Do Angiotensin Receptor Antagonists Decrease hsCRP Independent of Blood Pressure—and Does It Matter?

C‐reactive protein measured by high‐sensitivity assays (hsCRP) has been demonstrated to be an independent predictor of subsequent cardiovascular (CV) events in a variety of patient populations. Additionally, strategies that have been shown to reduce CV events, including 3‐hydroxy‐3‐methylglutaryl coenzyme A reductase inhibitors (statins) and aspirin, have also been shown to reduce hsCRP. Many investigators have suggested that hsCRP is an important surrogate end point for predicting future CV events. A number of investigators have also argued that in patients with hypertension, agents that block the renin‐angiotensin‐aldosterone‐system (RAAS), such as angiotensin receptor blockers (ARBs), have CV benefits over and above blood pressure (BP) control. There has been a strong temptation to link these 2 findings—ie, that the purported incremental CV benefit of RAAS blockade is mediated by decreasing vascular inflammation in a BP‐independent manner.

Two recently published clinical trials have explored this potential link. In the Valsartan, Blood Pressure Reduction, and C‐Reactive Protein (Val‐MARC) trial, 1668 men and women with stage 2 hypertension were allocated to either valsartan or valsartan/hydrochlorothiazide (HCTZ). Baseline hsCRP was 2.11 mg/dL in the valsartan group and 2.08 mg/dL in the valsartan/HCTZ group. After 6 weeks of therapy, hsCRP decreased significantly in the valsartan group (−0.12 mg/dL), but despite better BP control, hsCRP did not decrease in the valsartan/HCTZ group (+0.05 mg/dL), a difference of 13.3% (P<.001). The authors concluded that 6 weeks of treatment with the ARB valsartan produced a BP‐independent reduction in hsCRP that was apparently negated with the concomitant use of HCTZ. Similarly, in this issue of The Journal of Clinical Hypertension, Kyvelou et al ¹ report that in 323 subjects with well‐controlled high‐renin hypertension, the use of any of the 6 available ARBs led to a statistically significant decrease in hsCRP and serum amyloid A (another potential marker of inflammation). Reductions in hsCRP were seen with ARB therapy with or without the use of a diuretic.

A number of key questions remain:

• •Do agents that block the RAAS provide additional CV protection over and above BP control?
• •If so, are the additional benefits seen with ARBs mediated through reductions in hsCRP?
• •Is hsCRP an important surrogate end point for clinical or research use in patients with hypertension?

The remainder of this commentary will focus on the current state of our knowledge regarding each of these questions.

First, do agents that block the RAAS provide additional CV protection over and above BP control? While it is clear that RAAS blockade provides BP‐independent benefits in patients with chronic kidney disease or chronic symptomatic heart failure due to decreased left ventricular systolic function, the majority of the data do not support significant additional benefit for RAAS blockade in most patients with essential hypertension. While studies such as the Heart Outcomes Prevention Evaluation (HOPE) and the Perindopril Protection Against Recurrent Stroke Study (PROGRESS) are frequently cited to demonstrate a unique benefit of RAAS blockade (and have led to the designation of compelling indications for angiotensin‐converting enzyme inhibitors [ACEIs] and ARBs in the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure [JNC 7]), it is important to remember that these studies used placebo rather than active control—and that in many of the trials, diuretics were used in combination with the RAAS inhibitor.

To prove unique advantages of RAAS blocking agents over other antihypertensive therapies, studies with active control groups are needed. The only large‐scale studies with actively treated control groups that have suggested a benefit to RAAS blockade despite equivalent BP reduction were the Anglo‐Scandinavian Cardiac Outcomes Trial (ASCOT) (in which most subjects randomized to the calcium channel blocker [CCB] amlodipine received add‐on therapy with the ACEI perindopril) and Losartan Intervention for Endpoint Reduction in Hypertension (LIFE) trials (using the ARB losartan), both of which used atenolol as a comparator. Given the recent controversy around the comparable benefits of atenolol in the treatment of the elderly hypertensive, it seems likely or at least plausible that the results of ASCOT and LIFE are attributable to a worse‐than‐expected outcome with atenolol, rather than significant incremental benefit from RAAS blockade. This suggestion is supported by a subgroup study of ASCOT, where despite equivalent brachial cuff BP in the 2 groups, the atenolol‐based strategy was found to be associated with significantly higher central BP.

Studies using other active comparators, like the Antihypertensive and Lipid‐Lowering Treatment to Prevent Heart Attack Trial (ALLHAT) (ACEI compared with thiazide‐type diuretic) or Valsartan Antihypertensive Long‐Term Use Evaluation (VALUE) (ARB compared with CCB) revealed no outcome advantage with RAAS blockade compared with other BP‐lowering strategies. Both ALLHAT and VALUE have well described and potentially substantial methodologic limitations. If there was a reasonably large, clinically significant beneficial effect to RAAS blockade over and above BP control, one would have expected to at least see a trend toward improved outcomes in these studies. Although this question will continue to be debated, to date there does not seem to be clear evidence supporting substantial benefit to RAAS blockade over and above BP control for most patients with essential hypertension. However, these agents should remain an essential part of combination therapy to obtain goal BP.

Second, if present, are the additional benefits seen with ARBs mediated through reductions in hsCRP? Although the study from Kyvelou et al ¹ and Val‐MARC suggest that ARBs decrease hsCRP in patients with hypertension, it is important to recall that both of these studies looked at highly selected, not necessarily representative, populations, and that there are some troubling inconsistencies between these studies. Val‐MARC was limited to patients with stage 2 hypertension, and reductions in hsCRP were demonstrated only in patients on ARB monotherapy. The study by Kyvelou et al investigated an even more restricted population, limited to those who had high‐renin hypertension and were able to reach goal with either an ARB or ARB/diuretic combination. Patients with low‐to‐moderate baseline renin levels or whose BP was not controlled with an ARB or ARB/diuretic combination were specifically excluded. Also, the reduction in hsCRP appeared to be confined to nonsmokers, and unlike in Val‐MARC, the concomitant use of thiazide‐type diuretics did not seem to negate the effect on hsCRP.

It remains unclear, therefore, whether ARBs as monotherapy or as part of combination therapy (particularly with HCTZ) reduce hsCRP in the vast majority of patients with stage 1 hypertension. Future studies are needed to test the hypothesis that ARBs, alone or in combination, reduce hsCRP more than other antihypertensive medications. Also, it remains unclear whether the effect on hsCRP, if present, is a feature of RAAS blockade in general or due to a unique property of ARBs. If future studies are to be undertaken, it would be interesting to examine the comparative effects on hsCRP of other agents that block the RAAS, including ACEIs, aldosterone antagonists, and investigational oral renin inhibitors.

Before we commit substantial resources to such investigations, however, it is imperative to address the third question: Is hsCRP an important surrogate end point for clinical or research use in patients with hypertension? Certainly, there are data demonstrating that baseline levels of hsCRP are independently associated with the subsequent risk of CV events in patients with and without hypertension, and there have been advocates for incorporating hsCRP into global risk prediction strategies, at least for patients at intermediate or indeterminate risk. Also, given our increasing knowledge of the inflammatory nature of the atherothrombotic process, there is substantial biologic plausibility to the importance of hsCRP and other markers of inflammation. These facts combined with the relative ease and low cost of measurement have led to increasing enthusiasm for using hsCRP as a potential surrogate marker for future CV events. The modest decrease in hsCRP seen with valsartan monotherapy in Val‐MARC (−13.3%) and with all ARBs in Kyvelou et al (−20%) approaches that seen with other CV agents with proven anti‐inflammatory and antiatherosclerotic effects, such as moderate‐dose statin therapy, which has generally demonstrated a 15%–18% reduction in hsCRP.

Nonetheless, the extent to which reductions in hsCRP contribute to the beneficial effects seen with statins and other CV risk reduction strategies remains unclear. We must keep in mind that hsCRP is a nonspecific marker of inflammation and that other strategies that decrease inflammation, such as cyclooxygenase‐2 inhibitors, have actually been shown to increase the risk of CV events. It therefore seems reasonable to conclude that although baseline levels of hsCRP may be predictive of future CV risk, we do not yet have enough data to conclude that on‐treatment changes in hsCRP are predictive of changes in risk with medications other than statins. Whether other more specific markers of vascular inflammation such as lipoprotein associated phospholipase A₂ will provide better predictive value than hsCRP is also unknown.

This discussion should not dampen enthusiasm for future research with hsCRP; rather it should encourage future investigations to prove whether or not a change in hsCRP is an appropriate surrogate marker for clinical and research use. Until we have a better understanding of the role of hsCRP as a surrogate marker and target of intervention, future studies comparing the effects of various antihypertensive medications on hsCRP (and other markers of inflammation) should be seen as hypothesis generating rather than yielding results that should change our treatment recommendations or clinical practice.