Outcome Results of the Fosinopril Versus Amlodipine - Diabetes Care

P a t h o p h y s i o l o g y / C o m p l i c a t i o n sN A L A R T I C L E

Outcome Results of the Fosinopril VersusAmlodipine Cardiovascular EventsRandomized Trial (FACET) in PatientsWith Hypertension and NIDDMPATRIZIO TATTI, MDMARCO PAHOR, MDROBERT P. BYINGTON, PHDPATRIZIA DI MAURO, MD

RLCCARDO GUARISCO, MDGLOVANNA STROLLO, MDFELICE STROLLO, MD

OBJECTIVE — ACE inhibitors and calcium antagonists may favorably affect serum lipidsand glucose metabolism. The primary aim of the Fosinopril Versus Amlodipine Cardiovascu-lar Events Randomized Trial (FACET) was to compare the effects of fosinopril and amlodip-ine on serum lipids and diabetes control in NIDDM patients with hypertension. Prospectivelydefined cardiovascular events were assessed as secondary outcomes.

RESEARCH DESIGN AND METHODS — Inclusion criteria included a diagnosis ofNIDDM and hypertension (systolic blood pressure of >140 mmHg or diastolic blood pressureof >90 mmHg). Exclusion criteria included a history of coronary heart disease or stroke, serumcreatinine >1.5 mg/dl, albuminuria >40 ug/min, and use of lipid-lowering drugs, aspirin, or anti-hypertensive agents other than beta-blockers or diuretics. A total of 380 hypertensive diabeticswere randomly assigned to open-label fosinopril (20 mg/day) or amlodipine (10 mg/day) and fol-lowed for up to 3.5 years. If blood pressure was not controlled, the other study drug was added.

RESULTS — Both treatments were effective in lowering blood pressure. At the end of follow-up, between the two groups there was no significant difference in total serum cholesterol, HDLcholesterol, HbAlc, fasting serum glucose, or plasma insulin. The patients receiving fosinoprilhad a significantly lower risk of the combined outcome of acute myocardial infarction, stroke,or hospitalized angina than those receiving amlodipine (14/189 vs. 27/191; hazards ratio = 0.49,95% CI = 0.26-0.95).

CONCLUSIONS — Fosinopril and amlodipine had similar effects on biochemical meas-ures, but the patients randomized to fosinopril had a significantly lower risk of major vascu-lar events, compared with the patients randomized to amlodipine.

Hypertension is a common conditionamong diabetic patients, and hyper-tensive diabetic patients are about

twice as likely to experience cardiovascularevents as nondiabetic counterparts (1).

NIDDM is accompanied by changes ininsulin sensitivity and lipid metabolismthat increase cardiovascular risk. Severalclinical trials suggested that, in addition tolowering blood pressure, the new antihy-

From the Centro Diabetico Ospedale di Marino (PT., PD.M., R.G.), Marino, Italy; the Department of Preven-tive Medicine (M.R), University of Tennessee, Memphis, Tennessee; the Department of Public Health Sciences(R.RB.), Wake Forest University School of Medicine, Winston-Salem, North Carolina; and the Unita' Opera-tiva Endocrinologica (G.S., FS.), Istituto Nazionale di Ricerca e Cura per gli Anziani (INRCA), Rome, Italy.

Address correspondence and reprint requests to Marco Pahor, MD, Department of Preventive Medicine,University of Tennessee, 66 North Pauline, Suite 633, Memphis, TN 38105. E-mail: [email protected].

Received for publication 24 September 1997 and accepted in revised form 26 November 1997.R.P.B. has received honoraria for speaking engagements and consulting fees from Bristol-Myers Squibb.

M.R has received honoraria for speaking engagements from Bristol-Myers Squibb and Merck.Abbreviations: ECG, electrocardiogram; FACET, Fosinopril Versus Amlodipine Cardiovascular Events

Randomized Trial; HR, hazard ratio; PAI, plasminogen activator inhibitor; SHEP, Systolic Hypertension in theElderly Program.

pertensive agents, ACE inhibitors and long-acting calcium antagonists, might exertdirect beneficial metabolic effects on glu-cose tolerance and serum lipids in patientswith hypertension, diabetes, or renal dys-function (2-7). However, these findingshave not been confirmed by others (8,9)and one comprehensive review suggestedthat ACE inhibitors, but not calcium antag-onists, reduce serum lipids (10). The lipideffects associated with antihypertensivetreatment in NIDDM patients remain con-troversial. It is not known which treatmentmight be most effective in improving theserum lipid profile and to what extentchanges in laboratory measures may trans-late into clinical benefits.

The primary aim of the FosinoprilAmlodipine Cardiovascular Events Trial(FACET) was to assess treatment-related dif-ferences in serum lipids and diabetes controlin hypertensive patients with NIDDM. Thepatients were randomly given an ACEinhibitor or a long-acting calcium antagonistas the first-line agent. If blood pressure wasnot controlled, the other study drug wasadded to the initial treatment. Cardiovascu-lar events were collected as secondary out-comes. Additional outcomes were bloodpressure control and renal function status.

RESEARCH DESIGN ANDMETHODS— The FACET was anopen-label, randomized prospective trialcomparing fosinopril to amlodipine inhypertensive people with NIDDM. Patientswere men and women recruited from 1January through 31 December 1992 froman outpatient diabetes clinic in Marino, Italy.NIDDM was defined as having fastingserum glucose >140 mg/dl when thepatient was untreated and having norequirement of insulin as an initial treat-ment for diabetes. All patients were seen inthe clinic three or more times during a 3-month period before randomization.Hypertension was diagnosed as systolicblood pressure (sBP) > 140 mmHg or dias-tolic blood pressure (dBP) >90 mmHgmeasured in at least three consecutive visits,

DIABETES CARE, VOLUME 21, NUMBER 4, APRIL 1998 597

Dow

nloaded from http://diabetesjournals.org/care/article-pdf/21/4/597/586069/21-4-597.pdf by guest on 07 January 2022

Outcome results of the FACET

or sBP >160 mmHg or dBP >95 mmHgmeasured in at least two consecutive ornonconsecutive visits. Duration of hyper-tension was less than 1 year. Exclusion cri-teria included a history of coronary heartdisease, stroke, or any other morbid condi-tion with poor prognosis; a serum creatininelevel of >1.5 mg/dl; a microalbuminurialevel of >40 ug/min; and the use of lipid-lowering drugs, aspirin, or antihyperten-sive agents other than diuretics andbeta-blockers. The patients treated withdiuretics or beta-blockers were evaluated forblood pressure inclusion criteria after awashout period of 2 weeks. Diuretics orbeta-blockers were not used in the trial.

At baseline, a physical exam was per-formed by a single study physician (PX).Blood pressure was measured in the morn-ing while the patient was sitting. Fastingblood samples were drawn by venipunc-ture with the patient in a sitting positionand analyzed in the central clinic labora-tory. The biochemical assays included fast-ing serum glucose, serum creatinine,plasma insulin, HbAlc, total cholesterol,HDL cholesterol, triglycerides, fibrinogen,and microalbuminuria. Electrocardiograms(ECGs) were performed during the clinicvisit. Medical history, review of medicalcharts, and ECGs were used to excludepatients with coronary heart disease.

By using a computer-generated randomnumber sequence obtained from an investi-gator who was not involved in patientrecruitment, patients were randomized toeither fosinopril 20 mg/day given in themorning or amlodipine 10 mg/day given inthe evening. The study drugs were admin-istered open-label and were part of thepatients scheduled treatment. The goalblood pressure was defined as sBP <140mmHg and dBP ̂ 90 mmHg, or a decrease>20 mmHg of blood pressure if sBP was> 160 mmHg or dBP was > 110 mmHg. Ifblood pressure was not controlled onmonotherapy, the other study drug wasadded at full dose. Compliance with treat-ment was verified by self-report and by apill count from the medication containersbrought to the clinic by the participants.When the patients were on the study treat-ments, the compliance rate averaged >80%and was the same in both groups. Bloodpressure control was used as an additionalindicator for verifying compliance.

During the first 6 months after ran-domization, clinic visits and blood pressuremeasurements were scheduled monthlyuntil blood pressure goals were reached.

The participants were followed for clinicvisits every 6 months thereafter. Completeblood biochemical examination tests andurine assays for microalbuminuria were per-formed at least annually in the same cliniclaboratory The trial was completed on 30June 1995. Events were monitored duringthe study by asking the patients if they werehospitalized or had any other events sincethe last visit. The patients who did notreturn to the clinic were contacted person-ally or through proxies. After the comple-tion of the study, all patients or their proxieswere recontacted to ensure the complete-ness of follow-up and event ascertainmentin both arms of the trial. Hospital and med-ical records were obtained, and all eventswere independently adjudicated by aninternist and a cardiologist who had nopatient contact or any other role in thestudy and who were blinded to the assignedtreatment. The adjudication process usedpredetermined standardized algorithmsbased on clinical history, laboratory exams,and procedures. Hospitalized angina wasdefined as follows: the patient must havebeen hospitalized overnight with an admis-sion diagnosis of chest pain and either 1)electrocardiographic evidence or positivethallium-201 (or equivalent) myocardialstress test or new coronary angiographicfindings of angiographic significant coro-nary disease or 2) chest pain was typical,reproducible, or similar to previous docu-mented episodes of myocardial ischemia.Acute myocardial infarction was definedaccording to a standardized algorithm thatused information from clinical history of theevent, predefined electrocardiographicchanges, cardiac enzymes, and autopsy.Stroke was defined according to a stan-dardized algorithm that used informationfrom clinical history of the event, comput-erized tomography scan, lumbar puncture,surgery, and autopsy. The prospectivelydefined events were categorized as follows:I) all-cause mortality, 2) fatal or nonfatalstroke, 3) fatal or nonfatal acute myocardialinfarction, 4) hospitalized angina, 5) anymajor vascular event described in 2,3, or 4,6) coronary artery bypass, 7) percutaneoustransluminal coronary angioplasty, 8) anymajor vascular event or proceduredescribed in 5, 6 or 7, 9) other cardiovas-cular events or procedures, and 10) cancer.Cancers were histologically documented.

Data analysesThe trial had 80% power (a = 0.05) todetect a 10% difference in total cholesterol

between fosinopril and amlodipine. Stop-ping rules were not warranted. Analysis ofvariance (ANOVA) and the x2 test wereused to assess differences between meansand proportions, respectively. Data meas-ured at the last scheduled annual visit whenthe patient was seen in the clinic were com-pared with baseline values and betweentreatments. Changes from baseline are givenas 95% CIs. This change is significant at P^0.05 when the CI does not contain zero.

In analyses of events, the time to thefirst event was used. Follow-up ended on30 June 1995 or at the time of death,whichever occurred first. The Kaplan-Meier method was used to plot survivaltime free of events, and the log-rank testwas used to test differences between sur-vival curves (11). The Cox proportionalhazards regression model was used to esti-mate the hazard ratio (HR) and 95% CI forthe association of type of treatment withevents and to adjust for potential baselineconfounding variables (12). The assump-tion of proportionality of hazards wasassessed with log-minus-log plots and bytesting for an interaction between treat-ment and time (13). All P values were two-sided and all analyses were intention-to-treat, unless otherwise stated.

RESULTS— A total of 189 patientswere randomized to fosinopril and 191patients to amlodipine (Fig. 1). To controlblood pressure, amlodipine was added in30.7% of the fosinopril group patients(58/189), and fosinopril was added in26.2% of the amlodipine group patients(50/191, P > 0.1). The patients random-ized to fosinopril and amlodipine had sim-ilar demographic, blood pressure, andblood biochemical characteristics at base-line (Table 1). Microalbuminuria wasslightly higher in the amlodipine group. Atotal of 55 patients in the fosinopril groupand 66 patients in the amlodipine groupwere on diuretics, and 20 patients in thefosinopril group and 18 patients in theamlodipine group were on beta-blockersbefore the study Diuretics and beta-block-ers were not prescribed during the study.

Table 2 summarizes the measurementsat the last annual visit, which occurred onaverage 2.5 years after randomization, andthe change from baseline. Both treatmentssignificantly decreased sBP and dBP Duringfollow-up in the fosinopril and amlodipinegroups, the sBP goal was reached by 58.5and 60.7% of the patients, respectively; thedBP goal was reached by 89.4 and 90.1%

598 DIABETES CARE, VOLUME 21, NUMBER 4, APRIL 1998

Dow


Tatti and Associates

1172 patients screened

792 patients excluded

380 patients randomized

189 fosinopril

131 fosinopril only

58 fosinopril plus amlodipine

173 followed up

16 did not return to any annual visit

mean follow-up time - 2.6 years

188 followed up

1 lost to follow-up

mean follow-up time - 2.9 years

Randomized treatment

Post-randomization

treatment

Primary outcome:

jlood examination test

Secondary outcome:

pre-defined major

vascular events

191 amlodipine

141 amlodipine only

50 amlodipine plus fosinopril

4>

171 followed up

20 did not return to any annual visit

mean follow-up time ~ 2.4 years

•i>

188 followed up

3 lost to follow-up

mean follow-up time : 2.8 years

Figure 1—Trial profile.

of the patients, respectively; and both sBPand dBP goals were reached in 55.6 and58.6% of the patients, respectively (P >0.1). At the last annual visit, the sBP was 4mmHg lower in the amlodipine groupcompared with fosinopril (P < 0.01). Thisdifference reflected the greater reduction insBP over the follow-up period in patientstreated with amlodipine than in thosetreated with fosinopril (—19 and —13mmHg, respectively). The two treatmentgroups had the same reduction in dBPcompared with baseline (—8 mmHg).

In both treatment groups, serum crea-tinine, HbAlc, and triglycerides did notvary significantly during follow-up. Fastingserum glucose, serum insulin, and microal-buminuria were significantly decreased atthe last annual visit, compared with base-line values by a similar magnitude in thefosinopril and amlodipine groups. Subcu-taneous insulin was administered duringthe trial to 19.0% (36/189) of the patientsin the fosinopril group and 17.8% (34/191)of the patients in the amlodipine group (P> 0.1). During follow-up, total serum cho-lesterol increased slightly but significantlyin both groups. The increase in total cho-lesterol was accompanied by a smallincrease in HDL cholesterol, which wassignificant only in the fosinopril group.Compared with baseline levels, fibrinogendecreased significantly in the patientsreceiving fosinopril and did not changesignificantly in those receiving amlodipine.The difference in fibrinogen levels at the

last annual visit between fosinopril andamlodipine groups was not significant.

All patients had a complete follow-upfor events through 30 June 1995, except forone patient in the fosinopril group and

Table 1—Patient characteristics at baseline

three patients in the amlodipine group whowere lost to follow-up. The follow-up rangewas 2.5-3.5 years. A total of 69 participantsexperienced predefined events, includingdeath, vascular events, procedures, andcancer. The patients randomized to fosino-pril were less likely to experience eventsthan those randomized to amlodipine(Table 3). The proportion of patients reach-ing the prospectively defined combinedend point of stroke, acute myocardialinfarction, or hospitalized angina wassignificantly lower in the fosinopril groupcompared with amlodipine (P = 0.030, Fig.2; HR = 0.49, 95% CI = 0.26-0.95, Table3). In a separate proportional hazardsmodel adjusted for baseline albuminuria,the results were unchanged. The HR (95%CI) of the nonprospectively defined com-bined end point of fatal or nonfatal acutemyocardial infarction or stroke for fosino-pril compared with amlodipine was 0.58(0.30-1.13) (P = 0.11). The HR (95% CI)of major vascular events in men, women,people age <65 years, and in people age>65 years was 0.53 (0.17-1.62), 0.55(0.25-1.22), 1.12 (0.43-2.90), and 0.24(0.09-0.65), respectively. The P value forthe interaction of treatment with sex was

Characteristic

nAge (years)Sex (% women)Duration of diabetes (years)Cigarette smoking (%)

Never smokedCurrent smokerFormer smokerPassive smoker

BMI (kg/m2)Blood pressure

Systolic (mmHg)Diastolic (mmHg)

Blood examinationnCreatinine (mg/dl)Glucose (mg/dl)Insulin (mU/ml)HbAlc (%)Total cholesterol (mg/dl)HDL cholesterol (mg/dl)Triglycerides (mg/dl)Fibrinogen (mg/dl)

Albuminuria (ug/min)

Fosinopril

18962.8 ±0.5

36.510.7 ±0.7

86.84.85.33.2

30.7 ±0.3

170 ± 195 ± 1

1891.0 ±0.01

171 ± 325 ±2

6.9 ±0.1222 ±2

47 ± 1153 ±6292 ±5

20 ± 1

Amlodipine

19163.3 ±0.4

44.510.5 ±0.6

81.76.85.26.3

30.5 ±0.4

171 ±194 ± 1

1911.0 ±0.01

174 ±422 ±1

7.0 ±0.1222 ±2

47 ± 1159 ±5291 ± 5

24 ± 1 *

Data are means ± SE or %. *P < 0.05 vs. fosinopril.


Dow



Table 2—Blood pressure and biochemical measures at last annual visit

Measure

Blood pressurenSystolic (mmHg)Diastolic (mmHg)

Blood examinationnCreatinine (mg/dl)Glucose (mg/dl)Insulin (mU/ml)HbAlc (%)Total cholesterol (mg/dl)HDL cholesterol (mg/dl)Triglycerides (mg/dl)Fibrinogen (mg/dl)

Albuminuria (ug/min)

Last visit

157 ±188 ±1

1.0 ±0.01153 ±419 ±1

6.8 ±0.1226 ±349 ±1

155 ±5283 ±4

13 ±1

Fosinopril

Change from baseline

179- 1 3 (-16 to-10)

- 8 ( -9 t o - 6 )

173-0.01 (-0.03 to 0.01)

-17 (-28 t o - 7 )- 5 ( -8 t o - 3 )

-0.1 (-0.3 to 0.1)4 (0 to 9)3 (1 to 4)

1 ( -8 to 11)- 9 (-16 t o - 1 )- 8 (-11 t o - 5 )

Last visit

153 ± I t86 ±1

1.1 ±0.05161 ±520 ±17.0 ±0.1

228 ±349 ±3

161 ±5290 ±4

13 ±1

AmlodipineChange from baseline

178-19 (-22 to-15)*

- 8 ( -9 t o - 6 )

1710.06 (-0.03 to 0.16)

-14 (-26 t o - 2 )-2(-4to-l)0 (-0.2 to 0.1)

6 (2 to 10)2 (-0 to 3)1 (-7 to 102 (-4 to 8)

-11 (-14 to-8)Data are means ± SE or means (95% CI). Changes from baseline are significant at P ^ 0.05 when the CI does not contain 0.at last annual visit.

*P < 0.05; TP < 0.01 vs. fosinopril

0.97 and for the interaction of treatmentwith age was 0.031.

In the worst-case scenario, in whichthe single fosinopril patient who was lost tofollow-up had an event and none of thethree amlodipine patients who were lost tofollow-up had events, the difference in riskof major vascular events between fosinopriland amlodipine would still be significant (P= 0.048 with log-rank test). Incident can-cer events tended to be less frequent in thefosinopril group, but the difference

between the two groups was not signifi-cant.

In crude analyses according to post-randomization treatment given to controlblood pressure, the patients who receivedfosinopril only (n = 131), amlodipine only(n =141), and the combination of fosinoprilplus amlodipine (n = 108) experienced 10,27, and 4 major vascular events, respec-tively. In the three post-randomizationgroups, the number of patients who experi-enced acute myocardial infarction was 7,

13, and 3, respectively; the number ofpatients who experienced hospitalizedangina was 0,4, and 0, respectively; and thenumber of patients who experienced strokewas 3, 10, and 1, respectively. Comparedwith use of amlodipine only, the risk ofmajor vascular events was significantlydecreased with the use of fosinopril onlyand with the combination treatment (HR0.37,95% CI 0.18-0.77, P = 0.008 and HR0.17,95% CI 0.06-0.50, P = 0.001, respec-tively). The event rates in the fosinopril

Table 3—Events during follow-up

Event

nAll-cause mortality

Fatal or nonfatal strokeFatal or nonfatal acute myocardial infarctionHospitalized angina

Any major vascular eventCoronary artery bypassPercutaneous transluminal coronary angioplasty

Any major vascular event or procedure*Other cardiovascular events or procedurest

Any death or any vascular event or any procedureNonfatal cancerFatal cancer

Any cancerAny event listed above

Fosinopril

1890.7 (4)0.7 (4)1.8(10)0(0)2.6 (14)0.5 (3)0(0)2.6(14)0.7 (4)3.6 (20)1.3(7)0.2(1)1.5(8)4.9 (27)

Amlodipine

1910.9 (5)1.9(10)2.4(13)0.7 (4)5.0 (27)0.4 (2)0.2(1)5.0 (27)0.9 (5)6.3 (34)1.7(9)0.4 (2)2.0(11)7.8 (42)

HR (95% CI)

—0.39 (0.12 to 1.23)0.77 (0.34 to 1.75)

0.49 (0.26 to 0.95)——

0.49 (0.26 to 0.95)—

0.56 (0.32 to 0.97)——

0.64 (0.25 to 1.65)0.59 (0.37 to 0.97)

P value

—>0.1>0.1

—0.030

——

0.030—

0.036——

>0.10.035

Data are rates expressed as number of events per 100 person-years (n). HR of fosinopril versus amlodipine calculated with proportional hazards models. *A11 coro-nary artery bypass and angioplasty procedures were performed in patients who previously have had a myocardial infarction; Tinclude three transitory ischemicattacks, one pericarditis, and five pacemaker implants; Hoo few events to analyze.


Dow



1.0-

.9-

.Q03

JQO

189191

p=0.03 log-rank

188

191183176

Fosinopril

:--< Amlodipine

74 Fosinopril

65 Amlodipine

0 1

Follow-up time (years)

Figure 2—Probability of remainingjree of stroke, acute myocardial infarction, or hospitalized anginaaccording to treatment. The number of participants at each time point according to treatment is indi-cated at the bottom of the graph.

only and in the fosinopril plus amlodipinegroups were not significantly different (P =0.19). It is important to note that the post-randomization analyses shown above breakthe intention-to-treat principle (14).

A total of 36 and 52 patients droppedout from treatment before the end of follow-up in the fosinopril and amlodipine group,respectively (19.0 and 27.2%, respectively,P = 0.06). Among the patients who droppedout, only one event occurred in theamlodipine group and no events occurredin the fosinopril group. In on-treatmentanalysis, the risk of major cardiovascularevents with fosinopril compared withamlodipine was virtually unchanged, withrespect to the intention-to-treat analysis (HR= 0.45, 95% CI 0.24-0.86, P = 0.016).

CONCLUSIONS— The FACET didnot demonstrate significant differencesbetween fosinopril and amlodipine in theprimary outcome measures of lipid profileand glucose metabolism. A greater sBPreduction was observed with amlodipine.Despite this blood pressure difference,patients randomized to fosinopril wereabout 50% less likely to experience majorcardiovascular events, a secondary out-come, than those randomized to amlodip-ine. These findings illustrate again thelimitations of blood pressure lowering as asurrogate marker of clinical efficacy of anti-hypertensive therapy.

No significant differences were foundbetween fosinopril and amlodipine inabsolute values or changes in biochemical

measures. These findings are in agreementwith a recent study that compared cilaza-pril with amlodipine in patients withNIDDM (9). Although the interpretation ofchanges compared with baseline is limitedby lack of a placebo reference group, thepresent results indicate that both fosinopriland amlodipine may favorably affect fastingserum glucose, plasma insulin levels, HDLcholesterol, and urinary protein excretion.These findings are in agreement with otherstudies conducted in patients with hyper-tension, renal dysfunction, or NIDDM(2-4,15). Some trials have shown that ACEinhibitors and amlodipine tend to decreasetotal serum cholesterol and triglycerides(3-7). However, this association was notseen in the FACET.

A plausible explanation for the greaterdecrease in sBP found in the amlodipinegroup compared with the fosinopril groupis that, to reduce side effects, amlodipinewas given in the evening and fosinopril inthe morning. Therefore, because bloodpressure was always measured in the morn-ing and both drugs are long-acting agents(16,17), the time of the blood pressuremeasurement was likely to be close to thepeak effect of amlodipine and to the trougheffect of fosinopril. Alternative explanationsare that the relative dose of amlodipine wasgreater than that of fosinopril, or thatamlodipine was more effective than fos-inopril in lowering blood pressure. In somestudies, calcium antagonists tended toachieve a greater blood pressure reductionthan ACE inhibitors (10,15,18).

To our knowledge, this is the first ran-domized trial comparing major clinicalevents in NIDDM patients receiving anACE inhibitor or a long-acting calciumantagonist for hypertension. The reductionin risk of major vascular events in patientsrandomized to fosinopril compared withamlodipine was statistically significant andwas consistent when all prespecified eventswere included. Such an effect remainedunchanged in analyses adjusted for poten-tial baseline confounders, in analysesaccording to treatment given after random-ization, and in worst-case scenario analysesthat account for patients lost to follow-up.Follow-up and ascertainment of events wasvirtually complete in both groups, and theprospectively defined events were adjudi-cated according to objective criteria by anindependent blinded committee. Nonethe-less, these results should be interpretedwith caution. This trial was not designedand powered to assess a difference betweenthe two treatments in vascular events.Additional methodological limitations arethe open-label and single-site design of thestudy and the 6-month interval of the fol-low-up visits. The intriguing findings ofFACET on vascular events need to be con-firmed in large prospective long-term trials,such as for example, the Antihypertensiveand Lipid Lowering Treatment to PreventHeart Attack Trial (ALLHAT) (19).

Recent studies have raised questionsabout long-term safety of calcium antago-nists (20-24). Two observational studiesin hypertension (21,22) and one meta-analysis of placebo-controlled trials in coro-nary heart disease (20) have found excessmortality or occurrence of coronary eventswith the use of certain short-acting cal-cium antagonists. One recent case-controlstudy in hypertensive subjects has found anincreased risk of coronary events primarilyin patients using short-acting calciumantagonists, but only a small nonsignificantincrease in those using long-acting formu-lations (25). In a randomized placebo-con-trolled trial in patients with severe heartfailure, amlodipine did not affect the pri-mary outcome of mortality or hospitaliza-tion for heart failure, but it increased therisk of pulmonary edema (25). More recentsecondary analyses suggest that the excessrisk of adverse events associated with theuse of calcium antagonists may be particu-larly important in diabetic patients (26).

The results of the FACET on vascularevents may be due to a special benefit offosinopril in patients with NIDDM. This


Dow



interpretation is suggested by the analysesaccording to the treatment that was givenpost randomization. In these analyses, therisk of events was not increased amongthose receiving fosinopril plus amlodipine,compared with those receiving fosinoprilonly. About a third of the participants wereusing the combination treatment of fosino-pril plus amlodipine, and as a result, in theintention-to-treat analyses, any treatmenteffect would be diluted. In crude analysesthat compare those who received fosinoprilonly to those who received amlodipineonly, the relative risk of major cardiovascu-lar events was lower and more significantthan in intention-to-treat analyses (HR0.37,95% CI 0.18-0.77, P = 0.008 and HR0.49, 95% CI 0.26-0.95, P = 0.030,respectively). The findings according topost-randomization treatment may reflectmore closely the real-life clinical practice.However, such post hoc analyses should beinterpreted with caution because theybreak the intention-to-treat principle (14).

If fosinopril is truly more effective thanamlodipine in preventing major vascularevents in NIDDM, what are the potentialmechanisms? The present findings are notexplained by differences in blood pressurecontrol, diabetes control, renal function,or lipid profile. It is unlikely that theevening versus morning dosage might haveimpacted on the results of cardiovascularevents. Potential sympathetic stimulationwith amlodipine might have accounted forthe increased risk in cardiovascular eventscompared with fosinopril (27). However, ithas been reported that adrenergic cardio-vascular response is reduced in diabeticpatients who often have sympathetic neu-ropathy and consequent regional auto-nomic denervation (28,29). In theory, ACEinhibitors might be more beneficial thanother antihypertensive treatments in dia-betic patients. The greater propensity tothrombotic events in people with diabeteshas been linked to increased levels of plas-minogen activator inhibitor-1 (PAI-1) (30).Recent evidence has shown that the ACEplays a key role in the activation of PAI-1(31). It has been shown that ACE inhibitorscan suppress the expression of PAI-1 (32)and thus, facilitate fibrinolysis. This mech-anism may account for the beneficial effectof fosinopril on both stroke and coronaryevents.

There is growing evidence that ACEinhibitors prevent major complications indiabetic patients, including those withnephropathy (33), myocardial infarction

(34), or noncomplicated hypertension asshown in the FACET. Secondary analysesin the Systolic Hypertension in the ElderlyProgram (SHEP) have shown that diuretic-based treatment is more effective thanplacebo in preventing cardiovascular eventsin both diabetic and nondiabetic patients(1). These new findings in SHEP highlightthe primary importance of blood pressurelowering in diabetic hypertensives. How-ever, to optimize treatment, long-term trialscomparing active agents are needed toassess which treatment is most effective inpreventing major complications. Theresults of FACET are in agreement withthose of the Multicenter Isradipine DiureticAtherosclerosis Study (MIDAS) (35). Inboth trials, hypertensive patients with dia-betes or impaired glucose metabolism whoreceived alternative treatments had asignificantly lower risk of cardiovascularevents, compared with those who receiveda calcium antagonist. The FACET showsthat both fosinopril and amlodipine effec-tively lowered blood pressure and hadcomparable effects on surrogate biochemi-cal measures, but the patients randomizedto fosinopril had a significantly lower riskof major vascular events, compared withthose randomized to amlodipine.

Acknowledgments— The study drugs, labo-ratory examination tests, and procedures werepart of the patients regular treatment and fol-low-up. Data analyses were supported by agrant from Bristol-Myers Squibb, Princeton, NJR.RB. has received grant support from Bristol-Myers Squibb and Pfizer.

We are indebted to Dr. Giuseppe Claps andDr. Alessio Legramante for adjudicating theevents. The study was presented at the Ameri-can Heart Association 70th Scientific Sessions,Orlando, Florida, 9-12 November 1997.

References1. CurbJD, Pressel SL, Cutler JA, Savage PJ,

Applegate WB, Black H, Camel G, DavisBR, Frost PH, Gonzalez N, Guthrie G,Oberman A, Rutan GH, Stamler J: Effect ofdiuretic-based antihypertensive treatmenton cardiovascular disease risk in older dia-betic patients with isolated systolic hyper-tension: Systolic Hypertension in theElderly Program Cooperative ResearchGroup. JAMA 276:1886-1892, 1996

2. Harano Y, Kageyama A, Hirosa J, Asakura Y,Yokota T, Ikebuchi M, Suzuki M, Omae T:Improvement of insulin sensitivity for glu-cose metabolism with the long-acting Ca-channel-blocker amlodipine in essential

hypertension subjects. Metabolism 44:315-319,1995

3. Zehetgruber M, Beckmann R, Gabriel H,Christ G, Binder BR, Huber K: The ACE-inhibitor lisinopril affects plasma insulinlevels but not fibrinolytic parameters.Thromb Res 83:143-152, 1996

4. Nielsen FS, Rossing P Gall MA, Skott P,Smidt UM, Parving HH: Impact of lisinopriland atenolol on kidney function in hyper-tensive NIDDM subjects with diabeticnephropathy. Diabetes 43:1108-1113,1994

5. Letizia C, De CiocchisA, Cerci S, Coassin S,Fisher H, Tarsitani P, Scavo D: Amlodipinein ambulatory hypertensive patients:humoral and haemodynamic effects. IntJQin Pharmacol Res 13:151-159, 1993

6. Schlueter W, Keilani T, Batlle DC: Metaboliceffects of converting enzyme inhibitors:focus on the reduction of cholesterol andlipoprotein(a) by fosinopril. Am J Cardiol72:37H-40H, 1993

7. Keilani T, Schlueter W, Levin B, Batlle DC:Improvement of lipid abnormalities associ-ated with proteinuria using fosinopril, anangiotensin-converting enzyme inhibitor.Ann Intern Med 118:246-254, 1993

8. Ferrier C, Ferrari P, Weidmann P Keller U,Beretta Piccoli C, Riesen WF: Swiss hyper-tension treatment programme with vera-pamil and/or enalapril in diabetic patients.Drugs 44 (Suppl. 1):74-84, 1992

9. Velussi M, Brocco E, Frigato F Zolli M,Muollo B, Maioli M, Carrara A, Tonolo G,Fresu P, Cernigoi AM, Fioretto P, NosadiniR: Effects of cilazapril and amlodipine onkidney function in hypertensive NIDDMpatients. Diabetes 45:216-222, 1996

10. Kasiske BL, Ma JZ, Kalil RSN, Louis TA:Effects of antihypertensive therapy onserum lipids. Ann Intern Med 122:133-141,1995

11. Kaplan EL, Meier P: Nonparametric esti-mation from incomplete observations. J AmStatAssoc 53:457-481, 1958

12. Cox DR: Regression models and life tables.JRoy Statist Soc 34B: 187-220, 1972

13. SPSS: SPSS Advanced Statistics ReferenceGuide, release 7.5. Chicago, SPSS, 1997

14. Friedman L, Furberg C, DeMets D: Funda-mentals of Clinical Trials. 3rd ed. St. Louis,MO, Mosby Yearbook, 1996

15. NeatonJD, Grimm RH, Prineas RJ, StamlerJ, Grandits GA, Emler PJ, Cutler JA, FlackJM, Schoenenberger JA, McDonald R,Lewis CE, Leibson PR, Treatment of MildHypertension Study Research Group: Treat-ment of Mild Hypertension Study: finalresults. JAMA 270:713-724, 1993

16. Fortini A, Cappelletti C, Cecchi L, LaureanoR: Fosinopril in the treatment of hyperten-sion: effects on 24h ambulatory blood pres-sure and on blood pressure response toexercise. J Hum Hypertens 8:469-474,1994

17. Bainbridge AD, Herlihy O, Meredith, PA,Elliott HL: A comparative assessment of


Dow



amlodipine and felodipine ER: pharmaco-kinetic and pharmacodynamic indices. EurJ Clin Pharmacol 45:425-430,1993

18. Materson BJ, Reda DJ, Cushman WC,Massie BM, Freis ED, Kochar MS, Ham-burger RJ, Fye C, Lakshmanan R, Gottdi-ener J, Ramirez EA, Henderson WG, theDepartment of Veteran Affairs CooperativeStudy Group on Antihypertensive Agents:Single-drug therapy for hypertension inmen: a comparison of six antihypertensiveagents in men. N Engl] Med 328:914-921,1993

19. Davis BR, Cutler JA, Gordon DJ, FurbergCD, Wright JT, Cushman WC, Grimm RH,LaRosa J, Whelton PK, Perry HM, Alder-man MH, Ford CE, Oparil S, Francis C,Proschan M, Pressel S, Black HR, HawkinsCM: Rationale and design for the Antihy-pertensive and Lipid Lowering Treatmentto Prevent Heart Attack Trial (ALLHAT).AmJ Hypertens 9:342-360,1996

20. Furberg CD, Psaty BM, Meyer JV: Nifedip-ine: dose-related increase in mortality inpatients with coronary heart disease. Circu-lation 92:1326-1331, 1995

21. Psaty BM, Heckbert SR, Koepsell TD,Siskovick DS, Raghunathan TE, Weiss NS,Rosendaal FR, Lemaitre RN, Smith NL,Wahl PW, Wagner EH, Furberg CD: Therisk of myocardial infarction associatedwith antihypertensive drug therapies. JAMA274:620-625, 1995

22. Pahor M, Guralnik, JM, Corti C, Foley DJ,Carbonin PU, Havlik RJ: Long-term sur-

vival and use of antihypertensive medica-tions in older persons. J Am Geriatr Soc43:1191-1197,1995.

23. Pahor M, Guralnik JM, Ferrucci L, CortiMC, Salive ME, CerhanJ, Wallace RB, Hav-lik RJ: Calcium channel blockade and inci-dence of cancer in aged populations. Lancet348:493-497, 1996

24. Pahor M, Guralnik JM, Furberg CD, Car-bonin PU, Havlik RJ: Risk of gastrointesti-nal haemorrhage with calcium antagonistsin hypertensive persons over 67 years old.Lancet 347:1061-1065, 1996

25. Alderman MH, Cohen H, Roque R, Mad-havan S: Effect of long-acting and short-act-ing calcium antagonists on cardiovascularoutcomes in hypertensive patients. Lancet349:594-598, 1997

26. Pahor M, Kritchevsky SB, Zuccala G, Gural-nik JM: Diabetes and risk of adverse eventswith calcium antagonists (Letter). DiabetesCare 21:193-194,1998

27. Terland O, Gronberg M, Flatmark T: Theeffect of calcium channel blockers on theH(+)-ATPase and bioenergetics of cate-cholamine storage vesicles. EurJ Pharmacol207:37-41, 1991

28. Turpeinen AK, Vanninen E, Kuikka JT,Uusitupa MI: Demonstration of regionalsympathetic denervation of the heart indiabetes: comparison between patientswith NIDDM and IDDM. Diabetes Care19:1083-1090,1996

29. Trovik TS, Jaeger R, Jorde R, Sager G:Reduced sensitivity to beta-adrenoceptor

stimulation and blockade in insulin depen-dent diabetic patients with hypoglycaemiaunawareness. Br] Clin Pharmacol 38:427-432,1994

30. Juhan Vague I, Alessi MC, Vague P: Throm-bogenic and fibrinolytic factors and cardio-vascular risk in non-insulin-dependentdiabetes mellitus. Ann Med 28:371-380,1996

31. Vaughan DE, Lazos SA, Tong K:Angiotensin II regulates the expression ofplasminogen activator inhibitor-1 in cul-tured endothelial cells: a potential linkbetween the renin-angiotensin system andthrombosis. J Clin Invest 95:995-1001,1995

32. Hamdan AD, Quist WC, Gagne JB, FeenerEP: Angiotensin-converting enzyme inhibi-tion suppresses plasminogen activatorinhibitor-1 expression in the neointima ofballoon-injured rat aorta. Circulation 93:1073-1078,1996

33. Lewis EJ, Hunsicker LG, Bain RP, RohdeRD: The effect of angiotensin-converting-enzyme inhibition on diabetic nephropa-thy: The Collaborative Study Group. N Engl] Med 329:1456-1462, 1993

34. Zuanetti G, Latini R: Impact of pharmaco-logical treatment on mortality after myocar-dial infarction in diabetic patients. J DiabetComplications 11:131-136,1997

35. Byington RP, Craven T, Furberg CD, PahorM: Isradipine, raised glycosylated hemo-globin, and risk of cardiovascular events.Lancet 350:1075-1076, 1997


Dow


Outcome Results of the Fosinopril Versus Amlodipine - Diabetes Care

Documents

Transcript of Outcome Results of the Fosinopril Versus Amlodipine - Diabetes Care