Smoldering Arteries?: Title and subTitle BreakLow-grade Inflammation and Coronary Heart Disease

C-reactive protein (CRP) is the classic "acute-phase reactant," the plasma levels of which can increase as much as 10,000-fold in response to tissue injury and infection.¹ C-reactive protein was discovered in the plasma of patients with acute pneumococcal pneumonia 70 years ago and was so named for its capacity to bind pneumococcal C-polysaccharide. Plasma CRP level has long been a widely measured marker of disease activity in inflammatory conditions, but recently, there has been increased interest in the possible relevance of low-grade inflammatory processes to cardiovascular disease and vascular risk factors.^{2 - 3}

A number of favorable biological characteristics make CRP a useful marker of such subtle inflammation: its plasma levels are determined mainly by the synthesis rate in the liver and closely reflect inflammatory activity; the protein is stable in frozen blood samples, allowing measurements in banked serum samples; and, despite sharp increases that occur during the acute-phase response, longer-term plasma CRP levels show about the same degree of year-to-year consistency within individuals as some more extensively studied risk factors (such as blood cholesterol levels and blood pressure).⁴ Moreover, highly sensitive assays for CRP are now available that can precisely measure values within the range less than 1.0 mg/dL and thereby detect low-grade inflammation that would not previously have been noticed.

In this issue of THE JOURNAL, the study by Visser and colleagues⁵ illustrates several advantages of studying CRP in large-scale epidemiological samples. Previous reports suggested an association between plasma CRP levels and obesity, but these studies were relatively small and unable to exclude some possible biases. By contrast, Visser et al studied 16,000 American adults in a cross-sectional community-based national survey. This sample included large numbers of nonsmokers and younger individuals, groups in whom some potential biases, such as confounding by cigarette smoking or preexisting inflammatory diseases, should be minimized. The study convincingly demonstrates that plasma CRP levels are substantially higher in obese and overweight people than in leaner people. As the authors point out, future studies of obesity should attempt to measure other inflammatory factors, particularly interleukin 6, a proinflammatory cytokine produced in adipose tissue and elsewhere that stimulates the liver to make plasma acute-phase proteins such as CRP.

Visser et al also have raised broader questions about low-grade inflammation and coronary heart disease (CHD) by suggesting that plasma CRP might at least partly mediate the effects of obesity on CHD.⁶ By the early 1990s, a number of short-term studies had reported CRP levels to be predictive of cardiac complications, such as the need for coronary revascularization, among patients recently hospitalized for acute coronary syndromes. These studies, however, provided limited information because increased CRP levels in such patients might be partly attributable to the severity of the disease that caused hospitalization, despite attempts to adjust for this (such as by adjusting for infarction size). In contrast, long-term prospective studies, in which CHD events have been recorded for several years after baseline blood collection, should be less prone to such biases because they limit the influence of preexisting disease on the factors being investigated. The first such report on CRP and CHD was published only in 1996, and by 1998, 6 more had appeared.³ In aggregate, the 7 studies involved 1053 cases of nonfatal myocardial infarction or death from CHD, with a weighted mean follow-up of 6 years. A combined analysis of the studies yielded a risk ratio for CHD of 1.7 (95% confidence interval, 1.4-2.1) for people with baseline CRP levels in the top third compared with those in the bottom third of the population, corresponding to usual CRP values of 0.24 vs 0.10 mg/dL.³

Although the strength of this association with CHD risk appears comparable with that for some more extensively studied risk factors (such as blood fibrinogen level), several uncertainties remain. The meta-analysis of prospective studies substantially reduced the scope for a false-positive result due to chance.³ Moreover, the results of the separate studies are consistent with the overall risk ratio, even when several additional prospective studies reported since 1998 are included.⁷ Doubts persist, however, about the extent to which inadequate adjustment for possible confounding factors (such as smoking) and the preferential publication of studies with more extreme results might have led to exaggerated estimates. Also, increased plasma CRP level may be chiefly an indicator of the extent and severity of subclinical atherosclerosis, even in studies that monitored apparently healthy people. Some of these uncertainties can be resolved by conducting larger observational studies (as well as by more detailed syntheses of available studies) that provide much more reliable estimates of the risk associated with CRP and CHD in a range of different circumstances (such as at different levels of other risk factors, at different times in the evolution of risk, and at different ages).

Even if an independent association is established between plasma CRP level and CHD, several questions would remain. Is CRP itself a direct mediator of vascular damage or merely a marker of a more fundamental cause of disease—or both? There is no strong evidence so far for a direct role for CRP in vascular disease,^{8 - 9} but additional experimental research is clearly needed, as even the physiological role of plasma CRP remains largely unknown. In apparently healthy people, there are close correlations between CRP levels and other plasma acute-phase proteins, such as serum amyloid A and albumin.¹⁰ Meta-analyses of prospective studies have established associations between CHD and other acute-phase reactants (such as fibrinogen, albumin, and leukocyte count).³ Preliminary studies have also reported associations of CHD with certain circulating cytokines and cell adhesion molecules, which are different factors in inflammatory cascades that appear to be associated with plasma CRP levels.¹¹ Hence, there is a generally consistent pattern of some underlying processes related to inflammation, rather than plasma CRP per se, that is relevant to CHD.

If so, what factors might be responsible for such inflammation? A number of suggestions have been made, ranging from factors within the arterial wall, such as oxidized low-density lipoprotein and local infection by agents such as Chlamydia pneumoniae or cytomegalovirus, to factors outside the circulation, such as chronic gastric infection with Helicobacter pylori and periodontal disease.¹² Large studies are in progress to find out whether persistent infection with any particular agents is really associated with changes in plasma inflammatory factors, but so far the evidence is weak.^{10 ,13} Some female sex hormones are produced in adipose tissue, and the idea that estrogens might promote low-grade inflammation has been suggested by a randomized trial that reported sustained increases in plasma CRP levels in women treated with hormone therapy regimens.¹⁴ However, high levels of estrogens alone would not satisfactorily explain a role for low-grade inflammation in men with CHD. A partial genetic basis for low-grade inflammation is suggested by a study of several hundred apparently healthy twins in whom monozygotic pairs had closer correlations of circulating levels of CRP (and serum amyloid A protein) than did dizygotic pairs.¹⁵

Despite incomplete knowledge about the factors responsible for persistent low-grade inflammation, several interventions have been proposed to prevent CHD by their presumed "anti-inflammatory" actions. One report has suggested that the vascular protective effects of aspirin increase with increasing baseline CRP levels,¹⁶ but that claim is not statistically convincing and is not supported by large syntheses of randomized trials that have excluded much greater vascular benefits in patients taking higher ("anti-inflammatory") dosages of aspirin than in those taking lower dosages.¹⁷ Small trials of macrolides (antibiotics with antichlamydial and, possibly, anti-inflammatory effects) have tentatively suggested reductions in certain plasma markers of inflammation, including CRP, and these suggestions are currently being tested in larger trials that, additionally, aim to assess any effects on CHD.¹⁸ Long-term use of statins (3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors) may produce sustained reductions in plasma CRP levels that are apparently unrelated to the degree of cholesterol lowering achieved.⁴ Both this finding and a preliminary report that the benefits of pravastatin in CHD prevention may increase with increasing evidence of baseline inflammation⁷ support the possibility that statins involve potentially relevant nonlipid effects.¹⁹ Larger trials are needed to test this claim, perhaps particularly in stroke prevention, for which statin use has been surprisingly effective despite the lack of a strong overall association between stroke and serum cholesterol concentrations in observational studies.²⁰ Such studies of inflammatory markers in clinical trials might also help identify any vascular benefits of angiotensin-converting enzyme inhibitors, such as ramipril, that are unrelated to blood pressure lowering.²¹

Thus, the evidence for the relevance of low-grade inflammation to CHD now derives from a variety of different sources, ranging from pathological studies of the vessel wall to epidemiological studies of circulating "inflammatory" factors to preliminary intervention studies. This hypothesis is scientifically attractive and potentially important, but much more work needs to be done before any clinical implications are established.