QUICKLINKS AND VIEW OPITONS
Blood Pressure: “Lower is better” is false
Monday, November 17, 2003 5:10 am Email this article
"Much of the worry about blood pressure--especially the ?lower is better'--is false," says mathematics and statistics professor Sidney S. Port from the University of California, Los Angeles. Prof. Port is the lead author of an excellent, ground-breaking, landmark, paradigm-shifting paper about the risks associated with systolic blood pressure. The paper reveals that the risk of death associated with systolic blood pressure is "greatly overestimated... for more than 70 percent of the population".
Their other discoveries include the following:
—“An older person can have a higher systolic pressure than a younger person with no increased risk.”
—A woman can have a higher systolic pressure than a man of the same age with no increase in risk.
—The current beliefs about the risks associated with rising systolic pressure are based on the wrong statistical model which has over-estimated the risk for much of the population and underestimated it for others.
—The risk of death from all-causes or cardiovascular events associated with systolic blood pressure does not increase up to at least the 70th percentile for a person of a given age and sex. The current model “greatly overestimated the risk for more than 70 percent of the population”.
—The risk of death increases sharply after the 80th percentile for a person of a given age and sex. The risk for this group is substantially greater than is currently recognized. This means that for these people more aggressive treatment may be necessary.
—“Treatment based solely on systolic blood pressure exceeding 140 mm Hg is not justified.”
First recognized in 1980
The idea that the risk of death associated with systolic blood pressure does not increase up to at least the 70th percentile, but increases sharply above the 80th percentile was first recognized back in 1980 by A. B. Keys by simple graphical analysis. Keys argued that the assumption that there was a constant increase in risk for a constant increase in pressure was wrong. Keys also argued that because of this incorrect assumption that the wrong statistical model was being used. Apparently no one took Keys seriously until Port’s group came along.
Re-analysis of Framingham Heart Study data
A reanalysis of the data from the National Institutes of Health’s National Heart, Lung, and Blood Institute long-running Framingham Heart Study by Prof. Port’s team confirmed what Keys had suspected. The assumption that there is a constant increase in risk associated with constant increase in pressure is wrong. The current cutoff for systolic hypertension of 140 mm Hg is based on the wrong statistical model because of this incorrect assumption. The current model “greatly overestimated the risk for more than 70 percent of the population” according to Port team. The current model “substantially overestimates the risks in the mid-range of pressures (about 130-170 mm Hg).”
Data supports their finding
Port’s team also found that contrary to widely cited interpretations, a reanalysis of the Framingham 18-year data showed that “the relation between systolic blood pressure and all-cause and cardiovascular mortality is not strictly increasing.”
Problem #1 with current advise: No studies showing pressure of 140-160 is harmful.
The results of previous studies support, rather than conflict with Port’s finding. “Despite the popular belief to the contrary, no prospective, randomized, placebo-controlled clinical trial shows significant survival benefit for people starting with systolic blood pressure between 140 mm Hg and 160 mm Hg,” said Port.
Problem #2: Most mortality studies based on diastolic, not systolic pressure.
Most studies which have related blood pressure to the risk of dying have been based on diastolic, not systolic blood pressure according to Port.
Problem #3: Studies draw conclusions based on wrong assumption.
Port also notes that many studies have drawn their conclusions and made recommendations based on the incorrect assumption that there is a constantly increasing risk with constantly increasing pressure.
Problem #4: No allowance for normal blood pressure.
The current model, which states that there is a constant increase in risk for a constant increase in pressure, does not allow for a “normal” systolic blood pressure of 120 mm Hg. According to the current model there is an increased risk associated with even “normal” pressure. According to this model the risk should be lower with a pressure of 115 mm Hg, even lower at a pressure of 110 mm Hg, and lower yet at a pressure of 100 mm Hg. The current cut-off of 140 mm Hg is “essentially arbitrary… with no real intrinsic justification,” according to Prof. Port.
Using the wrong formula gives the wrong answer
The reason that the relationship between blood pressure and risk appears to be linear, that is to say that the risk appears to gradually increase as blood pressure increases, is because the wrong statistical model has been used. This model, called a linear model, assumes that the data is linear as Port explains. “At a minimum, it must be demonstrated that the model actually ?fits’ the data and that it does not ?smooth away’ important features of the data,” says Prof. Port. But when Port’s group tested the data they found that it does not fit the model. They found that the assumption about the data being linear is wrong.
“Shockingly, we found that the Framingham data in no way supported the current paradigm to which they gave birth. In fact, these data actually statistically rejected the linear model,” said Port’s group. So Prof. Port and his colleagues introduced a new statistical model called the “logic splines”, that fits the data. Analysis of the data using this new model revealed that a number of commonly held beliefs about the risks associated with blood pressure are wrong. They found that:
—Although “systolic blood pressure increases at a constant rate with age… this increase does not incur additional risk.” “An older person can have a higher systolic pressure than a younger person with no increased risk.”
—A woman can have a higher systolic pressure than a man of the same age with no increase in risk.
—The risk of death from all-causes or cardiovascular events associated with systolic blood pressure does not increase up to at least the 70th percentile for a person of a given age and sex. The current model “greatly overestimated the risk for more than 70 percent of the population.”
—The risk of death increases sharply above the 80th percentile for a person of a given age and sex. This risk is substantially greater than is currently recognized. “Thus, these persons may require more aggressive therapy than was previously believed.”
—“Treatment based solely on systolic blood pressure exceeding 140 mm Hg is not justified.”
Port’s paper also points out that:
—“No randomized trial has ever demonstrated any reduction of the risk of either overall or cardiovascular death by reducing systolic blood pressure from our thresholds to below 140 mm Hg.”
—“Most trials relating blood pressure to mortality are based on diastolic, not systolic pressure.”
—“Most of the clinical trials combine all participants with systolic blood pressure higher than 140 mm Hg…” therefore, any benefit from decreasing systolic pressure may be from those people with a pressure above the 80th percentile “risk threshold”, who’s risk is greater than currently believe and would be expected to benefit.
Pressure of 140 mm Hg not justification for treatment
“Our results do not diminish the importance of antihypertensive drugs” that may lower other risks besides systolic blood pressure according to Port, but “treatment based solely on systolic blood pressure exceeding 140 mm Hg is not justified… the value should be age- and sex-dependent.”
Rules of thumb when treatment is advisable
The 70th and 80th percentiles are shown in the table. However Prof. Port’s group has come up with a rule of thumb for determining the 80th percentile where treatment is advisable. For men the formula is 120 + 2/3 5 Age; while for women the formula is 114 plus 5/6 5 Age. The 70th percentile, or the point at which there is no increase in risk, is about 10 mm Hg less than these values.
Little support from hypertension researchers
Port’s discovery seems to have received little or no support from the major players in hypertension research. In an accompanying editorial Michael H. Alderman of Albert Einstein College of Medicine in Bronx, New York said that “Somehow, logic and biological precedent make it hard to imagine any sharp and definitive cut-off between normal blood pressure and hypertension.”
“Logic is the art of going wrong with confidence.”
— Joseph Wood Krutch, (1893-1970), American critic, naturalist, and author of “The Modern Temper”
Dr. Edward Frohlich, editor-in-chief of the journal Hypertension was quoted in Heartwire as saying “The fact of the matter is that the higher the systolic and the higher the diastolic pressure, the greater is the risk.” [Provided that you make the wrong assumption about the data and use the wrong statistical model.}
“Generally the theories we believe we call facts, and the facts we disbelieve we call theories.”
The National Institutes of Health’s National Heart, Lung, and Blood Institute (NHLBI) issued a statement regarding Port’s findings saying that they found it “thought provoking” but “After careful review of this study, the NHLBI finds that it does not offer a basis for changing the current hypertension guidelines.” The statement goes on to say that the Sixth Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure “was prepared by a multidisciplinary expert panel that reviewed all of the available evidence… [and] found a clear linear relationship between systolic blood pressure, diastolic blood pressure and deaths.” What are they implying? That surely all of these smart people couldn’t be wrong?
“For every complicated problem there is a solution that is simple, direct, understandable, and wrong.
—Henry Louis Mencken (1880-1956), American editor, critic, and founder of “American Mercury”
National Heart, Lung, and Blood Institute backpedals?
The statement from the NHLBI seems to backpedal a bit saying “Treatment recommendations are not based solely on a patient’s blood pressure level.” But wait a minute. Prof. Port and his colleagues have simply said that “treatment based solely on systolic blood pressure exceeding 140 mm Hg is not justified.” That’s it. Whereas the Sixth Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure recommends that “normal” systolic blood pressure is less than 130 mm Hg and that an optimum pressure is less than 120 mm Hg. Clearly one of the two groups is wrong.
The statement from the NHLBI seems to continue to backpedal when it notes that other risk factors must be considered when deciding when to treat, and that “an older male patient who has a systolic blood pressure of 140 mm Hg, diabetes, and heart failure, is clearly managed in a different way from someone with a similar blood pressure level but no other risk factors.” But this is not something that Port’s group disputes. Their statement avoids the issue. The NHLBI should have addressed Port’s discovery that there is no increase in risk up to at least the 70th percentile.
Is it surprising that none of the major players have come out in support of Port’s discovery? Of course not. In order to admit that Port’s group is right, they would have to admit that they were wrong. What an embarrassment this would be. But why wouldn’t the NHLBI simply analyze all of the data from all of the studies that they have in order to either prove or disprove Port’s group’s finding? Jumping ahead from January to May seems to provide the answer.
The real reason the National Heart, Lung, and Blood Institute fought so hard
May is National High Blood Pressure Education Month. On May 4th the NHLBI and the National High Blood Pressure Education Program launched a campaign aimed at older Americans with new clinical advise regarding systolic blood pressure. The clinical advisory was published in the May issue of Hypertension: Journal of the American Heart Association. The cornerstone of the campaign is a recommendation that systolic blood pressure be emphasized in the diagnosis and treatment of hypertension in middle-aged and older adults. They state that “For many years, the importance of lowering systolic blood pressure to less than 140 mm Hg has been overlooked and under-emphasized.” They go on to say that “isolated systolic hypertension,” or ISH (systolic at or above 140 mm Hg and diastolic under 90 mm Hg)... for older Americans… is the most common form of high blood pressure. Sixty-five percent of all hypertensives over age 60 have ISH.” That explains it.
That is the reason why the NHLBI fought so hard to dismiss Port’s finding. This new campaign and new advisory was scheduled for release during National High Blood Pressure Education Month. The party had already been scheduled and they weren’t about to let some UCLA statisticians and mathematicians stop it. So instead of doing the right thing and postponing release of the new advisory apparently they decided “We have a schedule to keep”.
National Heart, Lung, and Blood Institute stops study with Cardura
On March 8th the NHLBI issued a press release stating that they had stopped one part of a large high blood pressure study known as the ALLHAT study (Antihypertensive and Lipid Lowering Treatment to Prevent Heart Attack Trial) involving the alpha-adrenergic blocker doxasosin (Cardura). The reason? Although people taking Cardura showed a similar reduction in blood pressure they had 25 percent more cardiovascular events than those who were taking the diuretic chlorthalidone and were twice as likely to be hospitalized for congestive heart failure! The press release from the NHLBI said “We cannot conclude that [Cardura] was harmful. Rather it didn’t work as well as the diuretic in reducing cardiovascular disease.”
Most interestingly, the investigators of the study concluded that “Blood pressure reduction is an inadequate surrogate marker for health benefits in hypertension.”
Shame on the National Heart, Lung, and Blood Institute
The NHLBI and the National High Blood Pressure Education Program should have postponed release of their “Campaign Aimed at Older Adults”. They should have postponed release of the corresponding clinical advisory. Before launching the scheduled campaign and publishing the new advisory they should have reanalyzed all of the data from all of the studies based on Port’s discovery. This way they could have either confirmed or denied Port’s group’s findings. Apparently they decided that preserving their reputation was more important than the lives of millions of Americans. Apparently they decided that preserving their reputation was more important than discovering the truth.
Shame on the National High Blood Pressure Education Program. And shame on the American Heart Association for publishing the clinical advisory before the issue had been resolved. Shame on all of them.
The door has been opened
At least the door has been opened thanks to Prof. Port and his colleagues. Eventually the NHLBI, the National High Blood Pressure Education Program, and the American Heart Association will have to face the truth. Hopefully it will happen sooner rather than later, but it could take years.
“A new scientific truth does not triumph by convincing its opponents, but rather because its opponents die, and the new generation grows up that is familiar with it.”
—Max Planck (1858-1947), German physicist and Nobel Prize winner
What about diastolic blood pressure?
The paper published in The Lancet only dealt with systolic, not diastolic blood pressure. So I emailed Dr. Port and asked if they had performed a similar analysis on diastolic blood pressure. “Yes,” Port replied. “The diastolic results have been submitted for publication. However, the paper is under review so I can’t discuss it further.” It seems that they have made a similar discovery with diastolic blood pressure according to one of the papers posted on Prof. Port’s website which says “Our investigations of the relation of cardiovascular risks to diastolic blood pressure in the Framingham data shows that the paradigm is false for those relations as well. We find that a model very similar to that with the systolic blood pressure prevails. (Port S, Demer L, Boyle N, Jennrich R, Garfinkel A. Diastolic blood pressure, cardiovascular risk, and mortality; in review.)”
2-22% of men and 3-40% of women at no increased risk.
The percentage of men with a systolic blood pressure between the current cutoff of 140 mm Hg and the seventieth percentile at which Port’s team found no increased risk is 2 percent at the age of 45-54-years-old, 20 percent at the age of 55-64-years-old, and 22 percent at the age of 65-74-years-old. The percentages for women in the same age groups are 3 percent, 24 percent and 40 percent, respectively. These people are at no increased risk even though the current model would have them to believe otherwise. An additional 10 percent of people fall below the eightieth percentile, the point at which treatment is advisable according to Port. Their paper also notes that half of the population has a blood pressure between 120 mm Hg and 140 mm Hg.
How to get a copy of this excellent paper
Although the paper is not specifically about obesity, I decided to include it anyway because of it’s extraordinary nature and the close connection between elevated blood pressure and body weight. I strongly suggest that anyone interested in hypertension get a copy of this paper. You can download a copy of the paper as well as opposing opinions from Dr. Port’s website at www.stat.ucla.edu/~scp/crrp/publications/PUBLICATIONS.html or you can request a copy from Dr. Port directly at:
Coauthors of the paper include Prof. Linda Demer, Ph.D. from the Depts. of Medicine (Cardiology) and Physiology; Prof. Robert Jennrich, Ph.D. from the Depts. of Mathematics and Statistics; Prof. Donald Walter, Ph.D. from the Dept. of Physiological Science; and Prof. Alan Garfinkel, Ph.D. from the Depts. of Medicine (Cardiology) and Physiological Science.
Port S; Demer L; Jennrich R; Walter D; Garfinkel A. Systolic blood pressure and mortality [see comments]. Lancet 2000 Jan 15;355(9199):175-80. Comment in: Lancet 2000 Jan 15;355(9199):159.
Raloff J. Salt trial provokes DASH of skepticism. Science News, 2000 May 27, 157(22):340-1.
Articles on the same subject can be found here:
Please feel free to share your comments about this article.
© Copyright 2003-2017 - Larry Hobbs - All Rights Reserved.