Introduction

High blood pressure (BP) is one of the most common modifiable causes of mortality in the USA, accounting for approximately one in five deaths [1]. In addition, the global prevalence of hypertension has been steadily increasing, leading to higher attributable mortality [2]. As such, national and international efforts to control high BP are of paramount importance for public health. Based on recent clinical trials, most notably SPRINT (Systolic Blood Pressure Intervention Trial) [3••], the 2017 American College of Cardiology (ACC)/American Heart Association (AHA) guidelines for BP management have recommended more intensive BP targets, changing the definition of hypertension from ≥ 140/90 mmHg to ≥ 130/80 mmHg [4]. While SPRINT suggested benefit for treatment to BP levels of 120/80 mmHg, due in part to the technique of BP measurement in SPRINT (often unattended) - which has been extensively debated and remains controversial [5, 6] - the more conservative target BP of < 130/80 mmHg was chosen in the 2017 ACC/AHA guidelines. As a consequence, approximately 46% of American adults now meet diagnostic criteria for hypertension—an increase in prevalence from 32% with the prior definition [7]. However, concerns have been raised regarding the possibility of adverse outcomes associated with more aggressive BP control in some circumstances, specifically with pharmacologic therapy [8, 9].

Is Lower BP and More Intense BP Therapy Always Better?

The benefit of hypertension treatment has been well-established for many decades [10]. Early studies definitively established antihypertensive drug efficacy, with initial systolic BP (SBP) treatment targets < 160 mmHg showing clinical benefit [11, 12], followed by The Antihypertensive and Lipid Lowering Treatment to Prevent Heart Attack Trial (ALLHAT) showing a benefit of SBP < 140 mmHg [13]. Similarly, early randomized placebo-controlled trials demonstrated the benefits of antihypertensive agents targeted towards patients with either Diastolic BP (DBP) ≥ 115 mmHg or DBP ≥90 mmHg. Treatment to the lower DBP target resulted in lower cardiovascular disease (CVD) events and mortality [14, 15]. These findings were then supported by much larger studies such as the Hypertension Detection and Follow-up Program (HDFP), targeting a DBP ≤ 90 mmHg and showing a significant reduction in CVD events and mortality in the intervention group compared to the usual care group [16]. Taken together, these studies firmly established a BP target of <140/90 mmHg, which had been mostly unchanged within previous BP guidelines for decades.

However, controversy has existed regarding the benefits of targeting even lower blood pressures. A 2002 meta-analysis of observational data from 1 million individuals participating in the Prospective Studies Collaboration showed that the risk of CVD increases above a SBP of > 115 mmHg or DBP > 75 mmHg, supporting the concept of more intense BP control to even lower targets than recommended in prior BP treatment guidelines [17]. In another meta-analysis of prospective cohort studies, SBP from 130 to 139 or DBP 85 to 89 mmHg had a higher risk of CVD events compared to individuals with SBP 120 to 129 or DBP 80 to 84 mmHg, respectively. In the same study, BP < 120/80 mmHg had better outcomes than those with higher BP [18]. These findings suggest that a target BP of < 140/90 mmHg is too liberal and could be associated with worse CVD outcomes compared to a lower BP target within the physiologically normal range. However, these are observational findings, and the risk of selection and publication bias is a serious concern. In addition, individual observational studies vary in their approach to adjusting for potential confounding risk factors.

Despite these concerns, further support for lower BP targets can also be found in the randomized trial data. For example, meta-analyses of randomized control trials reported that, compared to patients who achieved a SBP < 130 mmHg with more intensive BP therapy, those with on-treatment SBP ≥ 130 mmHg had increased stroke and all-cause mortality without a difference in cardiovascular mortality. Stroke was also reduced when DBP was <80 mmHg compared to when it was ≥80 mmHg [19••, 20]. The reduction in stroke was particularly significant in patients with high CVD risk [21]. In another meta-analysis, an on-treatment SBP of 120–124 mmHg, compared to 130–134 mmHg, was associated with a 29% decrease in CVD and a 27% decrease in all-cause mortality [22]. However, the collated randomized data are inconsistent and there are other meta-analyses supporting no benefit for stricter BP targets, including a recent one reporting no benefit to more intense SBP treatment below a target SBP of 140 mmHg [23].

Unfortunately, most of the trials included in these meta-analyses were not designed to examine the benefits of specific BP targets (rather the majority studied less versus more antihypertensive therapy). In studies supporting an intensive management approach without specific BP targets, one could argue that the lower BP achieved in the intensive arm may reflect a healthier population [24], and adults less prone to resistant hypertension (which is most commonly due to noncompliance) [25,26,27]. In addition, the study by Bundy et al. [22] included a wide variety of baseline populations including those with stroke, coronary artery disease, or heart failure all of which are different risk profiles compared to primary prevention patients who are usually healthier. In fact, each meta-analysis of randomized antihypertensive trials has included a different combination of studies and used a different approach to analyses, highlighting the inherent flaws of these types of studies and helping to explain differences in their findings.

Despite these concerns, the above data supported the need for a contemporary randomized control trial of intensive BP control. SPRINT [3••] was a trial of individuals with high CVD risk of 2.2% per year without diabetes who were randomized to an intensive target SBP of < 120 mmHg compared with the standard target of < 140 mmHg. The intensive therapy arm had a lower rate of the primary composite outcome of myocardial infarction, other acute coronary syndromes, stroke, heart failure, or CVD death. The benefit persisted regardless of baseline systolic BP (when stratified into tertiles) [28] and in patients > 75 years [29]. In addition, a recent report from a substudy of SPRINT (SPRINT Memory and cognition IN Decreased – SPRINT MIND) reported that individuals randomized to the intensive arm had a lower risk of mild cognitive impairment and dementia on follow-up. Although the full publication is still pending, this is consistent with observational study data of SBP [30].

One major concern for intensive SBP management was the adverse outcomes of lowering diastolic BP which may affect coronary perfusion [31]. However, the benefits of intensive SBP for the composite outcome in SPRINT were not attenuated among those with low baseline diastolic BP (though results for coronary outcomes specifically have not been reported according to baseline DBP) [32]. In addition, the SPRINT study excluded individuals who were < 50 years of age, estimated glomerular filtration rate < 20 ml/min/1.73m2, with diabetes, or had symptomatic heart failure within the past 6 months or left ventricular ejection fraction < 35%. Therefore, caution is required before applying the study results to these patient populations.

ACCORD was a similarly designed trial among diabetics that showed lower risk of stroke at an SBP target of 120 mmHg, though this trial did not reach statistical significance for the primary endpoint (nonfatal myocardial infarction, nonfatal stroke, or cardiovascular death) [33•]. One reason for this apparent discrepancy may be that ACCORD was underpowered with a smaller study sample than SPRINT. Looking closely at the ACCORD results shows a trend towards lower hazard for the primary outcome among those randomized to a lower BP target (hazard ratio of 0.88, 95% confidence interval 0.73–1.06). Potentially supporting this hypothesis is an analysis combining results from both ACCORD and SPRINT, which showed that intensive BP management was associated with a reduction in stroke, heart failure, and for each study’s primary outcome (which were different for each study as discussed above) [34•].

However, another reason for the discrepancy between SPRINT and ACCORD may be that the benefits of lower SBP achieved with drug therapy may not be universal and may differ according to personal clinical attributes of the patient under consideration. In the Heart Outcomes Prevention Evaluation (HOPE)-3 study of intermediate CVD risk individuals with a mean baseline SBP of 138 mmHg, aggressive SBP lowering with dual antihypertensives resulting in a mean SBP < 130 mmHg versus placebo did not result in any significant difference in CVD outcomes or mortality [35••]. Although HOPE-3 and SPRINT are often compared, the latter specifically had two BP target groups whereas the former used a more intensive antihypertensive regimen without a target BP group. Furthermore, the absolute BP lowering during the trial in HOPE-3 was far less than that seen in SPRINT. However, another reason for the difference may be that patients that are at highest risk benefit from more intensive SBP control, whereas lower risk individuals may not, as suggested by a recent meta-analysis of trials that showed that benefits from BP reduction appeared to be related to baseline estimated CVD risk [36]. As such, lower BP with drug treatment may not always be necessarily better and adults with low CVD risk may be exposed in excess to the toxicity of antihypertensive agents when treated intensively with little chance of deriving any clinical benefits, making them particularly prone to the dangers of overly aggressive BP control.

What Are the Dangers of Overly Aggressive BP Control?

With increasingly aggressive BP management, it is conceivable that SBP or DBP may become too low, resulting in more dangers than benefits. For example, at a certain SBP or DBP threshold, physiological perfusion and autoregulation of vital organs could be impaired resulting in adverse outcomes [37, 38].

J-Curve Phenomenon

Unlike SBP, targeting a more aggressive lowering of DBP appears to have thus far failed to show any benefit. In the Hypertension Optimal Treatment (HOT) trial, there was no difference in major CVD events between the groups who were allocated to a DBP target of ≤ 90, ≤ 85, or ≤ 80 mmHg, although actual reduction in DBP with treatment was lower than initially targeted [39]. Over the past three decades, data has accumulated suggesting that, not only may there not be a benefit for aggressive DBP lowering, but instead, there may be a J-curve relationship with a nadir in benefit of DBP lowering below which the risk of adverse outcomes increases. Such a J-curve could be explained by the impairment of coronary perfusion, which occurs predominantly during diastole. Supporting this idea, among patients undergoing coronary angiography and fractional flow reserve measurement, more patients had dangerously low coronary blood flow when DBP was found to be < 70 mmHg [40]. As a result, a very low DBP would be expected to increase coronary events, and have a less significant effect on other vascular beds such as cerebral or renal.

In one meta-analysis of observational studies, there was a trend for increased ischemic heart disease events among those with DBP < 70 mmHg [17], and in a second systematic review of 13 studies, CVD events were increased when patients were treated to DBP < 85 mmHg [41••]. In the Prospective Observational Longitudinal Registry of Patients with Stable Coronary Artery Disease (CLARIFY) study of > 22,000 individuals with stable coronary artery disease, after extensive adjustment, a DBP ≥ 80 mmHg or < 70 mmHg compared with 70–79 mmHg (reference) was associated with increased CVD events, but not stroke. A DBP < 60 mmHg was associated with a 2-fold increase in CVD events compared to the reference group [42••].

This increased risk of adverse outcomes has also been noted in trials. Among participants with prior ischemic heart disease enrolled in HOT, those treated to a DBP of 80 mmHg had a numerically higher rate of myocardial infarction than those treated to a DBP of 85 mmHg [43]. Further, in individuals with known CAD and hypertension enrolled in the International Verapamil-Trandolapril Study (INVEST), there was a reduction in CVD events when DBP was below < 90 mmHg but the risk increased if DBP was < 70 mmHg [44, 45•]. In addition, the recently published secondary analysis of the Ongoing Temisartan Alone and in combination with Ramipril Global Endpoint Trial (ONTARGET) ACEI inhibitors and Telmisartan Randomised Assessment Study of ACE Intolerant (TRANSCEND) randomized trials demonstrated that among patients greater than 54 years of age with CVD, DBP reduction to < 70 mmHg was associated with higher myocardial infarction and all-cause death but not stroke, whereas DBP ≥ 80 mmHg increased the risk of stroke and heart failure hospitalization [46•]. In this excellent study, in addition to adjustment for baseline comorbidities, analyses were also adjusted for competing events occurring during observation. In addition, the adverse outcomes of low DBP persisted in separate analyses where patients with recent nonfatal events, extremely low SBP and those not on antihypertensives were excluded to address some of the possible concerns for reverse causality as an explanation for the increased DBP outcomes [46•]. The effect of low DBP on increased CAD events is beyond what would be expected if poor health status alone was the explanation, as a more unwell patient population would also be expected to have an increase in other adverse events such as stroke.

Of note, another patient population who may also be at increased risk due to low DBP are patients with heart failure, a condition where the left ventricular end-diastolic pressure may be elevated. Lowering of the DBP would reduce the coronary perfusion pressure (related to the difference in left ventricular end-diastolic pressure and DBP) in heart failure even further [47]. In recent post hoc analyses from the Treatment of Preserved Cardiac Function Heart Failure with an Aldosterone Antagonist Trial (TOPCAT) of patients with heart failure with preserved ejection fraction, DBP was associated with increased risk CVD events when < 70 mmHg, and increased risk of heart failure hospitalization when < 60 mmHg [48, 49].

The potential dangers of low DBP may often be unrecognized. In the Atherosclerosis in Communities (ARIC) cohort in individuals without known CVD who had a DBP < 70 mmHg or DBP ≥ 100mHg had increased odds of elevated high-sensitivity troponin-T, using individuals with DBP of 80–89 mmHg as reference, suggesting asymptomatic subclinical myocardial damage in individuals with a low DBP. On a 21-year follow-up of ARIC participants, DBP < 60 mmHg was associated with increased CVD events and all-cause mortality [31]. Further, the dangers of DBP < 60 mmHg are also evident in individuals with asymptomatic subclinical coronary atherosclerosis, as evaluated by coronary artery calcium [50]. Thus, aggressive DBP lowering appears to increase the risk of CAD events particularly in individuals with underlying coronary artery disease. However, other studies have suggested that the DBP J-curve may be a non-causal phenomenon, and as such, this issue is not settled entirely [32]. Nonetheless, we still recommend that persons being treated to more intensive systolic BP goals who have CAD or LVH have close follow-up for diastolic BPs < 60 mmHg, which may in some circumstances justify consideration of de-intensification of BP therapy.

It also remains unclear if there is a SBP J-curve. In the INVEST study discussed above, patients with diabetes had increased risk of all-cause mortality if the achieved SBP was < 115 compared to SBP < 130 [51]. In the CLARIFY registry, SBP ≥ 140 mmHg and < 120 mmHg was associated with increased myocardial infarction, stroke, and CVD death [42••]. One explanation for the adverse outcomes with lower SBP may be that individuals achieving a lower SBP may be less healthy at baseline and have unadjusted confounders (such as vascular stiffness and elevated pulse pressure). In one important observational study of individuals above 85 years of age, comprehensive adjustment for health status and frailty attenuated the relationship between low SBP and all-cause mortality that was seen in models not adjusted for health status and frailty [52]. Similarly, in the Framingham Heart Study a U-shaped curve between CVD mortality and SBP was no longer significant after adjustment [53]. Although it may be true that patients with lower SBP may have other comorbidities, these findings nonetheless highlight the practical importance of identifying these high-risk patients for an individualized approach to management and avoiding excessive sudden drops in SBP where possible given their excess “real-world” levels of risk for bad outcomes.

In addition, there may also be other vulnerable populations where particular caution is required. In a small randomized trial of patients in South Korea with internal carotid artery stenosis and a subacute ischemic stroke (within the past 7–42 days), SBP < 120 mmHg compared with < 140 mmHg was associated with increased ischemic lesion volume and a trend towards higher frequency of new ischemic lesions [54•]. Of note, this study only reached a mean SBP of 124.6 mmHg in the intensive treatment arm; thus, the evaluation more closely reflects < 140 mmHg versus < 130 mmHg. Vascular stenosis related impairment of cerebral blood flow is likely exacerbated by the more aggressive BP management in this vulnerable population. In comparison, in the multicenter randomized Secondary Prevention of Small Cortical Strokes (SPS3) study of patients with a lacunar infarct, SBP target of <130 mmHg versus 130–149 mmHg resulted in a significantly lower rate of intracerebral hemorrhage (hazard ratio 0.37, 95% confidence interval 0.15–0.95) and a trend towards lower strokes [55]. In these patients, cerebral blood flow is not reduced with intensive BP lowering [56]. Therefore, the mechanism of stroke is an important consideration regarding BP management in this setting as the pathophysiology will reflect the risks.

Other Adverse Effects

Additionally, although there was no significant difference in all serious adverse events in SPRINT overall, hypotension, syncope, acute kidney injury, and electrolyte abnormalities were significantly more common in the intense arm, sometimes resulting in emergency department visits. SPRINT was also stopped early, and thus, it is unclear if there could be even more cumulative adverse effects with longer duration of intensive BP therapy [3••].

One of the long-held concerns has been the risk of falls associated with hypotensive episodes. In one study of Medicare patients, there was a transient increase in the risk of falls within the first 15 days of antihypertensive initiation or uptitration [57]. However, there was no increase in long-term risk of falls. In fact, several studies have failed to demonstrate an increase in risk of falls with the initiation or intensification of antihypertensive treatment [58,59,60, 61•]. The risk of falls with antihypertensives is associated more to individual indicators of frailty such as impaired mobility, cognitive impairment, depression, history of falls, and presence of orthostatic hypotension [58, 59, 61•]. This is not surprising, as these are the same risk factors for falls for the general population [62]. This is consistent with the ACCORD study where, although hypotension and syncope was more common in the intensive arm, there was no increase in falls or non-spine fractures [33, 63]. Further, undertreated hypertension itself is a risk for falling [59, 64].

The renal risk of severe hypotension has also been shown in metanalyses [65]. Some of these effects may persist long-term, with one analysis of SPRINT and ACCORD at 3 years showing that the incidence of chronic kidney disease was 10% in the intensive arm compared with 4.1% in the standard arm [66]. Adverse renal outcomes appeared to be more common in patients with a DBP < 70 mmHg [67]. Estimates suggest that applying the SPRINT eligibility criteria for intensive BP implementation to the 1999 to 2006 National Health and Nutrition Examination Survey (NHANES) would result in 56,100 episodes of hypotension, 34,400 episodes of syncope, and 88,700 excess cases of acute kidney injury per year. However, importantly, the study also projected a reduction in 107,500 deaths with the intensive SBP goal [68]. Thus, although the current data demonstrate a significantly higher risk of adverse outcomes, there are also clear benefits.

It is likely the dangers of the adverse outcomes of hypotension and falls affect those with multiple comorbidities disproportionately as discussed above. It is important to note that patients with a prior stroke, end-stage renal disease or eGFR < 20 ml/min/1.73m2, and symptomatic heart failure were excluded from the SPRINT study in addition to patients with diabetes. Many of these comorbidities increase the frailty of the patients and risk for adverse effects.

Polypharmacy

Intensive BP therapy inevitably results in greater polypharmacy. In the SPRINT study, the intensive arm required a mean of three medications compared with 1.9 in the standard arm [3••], and in the ACCORD trial the mean number of medications was 3.4 and 2.3 in the intensive and standard arms, respectively [33•]. These means are tied with significant variability in medication numbers at the individual level (some patients required far more medication to achieve the intensive goal). This not only increases the cost [68] but also increases the risk of medication errors, medication non-adherence, medication interactions, and medication related side-effects [69, 70]. Indicators of frailty such as psychiatric disorders, cognitive dysfunction, and audiovisual problems exacerbate these issues and remain another concern for aggressive BP treatment [71]. One potential strategy to reduce polypharmacy while also attaining more intensive BP targets is renal denervation, which is recently enjoying a revival in the literature [72, 73].

Pulse Pressure

Finally, pulse pressure is an underappreciated and infrequently discussed risk factor for adverse outcomes and may also be an important consideration in the intensive systolic BP treatment of patients. Patients with a low DBP but a high SBP (that is a large pulse pressure) may have worse outcomes [31, 46•]. In the international Reduction of Atherothrombosis for Continued Health (REACH) registry, a higher pulse pressure quartile was associated with an elevated incidence of adverse CVD outcomes [74]. However, pulse pressure remains a poorly studied parameter in randomized trials of management of BP and further thought is required in integrating pulse pressure with our understanding of the SPRINT results and the J-curve phenomenon.

Individualized Approach

What is clear from the discussion above is that there are well-demonstrated benefits to intensive BP management when applied in the right patient population, such as those with a high CVD risk. However, there are also potential risks of aggressive pharmacological BP treatment, such as an increased risk of CAD events if the DBP is lowered too severely, an increased risk of stroke if there is obstructive cerebral blood flow, or an increased risk of electrolyte and renal dysfunction in high-risk patients. These findings highlight the importance of taking an individualized approach to BP targets and management strategies as opposed to blanket guidelines for a target BP for the general population (Fig. 1). In addition, future treatment strategies may integrate a complex approach taking into account, for example, genomics, proteomics, and pharmacokinetics [75]. However, what remains consistent and clear is that BP reduction with lifestyle modifications such as smoking cessation, avoidance of excess alcohol, exercise, and/or dietary changes is always beneficial [4]. Further, as discussed above, non-adherence is common among patients being treated for hypertension [25,26,27], and instead of focusing on lower targets (130/80 mmHg), providers often need to focus on adherence to ensure even old targets (140/90 mmHg) are achieved in their patients. Only then, and after a consideration of individual risk:benefit and discussion with the patient regarding goals and preferences, should the more intensive BP target be considered.

Fig. 1
figure 1

Clinicians should take an individualized approach to management of blood pressure in patients and consider assessing the specific benefits and risks of pharmacological treatment

Full size image

Limitations

As the discussion above shows, there are many reports of the effects of intensive BP management (usually referred to as target < 130 mmHg or < 120 mmHg) versus standard BP management (target < 140 mmHg). However, most of these studies are meta-analysis of observational studies or nonrandomized post hoc data from previous randomized control trials where randomization was not based on an intensive versus standard BP target as defined above. The most important study for aggressive BP management is the SPRINT study among high cardiovascular risk patients without diabetes or stroke. Thus, caution is necessary when applying the SPRINT data more broadly such as in lower risk patients, as well as patients with significant obstructive cerebral blood flow as discussed above. Further, it is worth considering that the absolute benefits of intensive SBP lowering is related to the patients underlying CVD risk, for example, as shown among patients with subclinical atherosclerosis [76].

Conclusions

The current guidelines recommend a target BP of < 130/80, which appears to be safe in the majority of patients. However, an individualized approach is required to management of patients by assessing specific risks for each patient. For example, a patient with advanced renal disease may be more vulnerable to the electrolyte abnormalities associated with aggressive BP management, or a patient on triple anti-thrombotic therapy may be more vulnerable to complications of a syncopal or hypotensive episode that appears more common with intense BP management, as described above. Caution is also required when aiming for more aggressive approaches among lower risk individuals as there is a lack of randomized control trials in that patient population. Further, an abundance of data support that patients whose DBP is lowered < 60 mmHg especially if they have underlying coronary artery disease may suffer from higher CAD outcomes. In addition, in our approach to pursue guidelines recommendations, clinicians often forget to develop a doctor-patient agreement to a therapeutic goal. This likely contributes to the poor adherence of patients, and the failure to reach targets that is commonly seen; leading to a reflex addition of more agents [77]. These settings are likely to increase the dangers of aggressive blood pressure management strategies.