tighter blood pressure control? the SPRINT trial

so, the SPRINT study (Systolic Blood Pressure Intervention Trial) was finally released at the Am Heart Assn meeting (see prior blog http://gmodestmedblogs.blogspot.com/2015/09/aggressive-blood-pressure-management.html  ), along with all of its details (see htn SPRINT trial nejm2015 in dropbox, or DOI: 10.1056/NEJMoa1511939). This was a large study looking at 2 different blood pressure targets and the associated clinical outcomes. The study was sponsored by the NIH/NHBLI (as opposed to the more usual drug company sponsors of studies). details:

--9361 patients (mean age 67.9, mean age of those >75 was 79.8; 30% non-Hispanic black, 11% Hispanic, 58% white, baseline BP 140/78, total chol 190 mg/dl, HDL 53, 52% on aspirin, 14% current smokers, BMI 30, 1.8 BP meds used) with systolic blood pressure (SBP) of 130-180 mmHg and an increased risk of cardiovascular disease (defined as one or more of: clinical or subclinical cardiovascular disease such as coronary artery calcium score, LVH or low ankle-brachial index, but excluding those with stroke; chronic kidney disease CKD with eGFR 20-59; 10-year risk of CVD of >15% by Framingham risk score; or >75 yo -- 61% qualitfied by Framingham risk score, 28% each by age>75 or CKD), study in 102 clinical sites in the US, including Puerto Rico. All of the elderly in the study were from the community, not assisted living or nursing homes

--the actual inclusion criteria for blood pressure were a tad more complex: SBP 130-180 mmHg on 0-1 med, 130-170 on up to 2 meds, 130-160 on up to 3 meds, or 130-150 on up to 4 meds. no diastolic inclusion criteria

--patients with diabetes, stroke and other conditions mentioned below were excluded

--randomized to achieve a systolic of <120 mmHg vs a target of <140mmHg.

--the goal of the trial was not to test different med strategies, but to allow flexibility in the meds used. but, they did encourage using meds with the strongest CVD outcome data (thiazide-type diuretics, calcium blockers, ACE-I/ARBs). thiazides-type diuretics were preferred as first line agents (preferentially chlorthalidone), loop diuretics in those with advanced kidney disease, b-blockers for those with coronary artery disease. amlodipine was the prefered calcium channel blocker. they also reinforced lifestyle changes

--primary composite outcome was: MI, other acute coronary syndromes, stroke, heart failure or death from cardiovascular causes

--proposed followup up of 6 years

--results:

    ​--the achieved SBPs were 121.4 mmHg and 136.2 in the intensive and standard treatment groups. mean diastolic was 68.7 mmHg in the intensive and 76.3 in standard groups

    --mean number of BP meds used was 2.8 in intensive and 1.8 in standard groups

    --primary composite outcome occured in 1.65% per year in the intensive group and 2.19%/year in the standard group, with an HR=0.75 (0.64-0.89), p<0.001

        --heart failure HR=0.62 (0.45-0.84), p=0.002

        ​--death from cardiovasc cause HR=0.57 (0.38-0.85), p=0.005

        --all-cause mortality HR=0.73 (0.60-0.90), p=0.003

        --no significant difference in MI, acute coronary syndrome, stroke

        ​--in those with normal baseline renal function, >30% reduction in eGFR to <60 was more common in those on intensive therapy, HR=3.49 (2.44-5.10), p<0.001

        --and, in general the curves showed improvement beginning about 1 year after therapy started, increased over the next 3 years, then plateaued

    --subgroup analysis: basically no statistically significant difference by prior hx of CKD, age < or >75, sex, race, previous hx of cardiac disease, or baseline systolic pressure

    --rates of serious adverse events were higher in those in the intensive group: hypotension (2.4 vs 1.4%), syncope (2.3 vs 1.7%), electrolyte disturbances (10.1 vs 7.4%), acute kidney injury or failure (4.1 vs 2.5%), but not injurious falls, and orthostatic hypotension more common in standard treatment group (18.3 vs 16.1%)

    ​--study stopped early, after 3.26 years, because of statistically significant benefit

so, this is a really good study, showing pretty clearly that for the group overall, and several of the subgroups, there was a clear benefit of a lower blood pressure target. there are a few issues of note:

--one of the most significant exclusions in this study was patients with diabetes. this clearly undermines the generalizability of the results, since hypertension and diabetes tend to run together for physiologic reasons (50% of patients with hypertension have diabetes and 80% of diabetics have hypertension: hypertension is associated with insulin resistance, and in therefore part of the Metabolic Syndrome; insulin causes renal sodium retention and contributes to hypertension). so, excluding diabetics effectively excludes a very large % of patients with hypertension. other important exclusions affecting the study's generalizability include: all participants were >50yo, none had SBP>180 mmHg, none had a prior stroke or had predominantly diastolic hypertension (though they did treat diastolic hypertension to <90 mmHg). other exclusions included those with urinary protein excretion>1g/d, and symptomatic heart failure or EF<35%.

--this study did not look at potentially subtle adverse effects of tighter blood pressure control, especially in the elderly. for example, an Italian study found that elderly with mild cognitive impairment had more rapid cognitive decline with lower achieved blood pressure (see http://gmodestmedblogs.blogspot.com/2015/04/too-low-blood-pressure-and-cognitive.html  ). also, many elderly have initial orthostatic hypotension (vs the usual one, which in this study is defined as blood pressure after standing one minute -- as described in the remarkably long and complete supplementary materials) and initial orthostasis is much more common than the usual one (see http://gmodestmedblogs.blogspot.com/2014/11/orthostatic-hypotension.html  ).  they did in fact find more hypotension and syncope in the intensive treatment group, though more usual orthostatic hypotension inthe standard group. this divergence in findings suggests that usual orthostatic hypotension is a less useful marker of the real clinical outcomes than initial orthostatic hypotension.  All of this raises the concern about falls, broken hips, even subtle cognitive deterioration -- huge quality-of-life changes which might be more important to the patient than the potential decrease in cardiac events.

--although it makes sense to stop a trial early because of clear benefit, there is a very important potential downside (which i personally have never seen stated): a longer-term study allows us to see if there are important adverse events that do not become manifest in the foreshortened study: perhaps renal disease related to more aggressive blood pressure control really becomes clinically evident after 5-8 years (and they did find a short-term increase in CKD in this study). or there are more falls/broken hips (which in this short study did not qualify as "injurious falls", though there was more hypotension and syncope). or there is more cognitive decline. so, stopping a study early as in this case focuses on the benefits of the intervention and minimizes the potential for finding longer-term risks (of course, some studies may be stopped early because of adverse events, but the point is that we are looking at potentially lifelong therapies and a short study may not really illuminate the long-term risks).

--there are clear concerns about basing treatment strategies on clinic blood pressures. the USPSTF has finally come out with strong recommendations about the use of home-based or ambulatory blood pressure monitoring as more relevant markers of future clinical events​  (see http://gmodestmedblogs.blogspot.com/2015/01/uspstf-recs-on-ambulatory-blood.html  )

--but, that all being said, i was concerned that the JNC-8 recommendations for BP treatment was surprisingly simplistic (though simple guidelines are much easier to adopt overall, i was pretty concerned about not incorporating reasonable data that diabetics, those with renal failure, etc, needed tighter control) -- see http://gmodestmedblogs.blogspot.com/2013/12/jnc-8.html for a review of JNC-8.  I do think that the pendulum swung a little far from the much more stringent goals in prior JNC guidelines. it is really great to have the SPRINT trial as a large, randomized intervention trial actually achieving a real difference in achieved blood pressures and showing a clear benefit (for example, one of the largest prior trials, the HOT, or Hypertension Optimal Treatment trial, only achieved a 2.9/3.1 BP difference between the groups, so was not so useful clinically).

--so, bottom line, I do think the SPRINT trial fills a large gap in our knowledge about blood pressure goals. but there are a few caveats: we should be very vigilant around issues of orthostatic hypotension (ie, regularly checking orthostatics, including initial orthostatic hypotension) and cognitive decline, especially in the elderly; and that we probably should use more home-based or formal ambulatory blood pressure monitoring more to make sure that the blood pressure is not a lot lower than it appears to be on clinic visit. but, given these issues, it seems pretty clear that we really should try to achieve lower blood pressure goals than in JNC-8.  My review of the supplemental materials indicate that at least some of these issues will be evaluated, including looking specifically at dementia, cognitive decline and MRI changes in the brain (eg, small vessel ischemic changes) and published in future articles.