Epidemiologic studies show a graded increase in risk at systolic blood pressure (BP) >115 mm Hg and diastolic BP >75 mm Hg. However, it remains unclear whether pharmacologic BP lowering to these levels reduces cardiovascular (CV) events and is safe.
Observational analyses raised concerns about aggressive BP lowering to levels <120/80 mm Hg, suggesting the existence of a J-curve phenomenon with increased risk especially for coronary events. However, such analyses are potentially confounded. More reliable data are provided by randomized trials.
Most BP-lowering trials have evaluated patients with high pretreatment BP and have rarely achieved average systolic BP <130 to 140 mm Hg, except in patients with high-risk CV disease, such as recent myocardial infarction (MI) or heart failure (HF) or those with chronic kidney disease. Three recent trials have specifically compared intensive systolic BP lowering to targets of <120 to 130 mm Hg versus traditional guideline recommended targets of generally <140 mm Hg,[1–3] while the older HOT trial (Hypertension Optimal Treatment) randomized patients to diastolic BP targets of ≤90, ≤85, or ≤80 mm Hg.[4] These trials differ significantly in their design, patients' characteristics and CV risk, BP regimens, duration of treatment, primary outcomes, BP measurement techniques, and results. The ACCORD trial (Action to Control Cardiovascular Risk in Diabetes) randomized 4733 patients with diabetes (mean 63 years of age, 34% with a previous CV event, BP 139/76 mm Hg) to intensive therapy to a systolic BP target <120 mm Hg or to a standard target <140 mm Hg.[1] Intensive therapy lowered BP by 14.2/6.1 mm Hg but did not reduce the primary outcome of CV death, MI, stroke, or secondary outcomes, including HF and overall coronary events. A significant reduction in stroke occurred, although the number of strokes was low. The SPS3 trial (Secondary Prevention of Small Subcortical Strokes) randomized 3020 patients (mean 63 years of age, BP 143/79 mm Hg) with a recent small subcortical stroke to a systolic BP target <130 mm Hg or between 130 and 149 mm Hg.[2] The difference in systolic BP was 11 mm Hg, but no significant reduction occurred in the primary outcome, recurrent stroke, or the secondary outcome, the composite of MI and vascular death. ACCORD and SPS3 may be underpowered, although the number of major CV events in ACCORD (including HF) is similar to that in SPRINT (Systolic Blood Pressure Intervention Trial).[3] The only trial that showed benefit for intensive systolic BP lowering <120 mm Hg is SPRINT, conducted in 9361 high-risk patients with chronic kidney disease, clinical or subclinical CV disease, or Framingham risk ≥15% (mean 68 years of age, BP 140/78 mm Hg; patients with diabetes or previous stroke were excluded).[3] The trial was terminated early. BP was lowered by 14.8/7.6 mm Hg. A 25% relative risk reduction occurred in the primary composite outcome of MI, acute coronary syndrome, stroke, HF, or CV death and reductions in HF, CV, and all-cause death but not in MI and stroke. The HOT trial evaluated 3 diastolic BP targets in 18 790 patients (mean 62 years of age, largely without previous CV events, BP 170/89 mm Hg) and found overall no significant reduction in CV events for more aggressive diastolic BP targets, except in the subset of 1501 patients with diabetes.[4]
Therefore, SPRINT is the exception among trials evaluating aggressive BP-lowering targets. Although it is well designed and conducted, several issues need to be considered: (1) Due to its early termination, the magnitude of benefit may be overestimated. It was suggested that after adjusting for this potential bias, the true relative risk reduction in major CV events is ≈15%, which is closer to estimates from the other trials and large meta-analyses of trials comparing more- versus less-intensive BP lowering; (2) No significant reduction in MI or stroke occurred. The lack of benefit on stroke is particularly surprising because this outcome is most closely associated with BP epidemiologically and generally most sensitive to BP lowering in trials. A major contributor to the reduction in the primary composite outcome is the reduction in HF, which may be related to higher use of diuretics in the intensive treatment arm; (3) There was a reduction in non-CV death, which is likely a chance finding; (4) More than 90% of patients were already receiving BP-lowering therapy at entry (on average 1.9 BP-lowering drugs), so their true untreated BP was higher than the reported high-normal baseline BP; (5) Intensive BP lowering resulted in a marked increase in side effects, including emergency room visits, hypotension, syncope, electrolyte imbalances, acute kidney injury or renal failure, and worsened eGFR in patients without chronic kidney disease at baseline. These side effects are similar to data from ACCORD; (6) Standardized BP measurements based on average of 3 largely-unattended measurements were used. It is expected that using the SPRINT targets in clinical practice without implementing the SPRINT BP measurement protocol and careful follow-up may result in lower BP levels and more side effects; (7) Two widely different BP targets were used, so it remains unclear whether an intermediate target (eg, 130 mm Hg) could have achieved the same benefit with fewer side effects; (8) Follow-up visits with titration of BP-lowering drugs, efforts to maintain adherence, and safety assessments were frequent. Such an approach may be difficult to implement in routine clinical practice; (9) Most important, the CV benefits in SPRINT far exceed those shown in other trials that evaluated aggressive BP targets. Considering data from all such trials, the magnitude of CV event reduction with aggressive targets is likely lower, ≈15%.
SPRINT was compared to the recent HOPE-3 trial (Heart Outcomes Evaluation-3).[5] These trials have different objectives and designs. HOPE-3 was conducted in an intermediate risk population without CV disease (yearly major CV event rates of 0.94% vs. 2.19% in SPRINT), 80% of participants were treatment naïve at baseline, BP measurements were attended by a healthcare provider, and a simple BP-lowering regimen with a fixed-dose combination of an angiotensin-receptor blocker and a thiazide diuretic (at low doses) was compared to placebo without titration to prespecified targets. Overall, no significant reduction in major CV events occurred. However, in a prespecified subgroup analysis, a significant 27% reduction in major CV events was shown in participants in the upper third of baseline systolic BP (>143.5 mm Hg, mean of 154.1 mm Hg) despite a more modest 6/3 mm Hg BP lowering, with no benefit in those with lower entry BP. These benefits were safely enhanced by simultaneous administration of a statin (40% relative risk reduction). These findings support antihypertensive therapy in mild uncomplicated hypertension with initiation of therapy at systolic BP >140 mm Hg. Specific targets of therapy were not tested in HOPE-3. However, results of this subgroup analysis suggest that aggressive targets in this intermediate risk population are not justified. Treatment was extremely safe and follow-up visits were infrequent and streamlined so that the HOPE-3 approach is widely applicable.
In summary, aggressive systolic BP targets may be used in selected, high-risk individuals treated in specialized settings using standardized automated BP measurements and careful and frequent monitoring, although further confirmation of the SPRINT targets would make such recommendations more persuasive. Limited data are available for diastolic BP targets <80 mm Hg, except possibly in diabetics (limited data from HOT). The simple, less aggressive treatment approach used in HOPE-3 appears reasonable for intermediate-risk persons with uncomplicated grade 1 hypertension, should be complemented by statin therapy, and can be readily incorporated into clinical practice in most settings.
Sources of Funding
The HOPE-3 trial was funded by the Canadian Institutes of Health Research and Astra Zeneca.
Circulation. 2016;134(18):1311-1313. © 2016 American Heart Association, Inc.
