM, Chae CU, Rexrode KM, Curhan GC, et al. Nonsteroidal antiinflammatory drugs, acetaminophen, and the risk of cardiovascular events. Circulation. 2006;113:1578–1587
   • View In Article
   • CrossRef
4. Hernández-Díaz S, Varas-Lorenzo C, García Rodríguez LA. Non-steroidal antiinflammatory drugs and the risk of acute myocardial infarction. Basic Clin Pharmacol Toxicol. 2006;98:266–274
   • View In Article
   • MEDLINE
   • CrossRef
5. Helin-Salmivaara A, Virtanen A, Vesalainen R, Grönroos JM, Klaukka T, Idänpään-Heikkilä JE, et al. NSAID use and the risk of hospitalization for first myocardial infarction in the general population: a nationwide case-control study from Finland. Eur Heart J. 2006;27:1657–1663
   • View In Article
   • MEDLINE
   • CrossRef
6. García Rodríguez LA, Taconnelli ST, Patrignani P. Role of dose potency in the prediction of risk of myocardial infarction associated with nonsteroidal anti-inflammatory drugs in the general population. J Am Coll Cardiol. 2008;52:1628–1636
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (699 KB)
   • CrossRef
7. Grosser T, Fries S, FitzGerald GA. Biological basis for the cardiovascular consequences of COX-2 inhibition: therapeutic challenges and opportunities. J Clin Invest. 2006;116:4–15
   • View In Article
   • MEDLINE
   • CrossRef
8. García-Dorado D, Castro-Beiras A, Díez J, Gabriel R, Gimeno-Blanes JR, Ortiz de Landázuri M, et al. Cooperative Cardiovascular Disease Research Network (RECAVA). Rev Esp Cardiol. 2008;61:58–65
   • View In Article
   • CrossRef
9. Lanas A, García-Rodríguez LA, Arroyo MT, Gomollon F, Feu F, Gonzalez Perez A, et al. Risk of upper gastrointestinal ulcer bleeding associated with selective COX-2 inhibitors, traditional non-aspirin NSAIDs, aspirin, and combinations. Gut. 2006;55:1731–1738
   • View In Article
   • MEDLINE
   • CrossRef
10. Clive DM, Stoff JS. Renal syndromes associated with nonsteroidal antiinflammatory drugs. N Engl J Med. 1984;310:563–572
    • View In Article
    • MEDLINE
    • CrossRef
11. Johnson AG, Nguyen TV, Day RO. Do nonsteroidal anti-inflammatory drugs affect blood pressure? (A meta-analysis). Ann Intern Med. 1994;121:289–300
    • View In Article
    • MEDLINE
12. Page J, Henry D. Consumption of NSAIDs and the development of congestive heart failure in elderly patients: an unrecognized public health problem. Arch Intern Med. 2000;160:777–784
    • View In Article
    • MEDLINE
    • CrossRef
13. Solomon DH, Glynn RJ, Rothman KJ, Schneeweiss S, Setoguchi S, Mogun H, et al. Subgroup analyses to determine cardiovascular risk associated with nonsteroidal antiinflammatory drugs and coxibs in specific patient groups. Arthritis Rheum. 2008;59:1097–1104
    • View In Article
    • CrossRef
14. Patrono C, Coller B, Dalen J, Fuster V, Gent M, Harker L, et al. Platelet-active drugs: the relationships among dose, effectiveness, and side effects. Chest. 2004;126:234S–264S
    • View In Article
    • MEDLINE
    • CrossRef
15. Capone ML, Tacconelli S, Sciulli MG, Grana M, Ricciotti E, Minuz P, et al. Clinical pharmacology of platelet, monocyte, and vascular cyclooxygenase inhibition by naproxen and low-dose aspirin in healthy subjects. Circulation. 2004;109:1468–1471
    • View In Article
    • CrossRef
16. Antman EM, Bennett JS, Daugherty A, Furberg C, Roberts H, Taubert KA American Heart Association. Use of nonsteroidal antiinflammatory drugs: an update for clinicians: a scientific statement from the American Heart Association. Circulation. 2007;115:1634–1642
    • View In Article
    • CrossRef
17. Watson DJ, Rhodes T, Cai B, Guess HA. Lower risk of thromboembolic events with naproxen among patients with rheumatoid arthritis. Arch Intern Med. 2002;162:1105–1110
    • View In Article
    • MEDLINE
    • CrossRef
18. Rahme E, Pilote L, LeLorier J. Association between naproxen use and protection against acute myocardial infarction. Arch Intern Med. 2002;162:1111–1115
    • View In Article
    • MEDLINE
    • CrossRef
19. Capone ML, Tacconelli S, Sciulli MG, Anzellotti P, Di Francesco L, Merciaro G, et al. Human pharmacology of naproxen sodium. J Pharmacol Exp Ther. 2007;322:453–460
    • View In Article
    • CrossRef