Statins are associated with lower risk of gastrointestinal bleeding in patients with unstable coronary syndromes: Analysis of the Orbofiban in Patients with Unstable coronary Syndromes–Thrombolysis In Myocardial Infarction 16 (OPUS-TIMI 16) trial

Article Outline

• Abstract
• Methods
  • Study population
  • Trial design
  • Statistical analysis
• Results
  • Overall GI bleeding
  • Severe and major GI bleeding
  • Post-discharge GI bleeding
  • GI bleeding during the index hospitalization
• Discussion
  • Study limitations
  • Conclusions
• References
• Copyright

Background

It has recently been shown that statins increase the myocardial content of prostaglandin (PG) I₂ (prostacyclin) and PGE₂. A systemic increase of PG production may protect the gastric mucosa and prevent gastrointestinal (GI) bleeding. We hypothesized that statins would lower the risk of GI bleeding associated with antiplatelet therapy in patients with acute coronary syndromes (ACS).

Methods

We retrospectively analyzed data on 10288 patients with ACS included in the OPUS-TIMI 16 trial and received aspirin and either the oral IIb/IIIa inhibitor orbofiban or placebo.

Results

Inhospital GI bleeding rate was significantly lower in patients who were receiving lipid-lowering drugs before admission compared with those who were not (0.2% vs 0.6%, P = .031). Throughout 10 months of follow-up, GI bleeding occurred in 1.8% of non–statin users compared with 1.0% of statin users (P = .001). Statin use was associated with less overall bleeding in both the orbofiban (1.4% vs 2.4%, P = .006) and the placebo groups (0.2% vs 0.8%, P = .047). Severe and major bleeding occurred less frequently with statin use (0.8% vs 1.5%, P = .001) in both the orbofiban (1.1% vs 2.0%, P = .006) and the placebo groups (0.1% vs 0.5%, P = .119). Logistic regression analysis showed that age >65 years, orbofiban treatment, Killip class >1, history of cerebrovascular disease, and calcium-channel blocker use were associated with higher risk of GI bleeding, whereas statin therapy was associated with a lower risk (odds ratio 0.68, 95% CI 0.45-1.04, P = .079).

Conclusions

Statins may exert protective effect against GI bleeding in patients with ACS. Additional studies are warranted to explore this additional potential benefit of statins.

The effectiveness of 3-hydroxy-3-methylglutaryl-coenzyme A reductase inhibitors (statins) have been proven in the management of hyperlipidemia and for the prevention of atherosclerotic vascular disease and cardiovascular events.¹, ² However, the therapeutic benefits of statin therapy seem to extend beyond their lipid-lowering effect.³, ⁴, ⁵

In a well-established model of coronary occlusion and reperfusion in rats, we have shown that atorvastatin reduces myocardial infarct size through a prostanoid-mediated mechanism. Atorvastatin increased the myocardial levels of both PGE₂ and 6-keto-prostaglandin (PG) F1α, the stable metabolite of prostacyclin (PGI₂). This effect is achieved through an increase in the expression and activity of c-phospholipase A₂, cyclooxygenase-2, PGI₂, and PGE₂ synthases.⁶ Although a systemic increase in PG production by statins has not yet been demonstrated, we hypothesized that an increase in PG content and activity in the gastrointestinal (GI) tract may exert a protective effect on GI bleeding in patients with acute coronary syndromes (ACS). The OPUS-TIMI 16 was a randomized placebo-controlled study that evaluated the effect of long-term treatment of patients with ACS with orbofiban, an oral IIb/IIIa inhibitor, with all patients receiving aspirin.⁷ We retrospectively analyzed the OPUS-TIMI 16 trial data to determine a possible protective effect of statins on GI bleeding.

Back to Article Outline

Methods 

Study population 

Between October 1997 and November 1998, 10288 patients were enrolled in the OPUS-TIMI 16 study.⁷ Inclusion criteria were ischemic discomfort at rest lasting >5 minutes with onset within 72 hours of randomization and associated with ≥1 of the following features: new or presumably new ST-segment deviation >0.5 mm; T-wave inversion >3 mm in 3 leads or left bundle branch block; positive cardiac markers; or (for the first 3000 patients only) history of myocardial infarction (MI), percutaneous coronary intervention (PCI), coronary artery bypass grafting (CABG), coronary stenosis >50%, age >65 years, and a history of angina or positive stress test, prior peripheral arterial or cerebrovascular disease (CVD), or diabetes mellitus.

Exclusion criteria were age <18 years; pregnancy; life expectancy <2 years; prior PCI within the previous 6 months (other than for the index event) or CABG within 2 months; hemostatic disorder; thrombocytopenia; significant systemic bleeding; documented peptic ulcer within 6 months; prior intracranial hemorrhage or vascular abnormality; current or future need for warfarin; need for chronic daily nonsteroidal anti-inflammatory drugs; receipt of >2 doses of ticlopidine/clopidogrel within 48 hours or abciximab within 24 hours; creatinine >1.6 mg/dL (142 μmol/L) or a calculated creatinine clearance of <40 mL/min; hypersensitivity to aspirin or a glycoprotein IIb/IIIa inhibitor; uncontrolled hypertension; participation in another investigational drug or device trial within 30 days; or prior treatment with orbofiban.

Trial design 

Eligible patients were treated with 150 to 162 mg of aspirin daily and were randomized to orbofiban or placebo as previously described.⁷ Patients were seen in follow-up at 14 and 30 days and every 3 months thereafter, with complete blood counts performed at each visit. Treatment and follow-up were planned to continue for an average of 1 year (minimum of 6 months). However, the study was prematurely terminated due to an increased 30-day mortality rate in the active treatment group.

Gastrointestinal bleeding was defined as blood loss from the GI tract of new onset, hematemesis, or melena with no other apparent source of acute blood loss. Severe or life-threatening GI bleeding was defined as bleeding associated with severe hemodynamic compromise. Major bleeding was defined as ≥15% absolute reduction in hematocrit or need for a transfusion. All GI bleedings were prospectively evaluated and adjudicated during the trial.⁷

Statistical analysis 

All analyses were performed using Stata version 8 (College Station, TX). Gastrointestinal bleeding event rates overall (inhospital plus post-discharge) as well as post-discharge bleeding events were evaluated based on the presence or absence of statin treatment post-discharge. Inhospital events were evaluated based on the use of any lipid-lowering medications at least 2 weeks before enrollment. Specific data on the use of statins before admission are not available. The χ² test was used for the analysis of categorical variables when sample size was >5 for all cells in a table. When the sample size was <5 in a given cell of a table, Fisher exact test was used. The Student t test was used for continuous variables. A P value <.05 was considered statistically significant. Multivariate logistic regression analysis that included all baseline covariates with P value <.05 was performed. Event rates are Kaplan-Meier estimates.

Back to Article Outline

Results 

Of 10288 patients enrolled, 6867 (66.7%) patients received one of two doses of orbofiban and 3421 (33.3%) patients received placebo. Follow-up data were available on 10230 patients, with a mean follow-up of 6 months and a range of 3 to 12 months. In the total cohort, from randomization to end of follow-up, 161 (1.6%) patients had episodes of GI bleeding (57 patients experienced GI bleeding during the index hospitalization and 104 post-discharge), and 3304 (32%) patients were discharged on statins.

The patients' characteristics, medication use before admission, and statin use after discharge are presented in Table I. Patients who had GI bleeding were older; were more often women; had more often history of hypertension, heart failure, CVD, and Killip class >1 on admission; were less often smokers; and had less often hypercholesterolemia. These patients received more often calcium-channel blockers (CCBs), diuretics, and angiotensin-converting enzyme inhibitors (ACEIs) before admission. Use of lipid-lowering drugs before admission was not different between those who experienced and those who did not experience GI bleeding. However, patients who experienced GI bleeding were less often discharged on statin therapy.

Table I. Characteristics of patients according to GI bleeding status

	GI bleeding (n = 161)	No GI bleeding (n = 10069)	P
Age (y)	71 ± 10	60 ± 12	<.0001
Male	40	60	<.001
Diabetes mellitus	24	21	.424
Hypertension	55	43	.002
Hypercholesterolemia	18	28	.004
Smoking	27	36	.012
Hx PCI	12	11	.685
Hx CABG	15	11	.081
Hx CHF	11	5	.001
Hx MI	29	27	.718
Hx CVD	7	3	.001
Killip >1	13	8	.001
Prior aspirin	44	39	.211
Prior β-blockers	29	27	.627
Prior Ca++-blockers	38	24	<.001
Prior ACEI	29	20	.012
Prior diuretics	24	13	<.001
Prior lipid-lowering drugs	15	20	.136
Discharge on statins	20	34	<.001

Values are presented as mean ± SD or percentage. Hx, History of.

Overall GI bleeding 

Overall, there were 129 GI bleeding episodes in patients who did not receive statins on discharge (1.9%), compared with 32 episodes (1.0%) in patients who received statins (P = .001). Gastrointestinal bleedings of any severity were significantly more frequent in patients who received orbofiban (2.1% vs 0.6%, P < .0001). As shown in Figure 1, statin use was associated with lower rates of GI bleeding in both the orbofiban (1.4% vs 2.4% without statins, P = .006) and the placebo groups (0.2% vs 0.8% without statins, P = .047).

• 
  • View Large Image
  • Download to PowerPoint

• Figure 1. 

  The effect of statin use on the overall incidence of GI bleeding (presented as percentage) in patients on orbofiban or placebo.

Overall, GI bleeding was more frequent in patients >65 years (3.2% vs 0.6%, P < .0001); in females (2.2% vs 1.3%, P < .0001); in patients with a history of CVD (4% vs 1.5%, P = .001), heart failure (3.4% vs 1.5%, P < .0001), and hypertension (2% vs 1.2%, P = .002); in patients who used CCBs within 2 weeks before admission (2.4% vs 1.3%, P < .0001); and in those with Killip class II-IV on admission (3% vs 1.4%, P < .0001). Gastrointestinal bleeding was significantly less frequent in patients with a history of hypercholesterolemia (1% vs 1.7%, P = .004) and in smokers (1.1% vs 1.8%, P = .012).

Logistic regression analysis, presented in Table II, showed that age >65 years, history of CVD, history of hypercholesterolemia, prior CCB use, and orbofiban treatment were independently associated with the risk of bleeding. Killip class >1 and statin use at discharge were of marginal significance.

Table II. Multivariate analysis of the likelihood of GI bleeding of any severity

OR, Odds ratio.

Severe and major GI bleeding 

As shown in Figure 2, statin treatment was also associated with significantly less severe and major GI bleeding in both the control and the orbofiban group. Logistic regression analysis, presented in Table III, showed that age >65 years, history of CVD, Killip class >1, and orbofiban therapy were independently associated with an increased risk of severe or major GI bleeding, whereas history of hypercholesterolemia was associated with reduced risk. Statin use was associated with a 29% relative risk reduction; however, the effect was not statistically significant.

• 
  • View Large Image
  • Download to PowerPoint

• Figure 2. 

  The effect of statins use on the overall incidence of severe and major GI bleeding (presented as percentage) in patients on orbofiban or placebo.

Table III. Multivariate analysis of the likelihood of severe or major GI bleeding

Post-discharge GI bleeding 

When we analyzed the data for GI bleeding of any severity that occurred only after discharge, patients who were discharged on statin treatment had significantly less GI bleeding events (0.7% vs 1.2%, P = .028). Moreover, as presented in Figure 3, in the orbofiban group, patients who were on statins after discharge had a trend toward less GI bleeding episodes than patients who did not receive a statin (1.0% vs 1.5%, P = .058). In patients who were on placebo, there was no significant difference in the post-discharge GI bleeding rates between those who did and those who did not receive statins (0.2% vs 0.4%, P = .358). Similar results were obtained when we analyzed the data for post-discharge episodes of severe or major GI bleeding (0.6% vs 1.0%, P = .077, for all patients; 0.9% vs 1.3%, P = .11, for the orbofiban group; and 0.1% vs 0.2%, P = .54, for the placebo group).

• 
  • View Large Image
  • Download to PowerPoint

• Figure 3. 

  The effect of statin use on the incidence of any post-discharge GI bleeding (presented as percentage) in patients on orbofiban or placebo.

Logistic regression analysis, presented in Table IV, demonstrated that age>65 years, history of congestive heart failure (CHF), prior use of CCBs, and use of orbofiban were independently associated with post discharge GI bleeding. Again, statin use was associated with a trend toward lower risk of bleeding; however, the effect was not statistically significant.

Table IV. Multivariate analysis of the likelihood of having any GI bleeding post-discharge

GI bleeding during the index hospitalization 

There were 57 episodes of GI bleeding during the index hospitalization (35% of the 161 total GI bleeding episodes). Patients with GI bleeding were significantly older than the patients without GI bleeding (71 ± 10 vs 60 ± 12 years, P < .001). The percentage of men was higher in the group with than the group without GI bleeding (40% vs 28%, P = .035). Patients with GI bleeding had less often history of hypercholesterolemia (14% vs 28%, P = .018) and were less often current smokers (23% vs 36%, P = .034). However, they had higher prevalence of prior CVD (7% vs 3%, P = .043). In contrast, there were no differences in the prevalence of diabetes, hypertension, CHF, prior MI, prior PCI, and prior CABG between patients with or without GI bleeding during the index admission. There was no difference between the patients with and without GI bleeding in the prevalence of prior use of aspirin, β-blockers, CCBs, and ACEIs. Among those who experienced GI bleeding, the rate of prior use of diuretics was higher (26% vs 13%, P = .005). In contrast, prior use of lipid-lowering medications was significantly lower among those with than without GI bleeding (8% vs 20%, P = .031). More patients who had GI bleeding had Killip class >1 on admission (18% vs 9%, P = .017). Systolic blood pressure on enrollment was higher among those who had GI bleeding (138 ± 27 vs 129 ± 21 mm Hg, P = .0012); however, mean diastolic blood pressure, heart rate, and body weight were comparable. Most of the GI bleeding episodes occurred in the orbofiban group (86% vs 14%, P = .002).

Inhospital GI bleeding rate was significantly lower in patients who were on lipid-lowering medications for at least 2 weeks before admission (0.2% vs 0.6%, respectively, P = .031). Figure 4 shows the rates of GI bleeding during the index hospitalization according to the treatment group and prior lipid-lowering medication use.

• 
  • View Large Image
  • Download to PowerPoint

• Figure 4. 

  The effect of prior lipid-lowering medication use on the incidence of inhospital GI bleeding (presented as percentage) in patients on orbofiban or placebo.

Predictors of inhospital GI bleeding, as presented in Table V, were age >65 years, elevated systolic blood pressure on admission, and treatment with orbofiban, whereas prior lipid-lowering medications use was associated with significantly lower risk.

Table V. Multivariate analysis of the likelihood of inhospital GI bleeding

BP, Blood pressure.

Back to Article Outline

Discussion 

This is the first study to indicate that statin use may be associated with a lower incidence of GI bleeding in patients with ACS. Univariate analysis showed that pre-admission use of lipid-lowering medications was associated with lower inhospital GI bleeding rate, whereas patients who were discharged with statin therapy had lower overall and post-discharge GI bleeding rate over a mean follow-up of 6 months. The effect was mainly evident in patients who were on orbofiban and aspirin and also in those on aspirin alone.

Bleeding complications in patients with ACS treated with aspirin are not uncommon, with the most common source of bleeding being GI.⁸ The reasons are not well elucidated, yet these patients have a worse prognosis than patients who do not bleed.⁸ The high bleeding rates may be due to an increasing use of invasive diagnostic and therapeutic procedures along with the use of potent anticoagulant, antithrombotic, and antiplatelet drugs and “stress-related” conditions such as gastritis.

It is not clear why men had more GI bleeding during the index hospitalization, whereas women had more GI bleeding overall. This might be related to the fact that men tend to undergo more invasive procedures and to receive more intense anticoagulation therapy during the index admission.

It has previously been suggested that the reduction in cardiovascular events with statins can be attributed to their lipid-lowering effect as well as their “pleiotropic effect.”⁵ It is currently suggested that through their anti-inflammatory and antithrombotic effects, statins stabilize the atherosclerotic plaque, improve endothelial function, and enhance fibrinolysis.⁹, ¹⁰ The antithrombotic effects, including decreasing tissue factor expression, factor XIII and V activation, thrombin generation, and fibrinogen cleavage, may potentially increase the overall bleeding risk, especially GI bleeding. On the other hand, we have recently demonstrated that atorvastatin increases myocardial content of PGI₂ and PGE₂ through increased expression and activity of phospholipase A₂, cyclooxygenase 2, and PGI₂ and PGE₂ synthases.⁶ Chen et al¹¹ have shown that statins induce cyclooxygenase 2 gene expression and PGE₂ release in murine macrophages. Prostaglandins are known to protect the gastric mucosa and reduce the risk of GI bleeding.¹² Assuming that an increase in systemic PGI₂ and PGE₂ production occurs with statins, there may be a protective effect exerted on GI bleeding by statins. It remains to be studied whether statins increase systemic prostaglandin synthesis in humans.

Interestingly, we found that a history of hypercholesterolemia was independently associated with less overall GI bleeding (any severity as well as major or severe) and inhospital bleeding (35% of all bleeding episodes), but not post-discharge GI bleeding. This may represent a confounding factor (prior use of statins) that affected mainly GI bleeding in the first few days after admission. Indeed, we demonstrated that prior use of lipid-lowering medications was an independent predictor of less inhospital bleeding. Unfortunately, we have only data on lipid-lowering medications but not specific data on statin use before enrollment. However, it is plausible that most of these patients received statins. Post-discharge statin use, but not a history of hypercholesterolemia, remained an independent predictor of lower GI bleeding rate after discharge, supporting our hypothesis.

Study limitations 

Our study may be biased for the effect of statins on outcome because this is a retrospective analysis and statin administration was not randomized. However, except for a younger age, there were no major differences in patient characteristics between the statin and the no-statin groups. The large study size minimizes the selection bias effect.

We do not have specific data on statin use before enrollment and on the timing of initiation of statins during hospitalization. However, it is plausible that most of the patients who were on lipid-lowering medications before admission received statins. Also, we do not have specific data as to the time relations of statin initiation and the occurrence of GI bleeding. Assuming that the protective effect of statins is not immediate, several days may elapse before the effect is fully manifested. This assumption is supported by our findings that history of hypercholesterolemia was independently associated with lower risk of inhospital, but not post-discharge, GI bleeding.

This study is exploratory because the effect of statins was modest and the overall incidence of GI bleeding was low. In addition, we studied patients enrolled in a clinical trial who may not be representative of patients in clinical practice, where the risk of GI bleeding is likely to be greater and therefore the mitigating effect of statins may also be greater.

Conclusions 

Our study suggests that statin treatment in patients with ACS may reduce GI bleeding rates. Analyses from large randomized placebo-controlled statin trials and/or prospective randomized trials are warranted to evaluate this additional potential benefit of statins.

Back to Article Outline

References 

1. Law MR, Wald NJ, Rudnicka AR. Quantifying effect of statins on low density lipoprotein cholesterol, ischaemic heart disease, and stroke: systematic review and meta-analysis. BMJ. 2003;326:1423
   • View In Article
   • CrossRef
2. LaRosa JC, He J, Vupputuri S. Effect of statins on risk of coronary disease: a meta-analysis of randomized controlled trials. JAMA. 1999;282:2340–2346
   • View In Article
   • MEDLINE
   • CrossRef
3. Davignon J. Beneficial cardiovascular pleiotropic effects of statins. Circulation. 2004;109(Suppl 1):III39–III43
   • View In Article
4. Wainwright CL. Statins—is there no end to their usefulness?. Cardiovasc Res. 2005;65:296–298
   • View In Article
   • MEDLINE
   • CrossRef
5. Undas A, Celinska-Lowenhoff M, Kaczor M, et al. New nonlipid effects of statins and their clinical relevance in cardiovascular disease. Thromb Haemost. 2004;91:1065–1077
   • View In Article
   • MEDLINE
6. Birnbaum Y, Ye Y, Rosanio S, et al. Prostaglandins mediate the cardioprotective effects of atorvastatin against ischemia-reperfusion injury. Cardiovasc Res. 2005;65:345–355
   • View In Article
   • MEDLINE
   • CrossRef
7. Cannon CP, McCabe CH, Wilcox RG, et al. Oral glycoprotein IIb/IIIa inhibition with orbofiban in patients with unstable coronary syndromes (OPUS-TIMI 16) trial. Circulation. 2000;102:149–156
   • View In Article
8. Moscucci M, Fox KA, Cannon CP, et al. Predictors of major bleeding in acute coronary syndromes: the Global Registry of Acute Coronary Events (GRACE). Eur Heart J. 2003;24:1815–1823
   • View In Article
   • MEDLINE
   • CrossRef
9. Undas A, Brummel-Ziedins KE, Mann KG. Statins and blood coagulation. Arterioscler Thromb Vasc Biol. 2005;25:287–294
   • View In Article
   • CrossRef
10. Schonbeck U, Libby P. Inflammation, immunity, and HMG-CoA reductase inhibitors: statins as antiinflammatory agents?. Circulation. 2004;109(Suppl 1):II18–II26
    • View In Article
11. Chen JC, Huang KC, Wingerd B, et al. HMG-CoA reductase inhibitors induce COX-2 gene expression in murine macrophages: role of MAPK cascades and promoter elements for CREB and C/EBPbeta. Exp Cell Res. 2004;301:305–319
    • View In Article
    • MEDLINE
    • CrossRef
12. Russell RI. Protective effects of the prostaglandins on the gastric mucosa. Am J Med. 1986;81:2–4
    • View In Article
    • Abstract
    • Full-Text PDF (255 KB)
    • CrossRef