Manhattan Institute for Policy Research.
Subscribe   Subscribe   MI on Facebook Find us on Twitter Find us on Instagram      

Event Transcript
March 19, 2003

The NHLBI/ALLHAT Blood Pressure Drug: Science or Politics?

ROBERT GOLDBERG: Today we are going to talk about the ALLHAT (The Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial ) study. Our briefing is designed to address the claim put forth by the ALLHAT Steering Committee that millions of Americans currently on an existing drug or combination of drugs that stabilizes their blood pressure should switch to a diuretic. In the case of patients in Medicaid or managed care plans, they are encouraged to switch to a diuretic and then, if that does not work, they can request that their physician authorize other medications.

The experts we have assembled today will explain that once the study results are considered in context, serious questions emerge regarding this recommendation. This is a critical time to raise these questions since the official hypertension guidelines for physicians will be issued some time in the next couple of months. 

We are not here to advocate the use of one drug over another, per se. The mission of the Center for Medical Progress is to ensure that medical policy and physician decisions that affect the health of the American people are based on the best science and best medicines available, and not on shortsighted monetary considerations. 

We are fortunate today to have four of the nation’s leading experts on hypertension and hypertension medications on our panel. Leading off our panel discussion today will be Mark Houston, Associate Clinical Professor of Medicine at Vanderbilt University Medical School and Director of the Hypertension Institute at Saint Thomas Hospital in Nashville. Dr. Houston was also one of the principal investigators in the ALLHAT study and one of the first physicians to voice his concerns about the ALLHAT recommendations to the clinical community.

Following Dr. Houston will be Dr. Franz Messerli. Dr. Messerli has a stellar reputation as a leader in the treatment of heart disease and is also the Director of Clinical Hypertension at the Ochsner Clinic Foundation Hospital.

Dr. Ralph Hawkins will be our third presenter. Dr. Hawkins formerly practiced in Canada, and is now professor of medicine at the University of Tennessee. Dr. Hawkins will discuss the comparative cost of diuretics to other medications considering the many complications associated with diuretic therapy. 

Dr. Michael Weber is our fourth and final panelist. Dr. Weber is one of the world’s leading experts on the treatment of hypertension and was also a principal investigator in the ALLHAT study. He is the Associate Dean and Professor of Medicine at the SUNY Health Science Center, a past president of the American Society of Hypertension, and current editor of the American Journal of Hypertension. Dr. Weber has also been quoted in a number of mainstream media outlets, including the AP Wire Service and the Wall Street Journal.

DR. HOUSTON: The concern that brings all of the panelists here today is patient care. As clinicians and researchers, we are very concerned about studies that misrepresent the efficacy of individual treatments, or combinations of treatments, for our patients.

My goal today is to put the ALLHAT study in perspective through an overview of the copious research that preceded it, and to express my concerns with the ALLHAT study design and recommendations.

In broad terms, the ALLHAT Steering Committee has clearly misinterpreted and misquoted some data—and produced findings that are contradictory to an enormous number of scientific studies on hypertension that preceded the ALLHAT report.

Diuretics, very specifically, Chlorthalidone the study diuretic used in ALLHAT (and Hydrochlorothiazide), are probably not the preferred course of hypertensive therapy in most North American hypertensive patients. In reality, they may not be cheaper or more economical over the long term for the treatment of hypertension due to a variety of complicating factors. For instance, long term compliance and prescription refills with diuretics is less than what has been observed in most of the other classes of hypertensive drugs, particularly calcium channel blockers and ACE-inhibitors, which were the two used in conjunction with the ALLHAT trial.

When considering the true cost of diuretics, the cost of complementary drugs must be considered as well; for instance, potassium supplementation is required for diuretics. There are also a variety of metabolic concerns that have to be taken into account with diuretic therapy, which can trigger the onset of diabetes. There is evidence that the ALLHAT study did not critically analyze the data related to these metabolic concerns or the target organ damage, including heart attacks, stroke, and kidney disease. Consequently, the ALLHAT study has to be compared to other studies that have shown opposing data trends.

One other very practical point that was lost in the ALLHAT study was that optimum blood pressure control requires at least three to four anti-hypertensive drugs. A diuretic may be part of that regimen, but not necessarily the initial or the preferred treatment in the majority of hypertensive patients.

My observations are based on many published hypertension trials that show contradictory findings to the ALLHAT study. These studies include approximately 140,000 patients worldwide, compared to about 40,000 in the ALLHAT trial. These studies all document the linkage between cardiovascular disease and renal mortality during the use of diuretic therapy. Again, these studies all underscore the message that diuretics may not be the preferred initial therapy in hypertensive patients—yet this is the very premise set forth by the ALLHAT Steering Committee.

The first group of studies, 23 different clinical trials in all, looked at diuretics compared to other drugs used in conjunction with beta-blockers. During these studies there were huge reductions observed in heart failure and stroke, but reductions in heart attack occurrences were sub-optimal: only about 16% compared to an expected 25%. What is more disturbing is that none of these studies have shown that the risk of kidney disease was prevented or decreased with a diuretic-based regimen. In fact, as Dr. Hawkins will tell you, diuretics may produce an increased incidence of renal disease and renal cancer.

These earlier studies also raised concerns about dose of the diuretic. With high dosages, there were increasing rates of metabolic events, like diabetes, Hyperlipidemia and low potassium, which can negate some of the beneficial effects of lowering the patient’s blood pressure. Dr. Messerli was actually one of the first to describe that beta-blockers by themselves have not been shown to lower heart disease and heart attack as a primary treatment for hypertension. This contradicted a number of earlier trials suggesting that it had, findings which were very similar to those from the ALLHAT study. There is agreement now that these findings were misinterpreted for approximately 20 years.

Calcium channel blockers are the other group of drugs studied in the ALLHAT study, specifically Amlodipine. There is a very large quantity of data available on calcium channel blockers and hypertension. The bottom line with calcium channel blockers related to the treatment of hypertension is that they are better at reducing strokes than diuretics and beta-blockers. Furthermore, they are at least equal to or better than diuretics in terms of cardiovascular mortality, coronary heart disease, and especially kidney disease related to end state renal disease and hemodialysis. The best data available on calcium channel blockers indicates that these drugs are clearly superior to beta-blockers as mono-therapy in virtually all respects for cardiovascular disease. This finding was repeated again in an article published just this week in The Journal of Hypertension. The researcher in this study, Giuseppe Mancia, a hypertension specialist in Milan, Italy, writes “calcium channel blockers, specifically Nifedipine GITS, is a preferred first line therapy for hypertensive diabetics.” Dr. Mancia’s study, like many others, totally contradicts the ALLHAT study’s conclusions.

Dr. Mancia’s work tracked Nifedipine GITS versus a diuretic and showed that Nifedipine GITS reduced heart attack, heart failure, stroke, and cardiovascular disease, equal to a diuretic. But even more importantly, Nifedipine GITS was found to be superior to a diuretic in all cases in reducing mortality from any vascular or non-vascular disease, and reduced the onset of diabetes and frequency of renal disease compared to diuretic monotherapy. Again, these findings are totally different from those put forth in the ALLHAT study.

An ACE-inhibitor was the other drug used in the ALLHAT trial. There are large numbers of studies on ACE-inhibitors. The literature indicates that they are superior to diuretics and beta-blockers for total mortality, fatal and non-fatal cardiovascular disease and heart attacks, congestive heart failure and kidney disease of any type. They are better in reducing the onset of new diabetes and they are equal to beta-blockers and diuretics for prevention of stroke. Supporting these findings are twenty-eight published and pending studies that contradict the ALLHAT recommendations.

There was another foreign study published last month, the Australian National Blood Pressure Study (II) on an ACE-inhibitor , Enalapril versus Hydrocholorothiazide, a diuretic very similar to the ALLHAT drug Chorthalidone.

Despite equal reductions in blood pressure, 26/12 HG, the ACE-inhibitor was superior to the diuretic . There was a significant--11%--greater reduction in cardio vascular events (heart attacks) with ACE-inhibitors compared to diuretics.

The ALLHAT study did not examine ARBs (angitenson receptor blockers) in any of its clinical drug studies, but there are a number of other studies showing that ARBs may, in fact, have equal, or better, efficacy in stroke prevention than diuretics. Now, as you can see, once we survey these other trials, comprising 140,000 patients, the ALLHAT study findings are flatly contradicted.

But let us look at what ALLHAT in fact showed. As most of you know, the ALLHAT study was a double-blind, multi-center trial with one primary question: “what happens to the incidence of heart attack in hypertensive patients going forward, prospectively, using a calcium blocker, an ACE-inhibitor, an alpha-blocker or a diuretic, as preferred primary therapy and then adding on secondary drugs.” ALLHAT also recorded data on secondary outcomes, which included stroke, heart failure, and end stage renal disease. ALLHAT tracked 42,000 patients for approximately 4.5 to 5 years.

It is very important to recognize that all the patients in the ALLHAT study were over the age of 55. The patients who entered the ALLHAT study had to have a certain level of blood pressure; but it is also important to note they were mostly treated patients who had drugs withdrawn early on, during the first stages of the study. Their average age was 67. Clearly, given the parameters of the study, you cannot necessarily make clinical correlations from an older, high-risk, hypertensive population to a younger, no-risk, hypertensive population. If you are under 55, with no risk factors, the clinical ability to make correlations from the ALLHAT study is very low. This was an elderly hypertensive population with risk factors.

The ALLHAT study documented blood pressure reduction of about 14/11 HG similar to many other published studies. But it also required at least two drugs to achieve that goal. This is not surprising—we already know that may take at least two, perhaps even four, drugs in most patients reach the new blood pressure goals, which are probably going to be lower than present recommendations. 

This is brings us to the crucial issue. When you design a study, if it is not designed properly at the beginning—if it is flawed—the results cannot be interpreted appropriately.

First of all, ALLHAT follows an elderly population. It is also a high-risk population. Drugs were withdrawn early in the protocols that may in fact have been causally related to the events described later in the study, especially congestive heart failure as opposed to the drugs subsequently prescribed for patients.

The second design flaw is that the study mandated an artificial drug combination. Clinicians will tell you that this is not at all what we do for patients today. The entire study is artificial in terms of the recommended drug combinations.

Finally, the study utilizes a flawed end point definition of heart failure, which even the ALLHAT investigators concede. The definition of congestive heart failure driving most of the ALLHAT end points is misguided. In fact, there are a lot of crossovers to other therapies that makes design issues difficult.

What about the protocols in the study itself? Did they measure blood pressure accurately? Was there good data suggesting that they had 24-hour blood pressure control? Standard procedure is that, during study planning, you must correct blood pressure differences to correlate with the prevention of target organ damage. And in the ALLHAT study, Lisinopril, the ACE-inhibitor, did not control blood pressure as well as it might have because it was dosed once a day. This is inappropriate, because it should be taken twice a day, meaning that for eight hours every day the patients who took Lisinopril did not have adequate medication to ensure blood pressure control.  African Americans in particular may not have achieved a reduction in blood pressure during the time of day when the highest incidence of stroke and other cardiovascular events occurs, which is usually between 3:00 A.M. and 10:00 A.M. in the morning, when blood pressures are higher. They probably did not achieve what is called a “nocturnal dip” in blood pressure. This is a major flaw in the study dosing.

Overall, the Lisinopril treated patients had 2mm greater systolic blood pressure, but in the African Americans group, it was 4mm higher. Once we correct for blood pressure, most of the differences in cardiovascular events for this group can be accounted for.

With Amlodipine there was a minimal difference in systolic and diastolic blood pressure. (The diastolic being the bottom number, systolic being the top number.)  But, related to systolic, it was slightly higher with the Amlodipine group. When one looks at the diastolic pressure, however, most of the drugs were pretty much the same.

Again, the evidence shows that ALLHAT did not really achieve effective 24-hour blood pressure control with Lisinopril. In fact, virtually all the difference in heart failure in CVA between the ACE-inhibitor Lisinopril and the diuretic Chlorthalidone can be explained by the differences in blood pressure. The heart attack incidence for the ACE-inhibitor would have been superior to the diuretic, after making that correction, by 3% to 6% in favor of the ACE inhibitor—if blood pressure had been properly controlled.

The diuretics also induce large numbers of biochemical abnormalities. Cholesterol was higher in the diuretic group than with the other two drugs. Potassium was lower, requiring supplementation. The estimated Glomerular filtration rate, a measure of kidney function, was worsened in the diuretic-treated patients. The incidence of diabetes was increased significantly; it was somewhere between 30% and 60% greater in the diuretic-treated patients versus those who received Amlodipine or Lisinopril. The primary end point in the ALLHAT study was heart attack. But there was no reported difference in heart attack between the three drugs.

The incidence of stroke was different. The Amlodipine group tended to have the lowest incident of stroke, Lisinopril the highest, with Chlorthalidone in the middle. But this data is based on poor blood pressure control. Even in the original data we see that only black hypertensive patients had an increase in stroke risk that was significant. Caucasians had no difference among the three drugs. But when you correct for the blood pressure level, most of the stroke incidence go away even in the African American treatment group.

Heart failure, which drove most of the secondary end points in the ALLHAT study, was noted very early on in the study—when the original medications were withdrawn. This indicates that the incidence of heart failure probably had more to do with the drugs that were withdrawn than subsequent treatments. At the beginning and at the end of the four-year study, the curves of Lisinopril and Chlorthalidone for heart failure come back together, indicating that no difference in heart failure rates would have been observed had the ALLHAT investigators focused on the withdrawal issue as opposed to the treatment issue.

The overall mortality was the same between the three drugs. The relative risk of end-stage renal disease, as Dr. Hawkins will discuss, indicates that the diuretics may have increased end stage renal disease and renal insufficiency. The measurement for that is called the GFR and it declined significantly in the diuretic-treated group compared to both Amlodipine and Lisinopril groups.

If blood pressure had been controlled for properly we find very different results. The reported reduction of stroke on diuretic therapy for African American patients was a difference of 40%. However, if you correct for the blood pressure difference, which was 4mm systolic for African Americans, that number drops to 24%.

If you look at the entire group of patients the difference was 15%, but if you correct for blood pressure it drops to 7%, for a 2mm change, and that is insignificant. The incidence of heart failure for the entire group was reported to be 20%. If you correct for blood pressure, it is actually better with the ACE-inhibitor, by about minus 1%., not a significant difference. And if you correct for heart disease and heart attack, ALLHAT reported a 0% difference between the three drugs. If you look at the Lisinopril treated patients, it actually was better by 3% to 6%, in favor of the ACE inhibitor correcting for blood pressure. In other words, the ACE-inhibitor outperformed the diuretic, for heart attack, which is not what the ALLHAT committee reported.

Now we can put the ALLHAT findings in perspective with the published hypertension literature: 140,000 patients showing significant reductions in heart failure, stroke, and heart disease, and where calcium channel blockers were shown to be superior in reducing stroke by 15%--a very significant finding. In that same body of literature, ACE-inhibitors were shown to be better than diuretics in reducing heart attack by 12% to 20%. These findings are totally contradictory to ALLHAT’s conclusion that diuretics outperformed those other drugs.

In short, controlling blood pressure is paramount and may require multiple drugs, at least three to four; diuretics may be part of the regimen, but are not the initial or preferred therapy. They are, in fact, contraindicated in many patients. Diuretics have a lower compliance and a lower refill rate. The ALLHAT study design limits the assessment of the role of diuretics in younger populations, such as those without significant risk factors, and in patients given more appropriate or scientifically-based combinations of drugs to reach our new blood pressure goals.

Ed Frolich, in a recent article in the The New England Journal of Medicine wrote, “health care providers must deal with a specific relationship between the physician and the patient. The relationship is where the therapeutic tire meets the road and there is no place for absolute or categorical answers. Population based studies of therapies help to point the way, but are not analogous to the care of individual patients.”

My original premise, that ALLHAT’s study design is flawed, seems to be supported by the weight of the evidence. Consequently, we need to think very seriously and critically about the ALLHAT study recommendations.

DR. FRANZ MESSERLI: These are the principal findings of the ALLHAT study regarding the primary outcome which was coronary heart disease, death and non-fatal M.I. Now, as you can see, there is absolutely no difference between the chlorthalidone, Amlodipine and Lisinopril arm.

Dr. Ed Frolich was privileged to write an editorial on the ALLHAT study stating, “The major finding of ALLHAT was a striking and unequivocal null; namely, that the occurrence of coronary heart disease, death, non-fatal M.I. was identical.” Again, null result, identical outcome.

Before the authors started in 1996, they reassured us by stating that secondary endpoints will be regarded only as “soft data” that will attest, confirm or supplement the primary endpoint. How come, all of a sudden, in that very same editorial diuretics are being touted as superior? Needless to say, the press picked up on this conclusion very quickly. The ALLHAT study was quoted in The Wall Street Journal and The New York Times, and it was argued that diuretics could save Americans billions of dollars if all patients were switched to diuretics. Perhaps most challenging was the statement that patients and doctors are now on notice that diuretics should almost always be first choice treatment with other drugs added later if needed.

Well, for beginners, hydrochlorothiazide is not chlorthalidone. Physicians are very familiar with hydrochlorothiazide but not familiar with chlorthalidone. The two diuretics are, in fact, different.

Also, when we closely look at the ALLHAT data we observe that patients who were on chlorthalidone had a between 40 and 65% higher risk of developing new onset diabetes than patients on the ACE inhibitor. Patients on chlorthalidone had a between 18% and 35% higher risk to develop diabetes than those on Amlodipine. This is hardly surprising. We have known for years that diuretics and beta-blockers in monotherapy and also in combination increase the risk of developing diabetes. The authors of ALLHAT attempt to assure us by stating that these metabolic differences did not translate into more cardiovascular events or did not cause excessive mortality. Again, hardly a surprise, the follow-up of ALLHAT is way too short for the development of diabetic complications.

In January of this year, the Lancet published a study on patients who have been diabetic for more than 20 years. This group had a twenty to thirtyfold higher incidence of sight threatening retinopathy than patients who had diabetes for less than ten years. Michael Weber and myself have written a letter to JAMA, pointing out that hypertension treatment is lifelong and that a study of two to four years is unlikely to give us any sense as to the long-term risk of diabetes with chlorthalidone. As you just heard a few minutes ago, the relationship between diabetes and diuretics is well known. In 1989 The New York Times published a study documenting the heart attack risk associated with diuretics. In the 15 years since that study, the prevalence of diabetes in the U.S. has doubled. I think it is irresponsible to suggest that every hypertensive patient, regardless of whether or not they are at risk for diabetes, should be started on a diuretic.

Of note, calcium antagonists have a very mixed reputation. As some of you may remember, there was a so-called calcium antagonist controversy that started in the early 1990's, but at late as August 2000, The New York Times stated that the use of such drugs as calcium antagonists has led to nearly 85,000 unnecessary heart attacks each year worldwide. Clearly, the ALLHAT study puts these concerns to rest since again, there was no difference in the primary endpoint (coronary artery disease) in these 40,000 patients among Amlodipine, Lisinopril and chlorthalidone.

Perhaps even more disturbing was the hypothesis that calcium antagonists may increase the risk of cancer, as was suggested by numerous papers from the same group. Quite in contrast, the ALLHAT study puts these concerns to rest. There was no cancer risk with Amlodipine. To the contrary, the mortality from non-cardiovascular causes was significantly lower for Amlodipine than it was for chlorthalidone.

However, I would like to note that hypertension is itself a cancer risk. This is not surprising, because hypertension on one hand, and cancer on the other, share certain risk factors, such as smoking. Smokers are at high risk for hypertension and cancer. Therefore, it is not surprising that patients with hypertension have a greater cancer risk. We published this finding last year in the American Journal of Medicine.

What bothered us particularly about the ALLHAT study is that the form of cancer associated with diuretics is renal cell carcinoma. This is very, very common. There is one study in which diuretics, when taken by women, conferred a cancer risk of 4.9. In other words, women who use diuretics have a five-fold greater risk of renal cell cancer than women who did not.

This is just one study. It could be anecdotal. But when we looked at this more carefully, we discovered not just one study. We found ten independent studies. In every single one of them, the risk of renal cell carcinoma with diuretic therapy was elevated.

Not only do we have ten independent case control studies, but there are also three cohort studies showing that the risk ratio of renal cell carcinoma with diuretic diabetes is distinctly elevated—more so in women than in men. What we see in all these studies is that renal cell carcinoma was related to the cumulative use of diuretics in a manner very similar to cigarette smoking. The longer you smoke, or the longer you take a diuretic, the greater the cancer risk.

There is some experimental data from Vanderbilt that shows that continuous treatment of rats with hydrochlorothiazide provokes degenerative alterations and cell death in the distal tubule. A personal communication from Vanderbilt indicated that the distal tubular cells treated with hydrochlorothiazide looked like tumor cells and had markers of tumor cells. Although these findings were in rats, we should consider that they, in the face of ten independent case control studies and three cohort studies in humans should not be taken lightly. The relationship between diuretic therapy and renal cell carcinoma is of concern and it casts a shadow on the rosy pictures of diuretics reducing cardiovascular morbidity and mortality.

Diuretics have been, are, and will remain, a cornerstone in the antihypertensive arsenal. However, ALLHAT very simply documents that chlorthalidone, Lisinopril and Amlodipine are equally effective in reducing coronary heart disease. Moreover, calcium antagonists are safe for the treatment of hypertension and ACE inhibitors (such as Lisinopril) are antihypertensive drugs — no more, no less. Thus, we conclude that the verdict of the ALLHAT study that diuretics should be preferred for the treatment of hypertension is inappropriate.

DR. RALPH HAWKINS: The justification—or lack thereof—for the medical benefits of diuretic therapy has been discussed by my fellow panelists. I am going to talk about the purported economic benefits. The ALLHAT investigators put forward a claim that diuretic therapy is economically more efficient than those of these other drugs, but the justification for that finding comes solely from a superficial analysis of drug acquisition costs.

This graph reflects therapy costs for each of the ALLHAT drugs. This data was very recently published in The Harvard Heart Letter. The drug costs can be found, by anyone who’s interested, at I have created a category for drug ratio costs, identifying chlorthalidone arbitrarily as a drug cost ratio of 1, and then extrapolating from that to the drug cost ratio for the other treatments.

Diuretic therapy, particularly Chlorthalidone therapy is known to cause depletion of an important mineral: potassium. A very sizeable number of patients in the ALLHAT study required potassium supplementation therapy and a very large proportion had dangerously low potassium levels in spite of that supplementation.

Because of the vital importance of maintaining normal blood potassium, these supplements are commonly prescribed and, as you can see, they are not cheap. The cost ratio for one tablet of brand name potassium per day or for two tablets daily of generic potassium is factored into the drug cost ratio. Please note that supplementation with two potassium pills a day, which is the usual North American dose, effectively obliterates the favorable drug acquisition costs for chlorthalidone.

When the ALLHAT paper was first reported in December, there were no significant differences in end stage kidney disease reported. However, they only dedicated a single sentence to what I think is the most important finding of the ALLHAT study. Perhaps I am a biased kidney specialist, but that one sentence said that there was a slower decline in kidney function with the newer drugs than in patients treated with diuretics. The ALLHAT study did not elaborate on that finding at all, and it did not elaborate on the fact that the rate of decline of kidney function in the diuretic treated patient population was faster than what you would expect for patients of that age, during that duration of time. Now, none of the patients developed end stage kidney failure, but the patient-follow-up in that study was only an average of 4.9 years.

I have subsequently researched several databases to examine the relationship between diuretic use and end stage kidney failure in North America. Based on these figures, we can predict that there will be 6,000 more patients starting on end stage renal failure treatment (dialysis) in 2003 that can be attributed to the increased use of diuretics. Each patient on dialysis costs Medicare $45,000. Those 6,000 patients’ additional patients in dialysis treatment will cost Medicare $272 million--attributable to diuretic therapy. In other words, spending $48 million on diuretic therapy in 2001 will generate $272 million of additional costs in 2003. That is a cost ratio of 5.6:1.

Dialysis patients also have non-Medicare compensated health costs, and if we include those as well, we add $111 million to the total cost and generate an 8:1 cost ratio. Furthermore, once a patient starts on dialysis, they have—on average—a 3-½ year life expectancy. For every dollar we spend on new diuretic therapy this year, we would be responsible for causing $28 of lifetime total health care costs for patients thereafter. Please note—Medicare pays both the up front expenditures for diuretic therapy as well as an additional $28 for complications. On balance, the lifetime Medicare cost (including complications) is probably closer to $20 than $28, but we are still paying an additional $20 for every dollar we invest in diuretic therapy.

MICHAEL WEBER: Like Mark Houston, I was a principal investigator for the ALLHAT study. I was responsible for the trial conducted at the Veterans Affairs hospital in Long Beach, California. When I relocated to Brooklyn a few years ago, I handed over the trial to one of my associates. The VA hospital did an excellent job and recruited a great many patients.

That is why I was disappointed when I read the ALLHAT report and the accompanying press releases. I felt that this was a scientific study that was being portrayed very much in political and economic terms, and in conflict with the results that were actually generated.

It critical to remember that the “HAT” in ALLHAT stands for Heart Attack Trial. That was the focus of the study. It is almost forgotten that the three drugs ALLHAT studied had identical effects on heart attacks, even though the diuretic had a strong blood pressure advantage due to the study’s design flaws. I think it is perfectly legitimate to infer that had blood pressure been treated appropriately, and had the blood pressure effects for the other drugs been comparable to those of the diuretic (which was the intent of the study), the ACE-inhibitor might have proved superior as far as heart attack incidence and mortality was concerned.

I was astonished when I saw that the ALLHAT officials had published a flat, assertive statement that diuretics are superior. How can diuretics be said to have strong advantages when mortality, if anything, is slightly higher as result of using them in non-black patients? The ALLHAT conclusion, to me, just does not make sense.

As far as other studies are concerned, two very important events have taken place in the few weeks since ALLHAT was published. First of all, as Mark Houston told you, an Australian trial was published just a few weeks ago in The New England Journal of Medicine that compared an ACE-inhibitor and a diuretic head to head. They achieved identical blood pressure readings in the two groups and, as Dr. Houston pointed out, the ACE-inhibitors had an 11% advantage in terms of mortality and cardiovascular events. [I would add one important caveat to the Australian study. The patients in Australia were about 90% Caucasian, and 10% Asian, as opposed to the ALLHAT study where over 1/3 of the patients were African American. That is an important qualifier.]

The LIFE study was published in The Lancet in March last year and came to the FDA in January of this year for a public hearing. The FDA is just about to approve a labeling claim for an angiotensin receptor blocker called Losartan indicating significant protection against stroke. In the LIFE study Losartan, combined with a diuretic, was compared to a beta-blocker combined with a diuretic. This is significant because diuretics and beta-blockers were supposedly the winning combination in the ALLHAT study. The newer drug combination of an angiotensin receptor blocker with a diuretic outperformed the ALLHAT combination by 25%, in terms of reducing stroke. I think that this study will generate a lot of discussion in the very near future.

When we look at the totality of the ALLHAT study and compare it to the hypertension literature, particularly more recent studies, we are left with a very different impression than the press releases and statements made by the National Heart, Lung and Blood Institute.

I will say that I am not opposed to diuretic therapy, for the simple reason that diuretics are a necessary treatment for many patients. I believe very much that angiotensin receptor blockers, ACE inhibitors, and calcium blockers should be the baseline therapies for many of our patients. But, as both Franz and Mark pointed out, we cannot control hypertension with just one drug in most people and complementary therapies will often include a low dose of diuretics. We are not diuretic critics. We are just saying that diuretics should not be the universal foundation treatment for hypertension.

Finally, I want to reiterate that I am disappointed that a government agency, in this case a federal agency, which should exhibit the highest standards of integrity in reporting medical research, has taken an inconclusive and flawed study and used a very aggressive public relations campaign to put out what I and my colleagues believe to be a misleading statement.

The ALLHAT study has become an exercise in political and economic rhetoric and inappropriate because we are talking about patients whose lives depend on the standard of care we provide. Putting out an unsubstantiated message to physicians on how to treat people who are at risk for stroke and heart attacks is irresponsible.


Center for Medical Progress.


Center For Medical Progress Breakfast Forum

In December 2002, the National Heart, Lung, and Blood Institute (NHLBI) released the results of its 8-year Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT). The ALLHAT study claimed to show that the latest generation of blood pressure drugs were no more effective than traditional, and less expensive, diuretics. Consequently, the ALLHAT study encouraged doctors to use diuretics as a first-line treatment for hypertension. Critics of the study point out that diuretics carry serious side effects that make long term use much more expensive and problematic than may appear at first glance. The Center for Medical Progress brought together four of the nation’s leading experts on hypertension and hypertension medications to discuss the ALLHAT’s study’s protocols and recommendations.


Robert Goldberg, Ph.D., Senior Fellow, Manhattan Institute; Director, Center for Medical Progress


Mark Houston, MD

Franz Messerli, MD

Ralph Hawkins, MD

Michael Weber, MD

Home | About MI | Scholars | Publications | Books | Links | Contact MI
City Journal | CAU | CCI | CEPE | CLP | CMP | CRD | ECNY
Thank you for visiting us.
To receive a General Information Packet, please email
and include your name and address in your e-mail message.
Copyright © 2015 Manhattan Institute for Policy Research, Inc. All rights reserved.
52 Vanderbilt Avenue, New York, N.Y. 10017
phone (212) 599-7000 / fax (212) 599-3494