This single-center cohort study of more than 5000 hypertensive patients evaluated the impact of TBPC (SBP < 130 mmHg) on the incidence of new MI and stroke. We suggest that TBPC independently decreased the incidence of stroke compared to SBPC, but did not have an impact on MI incidence.
A retrospective study using the ACTION trial database found the greatest difference between SBP < 130 mmHg (2.5% incidence rate) and SBP < 140 mmHg (3.8% incidence rate) groups in the risk of stroke [17]. The cohort in the ACTION trial was different from ours in that their patients had stable coronary heart disease [18]. A meta-analysis performed by Lee et al demonstrated that SBP < 130 mmHg compared to SBP 130-139 mmHg showed additional protective effect on stroke in patients with cardiovascular risk factors but no previously established cardiovascular disease [7]. In another meta-analysis, Reboldi et al found that more-tight SBP control (< 120 mmHg), compared to less-tight SBP control (< 140 mmHg), significantly reduced stroke risk by 31%, but no significant risk reduction was noted in MI [8]. While these results are similar to ours, Reboldi et al’s analysis included diabetic patients and a different BP cutoff for the 2 groups. The ONTARGET trial provided further evidence on the significant impact of BP control on stroke but not on MI [9]. The trial found that more frequent achievement of target BPs (< 130/90 mmHg or < 140/90 mmHg) resulted in significant cerebrovascular protection but not cardiac protection [9]. More recent large European meta-analyses further strengthen the evidence that more intense blood pressure control reduces cardiovascular events [6, 19].
Why TBPC decreases the risk of stroke but not MI remains unclear. Several long-term prospective population-based studies have shown that hypertension increased the relative risk of stroke to a greater degree than MI [20-23]. One possible hypothesis is different blood flow physiology between the brain and heart. The brain receives a larger fraction of cardiac output by nearly 3-fold than the heart [24]. While most of the coronary blood flow occurs during diastole, the cerebral blood flow occurs during systole [25]. Therefore, cerebral blood flow might be more sensitive to a change in SBP than coronary blood flow, and coronary blood flow might be more sensitive to a change in DBP than SBP. A recent multinational study supports this hypothesis in that results showed increased risk of MI with increased DBP whereas increased SBP did not increase the risk of MI [11].
Other studies have reported different results. Verdecchia et al performed an open-label randomized trial and found no difference in the incidence of MI or stroke between TBPC (< 130 mmHg) and SBPC (< 140 mmHg) [26]. Their population was similar to ours in the exclusion of diabetic patients. Another large randomized study using the database of the Hypertension Optimal Treatment trial showed that TBPC (< 130 mmHg) in patients with diabetes and coronary artery disease did not improve cardiovascular outcomes compared to SBPC (< 140 mmHg) [27]. The ACCORD study conducted in 2010 also showed no difference in fatal and nonfatal cardiovascular events between TBPC (< 120 mmHg) and SBPC (< 140 mmHg) groups in diabetic patients. More importantly, the TBPC group demonstrated significantly higher rates of serious adverse events attributed to antihypertensive therapy [28]. Furthermore, a post hoc analysis using the Irbesartan Diabetic Nephropathy Trial database concluded that BP ≤ 120/85 mmHg may be associated with increased cardiovascular events [29].
Our study and the majority of cohort and prospective studies have shown that TBPC is associated with decreased risk of stroke. Nevertheless, the issue of TBPC must be addressed carefully in the clinical setting. Even though cerebral blood flow has an autoregulation mechanism, the mechanism can be lost if the mean arterial BP drops below 60 mmHg [30]. A study that targeted BP < 130/80 mmHg for diabetic patients found intensive BP control caused progressive reduction of cerebral blood flow velocity [31]. According to the meta-analysis by Thomopoulos et al, the BP-lowering treatment reduced cardiovascular events by 24% whereas it increased the risk of discontinuation attributed to adverse events by 89% [32]. Ferreira et al recently demonstrated significantly increased rates of cardiovascular death, heart failure hospitalization, MI, and stroke when SBP decreased below 125 mmHg [11]. In our study, only 0.001% (5/4530) in the TBPC group showed SBP < 90 mmHg.
An interesting finding in our study is that statin use did not independently affect MI or stroke incidence. Based on the 2013 ACC/AHA guidelines, absolute indications for statin use include clinical atherosclerotic cardiovascular disease, type 2 diabetes mellitus, and serum LDL ≥ 190 mg/dl [15]. Perhaps statin use did not affect the MI or stroke incidence in our study because we excluded patients with diabetes and those with a history of MI or stroke. Very few patients in our study had a serum LDL ≥ 190 mg/dl (Table 1). Another interesting finding is that increased number of antihypertensive medications significantly predicted both MI and stroke incidence while the usage of antihypertensive medication itself did not predict either MI or stroke incidence in the multivariate model. This can suggest that requiring more antihypertensive medications rather than taking the medications itself could predict higher cardiovascular risk, or those who were prescribed many antihypertensive medications would have not been adherent to the medications.
Limitations
The major limitation of our study is that this is an observational retrospective study. Given the fact that authors did not do any interventional measures for the patients, strictly speaking, it might not be very appropriate to use the terms “tight control” and “standard control”. However, as all the patients were formally diagnosed with hypertension, authors assumed that they had received some type of BP control intervention including lifestyle counseling from their providers. And this would have justified our using the terms “control.” Our study also has several other limitations. Because of the retrospective design, we were not able to completely address and remove all potential confounders. There were significant differences in age, race, BMI, statin use, and the number of antihypertensives between TBPC and SBPC groups. The TBPC group was younger and had statins prescribed more often than the SBPC group. Although these variables were adjusted at multivariate analysis, these differences might still have affected our findings, and we admit that this is potentially the most major limitation of our study, which should be criticized. Also, as we excluded other significant risk factors of MI and stroke, such as diabetes and history of MI or stroke, our study population was different from actual patients commonly seen in clinical practice, which might have resulted in the relatively rare outcome events in our study. Also, as we were not able to review each patient’s medical record given the large, retrospective, and electronic-based data analysis, we were unable to collect information on the adverse events that can potentially occur from TBPC such as dizziness, syncope, falls, acute kidney injury, or cerebral hypo-perfusion. However, given that only 0.001% of the TBPC patients had SBP less than 90 mmHg, it is less likely that many of our TBPC patients developed significant hypotension-related complications. Also, we were unable to assess each patient’s medication adherence data as we only counted the numbers of the medication prescriptions which can potentially be significantly confounding. It also would have been more informative if we could include the mortality and quality of life outcomes in either group, which should be considered for future research.