NT-proBNP Beats CRP at Predicting CV Risk in Older Patients

June 23, 2011

June 22, 2011 (Updated June 28, 2011) (London, United Kingdom) — N-terminal pro-brain natriuretic peptide (NT-proBNP) was better than C-reactive protein (CRP) at predicting major cardiovascular events in older men with and without preexisting cardiovascular disease in a new prospective study [1].

The study, published in the June 28, 2011 issue of the Journal of the American College of Cardiology, was led by Dr Goya Wannamethee (University College London, UK).

Wannamethee commented to heartwire : "Other studies have shown that BNP can predict cardiovascular events, but the novel thing in our study is that we have shown that it is useful in stratifying patients when added on top of the risk factors we already use, ie, the Framingham score."

She noted that this particular study's results were applicable only to the older population (over 60). "We are saying that NT-proBNP seems particularly useful in this older population, where blood pressure and cholesterol are not as reliable predictors as they are in middle-aged people. Also, our study was only in men and mostly Europeans, so we would like to see confirmation in other populations. But our results do appear to suggest that BNP is more useful than CRP for risk prediction in this older population. I would say it is the best marker of risk so far in this population."

This is the second study showing NT-proBNP to be a useful risk predictor to be reported this week, with an analysis from ASCOT presented at the European Society of Hypertension (ESH)European Meeting on Hypertension 2011 this past weekend showing the marker to predict cardiovascular events in a hypertensive population without preexisting cardiovascular disease.

In this study, 3649 men age 60 to 79 years were followed for a mean of nine years, during which there were 608 major cardiovascular events. Results showed that NT-proBNP was significantly associated with risk of all major cardiovascular outcomes after adjustment for cardiovascular risk factors in both men with and without cardiovascular disease, whereas CRP was associated with cardiovascular outcomes only in men without preexisting cardiovascular disease and was a weaker predictor than NT-proBNP.

Adjusted Hazard Ratios (95% CI) for Future Cardiovascular Events

Measure Men without preexisting CV disease Men with preexisting CV disease
NT-proBNP 1.49 (1.33–1.65) 1.52 (1.33–1.75)
CRP 1.22 (1.10–1.34) 1.00 (0.86–1.38)

When incorporated into a model including Framingham data, NT-proBNP significantly improved the ability to predict cardiovascular events, as shown by an improvement in the C statistic, whereas CRP did not improve prediction.

Net Reclassification Improvement in C Statistic (%) When Added to Framingham Data

Measure Men without preexisting CV disease Men with preexisting CV disease
NT-proBNP 8.8 8.2
CRP 3.8 0.6

Wannamethee said: "NT-proBNP has traditionally been thought of a marker of heart failure. But this study was done in the general population, so it is obviously picking up something else here--not just heart failure. It predicts CHD events and stroke as well as heart failure."

She noted: "It's still at an early stage. It's not going to be used routinely any time soon, but I would hope that this study will stimulate further research to look at its potential role in general practice. This would involve thinking about how it would fit in with other tests and who it would best be used in." She pointed out that NT-proBNP was also more specific for cardiovascular disease than CRP, which tended to be elevated in most major diseases.

Outcome Trials Needed

Commenting on the paper for heartwire , Dr Michaal Lauer (US National Heart, Lung, and Blood Institute, Bethesda, MD) said the study adds to a growing literature that BNP might be a reasonable predictor of risk. "But there are hundreds of studies like this with various markers, so I wouldn't put too much weight on this particular one."

Lauer pointed out that the major problem with all these studies is that they don't answer the question of whether use of such markers will lead to a reduction in clinical outcomes. "To do this, we need to perform large clinical trials where people are tested or not tested then followed and events compared between the two groups." He noted that the US National Cancer Institute has conducted trials like this that have shown that CT scans to detect lung cancer do lead to a reduction in death rates but that measuring prostate-specific-antigen (PSA) levels does not affect outcomes. "We need the same sort of trials for the cardiovascular biomarkers," he added.

He suggested that a less expensive alternative, but a less satisfactory trial, would be one that gave a treatment based on the biomarker result and then compared outcomes between those who got the treatment and those who didn't. "This is what JUPITER did with CRP. This has given us some information about CRP, but it is not ideal, because we don't know what would have happened to the patients who were not positive for raised CRP."

Lauer said he believed the most powerful biomarker for the prediction of cardiovascular risk is actually coronary calcium, but it is not possible to say whether it would actually be useful without an outcome trial.

Also commenting for heartwire , Dr Christopher DeFilippi (University of Maryland, Baltimore), said: "The current analysis provides important support to establish NT-proBNP as a potentially important risk stratifier in the large 'at-risk' population of older adults living in the community. An important secondary analysis in this study is identifying NT-proBNP level as a significant risk stratifier in subjects without known cardiovascular disease. These are the individuals who can be the most difficult to identify, but as the authors show, their risk of cardiovascular events isn't trivial. The next step now is to determine what actions an elevated NT-proBNP level can trigger to reduce progression to symptomatic cardiovascular disease or cardiovascular death."

Dr Scott Grundy (University of Texas Southwestern Medical Center, Dallas) noted that BNP seems to reflect myocardial failure or a predisposition to myocardial failure, and it is possible that people who have an underlying myocardial disease are more prone to MI. "It is interesting that the study included only older people, who may be more likely to have a high BNP," he said, adding that other research has shown that persons with left ventricular hypertrophy (LVH) have higher BNP. "I recall that the Framingham study showed that LVH is a predictor of CHD. So maybe if the authors had looked at LVH by ECG they would have found the same thing."

Ridker: Time to Move Beyond Framingham

CRP enthusiast Dr Paul Ridker (Brigham and Women's Hospital, Boston, MA) said the paper "provides yet more evidence that the time has come to move beyond the simple Framingham score." He added: "The finding that any biomarker predicts risk among those with preexisting heart disease has little clinical relevance: all such individuals should already be on maximal therapies. For CRP, we already have trial data indicating that those with elevated levels benefit from a therapy they otherwise would not have received (from JUPITER). We don't yet have those data for BNP, but someday we may, and physicians need to be open to advances in diagnosis and treatment to best take care of their patients."