Systolic Blood Pressure Intervention Trial (SPRINT)
Two years ago, the results of SPRINT were announced in the New England Journal of Medicine (NEJM) (1). In this trial, the blood pressure (BP) medication is adjusted to achieve a target systolic BP of 120 millimeters of mercury (mm Hg). The investigation continued for 8 years and the total cost was $157 Million. In the announcement by National Heart, Lung and Blood Institute (NHLBI), the funding body it was claimed that the treatment resulted in reduced rates of cardiovascular events, such as heart attack and heart failure, as well as stroke, by almost a third. The risk of death was reduced by almost a quarter, as compared to the target systolic pressure of 140 mm Hg (2). According to Dr Lawrence Fine, Chief of the Clinical Applications and Prevention Branch at NHLBI:
“Our results provide important evidence that treating blood pressure to a lower goal in older or high-risk patients can be beneficial and yield better health results overall.”
It was concluded that the results of the SPRINT study reaffirm the critical importance of blood pressure control as the best approach to reduce complications of high blood pressure like heart attacks and strokes (3).
However within days of the publications, there were criticisms levelled at the conduct of the trial and the implications drawn from the results. In an article in the New York Times, Professor Harlan Krumholz (4) pointed out that:
- The study excluded people with diabetes or those who had experienced a stroke. This meant that the results applied to only about one in six of those already being treated for high BP.
- The potential benefits of lowering BP must be weighed against harms. The results indicated that one death would be avoided for every 300 people treated per year. On the other hand the intensive treatment increased the risk of kidney failure. Over the 3.3 years of follow-up, for every 100 people treated to achieve the lower BP, one more person suffered life-threatening low BP, one more fainted and two more had severe kidney problems.
- In the light of this information about harms it is important that patients are informed about the trade-offs so that they can decide the most appropriate treatment. I suspect most people would be amazed to discover that the benefits are minimal.
Dr Richard Amerling joins the critics
Further criticisms were made by Dr Richard Amerling (5) on the grounds that:
- It was not blinded, which means that patients and physicians knew what treatment group they were in. It is well established that in the absence of blinding there a genuine possibility of bias on the part of the investigator, especially when the effect of the treatment is small.
- The trial used a composite endpoint, which is a laundry list of cardiovascular outcomes, as opposed to a single endpoint. As Dr Amerling says:
“This is done to magnify the effect of an intervention where the actual incidence of individual outcomes is low. In other words, it’s a gimmick to get a “statistically significant” result.”
- The committee monitoring the trial stopped the study years ahead of schedule, presumably because the intensive treatment arm achieved a statistically significant reduction in the composite outcome. It is widely accepted that best practice is to comply with the original protocol. We are left with the suspicion that the investigators did not have confidence the effect observed would have persisted if the trial had continued as agreed at the outset.
- Patients with diabetes were excluded. However in the ACCORD trial (6), it was found that intensive BP treatment did not benefit diabetics.
- The intensive treatment group achieved the composite endpoint at a rate of 1.65% per year, as opposed to 2.19% in the standard treatment arm. According to the New York Times this meant that the intensively treated patients “had their risk of heart attacks, heart failure, and strokes reduced by a third”. In absolute terms this is a reduction of 0.54%, which is not exactly a big deal.
- Serious side effects such as kidney failure, low BP, electrolyte disturbances, and fainting, were more common in the intensively treated patients.
NEJM readers less than impressed
It is also interesting to consider some of the responses published by the NEJM (7). Here is a selection:
- “SPRINT is the type of trial that should not receive funding. It is poorly designed; the results are likely to be misinterpreted and cause more harm than benefits.”
- “Why did NIH without peer review post on its web site only the relative risk reduction outcomes that do not provide a full a full picture of the outcomes? Why was the trial rushed through peer review?”
- “We read with Interest this article which has a sound Methodology but having 14 exclusion criterion is not consistent with general thinking that having too many exclusion criterion creates difficulty…”
- “I am interested to know why this design was chosen. One guess would be to ensure a significant enough difference between the two groups. Yet wouldn’t simple logic, let alone rigorous scientific method, suggest this goal itself not a sufficient reason to create, in effect, an artificial “standard” approach?”
- “As a medical student I once shared a plane ride with a retired family physician of many years’ experience. His only advice to a young physician embarking on his career was to never start any of my patients on a new medication in the first two years after its approval due to concerns over side effects and other outcomes unforeseen at the end of phase III studies. I wonder if perhaps that advice should be extended to the outcome advantages derived from large RCTs.”
This is an all too familiar story. A trial produces some results that have little importance for the treatment of patients but are presented as a major breakthrough. The harsh reality is that it is in the interests of the investigators, the funding body and any commercial interests to do so. There are few kudos in negative results. As it happens in this case, there were also fundamental flaws in the design of the trial. Unfortunately this episode illustrates the weaknesses that are inherent in the checks and balances that are supposed to ensure high standards are maintained. The NHBLI does not cover itself in glory. A huge amount of taxpayers’ money was spent on the project yet it was allowed to proceed despite the limitations in the investigation. Presumably the senior staff felt obliged to praise the study and draw implications from it in order to justify the expenditure. Even more worrying is the failure of the NEJM to ensure that the submitted paper was thoroughly vetted during the review process. All of this is symptomatic of the bad science that affects so many of our institutions today, which is another story in itself (8).
Finally, it is somewhat ironic that despite these criticisms, the results of this study were apparently used to justify lowering target BP in recent guideline revision.