Can SPRINT Be Used To Inform Hypertension Treatment?

–The landmark trial results are not easy to apply to clinical practice.

Since the first breathless announcement of its preliminary results, SPRINT trial has been viewed as a landmark trial serving to establish a more aggressive approach to blood pressure management. It is now becoming increasingly clear, however, that the application of SPRINT in the real world will be much more limited and modest. Even trial leaders now acknowledge that the main message of the trial, setting a new standard goal of 120 mm Hg systolic blood pressure, requires very careful application.

At a debate at the recent American College of Cardiology meeting Michael Weber (SUNY Downstate College of Medicine) said that SPRINT was a “true landmark study” that “contributes to our understanding” but, he emphasized, “big questions remain” before its results can be applied to clinical practice.

In addition to the still unresolved questions surrounding blood pressure measurement in the trial (see below), Weber presented multiple reasons why the trial needs to be interpreted with caution.

Risk level: Weber pointed out that the benefits observed in SPRINT, which was performed in patients at medium and high cardiovascular risk, were not replicated in HOPE-3 in patients at lower risk. In fact, there was a troubling signal of increased risk in the group of patients who achieved the lowest, SPRINT-like blood pressure levels, Weber said. The possibility that there is a J-curve in important groups of people limits efforts to generalize the results.

Number of drugs used: Another point raised by Weber is that the benefits in SPRINT may not be a reflection of blood pressure level but, more simply, of the increased usage of drugs with established cardiovascular benefits in the aggressive treatment group. People in the aggressive treatment group took 3 drugs compared with 2 drugs for people in the control group. As a result, many more patients in the treatment group took diuretics, inhibitors of the renin-angiotensin system, beta blockers, and other cardiovascular drugs. Diuretics were received by a thousand more patients in the treatment group than in the control group, and there was a similarly sized imbalance in RAS blockers. Weber said it was possible that the imbalance in diuretics by itself could well explain the fact that there were 38 fewer heart failure events per 1,000 patients in the treatment group.

Diabetics: Another important limitation to SPRINT is that it did not include patients with diabetes, said Weber. ACCORD, which tested a question similar to SPRINT in patients with diabetes, did not find a benefit with aggressive treatment with targets below 140 mm Hg, and analysis of the ACCOMPLISH trial suggests that that diabetic patients may be at increased risk for coronary events if their blood pressure target is too low.

Younger patients: Weber also bemoaned the lack of information about how best to treat younger patients. “SPRINT, like virtually all other outcomes trials in hypertension, was performed in aging patients,” leaving practitioners with no good information on “how best to treat and optimize life-long care for younger hypertensive patients.”

Measuring Blood Pressure

A key point raised by Weber was the difficulty in applying SPRINT to clinical practice because of persistent questions about the way blood pressure was measured in the trial. Furthermore, in addition to the specific questions regarding SPRINT, hypertension researchers are well aware that the methods used to measure blood pressure in clinical trials are quite different from the way blood pressure usually is measured in clinical practice, and they are finding out that these differences may have important implications.

Most contemporary clinical trials follow a rigorous protocol, taking multiple readings over time with an automated blood pressure measurement machine. At the ACC meeting William Cushman (University of Tennessee), a member of the SPRINT steering committee, shed some light for the first time on the SPRINT BP measurement controversy. He acknowledged that BP measurement was not obtained in a consistent manner during the trial. As in other recent trials, blood pressure was “the average of 3 seated office BP measurements obtained using an automated measurement device,” but in some cases this was obtained with a healthcare practitioner in the room and in other cases the measurement was unattended. It is now well established that unattended measurements are lower than attended measurements. At a separate session on blood pressure measurement John Kostis (UMDNJ) estimated that the difference between attended and unattended measurements could lead to differences of 8 mm Hg. Cushman said that the SPRINT investigators will publish a paper fully exploring the discrepancy in methods used in SPRINT.

The other key point about blood pressure measurement is that the automated BP measurement technique used in trials– whether attended or unattended– differs importantly from the way BP is most often measured in clinical practice, where it is taken by a healthcare practitioner with a sphygmomanometer. When using a sphygmomanometer there may be large differences between practitioners. In addition, blood pressure is generally higher when taken in the presence of a healthcare practitioner.

So blood pressure measurement in SPRINT was not internally consistent and it was not consistent with clinical practice. At the simplest level, this means a goal of 120 mm Hg in SPRINT should not be translated to 120 mm Hg in clinical practice. So what does it mean?

The practical implication of this discrepancy can not fully be known at this time. There is no simple way to convert SPRINT measurements to clinic measurements. One recent paper found that research-grade measurements were on average 12.7 mm Hg lower than the routine measurements. But this does not mean that clinicians can simply adjust their readings accordingly, since the study also showed an enormous variability in this difference, so that research grade measurements were as much as 46.1 mm Hg lower or 20.7 mm Hg higher than clinic measurements.

Taking a stab at a number, Weber said he suspects the 120 mm Hg goal in SPRINT would likely translate to about 130 mm Hg in clinical practice, but he pointed out that “influential European experts” think the number could well be 140 mm Hg, which is the current standard, or, as he said, “nothing new.”

At the ACC Cushman acknowledged that “to replicate what was done in those trials, the assumption was that we were measuring blood pressure correctly in the clinic.” Cushman said that “you can’t just take a poorly done clinic BP and subtract 12 mm Hg or something like that and assume you have what we found in SPRINT.” Cushman’s concluding words were a recommendation to “be very careful about how we measure BP!”

Previous SPRINT Coverage:



  1. H. Holzgreve says

    There is only one single problem in SPRINT: the use of a method (unattended automatic blood pressure measurement) for a main component (blood pressure) in a major study which was and is not commonly used in clinical trials and in routine care of hypertensive patients. Indeed, a landmark trial, however, without clinical significance.

  2. James H Stein says

    Nice article. SPRINT helps us understand lower is better, especially in carefully selected patients. Even I have come around to believing we can’t pick an exact number to target and that SPRINT’s enthusiasm needs tempering. However, with so much of HTN undertreated, a message to be more aggressive in high CVD risk patients is a good one. It is amazing that after 100’s of thousands of patients being randomized over 50 decades of research in HT we still don;’ have a good handle on these things. Our research mission is like a random walk that moves towards the right answer, but is all of the place in the meantime.

  3. “lower is better”: what, always and for everyone? Lower without limit? This magic thinking is the opposite of scientific thinking.

    “It is amazing that after 100’s of thousands of patients being randomized over 50 decades of research in HT we still don’t have a good handle on these things.” Then a bit of intellectual humility is called for.

    “Our research mission is like a random walk that moves towards the right answer”: how the devil can you tell that it’s moving towards the right answer? Why should there be a single right answer?

    “but is all [over] the place in the meantime”: you said it.

Speak Your Mind