European Heart Journal (2000) 21, 457465

doi: 10.1053/euhj.1999.1874, available online at http://www.idealibrary.com on

Randomized comparative trial of triflusal and aspirin

following acute myocardial infarction
J. M. Cruz-Fernandez1, L. Lopez-Bescos2, D. Garca-Dorado3,
V. Lopez Garca-Aranda1, A. Cabades4, L. Martn-Jadraque5, J. A. Velasco6,
A. Castro-Beiras7, F. Torres8, F. Marfil9, E. Navarro10 and Triflusal in Myocardial
Infarction (TIM) Investigators*
1
Department of Cardiology, Hospital Universitario Virgen de la Macarena, Sevilla; 2Department of Cardiology,
Hospital Universitario de Getafe, Madrid; 3Department of Cardiology, Hospital Vall DHebron, Barcelona;
4
Department of Cardiology, Hospital La Fe, Valencia; 5Department of Cardiology, Hospital La Paz, Madrid;
6
Department of Cardiology, Hospital General, Valencia; 7Department of Cardiology, Hospital Juan Canalejo,
La Coruna; 8Laboratorio de Bioestadstica y Epidemiologa, Facultad de Medicina, UAB, Barcelona; 9Department
of Cardiology, Hospital Torrecardenas, Almera; 10Department of Cardiology, Hospital Miguel Servet, Zaragoza,
Spain

Aims To compare the ecacy and tolerability of the P=0278), non-fatal reinfarction (OR [95% CI]: 1577
antiplatelet agent triflusal with aspirin in the prevention [08732848]; P=0131) or revascularization (OR [95% CI]:
of cardiovascular events following acute myocardial 0864 [06441161]; P=0334). Overall, both drugs were
infarction. well tolerated, although there was a trend towards
fewer bleeding episodes with triflusal; significantly fewer
Methods and Results In this double-blind, multicentre, central nervous system bleeding episodes were observed in
sequential design study, patients were randomized within triflusal-treated patients (027% vs 097%; P=0033).
24 h of acute myocardial infarction symptom onset to
receive triflusal 600 mg or aspirin 300 mg once daily for 35 Conclusion Triflusal and aspirin have similar ecacy in
days. The primary end-point was death, non-fatal myocar- preventing further cardiovascular events after acute myo-
dial reinfarction or a non-fatal cerebrovascular event. The cardial infarction, but triflusal showed a more favourable
incidences of these individual outcomes and urgent revas- safety profile. Triflusal significantly reduced the incidence
cularization were secondary end-points. The null hypoth- of non-fatal cerebrovascular events compared with aspirin.
esis of no dierence between treatments in the primary (Eur Heart J 2000; 21: 457465)
combined end-point was accepted with 80% power after  2000 The European Society of Cardiology
recruiting 2124 validated patients (odds ratio (OR) for
failure [95% confidence interval (CI)]: 0882 [06341227]). Key Words: Triflusal, aspirin, cardiovascular events,
Non-fatal cerebrovascular events were significantly less cerebrovascular events, acute myocardial infarction,
frequent with triflusal (OR [95% CI]: 0364 [01460908]; sequential trial.
P=0030). There was no significant dierence between
treatments for death (OR [95% CI]: 0816 [05641179]; See page 430 for the Editorial comment on this article

Introduction provided the first strong evidence of the therapeutic role

of early aspirin treatment, with the risk of vascular death
Aspirin demonstrates considerable clinical benefits in within the first 5 weeks of an acute myocardial infarction
reducing the incidence of cardiovascular events follow- being reduced by 23% in comparison to placebo; non-
ing acute myocardial infarction[1]. The ISIS-2 study fatal cardiovascular events were also significantly
reduced[2].
Revision submitted 5 July 1999, and accepted 21 July 1999. Aspirin is a non-selective inhibitor of cyclo-oxygenase
*See Appendix for a full list of participants. and thus prevents synthesis of thromboxane A2, an
inducer of platelet aggregation[3]. However, aspirin does
Correspondence: Professor J. M. Cruz-Fernandez, Servicio de
Cardiologa, Hospital Universitario Virgen de la Macarena, Avda not target a number of relevant mechanisms of platelet
Dr. Fedriani, s/n, 41009 Sevilla, Spain. aggregation; for example, it does not aect platelet

0195-668X/00/060457+09 $35.00/0  2000 The European Society of Cardiology

458 J. M. Cruz-Ferna ndez et al.

Table 1 Inclusion and exclusion criteria

Inclusion criteria Patients of any gender, aged 1880 years

Signs and symptoms suggestive of AMI appeared less than 24h before inclusion and at least one of the two
following criteria:
(1) elevation of ST >1 mm in two limb leads, or >2 mm in two consecutive precordial leads, or Q wave or new
Q wave (longer than 004 s or wider than 25% of QRS) in at least one ECG performed within first 24 h of
symptoms, or
(2) rise of serum CPK to at least twice the upper limit of normality within first 24 h of symptoms.
Patients written informed consent.
Exclusion criteria Intake of drugs with antiplatelet activity in the previous 15 days
Unfeasibility of oral route for drug administration.
Onset of symptoms more than 24 h before inclusion.
History and/or presence of active peptic ulcer or blood dyscrasia.
Hepatic and/or renal disorders.
History of cerebrovascular haemorrhage.
Clinical record of hypersensitivity or severe adverse drug eects in general, and to salicylates in particular.
Severe, uncontrolled arterial hypertension (diastolic blood pressure >115 mmHg).
Pregnancy, breast-feeding or women not using reliable contraception.
Life expectancy of less than 1 year due to a cause unrelated to the study.
Pathological processes which require long-term treatment with NSAIDs.
History of drug addiction.
Patients aected by AIDS (ARC or known HIV infection).
Participation in other clinical research within previous 3 months.
Any sign of lack of collaboration in adhering to protocol norms.

AMI=acute myocardial infarction; CPK=creatinphosphokinase; NSAIDs=non-steroidal anti-inflammatory drugs; AIDS=acquired

immunodeficiency syndrome; ARC=AIDS related complex; HIV=human immunodeficiency virus.

adhesion or thrombin- or ADP-mediated platelet acti- embolic disease[19,20], as well as coronary angioplasty[21]
vation pathways. In addition, aspirin is associated with and bypass surgery[22]. Clinical data suggest that triflusal
tolerability and side eect concerns, particularly an may be more eective than aspirin in a number of these
increased susceptibility to cerebral and gastrointestinal indications, and that haemorrhagic events may be less
haemorrhage[47]. There is therefore a requirement for frequent than with aspirin[17,19,20,22]. The fact that triflu-
new and more powerful antiplatelet drugs with a more sal is a weaker cyclooxygenase inhibitor than aspirin
favourable tolerability profile. may partly explain this dierence[10].
Triflusal (Disgren, Uriach, Spain) is an antiplatelet The present study was undertaken to determine
drug that is structurally related to aspirin but has several whether triflusal provides an ecacy advantage over
notable dierences in its mechanism of action. Triflusal aspirin in the prevention of vascular complications
exerts multiple eects on platelet aggregation[8]. (death, reinfarction or cerebrovascular events) following
Although triflusal irreversibly blocks platelet cyclooxy- acute myocardial infarction, and to quantify the relative
genase in the same manner as aspirin[9], it inhibits safety benefits of the two drugs. The study used a
endothelial cyclooxygenase only slightly, so that sequential design, which involves continuous monitoring
prostacyclin synthesis is not significantly reduced[10]. of clinical data until evidence for the superiority of one
Both triflusal and its long-lasting active metabolite, drug over the other is attained, or until it is concluded
3-hydroxy-4-trifluoro-methylbenzoic acid, inhibit degra- that a treatment dierence is statistically unlikely. This
dation of platelet and endothelial cell cAMP, thereby design allows patient recruitment to be discontinued as
increasing cAMP levels and blocking intracellular soon a decision based on statistical findings can be
calcium mobilization and platelet-endothelial cell reached.
interactions[8,11]. Triflusal and 3-hydroxy-4-trifluoro-
methylbenzoic acid also inhibit cyclooxygenase-2 Methods
(COX-2) expression, activation of the transcription fac-
tor NF-kappa B, and NF-kappa B-induced inflamma- This double-blind, randomized, sequential, parallel
tory mediators such as the vascular cell adhesion group study was conducted in patients with confirmed
molecule 1 (VCAM-1)[12,13]. In addition, triflusal in- acute myocardial infarction within 24 h of onset of
creases nitric oxide synthesis in neutrophils, resulting in symptoms.
increased vasodilatory potential[14]. Consequently, it can
be suggested that triflusal exerts its antithrombotic eect Study population
by acting on dierent targets involved in the platelet
aggregation and vascular inflammatory processes[11]. The study was carried out at 29 centres in three countries
Triflusal has demonstrated ecacy in a number of (Spain, Portugal and Italy) between February 1993 and
cardiovascular indications, including unstable angina[15], March 1997. Inclusion and exclusion criteria are detailed
peripheral vascular disease[16], stroke[17,18], thrombo- in Table 1.

Eur Heart J, Vol. 21, issue 6, March 2000

 Trial of triflusal and aspirin following AMI 459

6615 Screened Reasons*:

1704 Antiplatelet treatment
598 Outside age range
4340 Not included 500 Not within 24 h of AMI
418 AMI not confirmed
385 Peptic disease
1175 Other reasons
2275 Included

1140 Randomised to aspirin 1135 Randomised to triflusal

1139 Received aspirin 1131 Received triflusal

71 Major protocol violations: 75 Major protocol violations:
53 Lack of inclusion criteria 54 Lack of inclusion criteria
9 Loss of follow-up at 35 days 11 Loss of follow-up at 35 days
9 Other 1068 Validated for 1056 Validated for 10 Other
sequential monitoring sequential monitoring

Figure 1 Trial profile. AMI: acute myocardial infarction. *384 patients had d2 exclusion criteria.

Written informed consent was obtained for every cerebrovascular events and urgent revascularization pro-
patient. The Institutional Ethics Committee for each cedures occurring within the first 35 days of an acute
participating centre approved the protocol. The study myocardial infarction.
was conducted in accordance with the Declaration of Drug tolerability and safety, with particular regard to
Helsinki. bleeding events, were also assessed. Safety was assessed
on the basis of spontaneously reported and solicited
adverse events which were categorized according to their
nature, severity, duration and outcome. In addition,
Drug treatment tolerability was assessed using a patient questionnaire
focussing on signs and symptoms known to be
Patients were randomized to receive triflusal 600 mg or
associated with either treatment. All adverse events
aspirin 300 mg orally once daily for 35 days. Medication
identified by the questionnaire were considered to be
was administered in matched capsules to ensure study
treatment-related.
blinding and was given with the main meal. Treatment
Patients were evaluated at each of three visits: at study
was started as soon as possible after study inclusion and
inclusion, day 15 or discharge (whichever occurred
within 24 h of symptom onset.
sooner) and day 35. At study entry, acute myocardial
Permitted concomitant medications comprised those
infarction was confirmed by ECG and creatine phos-
routinely administered following acute myocardial
phokinase assay, and patients underwent a physical
infarction, including thrombolytics, and these were
examination focusing on the cardiovascular system, as
recorded for each patient. Use of non-steroidal anti-
well as haematological and blood biochemistry analyses
inflammatory drugs and antiplatelet or platelet-active
and chest X-ray. The clinical history and details of the
drugs apart from the study medication was not allowed
current episode were also recorded. At the second and
during the study period.
third visits, clinical interview, physical examination and
follow-up ECG were carried out, and drug ecacy,
tolerability and safety were assessed.
Ecacy and safety measurements Drug capsules were counted at the second and third
visits to ascertain treatment compliance. Interruption of
The primary study end-point was a composite measure the study medication for d3 days was considered a
reflecting death or occurrence of non-fatal myocardial protocol deviation and necessitated withdrawal of the
infarction or non-fatal cerebrovascular events within the patient.
first 35 days after an acute myocardial infarction. The
presence of any individual component of this end-point
was considered as a failure of treatment. Trial design
Non-fatal reinfarction was defined as recurrence dur-
ing the treatment period of signs and symptoms sugges- Estimation of patient numbers required to demonstrate
tive of acute myocardial infarction, confirmed by ECG a clinically relevant dierence between treatments was
and/or enzyme analyses within the following 72 h. A cer- based on findings from the ISIS-2 study, in which the
ebrovascular event was defined as an abnormal neuro- combined incidence of reinfarction, cerebrovascular
logical finding due to an ischaemic or haemorrhagic event or vascular death in the 5 weeks following acute
cerebral event, confirmed by computed tomography. myocardial infarction was 106% with aspirin (compared
Secondary end-points were defined individually as with 144% for controls)[2]. We assumed a 10% incidence
the incidence of death, non-fatal reinfarction, non-fatal of the primary end-point in the aspirin group, and

Eur Heart J, Vol. 21, issue 6, March 2000

460 J. M. Cruz-Ferna ndez et al.

considered a 25% relative reduction in the incidence of Table 2 Baseline characteristics of validated patients*
the primary end-point to be a clinically relevant dier-
ence between treatments. The study was anticipated to Aspirin Triflusal
Characteristic
have an 80% power to detect a treatment dierence of (n=1068) (n=1056)
this magnitude at the 5% significance level (two-tailed).
A sequential probability ratio test design with trunc- Mean ageSD (year) 60961167 60741165
ation at 5305 patients[23] was selected for the study. Male (%) 8099 8267
Mean weightSD (kg) 74811084 74691126
According to statistical considerations (type I and II Clinical presentation (%)
errors and estimated risk reduction) the boundaries of Pain suggestive of AMI 9560 9678
the continuation region were established using Planning CPK >2 upper limit of normal 9501 9620
and Evaluating Sequential Trials software (PEST ver- New Q wave in ECG 7303 7311
sion 2.0 and 3.0)[2325]. As the study progressed, an Confirmed AMI diagnosis 10000 10000
Risk factors (%)
estimator of the accumulated dierence between treat-
Current smokers 5642 5389
ments (Z) was plotted against a measure of the accumu- Hypercholesterolaemia 3850 3702
lated information (V). If the plot of the accumulated Diabetes 2144 2257
data crossed the inner or outer boundaries of the Hypertension 3932 3725
continuation region the trial was terminated and the null Previous cardiovascular disease (%)
hypothesis accepted or rejected, respectively. Myocardial infarction 614 563
Heart failure 650 692
Sequential monitoring was carried out by planned Angina 4042 3939
inspections of data, which were defined a priori after Cerebrovascular event 253 275
recruitment of the first 1000 patients and then after Revascularization 047 066
additional increments of 250 patients, allowing for slight AMI localization (%)
deviations by calendar. A Christmas tree correction Anterior 3502 3570
Inferior 4850 4905
of the boundary limits, allowing adjustment for Other 655 720
discontinuous monitoring, was applied[23]. Non-Q wave 993 805
At the end of the study, a maximum likelihood
estimator (
|) of the dierence between treatments was AMI=acute myocardial infarction; SD=standard deviation;
derived from the final value of (Z/V), and adjusted for CPK=creatine phosphokinase.
bias (
x) associated with the sequential design. *No significant dierence between aspirin and triflusal groups for
any baseline characteristic.
The secondary end-point components of the primary
end-point were also adjusted. All subsequent reference
to proportions and odds ratios relates to the adjusted Tolerability and safety were quantified in terms of the
data. relative frequency and intensity of secondary eects and
adverse events in all patients who received at least one
dose of study medication.
Inter-group comparisons were performed using
Data handling and statistical analyses Students t test for quantitative variables, the
MannWhitney U test for ordinal variables and
Prior to statistical analysis, data obtained at each centre either the Chi-squared test or Fishers exact test for
were confirmed by a supervisory committee, which con- categorical variables. All analyses were two-tailed, with
tinuously validated all cases included in the sequential a significance level of 5%.
analysis as they were completed, and verified all reported
end-points as soon as these were achieved.
Treatment ecacy was determined by applying the Results
sequential approach to the validated patient population,
which was defined as all correctly diagnosed and ran- Study population
domized patients with no exclusion criteria, who com-
menced the study medication within 24 h of symptom Of 6615 patients screened, 4340 were not included in the
onset and for whom outcome data at day 35 were study (Fig. 1). The most frequent reasons for non-
available. Results were expressed as the risk estimate of eligibility were current treatment with drugs aecting
failure for the triflusal group relative to the aspirin platelet function, age >80 years and delay in acute
group (odds ratio [OR] with 95% confidence interval myocardial infarction diagnosis >24 h from symptom
[CI]). For all secondary variables, hypothesis testing was onset. Inclusion criteria were fulfilled by 2275 patients,
carried out using the conventional Wald test (estimator/ who were randomized to receive either aspirin (1140
standard error). A simulation of the sequential monitor- patients) or triflusal (1135 patients). Drug safety and
ing was performed on all randomized patients to identify tolerability was assessed from the 2270 patients who
the eects of any case selection bias on the primary received at least one dose of study medication. The
ecacy assessment. For the purposes of this simulation, supervisory committee excluded from sequential moni-
patients lost to follow-up at day 35 were considered toring those patients with major protocol deviations
event-free. (146 cases), leaving a validated population of 2124

Eur Heart J, Vol. 21, issue 6, March 2000

 Trial of triflusal and aspirin following AMI 461

Table 3 Concomitant cardiovascular drugs* administration of prohibited medication (1429%),

adverse events (1315%), non-compliance (1179%) or
Aspirin Triflusal other causes (1633%). No dierence between treatment
Treatment (%) (%) groups was observed in the reason for withdrawal.
(n=1068) (n=1056)
However, information on study end-points at day 35
was available for all validated patients.
Heparin 8801 8892
Nitrates 7921 8021
Fibrinolytic agents 7079 7074
Beta adrenergic blocking agents 4199 4091
ACE inhibitors 3065 2718 Ecacy
Antiarrhythmic drugs 2240 2112
Diuretics 1846 1667
Calcium channel blocking agents 1138 1184 Sequential monitoring stopped after the fifth sequential
Non-glycoside inotropic agents 909 748 data inspection. At this point the study trajectory
Digitalis glycosides 582 445 crossed the inner limit of the continuation region, per-
mitting acceptance of the null hypothesis (no significant
*No significant dierence between aspirin and triflusal groups for treatment dierence) with a total of 2124 validated
any concomitant treatment. patients recruited to the study (Fig. 2, Table 4).
The primary end-point occurred in 105 of 1068
patients that was used to determine primary and patients receiving aspirin and 99 of 1056 patients receiv-
secondary study end-points. ing triflusal, giving an adjusted OR estimate [95% CI]
The baseline characteristics of the patient population for treatment failure of 0882 [06341227] (P=0582)
are summarized in Tables 2 and 3. Female patients were (Table 5). The risk of non-fatal cerebrovascular events
significantly older than male patients (mean 6682 vs was significantly lower with triflusal than with aspirin
5952 years; P<00001). There were no significant (P=003). There were no significant dierences
dierences in baseline characteristics between the two between treatments in risk of death, non-fatal MI or
treatment groups. revascularization procedures (Table 5).
A total of 441 validated patients failed to complete the Results of analyses of data from all randomized
35-day treatment period for reasons other than death, patients were consistent with those from validated
the reasons being revascularization procedures (4444%), patients only; sequential monitoring would have been

30 Aspirin better

20

10
Z

20 40 60 80 100
V
10

20

30 Triflusal better

Figure 2 Sequential monitoring trajectory. The ascending and descending solid lines
represent the predefined boundary limits for significant superiority (P=005) of aspirin
or triflusal, respectively. The innermost, dashed segments represent the predefined
boundary limits for accepting the null hypothesis of no therapeutic dierence with a
power of 80%. Five data inspections (x) were performed. The five data points were used
to define the Christmas tree corrections for boundaries (dotted lines), which adjust for
discontinuous monitoring. The corrected innermost, descending boundary was crossed
on the fifth data inspection (V=461; Z=
242). The study accepted the null
hypothesis at this stage, after inclusion of 2124 validated patients.

Eur Heart J, Vol. 21, issue 6, March 2000

462 J. M. Cruz-Ferna ndez et al.

Table 4 Sequential monitoring of end-points

End-points Validated Christmas tree

Data
achieved patients V Z corrected Stop?
inspection
(cumulative) (cumulative) boundary

1 92 889 2062
664 131 No

2 116 1177 2614
505 135 No
3 130 1365 2941
386 094 No
4 158 1618 3564
512
086 No
5 204 2124 4610
242
363 Yesa

a
Null hypothesis accepted.

Table 5 Incidence of primary and secondary end-points in patients treated with aspirin 300 mg or triflusal 600 mg once
daily

Proportion (%)
Incidence
[95% confidence interval]a Odds ratio
Variable P
Aspirin Triflusal Aspirin Triflusal [95% confidence interval]c
(n=1068) (n=1056)

Primary end-pointb 105 99 1015 [8721154] 906 [7641049] 0882 [06341227] 0582
Secondary end-points
Death 79 69 762 [644877] 631 [515750] 0816 [05641179] 0278
Non-fatal AMI 18 30 177 [120241] 276 [211334] 1577 [08732848] 0131
Non-fatal cerebrovascular event 14 5 131 [094155] 048 [023085] 0364 [01460908] 0030
Revascularization 105 91 983d 862d 0864 [06641161] 0334

a
Corrected for sequential analysis.
b
Combined end-point of death, non-fatal AMI or non-fatal cerebrovascular event.
c
Triflusal relative to aspirin; corrected by maximum likelihood estimator.
d
Not adjusted for sequential analysis.
AMI=acute myocardial infarction.

stopped in the same inspection and would also have led Discussion
to acceptance of the null hypothesis (OR [95% CI]: 1011
[07191422]; P=0602). The most eective means of limiting infarct size is to
restore vessel patency rapidly, fully (TIMI 3 flow) and
without re-occlusion. This is usually achieved through
intravenous fibrinolysis and, increasingly, primary angi-
Tolerability oplasty; in all cases an anti-thrombotic regimen of
anti-thrombin and antiplatelet agents is of the greatest
Adverse events possibly related to study treatment were importance. Many clinical trials have been carried out to
reported in 771 patients. Considering all treatment- investigate dierent drug combinations to enhance
related adverse events, the most frequently aected body reperfusion, but treatment intensity has always been
systems were the gastrointestinal tract (primarily dys- limited by the increased risk of haemorrhagic complica-
pepsia, constipation and flatulence) and the central and tions, particularly cerebral haemorrhage. Following the
peripheral nervous system, with similar incidences in results of the ISIS-2 study[2], aspirin is now widely
both treatment groups (Table 6). However, triflusal was established in such treatments. However, to date no
associated with a significantly lower incidence of central clinical trials have been published comparing the ecacy
nervous system-associated bleeding (P=0033), and with of aspirin and other antiplatelet drugs directly in the
a non-significant trend towards less frequent bleeding in acute stage of acute myocardial infarction.
all organs (P=0090) (Table 6). The present study was undertaken to compare the
A total of 60 patients were withdrawn due to adverse ecacy and tolerability of triflusal vs aspirin in this
events (34 patients in the aspirin group, 26 in the triflusal setting, and showed that following an acute myocardial
group). Twenty-six patients (14 in the aspirin group and infarction there was no significant dierence in the
12 in the triflusal group) experienced severe adverse primary combined end-point (death, non-fatal acute
events that were possibly related to the study medication myocardial infarction or non-fatal cerebrovascular
in terms of temporal sequence. events) between the triflusal and aspirin treatment

Eur Heart J, Vol. 21, issue 6, March 2000

 Trial of triflusal and aspirin following AMI 463

Table 6 Incidence of adverse events possibly related to treatment

Number of patients (%)

Adverse event P
Aspirin (n=1139) Triflusal (n=1131)

Any 386 (3389) 385 (3404)

Gastrointestinal 285 (2502) 303 (2679) ns
Central and peripheral nervous system 94 (825) 101 (893) ns
Body as a whole 62 (544) 62 (548) ns
Respiratory system 41 (360) 43 (380) ns
Platelet, bleeding and clotting disorders 22 (193) 16 (141) ns
Skin and appendages 21 (184) 17 (150) ns
Hearing and vestibular disorders 13 (114) 14 (124) ns
Bleeding events
Any bleeding* 41 (360) 27 (239) ns
Gastrointestinal 17 (149) 10 (088) ns
Central nervous system 11 (097) 3 (027) 003
Cutaneous 10 (088) 7 (062) ns
Urinary tract 4 (035) 2 (018) ns

ns=not significant: (P>005).

*Four patients in aspirin group and one patient in triflusal group had two bleeding episodes.

groups at day 35. The adjusted proportions of end-point A feature of sequential clinical trial design is close and
occurrence (906% for triflusal and 1015% for aspirin) rigorous monitoring of patient data on a continual basis.
corresponded to an adjusted risk reduction of 118% This has the advantage of allowing study termination as
with triflusal treatment. However, the risk of a non-fatal soon as sucient information has been gathered to
cerebrovascular event was 63% lower with triflusal than permit a statistically based conclusion on treatment
with aspirin (05% vs 13%). There was also no signifi- outcome. The current study fulfilled the recommended
cant dierence between treatment groups with regard to criteria for adoption of a sequential design: a large
three of the secondary end-points: death, non-fatal acute sample size (sample size required by classical approach:
myocardial infarction and revascularization 4170 patients), gradual recruitment of patients over a
The patient population included in the present study long period of time, a single and well-defined primary
consisted overwhelmingly of patients experiencing their ecacy measure, and a short follow-up period for each
first acute myocardial infarction (941%), and these patient. This design has proved useful in the present
patients would be expected to have a lower rate of study, as reflected by the fact that the trial was termi-
complications and a better prognosis than the typical nated after recruiting 2124 validated patients. As confir-
infarct population. Indeed, a comparison of the TIM mation of the absence of bias in the patient selection
population with that of the RESCATE (Recursos procedure, a simulated sequential analysis of all ran-
Empleados en el Sndrome Coronario Agudo y Tiempos domized patients showed that the study would have
de Espera) study, which was specifically limited to been stopped at the same point under a strict intent-to-
patients presenting with first acute myocardial infarc- treat approach.
tion[26], reveals close similarities in terms of gender
balance, incidence of previous cardiovascular disease
and personal risk factors. Limitation of the TIM popu-
lation to those patients under 80 years of age with Conclusions
predominantly first acute myocardial infarction provides
for greater homogeneity and might be expected to The TIM study shows that treatment with triflusal in the
improve the accuracy of excess risk estimates. acute phase of myocardial infarction has a similar
The safety and tolerability of both drugs was good: ecacy to aspirin in the prevention of cardiovascular
only 264% of patients withdrew from the study due to events (death, non-fatal myocardial infarction or non-
adverse events, and the incidence of severe adverse fatal cerebrovascular event). However, triflusal is associ-
events was low. The most frequently reported adverse ated with a significantly lower incidence of non-fatal
events were those aecting the gastrointestinal and cen- cerebrovascular events and cerebral haemorrhage.
tral and peripheral nervous systems, a finding possibly Triflusal currently appears to constitute a valid alterna-
attributable in part to the use of a questionnaire focus- tive to aspirin in the acute phase of acute myocardial
ing on these known adverse eects. The incidence of infarction, particularly in patients at increased risk of
adverse bleeding events related to the central nervous stroke or haemorrhage.
system was over three-fold greater with aspirin than with This study was sponsored by Uriach (Spain), Poli Industria
triflusal, and there were fewer bleeding events in all other Chimica (Italy) and Tecnifar Industria Tecnica Farmaceutica
body systems with triflusal than with aspirin. (Portugal).

Eur Heart J, Vol. 21, issue 6, March 2000

464 J. M. Cruz-Ferna ndez et al.

References [18] Smirne S, Ferini-Strambi L, Cucinotta D et al. Triflusal and

prevention of cerebrovascular attacks: a double-blind clinical
study vs ASA [abstract 53]. J Neurol 1994; 241 (Suppl 1):
[1] Antiplatelet Trialists Collaboration. Collaborative overview
S130.
of randomised trials of antiplatelet therapy I: Prevention of
[19] Putz P, Buyse H, Delvaux D et al. Triflusal versus acetylsali-
death, myocardial infarction, and stroke by prolonged
cylic acid: a double-blind study for the prophylaxis of deep
antiplatelet therapy in various categories of patients. BMJ
vein thrombosis after hip surgery. Acta Chir Belg 1991; 91:
1994; 308: 81106.
26976.
[2] ISIS-2 Collaborative Group. Randomised trial of intravenous
[20] Monreal M, Lafoz E, Roca J et al. Platelet count, antiplatelet
streptokinase, oral aspirin, both, or neither among 17,187
therapy and pulmonary embolism a prospective study in
cases of suspected acute myocardial infarction: ISIS-2. Lancet patients with hip surgery. Thromb Haemost 1995; 73: 3805.
1988; ii: 34960. [21] Masotti M, Tura A, Crexells C et al. Antiplatelet agents and
[3] Patrono C. Aspirin as an antiplatelet drug. N Engl J Med their eect on complications during or soon after percu-
1994; 330: 128794. taneous transluminal coronary angioplasty. J Int Med Res
[4] Roderick PJ, Wilkes HC, Meade TW. The gastrointestinal 1991; 19: 4148.
toxicity of aspirin: an overview of randomised controlled [22] Guiteras P, Altamiras J, Aris A et al. Prevention of aorto-
trials. Br J Clin Pharmacol 1993; 35: 21926. coronary vein-graft attrition with low-dose aspirin and triflu-
[5] Steering committee of the physicians health study research sal, both associated with dipyridamole: a randomised, double-
group. Final report on the aspirin component of the ongoing blind, placebo-controlled trial. Eur Heart J 1989; 10: 15967.
physicians health study. N Engl J Med 1989; 321: 12935. [23] Whitehead J. The design and analysis of sequential clinical
[6] Hirsh Jy, Dalen JE, Fuster V, Harker L, Patrono C, Roth GJ. trials, 2nd edition. Chichester: Ellis Horwood, 1992.
Aspirin and other platelet active drugs: the relationship [24] Whitehead J, Brunier H. PEST 2.0 Operating Manual.
between dose, eectiveness, and side-eects. Chest 1995; 108 Reading, U.K.: Reading University, 1989.
(Suppl): S24757. [25] Brunier H, Whitehead J. PEST 3.0 Operating Manual.
[7] Boysen G, Boss AH, Odum N, Olsen JS. Prolongation of Reading, U.K.: Reading University, 1993.
bleeding time and inhibition of platelet aggregation by low [26] Marrugat J, Sala J, Masia R et al. Mortality dierences
dose acetylsalicylic acid in patients with cerebrovascular between men and women following first myocardial
disease. Stroke 1984; 15: 2413. infarction. JAMA 1998; 280: 14059.
[8] de la Cruz JP, Villalobos MA, Garcia PJ et al. Eects of
triflusal and its main metabolite HTB on platelet interaction
with subendothelium in healthy volunteers. Eur J Clin
Pharmacol 1995; 47: 497502. Appendix
[9] Albors M, Castellarnau C, Vila L, Sola J, Rutllant ML.
Inhibition of thromboxane production and platelet function TIM study organisation
by triflusal in healthy volunteers. Rev Pharmacol Clin Exp
1987; 4: 116.
[10] de la Cruz JP, Mata JM, Sa nchez de la Cuesta F. Triflusal
Chairman: J. M. Cruz-Ferna ndez
versus aspirin on the inhibition of human platelet and vascular Validation Committee: L. Plaza, J. Mora (Instituto de
cyclooxygenase. Gen Pharmacol 1992; 23 (3): 297300. Salud Carlos III, Madrid)
[11] Garca-Rafanell J, Ramis J, Go mez L, Forn J. Eect of Independent Statistical Centre: Laboratorio de Bioesta-
triflusal and other salicylic acid derivatives on cyclic AMP dstica y Epidemiologa. Universitat Auto`noma de
levels in rat platelets. Arch Int Pharmacodyn Ther 1986; 284:
15565.
Barcelona. F. Torres, V. Moreno, M. Horas
[12] Ferna ndez de Arriba A, Cavalcanti F, Miralles A et al. Independent Monitoring Centre: Phoenix International
Inhibition of cyclooxygenase-2 expression by Spain (ITEM-PHARMA)
4-trifluoromethyl derivatives of salicylate, triflusal, and its Medical Coordinator: J. C. Navas (Uriach)
deacetylated metabolite 2-hydroxy-4-trifluoromethylbenzoic Operations Group: I. Izquierdo, C. Pontes (Uriach)
acid. Mol Pharmacol 1999; 55: 75360.
[13] Bayo n Y, Alonso A, Sa nchez Crespo M. 4-trifluoromethyl
derivatives of salicylate, triflusal and its main metabolite
2-hydroxy-4-trifluoromethylbenzoic acid are potent inhibitors
of nuclear factor kB activation. Br J Pharmacol 1999; 126:
Collaborating clinical centres (by country
135966. and number of patients enrolled)
[14] Sa nchez de Miguel L, del Prado Sierra M, Lo pez Farre A
et al. Comparative eects of triflusal and acetylsalicylic acid Spain (2121): Hospital Universitario Virgen Macarena,
on nitric oxide generation by human neutrophils. Eur J Sevilla (258): V. Lo pez G-Aranda, M. Gonza lez, J. T.
Pharmacol 1998; 343: 5765.
[15] Plaza L, Lo pez-Besco s L, Martn-Jadraque L et al. Protective
Garca, J. C. Garca-Rubira, J. M. Leo n; Hospital
eect of triflusal against acute myocardial infarction in Torreca rdenas, Almera (226): J. F. Marfil, V.
patients with unstable angina: results of a Spanish multicentre Rodriguez, J. M. Gonza lez-Gay, J. C. Martn, R.
trial. Grupo de Estudio del Triflusal en la Angina Inestable. Martos; La Fe, Valencia (173): A. Cabade s, I. Ceniceros,
Cardiology 1993; 82: 38898. L. Go mez, R. Pitarch, L. Almenar, M. Palencia; Clnico
[16] Auteri A, Angaroni A, Borgatti E et al. Triflusal in the
treatment of patients with chronic peripheral arteriopathy:
Universitario, Salamanca (167): P. Pabo n; La Paz,
multicentre double-blind clinical study vs. placebo. Int J Clin Madrid (165): L. Martn Jadraque, I. Gonza lez, I.
Pharmacol Res 1995; 15: 5763. Rolda n, E. Armada, M. Mateos, J. M. de Alba, J. A.
[17] Matas-Guiu L, Alvarez-Sabn J, Codina A. Comparative Go mez, J. Marcos, P. Robles, G. Galeote, J. Fuertes;
study of the eect of low-dosage acetylsalicylic acid and Hospital General Universitario, Valencia (160): J. A.
triflusal in the prevention of cardiovascular events among
young adults with ischaemic cerebrovascular disease. Long-
Velasco, I. Echanove, J. V. Vilar, F. Atienza, J. L.
term outcome of young stroke patients. Rev Neurol 1997; 25: Pe rez-Bosca , R. Paya, J. L. Colomer; Hospital Universi-
166972. tario de Getafe, Madrid (127): L. Lo pez-Besco s, C.

Eur Heart J, Vol. 21, issue 6, March 2000

 Trial of triflusal and aspirin following AMI 465

Kallmeyer, J. M. Grande, L Melgares, M. A. Montero, Lesma, P. Castro; Complejo Hospitalario del INSALUD,
P. L. Talavera, J. Palma, F. de la Concepcio n; Hospital Toledo (19): J. Sa nchez-Domnguez, L. Martos, M.
General de Galicia, Santiago (126): A. Varela, M. Martnez; Hospital Clnico Universitario, Zaragoza (10):
Pedreira, A. Virgo s, M. A. Bandn, F. Ferna ndez, M. I. Ferreira, A. Millastre, G. Rodrigo, J. L. Calderero,
Jaquet, M. Gil de la Pen a; Hospital Universitario Puerto M. A. Suarez, R. Fajardo, A. del Ro, G. Pacheco, R.
Real, Ca diz (118): V. Gonza lez (), P. Martnez, A. di Izquierdo, J. I. Sa nchez, F. Roncale s, E. Civeira; Hospi-
Mateo, J. Gil; Hospital Vall dHebro n, Barcelona (112): tal Arnau de Vilanova, Valencia (8): M. France s, C.
J. Soler-Soler, D. Garca-Dorado, J. Figueras, R. M. Tormo, F. Fajarne s, M. Garca, A. Herva s, L. Corte s,
Lido n, M. T. Santos, E. Nieto; Miguel Servet, Zaragoza J. Flores; Hospital Universitario, Valladolid (7): F.
(112): E. Navarro, A. I. Loriz, F. Civera; Hospital de Ferna ndez-Avile s, J. Bermejo; Hospital de Mo stoles,
Navarra, Pamplona (89): E. de los Arcos, M. A. Imizcoz, Madrid (7): P. Galdo s, M. Ma s, N. Franco, J. Goizueta,
J. Berjo n, J. R. Cardona; Complejo Hospitalario del J. Lucas, A. Alcala ; Hospital Severo Ochoa, Madrid (6):
INSALUD, Leo n (72): E. Simarro, M. L. Fidalgo, I. F. del Nogal, M. J. Jime nez, S. Temprano, J. Suarez, A.
Iglesias, P. Prieto, J. Bayo n, N. Alonso, J. C. Obaya, M. Algora; Hospital Dr Peset, Valencia (5): F. Valls, L. L.
A. Rodrguez; Hospital Infanta Cristina, Badajoz (57): Miralles, V. Valentn, F. Garce s, G. Alama n; Hospital
M. del Aguila, A. Pijierro, L. Martnez, J. Arrobas, A. General de Especialidades, Jae n (4): L. Rucabado,
Zimbora, J. R. Lo pez, J. Ferna ndez, J. M. Garca, J. M. F. Ruiz, E. Castillo, C. Pagola. Italy (95): University
Alvarez, G. Ferna ndez, J. C. Altozano; Hospital Juan Hospital St Orsola, Bologna (56): P. Puddu, A. Muscari,
Canalejo, La Corun a (50): A. Castro-Beiras, H. Alvarez, G. Massarelli; Policlinico, Padua (39): S. Dalla Volta,
P. Lado, R. A. Calvin o, R. Manteiga, S. Calvo, R. F. Maddalena, E. Bertaglia, A. Bianchi. Portugal (59):
Va zquez, A. Otero, J. M. Gulas, P. Jime nez, J. M. Hospital Distrital, Evora (47): A. Jara, F. Araujo;
Va zquez; Hospital Universitario Valme, Sevilla (43): L. Hospital St. Marta, Lisboa (11): A. Mata-Antunes;
Pastor, R. Carillo, R. Va zquez, J. Beltran, C. Leo n, A. Hospital St Cruz, Lisboa (1): R. Seabra-Gomes.

Eur Heart J, Vol. 21, issue 6, March 2000