Blood Pressure Targets in Diabetes: Is This the Time for Change?
David M. Kendall, MD, American Diabetes Association, 1701 North Beauregard Street, Alexandria, VA 22311
Following the publication of the Action to Control Cardiovascular Risk in Diabetes (ACCORD)1 blood pressure (BP) trial results in early 2010, many were contemplating whether the longstanding recommendations for BP targets in diabetes were in need of modification. Prior to the ACCORD trial, there was a general consensus that BP targets in diabetes should be lower (<130/80 mm Hg) than those recommended for the general population with hypertension. For many, the results of the ACCORD trial were a wakeup call and represent a time for change. The results (at least for some) were evidence for many that the prior targets were inappropriately low. Without question, the results of the ACCORD trial add substantively to our understanding of hypertension treatment in diabetes. But does the ACCORD study alone provide enough additional information to establish new (and higher) BP targets for all patients with diabetes?
In this counterpoint, I provide perspective supporting the value of the current clinical practice recommendations and offer an interpretation of the ACCORD trial results and evidence that support maintaining the current treatment targets. In considering this position, it is critical to recall the nearly 2 decades of clinical trial data that guided current recommendations for optimal care in diabetes, including those of the American Diabetes Association (ADA),2 the National Kidney Foundation (NKF),3 and the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC).4
Several key elements warrant closer appraisal. In this summary, I provide (1) a brief historical review of the current treatment guidelines and the data that support these recommendations; (2) a review of and perspective on the ACCORD trial data; and (3) a viewpoint of the possible consequences—intended and unintended—that may result from re-establishing new (and higher) treatment targets for BP in diabetes.
BP Standards in Diabetes: A Historical Perspective
Over the past 2 decades, the standards of medical care for diabetes have evolved from a series of consensus-based recommendations on care delivery to a broad set of specific practice guidelines and treatment targets.2 These standards are developed (and periodically modified) based on review of several decades of clinical data including randomized clinical trials, large-scale observational studies and epidemiologic analyses. In its seventh report, the JNC (JNC 7) reaffirmed the central role of BP as a determinant of both macrovascular and microvascular disease risk in diabetes. The committee also reaffirmed that the BP target for patients with diabetes be maintained at <130/80 mm Hg based on evidence that lower levels of BP are associated with a significantly lower risk for both cardiovascular disease (CVD) and progression of renal disease.4
An overwhelming majority of national and international standards have, with few exceptions, supported these recommendations. The BP target of <130/80 mm Hg is a central feature of the clinical practice recommendations of the ADA2 and the NKF.3 Each of these clinical practice recommendations underscore that while the data supporting any specific BP target are limited,2,3 the preponderance of evidence support that lower BP values (and actively treated hypertension in general) are associated with a significant reduction in the risk of complications (including both CVD events and microvascular complications of diabetes).
A recent review by Zanchetti and colleagues5 provides a comprehensive summary of the large number of studies of BP lowering. The authors emphasize that the majority of clinical trials of BP lowering in diabetes have neither targeted nor achieved mean systolic BP (SBP) values <130 mm Hg. Indeed, active drug (or multidrug) therapy achieved mean SBP values over a broad range (from the low 130s to mid-140 range).5 Only the Hypertension Optimal Treatment Trial (HOT) has provided direct evidence of the benefit of targeting specific blood levels and demonstrated that a lower pressure target (in this case, lower diastolic rather than SBP) resulted in greater CVD protection in diabetes.6 In a number of other studies, lower values of SBP were associated with a lower rate of CVD and reduced rates of deterioration in renal function or progression of renal disease. Recent studies have reaffirmed the historical finding of improved preservation of renal function7 and underscore why Bakris and colleagues8 and recent consensus guidelines2–4 have maintained the lower BP target for patients with diabetes.
Even after completion of several dozen studies assessing the impact of ≥1 antihypertensive therapies on rates of both CVD and microvascular disease in diabetes, only indirect evidence support any specific treatment target for BP in the diabetes population. In general, clinical guidelines have established current targets at a single cutpoint (130/80 mm Hg) with the understanding that BP is a continuous variable for risk. The BP targets established have been provided in hopes of achieving a balance between rational treatment approaches (in a population with widely variable risk and baseline BPs) and maximal protection against all of the complications of diabetes. Given that few studies addressed the impact of targeting a specific BP, the ACCORD trial investigators embarked on the ACCORD BP study to assess whether near-normal BP values could further reduce CVD risk compared with an approach that targets BPs near those recommended by current guidelines.9
Implications of the ACCORD BP Trial
The ACCORD BP trial was designed to address a specific clinical question; namely, would a target-driven approach to BP treatment support those data used to develop current BP targets? The ACCORD trial protocol was designed to address whether, in the context of good glycemic control, a therapeutic strategy that targets SBP <120 mm Hg would reduce rates of major CVD events to a greater degree than a strategy that targets SBP <140 mm Hg?
In the ACCORD study, the intensive treatment group achieved a mean SBP of 119 mm Hg, while the standard treatment cohort achieved a mean SBP of 133 mm Hg.1 Importantly, the intensive treatment strategy achieved SBP well below current treatment targets while standard treatment achieved SBP levels nearer to the current target of 130 mm Hg (and well below the protocol defined target of <140 mm Hg). Despite the significant difference in mean SBP, more intensive treatment was not associated with a significant reduction in the rates of major CVD events.1 However, the more intensively treated cohort had a significantly lower rate of both nonfatal and total stroke and lesser rates of progression of albuminuria.1 In addition, more intensive treatment was associated with an increase in both minor and major adverse events, although neither mortality nor significant renal, laboratory, or cardiovascular complications differed between groups.
A number of features of the ACCORD trial should be carefully considered to inform the discussion at hand; namely, whether the ACCORD study provides adequate information to alter current BP targets. If the ACCORD trial is considered the definitive data set for establishing treatment targets, 4 important factors warrant further discussion:
- • Are data from the standard treatment cohort of the ACCORD trial evidence that a target of 140 mm Hg is sufficient to re-establish treatment targets or should these data be used to support a target that encourages achieving an SBP well below 140 mm Hg (as the attained SBP was 133 mm Hg)?
- • Do comparable rates of major CVD events in the ACCORD BP trial provide sufficient evidence for a change in BP standards even with evidence of reduced rates of stroke and progression of renal disease?
- • Were the adverse effects of lower SBP targets of significant concern to support the safety and benefit of higher BP targets?
- • Do the ACCORD trial data support a new threshold for BP targets or support establishing a ceiling for BP limits?
First, we should consider the treatment approach applied in the standard cohort of the ACCORD trial. If one accepts that this approach is optimal (or at least provides equivalent protection with fewer adverse effects), then one may argue that the treatment goal in diabetes should ensure that individual patients achieve BPs in the low 130-mm Hg range (as occurred in the ACCORD trial). In addition, such a recommendation would require a treatment approach that mimics that of the ACCORD trial, meaning additional BP medication would be necessary if any 2 SBP readings were >140 mm Hg. Such an approach would also have to encourage regular use of combination therapy, because, in general, two drugs were required to achieve “standard” targets in the ACCORD trial.
Second, we must consider whether comparable rates of major CVD events alone provide sufficient evidence for a change in BP treatment targets. As noted, the risk of nonfatal and total stroke was significantly lower with the more intensive treatment strategy. In addition, this strategy was associated with reductions in markers of progressive renal disease. If this is the case, is the current target not more acceptable? In a recent review,10 Mancia summarizes this point, stating that if:
maximal protection is achieved at BP values lower for the brain (and kidney) than for the heart, then an important question is what BP-lowering strategy should be recommended in clinical practice. One might argue that if an aggressive BP reduction leads to cerebrovascular protection without substantially increasing the risk of coronary events (as suggested by the ACCORD trial), this strategy deserves to be adopted.
Third, consider whether the adverse effects of more intensive therapy were substantial enough to support a significant change in BP targets. As noted in the original report,1 serious adverse events were twice as common with more intensive BP control (although this occurred in an effort to achieve SBP <120 mm Hg not <130 mm Hg). These adverse events included any episode of hypotension, syncope, bradycardia, hyperkalemia, angioedema, or renal failure. While each are arguably important clinical end points, they generally occurred in only a minority of patients. In addition, these adverse effects were generally therapy related (increase in serum creatinine with use of angiotensin-converting enzyme [ACE] inhibition or angiotensin receptor blocker [ARB] therapy, hypokalemia, and hyperkalemia as a consequence of diuretics and ACE inhibitor/ARB, respectively). Virtually all were reversible with changes in therapy and they did not result in other significant morbidity or an increase in mortality. The authors report that neither rates of combined CVD nor rates of end-stage renal disease requiring dialysis or transplant (commonly reported end points of severe renal disease) were higher with more intensive therapy, suggesting little if any serious long-term harm. While not discounting the importance of adverse effects of treatment, we must consider whether these events truly represent undue risk. And, if so, this again supports the less intensive strategy that achieved SBP values nearer 130 mm Hg than 140 mm Hg.
Finally, consider the practical intent of treatment targets. In most circumstances, targets are set in hopes that a population achieves outcomes at or near recommended treatment targets (eg, in the case of hemoglobin A1c, low-density lipoprotein cholesterol where one third to one half of individuals have values above current targets [If higher blood pressure targets are established a similar trend may occur with blood pressure control – where more than half of treated individuals would anticipated to have SBP above the new target.]). This could be anticipated for BP targets as well. If so, re-establishing an SBP target of <140 mm Hg may well result in SBP values ≥140 mm Hg rather that nearer those obtained with “standard” treatment in the ACCORD study. Such an occurrence would potentially result in a population that is less effectively treated than that in the ACCORD standard cohort. This is but one of several potential unintended consequences that might occur with the adoption of new and higher targets.
Unintended Consequences: What Might Establishing Higher Targets Mean?
Admittedly, we must accept that some could reasonably conclude that the ACCORD trial data support a change to the SBP target employed in that trial. The ACCORD study provides good evidence that a target-driven approach to hypertension designed to achieve SBP <140 mm Hg in the majority of individuals would reduce CVD risk to a similar degree as more intensive BP treatment. The data also suggest that this could be achieved with less clinical effort and fewer resources. In addition, other randomized clinical trials support the concept that any lowering of BP (regardless of target) is of benefit. And one could maintain that only minimal improvement in rare (stroke) or intermediate (renal disease markers) clinical events were favorably affected by more intensive BP control.
With such conclusions one would argue that more intensive BP control should be limited to patients at higher risk for stroke and in patients with established renal disease (a large number of individuals). At first glimpse, all of these are rational arguments for setting higher BP targets; however, let us consider the possible unintended consequences that could result from such a change, knowing that other data of reasonable quality would have to be dismissed to make such a change. A move to higher BP targets in diabetes must be carefully considered based on the preponderance of evidence, not solely on the basis of the ACCORD trial results.
We must consider that if higher targets are established, the following possibilities must be taken into account:
- • Would an increase in the general BP target result in higher BP readings even in patients at higher risk, such as those at risk for stroke or those with renal disease?
- • Treatment targets define a threshold for optimized care (not simply an acceptable level of care), and in any circumstance when a measure is a continuous rather than a dichotomous variable, would an “acceptable” level of treatment adequately serve individualized needs?
- • Does establishing higher BP targets based on results of a single trial (albeit the most specific study addressing the issue) give adequate weight to data from other populations and other treatment approaches (such as in HOT) or for patients with differing baseline BP?
- • Finally what is the impact of “backing off” on targets and how might this be perceived by clinicians and patients concerned about complications risk in diabetes.
Major treatment guidelines from the ADA, the NKF, and JNC 7 have established and reiterated the current treatment targets for BP in diabetes.2–4 These recommendations are based on the results of more than 2 decades of clinical trials, observational studies, and epidemiologic analyses, which have been developed based on data that demonstrate a reduction in the risk of ALL complications of diabetes not solely on evidence of reduced CVD events (the primary end point of the ACCORD trial). In addition, these guidelines provide perspective on optimized care, not simply acceptable care and, by their nature, may be more stringent than clinical trial data alone can support. This issue is of critical importance as guidelines and targets are generally applied across a population, with the understanding that clinical decisions for individuals must include some variation in practice due to individual differences in risk. By conceding that higher BP values are acceptable, we put all patients at risk for stroke in a position where they may be undertreated, based as well on the ACCORD trial data. Similarly, any patient who develops or is at risk for renal disease may (using the higher target) be undertreated, and rates of progression of the most common form of chronic kidney disease would increase.
Also, consider the impact of establishing the new and higher dichotomous cutpoint (assuming ACCORD trial data are considered conclusive). Such a change provides little additional advantage for those who argue that dichotomous cutpoints are of limited value. Noting that BP is a continuous risk factor, evidence better supports that any lowering of BP (or at least a 10- to 20-mm Hg reduction) would provide similar guidance and comparable protection from CVD, stroke, and renal disease. Similarly, personalized and individualized decisions regarding BP control will still be necessary, and, as noted above, lower thresholds would promote clinical discussions between patient and provider, wherein higher targets might create a false sense of security or indifference to BP readings.
In addition, do we assume that the ACCORD data trump all prior data sets as regards establishing guidelines? Current guidelines are based on decades of data and dozens of studies with large data sets and included varying populations with varying baseline characteristics. Similarly, should we consider the treatment target applied (as in ACCORD and HOT) or should guidelines be based on the actual mean BP achieved (which is the putative determinant of risk)? Should BP targets offer acceptable levels of control or optimal levels of protection from all complications? This short commentary will not resolve these issues, and indeed further consensus statements will reflect best available evidence but will be influenced by both the bias of the experts involved and the considerations of cost of care, use of resources, and the need for individualized targets in diabetes.
Finally, consider the impact of a new and higher BP target on practitioners and patients who consider this guidance but may be less familiar with the data used to construct the guidelines. Such a change would be put forth in a community that has used (and in general embraced) the current target of <130 mm Hg for many years. Would a change in targets lead practitioners and patients to assume that BP control is now less important or unimportant? How will practitioners react to an individual patient with modest elevations in BP that met prior criteria for treatment but now would not require treatment? And what of the individual patient at lower risk who develops early renal disease or experiences a transient ischemic attack? What target is best for such individuals?
To think that one (even high-quality) study such as ACCORD can and should change a long–agreed upon standard is not rational. Any change in treatment targets would alter the landscape of care for decades to come, and, based on the data reviewed above, would be of little near-term or long-term benefit to diabetes patients. Maximal protection against all complications of diabetes while limiting the risk of over-treatment could only be achieved by rigorous adherence to an ACCORD-like protocol. Similarly, one could argue that similar results would be achieved by maintaining the current standard until more data or further analysis suggest otherwise.
The current standards of medical care in diabetes2 state that goal SBPs of <130 mm Hg are appropriate for most patients with diabetes. Based on patient characteristics and response to therapy, higher or lower SBP targets may be appropriate. Patients with diabetes should also be treated to a diastolic BP <80 mm Hg. Based on a review of ACCORD trial data, the professional practice committee of the ADA maintains these targets in their 2011 report and have concluded:
The absence of significant harm, the trends toward benefit in stroke, and the potential heterogeneity with respect to intensive glycemia management suggests that previously recommended targets are reasonable pending further analyses and results. SBP targets more or less stringent than 130 mm Hg may be appropriate for individual patients, based on response to therapy, medication tolerance, and individual characteristics, keeping in mind that most analyses have suggested that outcomes are worse if the SBP is >140 mm Hg.11
Now is not the time for change. The burden of diabetes—and its complications—continues to grow. Optimized care of diabetes warrants careful consideration of all treatment targets. While the evidence base may provide more exacting targets over time, targets must also be considered for what they are: a reflection of optimal care, not a simple acceptable standard. More intensive lowering of BP in diabetes results in a significant reduction in stroke risk1 and limits the progression of renal disease in patients with significant chronic kidney disease. To take into full account the broad evidence base supporting these benefits, including the ACCORD trial results, supports maintaining current BP targets as established.
Disclosures: The views presented in this manuscript do not necessarily reflect those of the American Diabetes Association. Dr Kendall served as an investigator for the ACCORD trial from 1999 to 2005 and again from 2008 to 2009.
Dr Kendall receives no honoraria, research support, or other payments from outside interests. His spouse is currently employed by Genentech, Inc.