CardioBrief: More Controversy and Confusion Over SPRINT

— "Landmark" trial off by 5, 10, even 16 mm Hg from real-world?

by CardioBrief

MedpageToday

More questions and concerns are being raised about SPRINT, the NIH's "landmark" blood pressure-lowering trial.

In sharp contrast to the enormous amount of initial hype, many hypertension experts are now saying that the SPRINT trial is difficult to interpret and can't be readily applied to real-world practice or be used to change clinical guidelines.

Last week at a debate at the European Society of Cardiology meeting, the audience voted "thumbs down" against the use of SPRINT by guideline committees to lower blood pressure thresholds. Now, in the aftermath of that debate, more controversy and confusion have emerged over SPRINT, including basic questions about how the trial itself was conducted.

It turns out that a year after the initial SPRINT results were first announced, even veteran blood pressure experts and clinical trialists have not been informed about or have been unable to find out about crucial details about how the trial was performed. In particular, there is still considerable confusion about the precise method used in the trial to measure blood pressure (see below). Without a full and transparent understanding of this issue, these experts say, it is impossible to really understand or interpret the trial.

The problem is that the NIH and the SPRINT investigators have been working furiously -- even prior to the presentation and publication of the full results -- to sell the idea that the "landmark" trial offered powerful evidence that 120 was the new 140 -- that systolic blood pressure targets should be lowered for many patients to 120 mm Hg. Two months before the full trial results were published, the NHLBI director Gary Gibbons, MD, said, in a press release, that SPRINT "provides potentially lifesaving information that will be useful to health care providers... We are delighted to have achieved this important milestone in the study in advance of the expected closure date for the SPRINT trial and look forward to quickly communicating the results to help inform patient care and the future development of evidence-based clinical guidelines."

"Investigators designed SPRINT to determine the potential benefits of achieving systolic blood pressure of less than 120 mm Hg," the NIH said; but an array of hypertension experts outside the trial say that SPRINT can't support this claim. The problem is that the 120-mm Hg goal in SPRINT is not reflective of a 120-mm Hg goal in the real world due to major concerns over the precise details of blood pressure measurement in the trial.

All the hypertension experts I have consulted have agreed that the 120 mm Hg will need to be adjusted upward by a substantial amount in the real world. Unfortunately, to date the SPRINT investigators have not been fully aggressive about addressing this issue.

Confusion About BP Measurement in SPRINT

Blood pressure was measured in SPRINT using an automated measurement device, the Omron 907XL. Similar devices have been used in all the recent major hypertension trials. Automated BP measurements generally deliver readings significantly lower than readings in a typical doctor's office, in which a blood pressure cuff is slapped on a patient and the doctor or other healthcare professional obtains a reading. In a paper published online last week in Circulation, George Bakris, MD (University of Chicago), explained the importance of this discrepancy:

"BP recorded in research studies using the standard BP measurement guidelines, which mandate a rest period prior to measurement..., is on average 10/7 mm Hg lower than BP measured in routine clinical practice. Even when patients take their own BP using an automated sphygmomanometer while alone in an examining room, the mean systolic BP is still ~5 mm Hg higher than the corresponding awake ambulatory or home BP. Thus, systolic BP as measured in recent randomized trials, including SPRINT, is likely ~5-10 mm lower than that measured with traditional office BP measurement methodology. Consequently, targeting the systolic BP <120 mm Hg without using similar BP measurement methods as in the trials may increase the risk of serious adverse events by systematically overshooting the trial-based BP targets and potentially leading to hypotensive complications."

But it turns out that the actual discrepancy in SPRINT may be even larger than the discrepancy described by Bakris, as new details are emerging about the blood pressure measurement technique, or techniques, used in SPRINT.

Here is how the blood pressure measurement was described in the trial's original manual of procedures (MOP):

"During the 5 minute rest period, participants should be resting and should not be completing questionnaires or speaking with study staff. The staff member should leave the room during this 5 minute rest period. The following script can be used at this time.

"SCRIPT: 'I would like you to rest for 5 minutes before I begin taking your blood pressure. I will leave the room. When I return, I will not speak to you but will immediately begin to take your blood pressures. Do you understand?'

"With their agreement, leave the room and return in 5 minutes. Push the button on the machine and wait for the output. Record the systolic and diastolic blood pressure and pulse readings obtained at each of the three readings."

In other words, this is a variation of an "attended" blood pressure measurement, in which a healthcare professional remains in the room while the blood pressure is taken by the machine. Until now, however, it has not been known that the SPRINT manual of procedures was amended. In fact, as stated by SPRINT investigator William Cushman during the ESC debate, the BP readings were obtained automatically without the presence of a healthcare professional. The difference in an attended and an unattended BP reading may be substantial, since it has been well documented that blood pressure is higher when a healthcare professional is present.

In a letter in the journal Hypertension, Martin Myers, MD (Sunnybrook Health Sciences Centre, Toronto), described the actual method used in SPRINT, as related to him by Cushman:

"In SPRINT, study staff were trained to program an Omron 907XL (Omron Healthcare Inc, Lake Forest, IL) to wait 5 minutes and then record 3 readings at 1-minute intervals. After the device was activated, research staff left the examining room, with the patient then being alone during the 5 minute rest period and while the 3 readings were recorded automatically (W.C. Cushman, personal communication, 2016)."

A Conversion Factor to the Real World?

Sverre Kjeldsen, MD, PhD (Oslo, Norway), the participant in the ESC debate who has been critical of SPRINT, told me that he believes that this change in the protocol "was done on purpose in order to maximally-standardize BP measurement." He quickly added that there's "nothing wrong with that" but that it also requires "that 'something' must be added to 120 mm Hg. He said he thinks the adjustment should be about 16 mm Hg.

But the situation may be even more complicated. Darren McGuire, MD (University of Texas Southwestern), told me that he spoke with a research coordinator at a SPRINT site who confirmed that, at the coordinator's site, the blood pressure measurements were attended (performed in the presence of a healthcare professional). In other words, it is entirely possible that the change in procedures may not have been followed by all the centers.

Bryan Williams, MD (London), one of the moderators of the ESC debate, told me that "clear information on BP measurement methods should be published in the methods section so there is no ambiguity. SPRINT clearly shows an advantage of a lower BP -- the debate is how low is low according to the way BP is usually measured?" He said he suspects that the best goal is "probably nearer 130 than 120." He said it was "wrong" that the precise details of BP measures were "not clarified in the [New England Journal of Medicine] paper."

"We need less opinion and more information," Williams told me.

Given this uncertainty, Kjeldsen said that "there will be a proposal coming up to interview all participating 102 centers in the U.S. [on] how they actually did the BP measurements."

Previous coverage of SPRINT: