Blood pressure news has set the national media pulse racing. Two weeks ago the National Institutes of Health halted a major clinical trial early because, as The New York Times explained, the study had already “conclusively answered a question cardiologists have puzzled over for decades: How low should blood pressure go?” A lot lower than many doctors had thought, apparently: At-risk older adults who reduced their systolic blood pressure (the top number in the blood pressure ratio) to 120, down from the recommended upper threshold of 140, also dropped their risk for heart attacks, heart failure, strokes and death. Simply put, “aiming lower saves more lives,” ABC News concluded.
Strong words for a trial for which no one has seen the actual, detailed results. The scientists involved have not even fully analyzed the data. This study, the Systolic Blood Pressure Intervention Trial (SPRINT), could change recommendations for blood pressure management. But the “bottom line is that it is hard to say much without more information,” says Harlan Krumholz, a cardiologist at the Yale University School of Medicine, whose research focuses on improving patient outcomes. Doctors and reporters only had the slick NIH press release, which glistened with words such as “landmark” and “life-changing” and provided hype but little substance. It did not emphasize that the findings may only apply to a limited segment of the population or attempt to explain why these new results contradict those obtained in other trials; and it did not mention potential risks associated with taking multiple blood pressure drugs, which most trial subjects had to do.
The only numbers that the NIH provided in its press release about the trial, which involved 9,300 patients, were these: Compared with a systolic blood pressure of 140, a blood pressure of 120 reduced the risk of cardiovascular events such as heart attacks and strokes in adults over 50 by “almost a third” and reduced the risk of death by “almost a quarter.” (It’s unclear what “almost” means.) Although these drops in risk sound good, they may only apply to people like those enrolled in the study. These patients started out at high risk. In addition to having a systolic blood pressure over 130, they had one or more of these risk factors: established heart disease, chronic kidney disease, were 75 or older or had a 15 percent chance of having a heart attack or other cardiovascular event in the next 10 years as predicted by the NIH’s Framingham risk calculator, which is based on a long-running heart health study. Lower-risk older adults may or may not experience these same risk reductions when they aggressively lower their blood pressure, but the NIH announcement did not make this distinction. Instead, it framed the findings this way: “Lower blood pressure target greatly reduces cardiovascular complications and deaths in older adults.”
Even if the intervention does reduce a healthier person’s risk, the change may be too small to make a real difference to their health. Here’s one example. According to the Framingham risk calculator, a 55-year-old woman with borderline high blood pressure and cholesterol has only a one in 100 chance of having a heart attack in the next 10 years—so compared with the SPRINT subjects, she is at extremely low risk. Considering that the SPRINT patients who reached the target number of 120 had to take, on average, three different blood pressure drugs to do so—diuretics, calcium channel blockers and ACE inhibitors, among other options—should this low-risk woman really take a bunch of drugs to get her systolic pressure down to 120 so that she can reduce her risk of a heart attack by 30 percent, from one in 100 to one in 143? Among those at low risk for heart problems, drug-related risks—which include kidney problems and abnormally slow heart rate—could trump potential drug benefits. This is not to say that aggressive drug treatment might not help many older people, but there may be exceptions that can only be uncovered with access to the data.
And drugs seem to be the only option here. In general, older adults cannot reach such a low number simply by eating better or exercising more. According to George Bakris, director of the Comprehensive Hypertension Center at the University of Chicago Medicine, adhering to a very low sodium diet of about 2,400 milligrams per day only lowers blood pressure the amount that a single drug does. What is more, an estimated 10 to 15 percent of Americans with hypertension have a resistant form and must take four or more medications to get their blood pressure under control. And with each new medication comes additional risk, although the NIH press release does not mention anything about that. “They told us the good without telling us the bad,” says Jeremy Sussman, a research scientist with the Center for Clinical Management Research at the Veterans Affairs Ann Arbor Health System.
SPRINT researchers did, however, find drug side effects. Principal investigator Suzanne Oparil, director of the Vascular Biology and Hypertension Program at the University of Alabama at Birmingham, told Scientific American that there were risks but that they have not been fully analyzed yet; she also said that she could not provide any further information because the study has not yet been published in a peer-reviewed journal. (According to the Ingelfinger rule, which was created by the editor of The New England Journal of Medicine in 1969, journals can refuse to accept papers whose results have already been publicized in too much detail. This is yet another problem with the NIH’s decision to publish a press release containing teasers about the results: When doctors, patients and the media have follow-up questions, they cannot get answers until all the data have been published in a journal.)
Oparil is, however, confident that the benefits largely outweigh the risks or the press release would have said so. “Death and heart attacks are more important than feeling dizzy in the morning,” she says. But aggressive blood pressure lowering can pose serious risks to some people. People with advanced heart disease, for instance, should keep their systolic blood pressure above 115 or they may experience unique and dangerous problems such as reduced blood flow to the heart. So even if the benefits did trump the risks in the SPRINT trial overall, “that doesn’t mean that everyone in the trial was safe,” Sussman says. Again, physicians must have access to the data to discern whether or not this is true, which participants may have been harmed more than helped and which drug combinations may have been better or worse.
Another major question is why the findings from SPRINT differ from those seen in a previous trial, the Action to Control Cardiovascular Risk in Diabetes (ACCORD) Blood Pressure Trial, which was published in NEJM in 2010. This trial reported that, among individuals with hypertension and type 2 diabetes, those who achieved a systolic pressure of 120 compared with 140 did not have a lower risk of cardiovascular events. Those who more intensively lowered their blood pressure were, however, 2.5 times more likely to report serious side effects such as abnormal heart rhythms. Oparil suspects that the trial—which had about half as many participants as SPRINT—simply was not large enough to find statistically significant benefits associated with more aggressive treatment. Alternatively, the conflicting results may be because patient characteristics in the two trials differed. “Biology is complex and what will probably emerge as the correct answer is that ACCORD looked at one group of patients and SPRINT looks at another—as simple as that,” says Vinay Prasad, a health policy researcher and oncologist at Oregon Health & Science University. His idea further supports the idea that there is not a one-size-fits-all approach to treating hypertension, as the NIH press release seems to suggest. Most of the media coverage of SPRINT failed to even mention ACCORD. New trials do not erase older ones; their differences and similarities can help solve the puzzle of who benefits from treatment and who does not.
Some might argue that because the trial seemed to get high-magnitude lifesaving results, it is irresponsible to sit on findings while all the data are analyzed and papers are readied for publication. Patients and doctors do, after all, have a right to lifesaving information as soon as possible. But details matter in medicine, especially when it involves the use of intensive multidrug therapy. And outside researchers and practicing doctors should be given the chance to independently analyze the data before it is applied en masse to clinical practice.
Prasad says that “it is okay to announce that a trial has closed or that a study was terminated due to results or stopped early for futility or benefit. But to offer a teaser of the final results, an unanchored relative risk reduction, without more information—that is a disservice to the dissemination of science.” He adds that “medical results are not teaser trailers for a forthcoming film. They affect the lives of thousands or millions of people.”
No one doubts that the NIH will release the data in due time—the agency says it should happen in a few months—but it may not happen soon enough, as all the hype may already be changing decisions about medical care.