The ALLHAT Officers and Coordinators for the ALLHAT Collaborative Research
Group. Major Outcomes in Moderately Hypercholesterolemic, Hypertensive Patients Randomized to Pravastatin vs Usual CareThe Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT-LLT). JAMA. 2002;288(23):2998–3007. doi:10.1001/jama.288.23.2998
ALLHAT Authors/Officers and Coordinators: Curt D. Furberg, MD, PhD; Jackson T. Wright, Jr, MD, PhD; Barry R. Davis, MD, PhD; Jeffrey A. Cutler, MD, MPH; Michael Alderman, MD; Henry Black, MD; William Cushman, MD; Richard Grimm, MD, PhD; L. Julian Haywood, MD; Frans Leenen, MD; Suzanne Oparil, MD; Jeffrey Probstfield, MD; Paul Whelton, MD, MSc; Chuke Nwachuku, MA, MPH; David Gordon, MD, PhD; Michael Proschan, PhD; Paula Einhorn, MD, MS; Charles E. Ford, PhD; Linda B. Piller, MD, MPH; J. Kay Dunn, PhD; David Goff, MD, PhD; Sara Pressel, MS; Judy Bettencourt, MPH; Barbara deLeon, BA; Lara M. Simpson, MS; Joe Blanton, MS; Therese Geraci, MSN, RN, CS; Sandra M. Walsh, RN; Christine Nelson, RN, BSN; Mahboob Rahman, MD; Anne Juratovac, RN; Robert Pospisil, RN; Lillian Carroll, RN; Sheila Sullivan, BA; Jeanne Russo, BSN; Gail Barone, RN; Rudy Christian, MPH; Sharon Feldman, MPH; Tracy Lucente, MPH; David Calhoun, MD; Kim Jenkins, MPH; Peggy McDowell, RN; Janice Johnson, BS; Connie Kingry, RN, BSN; Juan Alzate, MD; Karen L. Margolis, MD; Leslie Ann Holland-Klemme, BA; Brenda Jaeger; Jeffrey Williamson, MD, MHS; Gail Louis, RN; Pamela Ragusa, RN, BSN; Angela Williard, RN, BSN; R. L. Sue Ferguson, RN; Joanna Tanner; John Eckfeldt, MD, PhD; Richard Crow, MD; John Pelosi, RPh, MS.
Context Studies have demonstrated that statins administered to individuals with
risk factors for coronary heart disease (CHD) reduce CHD events. However,
many of these studies were too small to assess all-cause mortality or outcomes
in important subgroups.
Objective To determine whether pravastatin compared with usual care reduces all-cause
mortality in older, moderately hypercholesterolemic, hypertensive participants
with at least 1 additional CHD risk factor.
Design and Setting Multicenter (513 primarily community-based North American clinical centers),
randomized, nonblinded trial conducted from 1994 through March 2002 in a subset
of participants from the Antihypertensive and Lipid-Lowering Treatment to
Prevent Heart Attack Trial (ALLHAT).
Participants Ambulatory persons (n = 10 355), aged 55 years or older, with low-density
lipoprotein cholesterol (LDL-C) of 120 to 189 mg/dL (100 to 129 mg/dL if known
CHD) and triglycerides lower than 350 mg/dL, were randomized to pravastatin
(n = 5170) or to usual care (n = 5185). Baseline mean total cholesterol was
224 mg/dL; LDL-C, 146 mg/dL; high-density lipoprotein cholesterol, 48 mg/dL;
and triglycerides, 152 mg/dL. Mean age was 66 years, 49% were women, 38% black
and 23% Hispanic, 14% had a history of CHD, and 35% had type 2 diabetes.
Intervention Pravastatin, 40 mg/d, vs usual care.
Main Outcome Measures The primary outcome was all-cause mortality, with follow-up for up to
8 years. Secondary outcomes included nonfatal myocardial infarction or fatal
CHD (CHD events) combined, cause-specific mortality, and cancer.
Results Mean follow-up was 4.8 years. During the trial, 32% of usual care participants
with and 29% without CHD started taking lipid-lowering drugs. At year 4, total
cholesterol levels were reduced by 17% with pravastatin vs 8% with usual care;
among the random sample who had LDL-C levels assessed, levels were reduced
by 28% with pravastatin vs 11% with usual care. All-cause mortality was similar
for the 2 groups (relative risk [RR], 0.99; 95% confidence interval [CI],
0.89-1.11; P = .88), with 6-year mortality rates
of 14.9% for pravastatin vs 15.3% with usual care. CHD event rates were not
significantly different between the groups (RR, 0.91; 95% CI, 0.79-1.04; P = .16), with 6-year CHD event rates of 9.3% for pravastatin
and 10.4% for usual care.
Conclusions Pravastatin did not reduce either all-cause mortality or CHD significantly
when compared with usual care in older participants with well-controlled hypertension
and moderately elevated LDL-C. The results may be due to the modest differential
in total cholesterol (9.6%) and LDL-C (16.7%) between pravastatin and usual
care compared with prior statin trials supporting cardiovascular disease prevention.
The important etiologic role of circulating levels of low-density lipoprotein
cholesterol (LDL-C) in the development of atherosclerotic coronary heart disease
(CHD) is well established. Numerous randomized trials in the 1970s and 1980s
affirmed that lowering LDL-C levels with diet and/or drugs, such as bile acid
sequestrant resins and fibrates, reduced CHD event rates.1 However,
the total cholesterol reductions attained in these trials were modest (approximately
10%), and the correspondingly modest reductions in CHD mortality were offset
by small increases in noncardiovascular mortality, with no net effect on overall
mortality.1 In the mid-1980s, a new potent
and well-tolerated class of drugs, the 3-hydroxy-3-methylglutaryl coenzyme
A (HMG-CoA) reductase inhibitors (statins) provided the means to conduct randomized
trials in which total cholesterol reductions of 20% and greater could be sustained
long-term. These trials also allowed questions about the overall benefits
and risks of cholesterol lowering to be effectively addressed.
The lipid-lowering trial (LLT) component of the Antihypertensive and
Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT)2 (ALLHAT-LLT)
was originally envisioned as a free-standing double-blind trial to evaluate
the effects of cholesterol lowering with a statin drug in a population that
was older and more inclusive than those studied in prior trials. After successful
completion of the Cholesterol Reduction In Seniors Program (CRISP),3 a 2-year feasibility study, the concept was modified
and incorporated into ALLHAT as a randomized, nonblinded trial comparing pravastatin
treatment with a usual care control group in a moderately hypercholesterolemic
subset of the planned 40 000 ALLHAT participants. The principal objectives
of the ALLHAT-LLT were to evaluate the impact of large sustained cholesterol
reductions on all-cause mortality in a hypertensive cohort with at least 1
other CHD risk factor and to assess CHD reduction and other benefits in populations
that had been excluded or underrepresented in previous trials, particularly
older persons, women, racial and ethnic minority groups, and persons with
diabetes.2 Emphasis on primary care settings
was deemed important because of the study's substantial implications for these
providers and their patients. Despite the publication of more than 20 long-term
statin trials4- 13 since
ALLHAT began in 1994 and the publication of the National Cholesterol Education
Program (NCEP) Adult Treatment Panel Guidelines (ATP III)14 in
2001, ALLHAT-LLT remains the second largest long-term statin trial and addresses
a unique population.
This article presents results of the pravastatin vs usual care comparison
for all-cause mortality and CHD end points in ALLHAT-LLT. Results of the ALLHAT
antihypertensive trial appear in an accompanying article.15
The design of ALLHAT, including the LLT, and its participant and clinical
site recruitment and selection have been described previously.2,16,17 Briefly,
ALLHAT-LLT was a randomized, nonblinded, large simple trial conducted from
February 1994 through March 2002 at 513 clinical centers in the United States,
Puerto Rico, US Virgin Islands, and Canada. The intervention was open-label
pravastatin (40 mg/d) vs usual care. Participants were drawn exclusively from
ALLHAT, a 4-armed antihypertensive trial in which a calcium channel blocker
(amlodipine), an angiotensin-converting enzyme inhibitor (lisinopril), and
an α-adrenergic blocking agent (doxazosin) were each compared with a
thiazide-like diuretic (chlorthalidone). The doxazosin arm of ALLHAT was discontinued
in March 2000.18 ALLHAT-LLT participants originally
assigned to doxazosin continued in the LLT with their original visit schedule
and were offered open-label chlorthalidone for antihypertensive treatment.
The specific eligibility criteria for the ALLHAT-LLT included prior
enrollment in ALLHAT (age ≥55 years and stage 1 or 2 hypertension with
at least 1 additional CHD risk factor); fasting LDL-C level of 120 to 189
mg/dL (3.1 to 4.9 mmol/L) for those with no known CHD, or 100 to 129 mg/dL
(2.6 to 3.3 mmol/L) for those with known CHD (the upper limit was 159 mg/dL
[4.1 mmol/L] prior to April 5, 1994, but was changed in light of 4S4 findings); and fasting triglyceride levels lower than
350 mg/dL (3.9 mmol/L). Participants were excluded who were currently receiving
lipid-lowering therapy, taking large doses of niacin, or taking probucol in
the last year; were known to be intolerant of statins or to have significant
liver or kidney disease (serum alanine aminotransferase [ALT] >100 IU/L or
serum creatinine >2.0 mg/dL [176.8 µmol/L]) or other contraindications
for statin therapy; or had a known secondary cause of hyperlipidemia. Enrollment
was discouraged for participants whose personal physicians recommended cholesterol-lowering
Eligibility for ALLHAT-LLT was based on the average of 2 fasting (calculated)
LDL-C measurements19 taken at the ALLHAT baseline
and 1-month follow-up visits. Enrollment in the LLT took place an average
of 88 days after randomization into ALLHAT, from March 1994 through May 1998.
By telephone, participants were randomly assigned to pravastatin or usual
care in a ratio of 1:1. The concealed randomization scheme was generated by
computer, implemented at the clinical trials center (CTC), stratified by center
and antihypertensive treatment arm, and blocked in random block sizes of 4,
6, and 8 to maintain balance. All participants signed an informed consent
form, and all centers received institutional review board approval.
Follow-up visits for the ALLHAT-LLT were scheduled to coincide with
follow-up visits for the ALLHAT parent trial, ie, at 3, 6, 9, and 12 months
following randomization into ALLHAT and every 4 months thereafter. At each
visit, participants were questioned about intervening events since the previous
visit and were provided refills of study medications. Baseline fasting lipid
profiles and electrocardiograms (ECGs) were performed. Total cholesterol measurements
and resting ECGs were also obtained at the 2-, 4-, and 6-year visits. At these
same visits, a fasting lipid profile was obtained in random preselected samples
of usual care (5%) and pravastatin (10%) participants. Levels of ALT were
obtained for all ALLHAT-LLT participants at baseline and during follow-up
in accordance with US Food and Drug Administration requirements. All blood
samples were shipped with a frozen refrigerant pack to be analyzed at the
ALLHAT Central Laboratory (Fairview-University Medical Center Clinical Laboratories,
Minneapolis, Minn), a Centers for Disease Control and Prevention Standardized
All ALLHAT-LLT participants were advised to follow the NCEP Step I diet.20 Initially, pravastatin participants began with a
dosage of 20 mg taken each evening. The dosage was increased to 40 mg/d as
needed to achieve at least a 25% decrease in LDL-C. After the first 1000 participants
had been enrolled, a uniform dosage of 40 mg/d was adopted for all participants
in the pravastatin group. Study practitioners retained the option to lower
the dose of pravastatin, discontinue the drug if significant adverse effects
occurred, or prescribe other lipid-lowering interventions, including cholesterol-lowering
drugs not supplied by the study.2
The usual care group was treated for LDL-C lowering according to the
discretion of their primary care physicians. However, vigorous cholesterol-lowering
therapy in the usual care group was discouraged unless warranted by a change
in clinical circumstances.
Originally, the sample size estimate of 20 000 provided 80% power
to detect a 12.5% reduction in mortality rate in the pravastatin vs usual
care group with a 2-sided α = .05.2 With
changing scientific and community standards of practice for persons with prevalent
CHD,4 evolving recruitment experience of the
ALLHAT-LLT indicated that a sample size of approximately 10 000 participants
was the largest that could be realistically enrolled within the constraint
of drawing exclusively from participants already enrolled in ALLHAT. Although
10 000 participants would not provide adequate power for the originally
assumed 12.5% reduction in mortality, this revised sample size was estimated
to provide 84% power to detect a 20% reduction in mortality, a degree of reduction
comparable to that observed in the 4S study.4 This
estimated power was considered sufficient to continue the study.
The primary outcome for the LLT was all-cause mortality. Secondary outcomes
included (1) composite of fatal CHD or nonfatal myocardial infarction (MI)
(CHD events), (2) cause-specific mortality, (3) total and site-specific cancers,
(4) Q-wave MI identified in the biennial centrally and blindly coded ECGs
(included in CHD events), (5) health-related quality of life, and (6) major
costs of medical care. The last 2 outcomes are to be addressed in subsequent
reports. Other end points of interest (though not specified a priori as secondary
end points) were total incidence of stroke and heart failure.
Study end points, ascertained at follow-up visits, were reported to
CTC by the site investigators, who submitted death certificates for each death
and hospital discharge summaries for each hospitalized study event. Outcomes
were primarily based on clinic investigator reports, and pathology reports
were requested for cancer diagnoses. Each event report along with its documentation
underwent medical review at the CTC to verify the investigator-assigned diagnosis
or cause of death. In addition, searches for outcomes were conducted through
the Center for Medicare and Medicaid Services, the Department of Veterans
Affairs, the National Death Index, and the Social Security Administration.
A death was ascertained by clinic report or by match with the aforementioned
databases plus a confirmatory death certificate. Death certificates with unspecified
causes of death were submitted to a nosologist for International
Classification of Diseases, Ninth Revision (ICD-9)21 coding.
In addition to the death certificates and hospital summaries, further documentation
was requested for a random 10% sample of episodes of fatal CHD, hospitalized
nonfatal MIs, and strokes (hospitalized and fatal) for quality control review.
Data were analyzed according to participants' randomized treatment assignments
regardless of their subsequent medication status (intention-to-treat). No
imputation was used for missing data. Cumulative event rates were calculated
using the Kaplan-Meier procedure.22 An individual's
duration in the study began at randomization to ALLHAT-LLT and ended at the
date of last known follow-up. The log-rank test and the Cox proportional hazards
model were used to evaluate differences between cumulative event curves and
to obtain 2-sided P values. Only the proportional
hazards results are presented because P values obtained
by both methods were essentially identical. Hazard ratios, hereafter referred
to as relative risks (RRs), and 95% confidence intervals (CIs) were obtained
from the Cox proportional hazards model.22 For
fatal and nonfatal CHD, fatal and nonfatal cancer, cause-specific mortality,
and stroke, the Cox model was also used. Heterogeneity of effect in prespecified
and other subgroups was examined by testing for treatment-covariate interaction
with the proportional hazards model, using P<.05.
For other outcomes, including cancer deaths and overall and site-specific
cancers, comparison of proportions was used to evaluate differences between
pravastatin and usual care. Analyses are presented for total follow-up unless
A data and safety monitoring board appointed by the National Heart,
Lung, and Blood Institute met at least annually to review the accumulating
data for safety and to monitor the trial for either superiority or inferiority
of pravastatin compared with usual care. The Lan-DeMets version of the O'Brien-Fleming
group sequential boundaries was used to assess treatment group differences,
and conditional power was used to assess futility.23,24 Data
analyses were performed using SAS version 8 (SAS Institute, Cary, NC) and
STATA version 7 (STATA Corp, College Station, Tex).
Numbers of individuals screened and enrolled, vital status, and losses
to follow-up are depicted by treatment group in Figure 1. Ultimately, 10 355 participants were enrolled in
ALLHAT-LLT after exclusion of 2 participants due to poor documentation of
informed consent. The mean (SD) duration of follow-up was 4.8 (1.3) years
(maximum, 7.8 years). At the end of the trial, 84.8% of participants were
known to be alive, 12.3% were confirmed dead, 0.5% were reported dead with
confirmation pending, and 2.4% had unknown vital status.
Baseline characteristics, including serum lipid levels, are shown in Table 1. Mean total cholesterol was 224
mg/dL (5.8 mmol/L); LDL-C, 146 mg/dL (3.8 mmol/L); high-density lipoprotein
cholesterol, 48 mg/dL (1.2 mmol/L); and triglycerides, 152 mg/dL (1.7 mmol/L).
Participants' mean age was 66 years; 49% were women, 38% were black, 23% were
Hispanic, and 35% had diagnosed type 2 diabetes. A history of previous CHD
diagnosis was reported by 13% of pravastatin participants and 15% of usual
care participants. Higher mean total cholesterol and LDL-C values in LLT participants
without a history of CHD reflect differences in eligibility criteria. Other
baseline characteristics were similar in the 2 treatment groups.
Visit adherence is shown in Table
2. The percentage refusing to continue participation during the
trial was 0.3% (15/5170 pravastatin) and 0.6% (31/5185 usual care). At the
close of the trial 2.2% (113) in the pravastatin and 2.7% (139) in the usual
care groups had unknown vital status.
Adherence to assigned treatment declined over time (Table 2). For those assigned to pravastatin, adherence dropped from
87% at year 2 to 80% at year 4 (and 77% at year 6, though the participant
number was small). Approximately 70% to 75% of the participants reported taking
80% or more of their assigned pravastatin. About half of those discontinuing
pravastatin did so without citing a specific reason, while the remainder cited
adverse effects and other medical and nonmedical reasons. Specific adverse
effects data were not collected. Elevation of ALT to levels greater than 3
times the upper limit of normal (>150 IU/L) occurred in 0.4% (21/5170) of
the pravastatin group.
In the usual care group, crossovers to statin treatment increased from
8% at year 2 to 17% by year 4 (Table 2).
This increase continued in year 6, but the number of participants was small.
Among usual care participants with CHD at baseline, 32% (251/780) started
lipid-lowering drugs at some time during the trial. For those without CHD
at baseline, 29% (1279/4405) started lipid-lowering drugs; of these, less
than 5% (61/1279) had a preceding CHD event (data not shown).
Lipid and lipoprotein changes during the trial are shown in Table 3 and Figure 2. After 4 years of follow-up, total cholesterol levels decreased
by 17.2% in the pravastatin group and by 7.6% in usual care (Figure 2A). The resultant total cholesterol differential was 9.6%.
At 4 years calculated LDL-C levels decreased by 27.7% in the pravastatin group
and by 11.0% in usual care (Figure 2B).
The resultant LDL-C differential was 16.7%. Mean total cholesterol differences
(usual care − pravastatin) were 25.3 mg/dL at 2 years, 21.6 mg/dL at
4 years, and 18.9 mg/dL at 6 years. Mean LDL differences (usual care −
pravastatin) were 23.8 mg/dL at 2 years, 24.2 mg/dL at 4 years, and 17.2 mg/dL
at 6 years. (To convert values to mmol/L, multiply by 0.0259.) High-density
lipoprotein cholesterol increased by 3.3% in the pravastatin group and 2.4%
in the usual care group (data not shown). Body weight data were not gathered
The effect of pravastatin treatment on clinical outcomes is shown in Table 4; Kaplan-Meier plots and subgroup
analyses for mortality and CHD events are shown in Figure 3 and Figure 4.
All-cause mortality, the primary end point, did not differ significantly between
the pravastatin and usual care treatment groups (Table 4 and Figure 3A).
There were 631 deaths in the pravastatin group and 641 deaths in the usual
care group (RR, 0.99; 95% CI, 0.89-1.11; P = .88).
The 6-year mortality rate for pravastatin was 14.9%, and for usual care, 15.3%.
The results were similar when the unconfirmed deaths (27 pravastatin vs 28
usual care) were included (data not shown). Numbers of cardiovascular deaths
were similar in the 2 groups. There were more cancer deaths and slightly fewer
other medical deaths with pravastatin than usual care. None of the differences
in cause-specific mortality was statistically significant (Table 4).
Rates of CHD (fatal CHD plus nonfatal MI; Table 4 and Figure 3B)
and stroke (Table 4) were somewhat
lower in the pravastatin than in the usual care group. There were 380 CHD
events in the pravastatin group and 421 in the usual care group (RR, 0.91;
95% CI, 0.79-1.04; P = .16). The 6-year incidence
rate was 9.3% for the pravastatin group and 10.4% for usual care. There were
209 total strokes in the pravastatin group and 231 in usual care (RR, 0.91;
95% CI, 0.75-1.09; P = .31). Heart failure rates
were similar in the 2 groups (Table 4).
The 6-year incident cancer rates (Table 4) were similar in the 2 groups. The largest differences for
cancers at specific sites were for lung cancer (63 pravastatin vs 78 usual
care) and colon cancer (46 pravastatin vs 38 usual care). The number of participants
who developed breast cancer was similar in the 2 groups (34 pravastatin vs
37 usual care). All comparisons were nonsignificant
An important secondary objective of the ALLHAT-LLT was to address the
generalizability of the effects of cholesterol lowering to population groups
that had been underrepresented in prior trials. Thus, the homogeneity of the
results for mortality and for CHD events was assessed in prespecified subgroups
by age (≥65 vs <65 years), sex, race (black vs nonblack), and presence
or absence of diabetes (Figure 4).
There was no significant heterogeneity for any of these outcomes with regard
to age, sex, or history of type 2 diabetes. However, pravastatin showed a
significantly more favorable effect on CHD events (RR, 0.73 vs 1.02) in blacks
than in nonblacks (P = .03). Parallel analyses for
stroke showed a significantly less favorable effect (RR, 1.12 vs 0.74) in
blacks than in nonblacks (P = .03). No difference
in effect was observed in a parallel analysis of combined cardiovascular disease
outcomes (data not shown).
No statistically significant heterogeneity of the pravastatin treatment
effect was observed across the 4 ALLHAT hypertensive treatment groups. For
mortality, the RR in the chlorthalidone group was 1.03; amlodipine, 1.06;
lisinopril, 0.95; and doxazosin, 0.91; for the interaction P = .77. For CHD the RR in the chlorthalidone group was 1.05; amlodipine,
0.79; lisinopril, 0.90; and doxazosin, 0.83; for the interaction P = .43. Similarly, no statistically significant heterogeneity was
observed for subgroups defined by CHD status and LDL-C levels (with CHD, without
CHD plus LDL-C ≥130 mg/dL [3.4 mmol/L], and without CHD plus LDL-C <130
mg/dL [3.4 mmol/L]) (Figure 4).
ALLHAT provided a diverse population base for ALLHAT-LLT. This study,
comparing pravastatin with usual care, assessed the value of cholesterol lowering
in a population underrepresented in prior cholesterol trials—individuals
with well-controlled hypertension, almost half women, 38% black, 35% with
a history of diabetes, 55% at least 65 years of age, and 25% with LDL-C lower
than 130 mg/dL (3.4 mmol/L). Adherence to pravastatin in ALLHAT-LLT, 80% at
4 years of follow-up, was comparable to adherence in other large statin trials4- 11,18 and
decreased levels of total cholesterol by 17% and LDL-C by 28% from baseline.
However, unlike other statin trials, our study found no significant reductions
in total mortality, CHD, or stroke with pravastatin vs usual care.
There are several possible explanations for the findings of ALLHAT-LLT,
including the smaller than expected differential in total cholesterol between
the 2 treatment groups; the trial's unique participant population; and the
study's nonblinded design.
The usual care group had reductions of 8% in total cholesterol and 11%
in LDL-C at 4 years, in contrast to other placebo-controlled statin trials,
which observed little or no cholesterol reduction in the placebo groups.4- 9,11 The
resulting 9.6% total cholesterol differential was less than half the average
for the 8 other long-term statin trials with at least 1000 participants4- 11 (Table 5) and comparable to the cholesterol
differential attained in prestatin trials using resins, niacin, diet, or fibrates.1 Under the assumption of no change from baseline total
cholesterol levels among participants in whom follow-up extended to 4 years
but whose cholesterol was not measured at their fourth annual visit, the true
total cholesterol differential might have been as low as 8.8% (14.9% in pravastatin
vs 6.1% in usual care), and the true LDL-C differential might have been as
low as 15.1% (24.0% in pravastatin vs 8.9% in usual care).
The effect of attaining only a modest total cholesterol differential
is best appreciated by plotting the natural log of the odds ratio (ln OR)
and the 95% CI for mortality (Figure 5A)
and CHD events (Figure 5B) in each
of the trials in Table 5 vs the
mean cholesterol differential between the treatment and control groups in
that trial. In addition, regression lines based on a prior meta-analysis of
45 cholesterol-lowering trials of 2 or more years' duration published before
the end of 200025 are plotted for comparison.
While the observed 95% CI of ln OR for all-cause mortality and CHD in ALLHAT-LLT
do not exclude the null value, they are also consistent with the predicted
OR for a 10% cholesterol reduction. However, because of the modest cholesterol
differential between pravastatin and usual care, ALLHAT-LLT lacked the power
to discriminate between the expected reductions in mortality and CHD events
and the null hypothesis.
The reduction in study power was not due to low mortality rates; the
number of deaths in the ALLHAT-LLT usual care group (641) differed only slightly
from the estimate (625) used in the revised power calculation for a sample
size of 10 000. Moreover, the numbers of participants and deaths in ALLHAT-LLT
were larger than in any other statin trial except the Heart Protection Study
(HPS).9 The lack of study power likely was
due to a failure to achieve a total cholesterol differential sufficient to
yield the anticipated 20% reduction in mortality, which would be about 20%
according to the regression model (Figure
Finally, ALLHAT-LLT did not test the widely advanced hypothesis that
statin treatment reduces CHD risk and mortality by mechanisms independent
of cholesterol lowering (eg, anti-inflammatory effects).26 Furthermore,
the observed differences in both CHD events and all-cause mortality in ALLHAT-LLT
were consistent with those predicted for a 10% total cholesterol differential
in a model based on trials using a wide array of cholesterol-lowering interventions.
ALLHAT-LLT is the only published statin trial, to our knowledge, conducted
exclusively in treated hypertensive participants. In a meta-analysis of 3
published pravastatin trials,26 treatment was
associated with only a 14% CHD event rate reduction (P =
.03) in 6568 hypertensive participants vs 33% (P<.001)
in 13 200 nonhypertensive participants. The difference in CHD event rate
reduction between hypertensive and nonhypertensive participants was statistically
significant and might help explain the modest 10% CHD event rate reduction
in ALLHAT-LLT. However, in the hypertensive subgroup of the HPS,9 simvastatin
treatment was associated with the same 24% reduction in CHD event rates (P<.001) as in nonhypertensive participants.
ALLHAT-LLT included larger proportions of older participants, women,
blacks, and Hispanics than any other statin trial completed. However, subgroup
analyses of ALLHAT-LLT, like those of prior statin trials,9,27 do
not show age- or sex-related differences in RRs for CHD event rates. The RR
for pravastatin vs usual care was significantly lower in blacks than nonblacks
for CHD events (Figure 4B) but was
higher for strokes, with no overall difference for combined cardiovascular
events (data not shown). In the absence of racial differences for the efficacy
of pravastatin regarding all-cause mortality or other end points, the biological
significance of the racial differences for CHD and stroke is unclear.
Although only a small proportion of ALLHAT-LLT participants had overt
CHD at entry, they were predominantly a cohort with multiple CHD risk factors,
considered "CHD equivalents" by the 2001 NCEP-ATP III.14 Other
than the HPS,9 which contained a different
mixture of participants with CHD, atherosclerotic cardiovascular disease,
diabetes, and treated hypertension, this category of participants—not
purely primary or secondary prevention—has not been explicitly addressed
by prior statin trials. While the 14% of LLT participants with overt CHD at
entry had higher event rates than those with comparable LDL-C levels (<130
mg/dL [3.4 mmol/L]) but without CHD, the pravastatin/usual care RRs for mortality
and CHD were similar in both groups. These RRs were also unaffected by LDL-C
level at baseline. By contrast, HPS9 reported
similar estimates of benefits with simvastatin at all levels of LDL-C, while
a pooled analysis of 3 large pravastatin trials27 suggested
benefit only in participants with LDL-C levels higher than 125 mg/dL (3.2
mmol/L). None of these unique subgroups, including blacks, seems a likely
explanation for the results of ALLHAT-LLT.
ALLHAT-LLT was a nonblinded trial, designed and carried out during a
period in which a series of landmark trials4- 11 and
guidelines14,20 stimulated the
prescription of statins and progressively broadened the indications for their
use in individuals targeted by ALLHAT-LLT. This may have contributed to the
use of open-label statins in the usual care group. Because the study was not
blinded, there may also have been greater use of nonpharmacologic cholesterol-lowering
interventions in usual care than in pravastatin, although changes in participants'
diets, exercise habits, and weight were not examined in ALLHAT.
Do the results of ALLHAT-LLT indicate a need to draw back from the widespread
use of statins? When viewed in context, the overall experience with statins
remains highly favorable (Table 5, Figure 5). In the 8 prior large long-term
statin trials,4- 11 a
mean 20% cholesterol reduction was associated with a 30% reduction in CHD
events (95% CI, 26%-33%) and a 17% reduction in all-cause mortality (95% CI,
12%-22%). After including ALLHAT-LLT, the 9 large long-term statin trials
now show a 27% reduction in CHD events (95% CI, 23%-31%) and a 14% reduction
in all-cause mortality (95% CI, 10%-18%) associated with an 18% reduction
in mean total cholesterol level. Both results remain highly significant statistically.
There remains little evidence in ALLHAT-LLT or elsewhere that statins specifically
increase any category of noncardiovascular mortality.
ALLHAT-LLT demonstrated no significant difference between pravastatin
and usual care groups in all-cause mortality or combined fatal and nonfatal
CHD. However, in the context of the modest cholesterol differential, the results
are consistent with the evidence from other large trials. Indeed, the overall
findings from the 9 large long-term statin trials (including ALLHAT-LLT) leave
little doubt regarding the broad efficacy and safety of this treatment in
the prevention and treatment of atherosclerotic cardiovascular disease. In
the absence of evidence for increases in any category of noncardiovascular
mortality, the ALLHAT-LLT results should be interpreted as consistent with
current recommendations for cholesterol control in the prevention and treatment
of cardiovascular disease. These results emphasize the need for obtaining
an adequate reduction in LDL-C in clinical practice when lipid-lowering therapy