The SPRINT Blood Pressure Study: Small Numbers, Questionable Significance
November 12, 2015
Finally the SPRINT numbers have been published, and the results are both unimpressive and of questionable significance. In a highly selected and very small (2500 people over age 70) group of patients with heart disease, there was a reduction in death of 3.5/1000/year in the tightly controlled hypertensive group compared to the less controlled group. See BRCT on the left.
CLICK HERE to see BRCTs of SPRINT endpoints and others.
The SPRINT numbers are in direct contradiction to several other studies that show increased mortality with aggressive blood pressure reduction. A 2013 VA study of 650,000 veterans with chronic renal disease not on dialysis, those with systolic blood pressure under 120 had a 1.5 times greater death rate than those with blood pressure of 140-160. While this is a relative risk, the overall mortality in the group was 73/1000, so the increase in death in the tightly controlled risk was likely much more than the decrease in death found in the SPRINT cohort (I have asked the authors to provide me with absolute numbers). In a 2010 JAMA study of 6500 people with diabetes and CAD, those whose blood pressure was controlled tightly (under 120) had a higher mortality of 8/1000 compared to those without tight blood pressure control.
Why is SPRINT not consistent with past research, and why is the media suggesting that this small study (2500 people over 70) should supplant all prior studies that include hundreds of thousands of people? It is hard to know. The very design of SPRINT is suspect, and the small reduction in death could be explained in ways that are not related to blood pressure reduction. To remind people about SPRINT, it looked at a highly screened group of older hypertensives at high risk of heart disease. For those people who had normal pressures above 120, a second agent was added to bring the pressure below 120 in the intervention group compared to keeping the blood pressure as is in the control group. The cohort was followed for 3.3 years. The intervention group experienced a reduction in death of 3.5/1000/year compared to the control group, a reduction of cardiac death of 2/1000/year, and a decrease in cardiac endpoints of 5/1000/year. There were side effects from intensive control, including an increase in life-threatening hypotension of 3/1000/year and an increase in severe in severe kidney disease of 6/1000/year in the intervention group. There was no measurement of how many people suffered fatigue, dizziness, and other common side effects of intensive blood pressure lowering other than a statement that about 30% of people in both groups experience serious side effects, although the numbers showed that 12/1000 more people had serious side effects in the intense group than the control group. Non-serious side effects, however, are often more significant in terms of an older person’s quality of life, and it is vital we know that number if we want to know the impact of intensive blood pressure treatment. Shared decision making requires more information than SPRINT delivers.
Why would SPRINT find any mortality benefit from pressure reduction in a group of people that prior studies suggest could be harmed from aggressive blood pressure management? In the study, if pressure was deemed too high, then another blood pressure medicine was added to bring the pressure below 120. But in this group of people with heart disease the addition of certain medicines (such as beta blockers or ace inhibitors) can reduce mortality independent of any effect they have on blood pressure. These medicines are known to reduce death rates in people with congestive heart failure and coronary artery disease. So how are we to know if the SPRINT results show us anything meaningful about blood pressure reduction in people with heart disease? Perhaps the addition of these drugs lowered mortality despite their effect on blood pressure. By designing SPRINT in this fashion, the authors were able to make assumptions about blood pressure reduction that are not necessarily cause and effect. This could be why the SPRINT data is not consistent with all the prior data; it asked a different question.
For now, SPRINT should not impact how doctors and patients manage blood pressure. There is enough good data to convince us that aggressive blood pressure management is not clinically effective and is likely deleterious. SPRINT does not alter what we already know. It is a small highly screened study with a very suspect design that could have achieved its results completely unrelated to blood pressure reduction. Even if the data is perceived to be accurate, its results are hardly impressive. A reduction of death of 3.5/1000 means that 997/1000 people with intensive blood pressure treatment do not benefit. We in primary care know that intensive blood pressure control in the elderly can cause dizziness, fatigue, falls, syncope, and many other side effects that impair quality of life. We also know from this study that there is an increase in kidney disease and life threatening hypotension in the treatment group. So let’s take SPRINT for what it is: one among many studies that does very little to help us care for our patients or to facilitate shared decision making.