点击显示 收起
第15届欧洲高血压会议
Fifteenth European Meeting on Hypertension
2005年6月17－21日
意大利米兰
June 17 - 21, 2005, Milan, Italy
Outcomes in Patients Treated With Monotherapy in the Valsartan Antihypertensive Long-term Use Evaluation (VALUE) Trial
Linda Brookes, MSc
An analysis of patients in the Valsartan Antihypertensive Long-term Use Evaluation (VALUE) who were on monotherapy with 1 of the 2 study drugs confirmed the main study findings of significantly less heart failure and new-onset diabetes with valsartan. However, unlike the main study, in which the incidence of myocardial infarction (MI) was significantly more frequent among the valsartan-treated patients, there was no difference in MI between the amlodipine and valsartan groups in the monotherapy analysis.[1]
The main randomized, double-blind, parallel-group VALUE trial was set up to compare the angiotensin receptor blocker (ARB) valsartan with the calcium channel blocker (CCB) amlodipine in hypertensive patients at high cardiovascular risk, based on the hypothesis that for the same level of blood pressure control, a valsartan-based regimen would reduce cardiac morbidity and mortality more than an amlodipine-based regimen.[2] However, the trial failed to achieve equivalent blood pressure control on the 2 treatment arms, confounding analysis of the results.[3]
Patients entered into the main VALUE trial stopped all previous antihypertensive medications and were immediately rolled over to either valsartan 80 mg or amlodipine 5 mg, without a placebo run-in period. Over the initial 6-month period of the trial, doses of either drug could be titrated upward (valsartan to 160 mg and amlodipine to 10 mg), and thereafter hydrochlorothiazide (12.5 mg to 25 mg) and other drugs could be added, at the discretion of the physician to achieve blood pressure control (< 140/90 mmHg).
Although both regimens reduced blood pressure, the blood pressure-lowering effect of the amlodipine-based regimen was greater, especially during the first 6 months of the study. Most (13%) of the (19%) relative increase in MI associated with valsartan was attributed by the VALUE investigators to the difference in blood pressure between the 2 treatment groups.
An attempt was made to overcome the blood pressure discrepancies in an initial post hoc analysis by applying a new statistical technique called serial median matching.[4] This was used in an analysis of outcomes at 6 months when treatment adjustments intended to achieve control of blood pressure were complete. By creating valsartan-amlodipine patient pairs matched for a range of characteristics, including systolic blood pressure, gender, age, and previous coronary disease, stroke, and diabetes, outcomes including MI, were shown to be identical in the 2 treatment groups except for a significantly lower rate of heart failure hospitalization with valsartan.[4]
The monotherapy analysis addressed the problems of blood pressure discrepancies in VALUE in a different way. According to the VALUE investigators, this allowed assessment of whether true pharmacological differences between drugs resulted in different clinical outcomes because it could be assumed the patients who remained on monotherapy were not given additional drugs because they had good blood pressure control.
The population for the monotherapy analysis comprised 7080 patients (46%) of the 15,313 patients randomized in the main VALUE trial who were still on monotherapy with valsartan (n = 3263) or amlodipine (n = 3817) by the end of the 6-month drug titration period. The average time on monotherapy was 3.6 years compared with a mean follow-up time of 4.2 years in the main study.
A censored analysis, in which events were counted throughout the entire study but only while patients were receiving monotherapy, was used to evaluate differences in outcome between the 2 drugs. Two sensitivity analyses by intention to treat (ITT), using outcomes throughout the entire study, addressed the question, more practically according to the investigators, of whether there were any differences after addition of other drugs and whether prolonged therapy attenuated or enhanced the effects of monotherapy.
In-trial mean systolic and diastolic blood pressures in the amlodipine and valsartan monotherapy groups were similar throughout the trial, allowing testing of the original hypothesis on which VALUE was based (see previous discussion).
As in the main VALUE trial, the primary cardiac composite endpoint did not differ between the 2 treatment groups (HR 0.883, 95% CI 0.728-1.071; P = .206). Also confirming the main trial result, significantly less heart failure and new-onset diabetes were seen with valsartan (Table), but unlike the main results there was no difference between the 2 treatment groups in rates of MI (P = .778).
ITT = intention to treat
Rates of stroke (P = .671) and all-cause death (P = .289) were also similar in the 2 groups. All these results were preserved in ITT analysis. The differences were seen despite much lower levels of absolute risk in the monotherapy groups.
A clear trend was seen as early as 12 months into the trial and the longer the patients were on therapy, the stronger the effect of valsartan on relative reduction of heart failure. No such trend was seen for MI. Further analyses, such as predictors of response, are being pursued before publication.
Direct evidence in favor of lowering systolic blood pressure (SBP) to a goal of < 140 mm Hg, as recommended in current guidelines,[1-3] has been available only for those hypertensive patients with diabetes or previous cardiovascular disease, but has principally excluded the influence of lowering SBP in hypertensive patients with other risk factors. A multicenter clinical trial performed in China, the Felodipine EVent Reduction (FEVER) study, is the first to show the benefits of reducing blood pressure below the recommended 140/90 mm Hg goal in a hypertensive population at high risk because of other risk factors, as well as diabetes or previous cardiovascular events.[4] In the FEVER study, a "more intensive" antihypertensive therapy regimen consisting of a low-dose diuretic plus low-dose calcium channel blocker (CCB) reduced blood pressure to < 140/90 mm Hg in a majority of patients and also brought about a significant reduction in cardiovascular events, including stroke; less-intensive therapy with a low-dose diuretic only did not achieve such benefits.
Two clinical trials of CCBs in hypertension were previously performed in China. The Shanghai Trial Of Nifedipine in the Elderly (STONE)[5] and Systolic Hypertension in China (Syst-China)[6] showed significant reductions in stroke and other cardiovascular outcomes by active reduction of blood pressure with a CCB vs placebo. Both trials enrolled patients with relatively high SBP, however, and goal SBP was 150 mm Hg, far above the presently recommended goal.
The randomized, double-blind, placebo-controlled FEVER study was designed to compare the effects on cardiovascular outcomes of adding a low dose of the CCB felodipine vs placebo in hypertensive patients at high cardiovascular risk whose blood pressure had already been reduced by a low-dose diuretic, hydrochlorothiazide (HCTZ), which was continued throughout the trial. The trial also compared more vs less intense blood pressure-lowering treatment to achieve currently recommended goals. The primary outcome of FEVER was stroke (fatal and nonfatal), which has a steeper relationship with blood pressure than other cardiovascular events in the Chinese population.
Eligible hypertensive patients were aged 50-79 years, male or female, with SBP/DBP < 210/115 mm Hg if they were on antihypertensive drug treatment or SBP 160-210 mm Hg or DBP 95-115 mm Hg if untreated. Patients aged 50-60 years had to have >/= 2 risk factors or clinical evidence of an event, and those aged 61-79 years had to have >/= 1 risk factor or clinical evidence of an event.
Qualifying cardiovascular disease risk factors included male gender, current smoker, total cholesterol >/= 240 mg/dL within the previous 1 year or on treatment for hypercholesterolemia, diabetes mellitus under glycemic control, left ventricular hypertrophy, proteinuria, and overweight (BMI > 27 kg/m2). Qualifying cardiovascular events included myocardial infarction (MI), stroke, stable angina or clinical evidence of coronary heart disease, congestive heart failure, peripheral ischemic artery disease, and transient ischemic attack.
Over 10,000 patients were recruited between 1999 and 2001 at 109 hospital and community clinics throughout China coordinated by the Fu Wai Hospital, Beijing, and the Chinese Hypertension League. Patients entered a 6-week run-in period on HCTZ 12.5 mg/day (during which other antihypertensive therapy was discontinued in the 89% of patients who had been previously treated), after which 9800 patients with DBP 90-100 mm Hg and/or SBP 140-160 mm Hg were randomized to receive HCTZ plus either felodipine 5 mg/day or placebo.
Add-on diuretics or other antihypertensive agents were allowed at the investigators' discretion. The drug classes most often added were diuretics, angiotensin-converting enzyme inhibitors, and CCBs, although 66.1% of patients on felodipine and 57.7% of those on placebo remained on study drugs only throughout the study. The study was terminated on June 30, 2004, when the required number of primary endpoints (400) was achieved.
Mean SBP/DBP at screening was 159/93 mm Hg. After 6 months of follow-up, mean SBP/DBP was 142.5/85 mm Hg in the placebo group vs 137.3/82 mm Hg in the felodipine group. This difference between the treatment groups remained constant throughout the study for a mean SBP/DBP reduction of 4/2 mm Hg. By the end of follow-up (60 months), more patients on felodipine plus HCTZ achieved SBP < 140 mm Hg and DBP < 90 mm Hg than those on placebo plus HCTZ (Table 1).
DBP = diastolic blood pressure; SPB = systolic blood pressure
The primary outcome of FEVER, stroke (fatal and nonfatal), was significantly reduced (-28%) in the felodipine group, a difference driven by nonfatal strokes (Table 2).
Significant reductions in the felodipine group vs placebo were also seen for secondary outcomes of cardiovascular events (-28%), all cardiovascular events (-28%), all cardiac events (-34%), all-cause death (-30%), cardiovascular death (-32%), coronary events (-32%), heart failure (-24%), and cancer (-40%). New-onset diabetes occurred in only 1% of patients in each group.
The regimens in FEVER were well tolerated, although the felodipine group experienced significantly more flushing (1.4% vs 0.2%, P < .001) and more ankle edema (1.0% vs 0.37%, P < .001), but significantly less fatigue (0.64% vs 1.05%, P = .037) than patients in the placebo group. There was no difference between the 2 treatment groups with regard to rates of dizziness, headache, or palpitations.
Commenting on the FEVER results, Prof. Alberto Zanchetti noted that even a 4/2 mm Hg difference in SBP/DBP -- as seen for the CCB regimen vs placebo -- can be associated with substantial reductions in most cardiovascular events, even in patients at lower overall cardiovascular risk than those in many recent trials (eg, Syst-Eur, Syst-China, HOPE, PATS, PROGRESS, ALLHAT, SCOPE, LIFE, VALUE, INVEST, EUROPA, ACTION). Prof. Zanchetti also noted the significant and marked reduction in incident cancer in the patients on felodipine, which is in contradiction to past claims that CCBs are associated with an increased risk of cancer.
Two recent posthoc analyses of patients in the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT) have compared the effects of the calcium channel blocker (CCB) used in the trial (amlodipine) with the angiotensin-converting enzyme (ACE) inhibitor (lisinopril) in patients with or without coronary heart disease (CHD) at baseline and in black patients vs nonblack patients. Although amlodipine and lisinopril resulted in equivalent blood pressure control and rates for most outcomes in CHD and non-CHD patients in ALLHAT,[1] the results of the analysis in black/nonblack patients suggested that a CCB may have advantages over an ACE inhibitor in black men and women and in nonblack women.[2]
The ALLHAT randomized, double-blind, active-controlled, clinical outcome trial was conducted between February 1994 and March 2002 in 33,357 hypertensive patients aged >/= 55 years with at least 1 other cardiovascular risk factor.[3] Patients were randomized to amlodipine 2.5-10 mg, lisinopril 10-40 mg, or a thiazide-type diuretic, chlorthalidone, 12-25 mg for a mean follow-up of 4.9 years. Other medications could be subsequently added to achieve goal blood pressures < 140/90 mm Hg.
For the overall trial results, the primary outcome (combined fatal CHD or nonfatal myocardial infarction [MI]) and a number of secondary outcomes did not differ between the 3 treatment groups. Compared with chlorthalidone, amlodipine was associated with higher rates of heart failure and lisinopril with higher rates of combined cardiovascular disease, stroke, and heart failure.
The first presentation looked at patients on amlodipine vs lisinopril regimens in the subset of ALLHAT patients with CHD; the second in the subset of blacks in ALLHAT, vs nonblacks, on these medication regimens.
Results from placebo-controlled trials such as the Heart Outcomes Prevention Evaluation (HOPE) study[4] and the European trial on Reduction Of cardiac events with Perindopril in stable coronary Artery disease (EUROPA),[5] although not the Prevention of Events with Angiotensin-Converting Enzyme inhibition (PEACE) trial,[6] indicated that treatment with ACE inhibitors results in improved outcome beyond what may be expected from the decrease in blood pressure. ALLHAT provided the opportunity to compare clinical outcomes in hypertensive patients with CHD at baseline, randomized to lisinopril vs amlodipine, although the overall results provided no evidence for specific benefits for ACE-inhibitor based treatment in this patient population. This current analysis also addresses issues raised by small trials and observational studies, suggesting that CCBs were associated with unfavorable cardiovascular outcomes, particularly in CHD patients.
From the total ALLHAT population, 4472 participants were identified as having had CHD at baseline. CHD was defined as "known prior myocardial infarction, angina, primary cardiac arrest, coronary artery stenosis, > 50% reversible perfusion defect, or major coronary revascularization procedure." Among these CHD patients, 2202 patients turned out to have been assigned to amlodipine and 2270 to lisinopril. Patient characteristics of these 2 treatment groups did not differ significantly.
Compared with 13,492 patients with no CHD (6777 randomized to amlodipine and 6715 to lisinopril), the patients with CHD at baseline tended to be older, have slightly lower blood pressure, were more likely to be on antihypertensive medication and/or aspirin, and were less likely to be female, black, or a current smoker. Fewer CHD patients than non-CHD patients had diabetes mellitus, a result of the 1-only risk factor requirement for entry into ALLHAT.
In the CHD patients, mean blood pressures were similar at baseline (146/83 mm Hg in the lisinopril group vs 145/83 mm Hg in the amlodipine group) and decreased similarly during follow-up, to 137/77 mm Hg vs 136/77 mm Hg, respectively, at Year 2 and to 133/75 mm/Hg vs 133/74 mmHg at Year 4. Compared with the CHD patients, blood pressure in the non-CHD patients was 1-2 mm Hg higher at baseline. Blood pressure was treated more successfully to a lower level in the CHD patients, but lower SBP and DBP were achieved on amlodipine in this group.
The primary outcome (combined fatal CHD or nonfatal MI) did not differ between lisinopril and amlodipine in the 2 arms in either the CHD or non-CHD patients (Table 1).
CHD = coronary heart disease; CI = confidence interval; RR = relative risk
Among the secondary endpoints, stroke was significantly increased for lisinopril vs amlodipine by 31% (P = .04) in the CHD patients and by 20% (P = .03) in the non-CHD patients. The increase in stroke with lisinopril was even greater (63%) in the subgroup of black patients with CHD.
A nonsignificant reduction of 9% in heart failure was seen for lisinopril vs amlodipine in heart failure in the CHD patients, but a 15% reduction for lisinopril vs amlodipine in the non-CHD patients was significant (P < .02). In contrast to the main ALLHAT results, which showed a significant 17% increase in angina for lisinopril vs amlodipine in non-CHD patients (P = .02), there was a 1% reduction for lisinopril vs amlodipine in the CHD group. Lisinopril was also associated with a 7% reduction in peripheral arterial disease in the CHD group, but a significant 39% increase (P = .002) in the non-CHD group.
There were no differences between lisinopril and amlodipine with regard to combined cardiovascular disease, all-cause mortality, or end-stage renal disease in either the CHD or non-CHD groups.
The ALLHAT investigators have concluded that in hypertensive patients with CHD, lisinopril- and amlodipine-based treatments provide equivalent blood pressure control and equivalent rates for most outcomes and that their analysis has failed to provide any evidence for specific benefits of ACE inhibitor-based treatment in patients with CHD.
Race was a prespecified subgroup in ALLHAT, which was the first large-scale trial to compare diuretics, CCBs, and ACE inhibitors as initial therapies in a population with a large number of black participants. A total of 11792 (35%) ALLHAT participants were black, 3213 of whom were randomized to amlodipine, 3210 to lisinopril, and 5369 to the thiazide-type diuretic, chlorthalidone. In a previously published analysis of outcomes for all 3 study drugs in hypertensive black patients compared with the 21,565 patients classified as nonblack,[7] the primary CHD endpoint did not differ between treatment groups and the results for the other outcomes mostly paralleled those of the main study. Both blacks and nonblacks had a higher risk of heart failure with amlodipine or lisinopril vs chlorthalidone, but for lisinopril vs chlorthalidone blacks also had a higher risk of stroke and combined cardiovascular disease. SBP control was also less well controlled with lisinopril compared with chlorthalidone in blacks.
The latest ALLHAT analysis has directly compared outcomes for lisinopril vs amlodipine in black and nonblack patients. It shows differences in blood pressure control in the 2 groups of patients, as well as a number of differences in outcomes.
For lisinopril vs amlodipine, SBP was less well controlled in the black patients, at a mean of 2-3 mm Hg higher at the end of follow-up. Final mean SBP was 0.5 mm Hg higher for lisinopril vs amlodipine in the nonblacks. In the nonblacks, blood pressure control was equivalent for the lisinopril and amlodipine groups. The number of add-on drugs in the blacks was 0.3 higher for lisinopril vs amlodipine, but similar in the nonblacks.
The ALLHAT primary outcome, combined fatal CHD or nonfatal MI, similar to the overall result of the trial, did not differ for lisinopril vs amlodipine in the black and nonblack groups (Table 2).
CI = confidence interval; RR = relative risk
Secondary endpoints of CHD, combined CVD, all-cause mortality, and end-stage renal disease did not differ between the 2 treatment groups in either the blacks or nonblacks. Stroke was similar in nonblacks but increased by 51% for lisinopril vs amlodipine in blacks (Table 3). Stroke was also increased for lisinopril vs amlodipine in women compared with men in both groups: +45% vs +37% in blacks and +46% vs -11% (P = .02) in nonblacks. Combined cardiovascular disease, similar between treatments in nonblacks, was increased by 13% with lisinopril vs amlodipine in blacks. Heart failure was increased with lisinopril in both groups: -15% in nonblacks and -11% in blacks vs amlodipine.
CI = confidence interval; CVD = cardiovascular disease; RR = relative risk
A surprising result in this analysis, since increased gastrointestinal (GI) bleed had previously been reported with CCBs, was a 16% increase in GI bleed associated with lisinopril (+16% in nonblacks and +28% in blacks, compared with amlodipine) (Table 4). The ALLHAT investigators believe that this may be due, although not entirely, to less-well-controlled blood pressure in the ACE inhibitor group. It would not be a reason to avoid ACE inhibitors, Dr. Cushman stressed. However, CCBs should be exonerated of this as well as of causing cancer, which was not increased in either treatment group in the black or nonblack patients.
CI = confidence interval; GI = gastrointestinal; RR = relative risk
As expected, lisinopril was associated with a higher incidence of angioedema than amlodipine in black and nonblack patients (P < .001).
Although ALLHAT was not designed to identify the best step-1 drug alternative if a thiazide-type diuretic cannot be used, considering these results, the ALLHAT investigators believe that amlodipine appears to have an advantage over lisinopril, particularly in blacks and nonblack women, and might be a reasonable first alternative to a diuretic in these groups.
The Blood Pressure Lowering Treatment Trialists' Collaboration (BPLTTC) was formed over a decade ago by members of several major clinical trial groups to undertake prospectively planned overviews (meta-analyses) of the results of antihypertensive drug trials. The group has published 2 major reviews so far, and in this most recent presentation, at the 2005 European Society of Hypertension (ESH) meeting in Milan, has addressed the question of whether there are any differences between the angiotensin-converting enzyme (ACE) inhibitor and angiotensin receptor blocker (ARB) drug classes.
Specifically, the BPLTTC compared treatment effects of these 2 drug classes and attempted to determine whether these are due to the blood pressure (BP)-lowering effects of the drugs, or if they have effects beyond BP lowering (ie, drug-specific). From their analysis of 21 large-scale randomized trials, the BPLTTC concluded that there are no clear differences between ACE inhibitors and ARBs for the outcomes of stroke and heart failure. However, although not found with ARBs, there was some evidence of drug-specific protection against coronary heart disease (CHD) events for ACE inhibitors.[1] They also concluded that for both ACE inhibitors and ARBs, the magnitude of the risk reductions seen in the trials were positively associated with the size of BP reduction.
These formal and systematic analyses represent the first such overview of ACE inhibitor and ARBs, and they involved a very large number of events (> 17,000). The BPLTTC investigators believe that the validity of their results is supported by sensitivity analyses and the results of previous BPLTTC overviews.
The second cycle of analyses of the BPLTTC,[2] which included 29 clinical trials involving over 160,000 patients, showed no significant differences in total major cardiovascular events between regimens based on ACE inhibitors, calcium channel blockers (CCBs), diuretics, or beta-blockers, although they did find that ACE inhibitor-based regimens reduced BP to a lesser degree.
In the earlier, placebo-controlled trials in the analysis, the relative risks of total major cardiovascular events were reduced by regimens based on ACE inhibitors (22%) or CCBs (18%). Greater risk reductions were produced by regimens that targeted lower BP goals (15%). ARB-based regimens reduced the risks of total major cardiovascular events (10%) compared with control regimens, although there was no conclusion as to whether benefits were greater than those expected given the reduction in BP.
Although there was evidence of some differences on cause-specific outcomes (nonfatal myocardial infarction [MI], stroke) among the different active regimens, for every outcome other than heart failure, the difference in achieved BP reduction was directly related to the observed difference in risk.
The latest BPLTTC overview involved 21 large-scale randomized clinical trials taken from the second cycle of analyses, including 16 trials encompassing a total population of 104,933 patients that included an ACE inhibitor arm:
Five trials with an ARB arm, involving 32,603 patients, were also included in the analysis:
In all of these trials, the ACE inhibitor or ARB arms were compared with a placebo arm, a diuretic plus beta-blocker arm, or a CCB arm. The analysis focused on 3 prespecified outcomes: stroke, CHD, and severe heart failure, and the Trialists applied meta-regression analysis to follow up BP reductions and corresponding risk reductions. For each trial, the difference in achieved SBP reduction (mm Hg) was plotted against the relative risk for each event.
The plotted line of regression showed that for ACE inhibitors there was a BP-independent protective effect of +1% (95% confidence interval, -9% to +12%), although since the confidence interval included unity, this was nonsignificant. It could therefore be concluded that the BP effect accounted for most of the effect of ACE inhibitors on stroke.
A similar result was seen for ARBs, with a similar BP-lowering effect and a nonsignificant effect of -2% (-32% to +3%), showing no evidence of a drug-specific effect for outcome of stroke.
Finally, formal comparison of the ACE inhibitors vs ARBs showed no significant difference (P = .6) between their effects on the outcome of stroke.
For ACE inhibitors, consistent with the results of the second cycle of analyses, larger BP reductions were associated with greater reductions in events. However, a statistically significant reduction (-9% [-14% to -3%]) in the risk of CHD provided clear evidence of a drug-specific protection against CHD events, in addition to BP-lowering effect. For ARBs, however, the effect was nonsignificant (+7% [-7% to +24%]) for CHD, suggesting that unlike in the case of ACE inhibitors, there is no BP independent effect.
Here, the comparison of ACE inhibitors vs ARBs was highly significant (P = .001), showing that ACE inhibitors confer a special protective advantage on the outcome of CHD compared with ARBs.
ACE inhibitors showed a borderline significant (-10% [-10% to 0%)] risk reduction for heart failure. When the trials with a CCB control treatment arm were excluded (due to evidence from the second-cycle analyses of an independent adverse effect for this outcome with CCBs), the result for ACE inhibitors became nonsignificant (-5% [-14% to +4%]). The result was also nonsignificant for ARBs (-16% [-36% to +5%]) and also when CCB trials were removed from the analysis (-18% [-95% to +1200%]), indicating that all the effect of ARBs on this outcome is BP dependent.
Comparison of the ACE inhibitors vs ARBs, with or without the CCB trials included, were both nonsignificant (P = .4 and P = .9), showing no difference between the 2 drug classes in treatment effects for the outcome of heart failure.
The power of the analysis was limited by the much smaller number of ARB trials conducted to date. This is expected to change as more results of large ARB trials become available, particularly large head-to-head trials of ARB and ACE inhibitors, such as the ONgoing Telmisartan Alone and in combination with Ramipril Global Endpoint Trial (ONTARGET).
Any real differences between drug classes are likely to be small compared with the benefit of maximizing BP reduction. Combination therapy with a regimen including an ACE inhibitor may optimize treatment benefits.
The idea that ACE inhibitors and ARBs are not interchangeable was recently raised in an editorial in the BMJ,[3] accusing clinicians of naiveté if they regard ARBs as being "like ACE inhibitors without the cough." Two Canadian physicians proposed that ARBs may increase, or at best not reduce, risk of MI, despite their beneficial effects on BP and urged a review of "the unexpected effects of ARBs on MI." Their argument, which was purportedly based on evidence from clinical trials, was severely criticized in responses to the BMJ from leading investigators worldwide, who called it "unbalanced" and "based on misinterpretations of data." All the respondents criticized the BMJ for publishing the editorial. Similar concern that the results of the BPLTTC analysis regarding ARBs should also not be misinterpreted was expressed at an investigator-led symposium hosted by the BPLTTC in Milan prior to the ESH meeting.
Defending ARBs against the accusation that they cause MI, incoming ESH president, Professor Sverre E. Kjeldsen, MD, PhD (Ullevaal University Hospital, Oslo, Norway), in a plenary session presentation, maintained that ARBs are "safe, exceptionally well tolerated, have target organ protective abilities, and prevent endpoints." He based his opinion on evidence from hypertension, diabetes, renal, heart failure, and post MI trials. The perceived association between ARBs and increased risk of MI in the Valsartan Antihypertensive Long-term Use Evaluation (VALUE) trial[4] could be explained by the difference in BP,[5,6] he maintained. He stressed that ARBs seem to be more effective than comparators in heart failure and prevention of diabetes and high tolerability may add to their beneficial effects over the years.
A post hoc analysis comparing patients with and without the metabolic syndrome who participated in the Antihypertensive and Lipid-Lowering treatment to prevent Heart Attack Trial (ALLHAT) trial reported no advantage for amlodipine or lisinopril.[1] Neither the calcium-channel blocker (CCB), amlodipine, nor the angiotensin-converting enzyme (ACE) inhibitor, lisinopril, was superior to the thiazide-type diuretic, chlorthalidone, in either group of patients with or without the metabolic syndrome, although the diuretic was more likely to induce new-onset diabetes in both groups, especially in those without the metabolic syndrome at baseline. Nevertheless, the ALLHAT investigators still maintain that thiazide-type diuretics are unsurpassed in preventing all, and superior in preventing one or more, of the major forms of cardiovascular disease.
ALLHAT was a practice-based, randomized, double-blind, active-controlled clinical trial of 33,357 hypertensive participants aged >/= 55 years who had at least one other risk factor for coronary heart disease (CHD).[2] Patients were randomized to receive either chlorthalidone 12.5 to 25 mg/day, amlodipine 2.5 to 10 mg/day, or lisinopril 10 to 40 mg/day, plus open-label step-up drugs (reserpine, atenolol, clonidine, hydralazine, or others at the physician's discretion to reach blood pressure goal). Planned follow-up was 4 to 8 years (mean achieved, 4.9 years). Although the primary outcome (combined fatal CHD or nonfatal myocardial infarction) and a number of secondary outcomes did not differ among the 3 treatment groups, compared with chlorthalidone, amlodipine was associated with 6-year higher rates of heart failure and lisinopril was associated with 6-year higher rates of combined cardiovascular disease, stroke, and heart failure. The ALLHAT investigators famously concluded that thiazide-type diuretics are "unsurpassed in lowering blood pressure, reducing clinical events, and tolerability, and less costly."
Because patients with the metabolic syndrome are at especially high risk for many hypertensive complications, the ALLHAT post hoc analysis was done to compare the efficacy of first-step antihypertensive therapy with a CCB or an ACE inhibitor with a thiazide-type diuretic in the nondiabetic patients with and without metabolic syndrome at baseline. (The diabetic patients, approximately 36% of the ALLHAT population, are subjects of a separate analysis.)
Definitions of the metabolic syndrome prepared by the International Diabetes Federation,[3] World Health Organization,[4] and the US National Cholesterol Education Program (NCEP) Adult Treatment Panel III (ATP III)[5] could not be applied to the ALLHAT population, because no measurements of insulin resistance or waist circumference measurement had been taken. For the purposes of this analysis, the metabolic syndrome was therefore defined as the presence of >/= 3 of the following:
High blood pressure (present in all patients as a condition of enrollment into ALLHAT);
Fasting glucose 100-125 mg/dL;
Body mass index (BMI) >/= 30 kg/m2;
Fasting triglycerides >/= 150 mg/dL; and
High-density lipoprotein (HDL)-cholesterol < 40 mg/dL in men, < 50 mg/dL in women.
Of the nondiabetic participants enrolled in ALLHAT, 8013 (40%) were classified as having metabolic syndrome and 9502 (47%) were classified as not having metabolic syndrome (Table 1); 2674 (13%) participants could not be classified either because they had no fasting glucose value available or because they had a nonfasting glucose value of 100-199 mg/dL.
The patients with metabolic syndrome were older than those without (mean age 66 vs 68 years, respectively), and were more like to be men (45.6% vs 43.7%), current smokers (30.0% vs 24.2%), and less likely to be black (24.1% vs 38.0%). BMI and fasting glucose and triglycerides levels were all significantly higher and HDL-cholesterol significantly lower in the patients with metabolic syndrome (all P < .05).
At 5 years, blood pressure was similar between the patients with and without metabolic syndrome for all treatment groups (Table 2). After 4 years, total cholesterol and fasting glucose levels were decreased in all treatment groups in all patients with or without metabolic syndrome. Serum potassium level remained constant.
*P < .05 vs chlorthalidone
DBP = diastolic blood pressure; SBP= systolic blood pressure.
Treatment group comparison outcomes, including CHD, all-cause mortality, stroke, and end-stage renal disease, were similar in patients with and without the metabolic syndrome at baseline (ie, there were no significant interactions). However, the risk of heart failure for lisinopril vs chlorthalidone was increased in patients with and without metabolic syndrome, and for amlodipine vs chlorthalidone in patients without metabolic syndrome (Table 3). Increasing the sample size by addition of the patients with diabetes mellitus resulted in an increased risk of heart failure for amlodipine vs chlorthalidone in patients with both the metabolic syndrome and diabetes.
*P < .05
CI = confidence interval; DM = diabetes mellitus; RR = relative risk
Rates of new-onset diabetes were markedly higher in patients with metabolic syndrome at baseline compared with those without metabolic syndrome (Table 4). New-onset diabetes was highest on the chlorthalidone arm in patients with and without metabolic syndrome at baseline, but in the amlodipine arm, the rate at 4 years in patients with metabolic syndrome approached that in the chlorthalidone patients.
Although this analysis showed that all 3 treatment arms had similar effects on outcomes in patients with or without the metabolic syndrome, the ALLHAT investigators' conclusion is that "thiazide-type diuretics should be the drugs of choice for first-step antihypertensive drug therapy in patients with and without metabolic syndrome." This did not go unchallenged at the meeting, particularly with respect to the lack of a difference in outcomes between the 3 treatment arms and the effect of diuretics on new-onset diabetes. It was also noted that diuretics no longer have the cost advantage over CCBs and ACE inhibitors that they did when the ALLHAT study was initiated.