Blood Test's Value Questioned

[Guest guest]

A study casts doubt on the belief that inflammation markers in the

bloodstream are a powerful predictor of heart attacks. The report in

today's New England Journal of Medicine suggests that inflammation is only a

moderate predictor, and that the blood tests are of little value.

Danesh of the University of Cambridge and colleagues used data

from an Iceland study that began in 1967. They compared 2,459 people

who had a heart attack or died of heart disease over 20 years of follow-up

with 3,969 participants who did not have a heart attack. The researchers

calculated that those with higher levels of a protein indicating

inflammation had a 45 percent increased risk of heart disease.

[Guest guest]

Hi All,

I am sorry.

Cheers, Al Pater.

New Engl J Med 2004;350:1387-97.

C-Reactive Protein and Other Circulating Markers of Inflammation in

the Prediction of Coronary Heart Disease

J. Danesh and Others

C-reactive protein is an inflammatory marker believed to be of value

in the prediction of

coronary events.We report data from a large study of C-reactive

protein and other circu-

lating inflammatory markers,as well as updated meta-analyses,to

evaluate their rele-

vance to the prediction of coronary heart disease.

methods

Measurements were made in samples obtained at base line from up to

2459 patients

who had a nonfatal myocardial infarction or died of coronary heart

disease during the

study and from up to 3969 controls without a coronary heart disease

event in the Reyk-versity

javik prospective study of 18,569 participants.Measurements were made

in paired

samples obtained an average of 12 years apart from 379 of these

participants in order to

quantify within-person fluctuations in inflammatory marker levels.

results

The long-term stability of C-reactive protein values (within-person

correlation coeffi-

cient,0.59;95 percent confidence interval,0.52 to 0.66)was similar to

that of both

blood pressure and total serum cholesterol.After adjustment for base-

line values for

established risk factors,the odds ratio for coronary heart disease

was 1.45 (95 percent

confidence interval,1.25 to 1.68)in a comparison of participants in

the top third of the

group with respect to base-line C-reactive protein values with those

in the bottom third,

and similar overall findings were observed in an updated meta-

analysis involving a total

of 7068 patients with coronary heart disease.By comparison,the odds

ratios in the

Reykjavik Study for coronary heart disease were somewhat weaker for

the erythrocyte

sedimentation rate (1.30;95 percent confidence interval,1.13 to 1.51)

and the von Wil-

lebrand factor concentration (1.11;95 percent confidence

interval,0.97 to 1.27)but

generally stronger for established risk factors,such as an increased

total cholesterol

concentration (2.35;95 percent confidence interval,2.03 to 2.74)and

cigarette smok-ing

(1.87;95 percent confidence interval,1.62 to 2.16).

conclusions

C-reactive protein is a relatively moderate predictor of coronary

heart disease.Recom-

mendations regarding its use in predicting the likelihood of coronary

heart disease may

need to be reviewed.

Discussion

.....We found that the decade-to-decade consistency of

values for C-reactive protein,the erythrocyte sedi-

mentation rate,and von Willebrand factor is similar

to that of values for blood pressure and total serum

cholesterol concentration,suggesting that these

inflammatory markers are sufficiently stable for

potential use in the long-term prediction of coro-

nary heart disease.Our findings — reinforced by an

updated meta-analysis — indicate,,however,that

the odds ratio for coronary heart disease in people

with elevated C-reactive protein values is lower than

that reported recently.Whereas a previous meta-

analysis 14 of studies published before 2000 (based

on 1953 cases of coronary heart disease)reported

an odds ratio for coronary heart disease of about

2.0 (95 percent confidence interval,1.6 to 2.5),our

updated meta-analysis,which adds 5115 cases of

coronary heart disease from a further 12 studies,

yielded an odds ratio of about 1.5 in a comparison

of people with base-line values in the top third with

those with base-line values in the bottom third for

the population.Moreover,in comparison with major

established risk factors (such as an increased total

serum cholesterol concentration and cigarette

smoking),the C-reactive protein concentration was

a relatively moderate predictor of the risk of coro-

nary heart disease and added only marginally to the

predictive value of established risk factors for coro-

nary heart disease.These findings suggest that re-

cent recommendations regarding the use of mea-

surements of C-reactive protein in the prediction of

coronary heart disease may need to be reviewed.3

The potential limitations of our study merit care-

ful consideration.The validity of our measurements

is demonstrated by the reasonably high decade-to-

decade consistency of C-reactive protein values re-

corded in paired samples from 379 participants (a

level of stability that was at least as high as those re-

corded in previous studies with sampling intervals

of just one to five years 35-38 ).Further validation is

suggested by the finding of the expected base-line

associations of C-reactive protein with other in-

flammatory markers and with established coro-

nary risk factors.

The mean values and the distributions of several

established coronary risk factors (and the strength

of their associations with the risk of coronary heart

disease)in our study were generally similar to those

reported in other western European populations.8

Therefore,although the relative homogeneity of the

Reykjavik population should have minimized cer-

tain residual biases (such as that due to differences

in socioeconomic status),the present findings

should have wider relevance.Only total serum cho-

lesterol concentrations were measured in the pres-

ent study (rather than those of its subfractions,

which have opposing effects on the risk of coronary

heart disease),thereby underestimating the predic-

tive ability of lipid concentrations (and potentially

overestimating the adjusted predictive value of the

C-reactive protein concentration).

No information was recorded on the use of aspi-

rin and statins,which,like hormone-replacement

treatment,may alter C-reactive protein values.How-

ever,fewer than 5 percent of the women in this

study reported the use of such hormonal treatment

during recruitment,and the use of aspirin and of

statins was similarly uncommon in the general

middle-aged population of Reykjavik between 1967

and 1991.We did not address the separate issues of

the predictive value of inflammatory markers with

respect to the risk of cardiac complications among

patients recently hospitalized for acute coronary

syndromes 39 or the long-term risk of coronary heart

disease in patients with a history of cardiovascular

disease.14

As suggested by the statement of the Centers for

Disease Control and Prevention and the American

Heart Association,3 further clarification of the pre-

dictive value of C-reactive protein in coronary heart

disease in general populations will require the pool-

ing of studies on the basis of data for individual

participants from each of the available prospective

studies.Such a strategy will permit more complete

adjustment for other risk factors and for within-

person fluctuations of C-reactive protein levels,

more precise quantification of the associations in

particular subgroups (such as age-,sex-,and dura-

tion-specific associations as well as assessments of

combinations of inflammatory markers),more reli-

able characterization of the shape of any dose –re-

sponse relation,and more detailed investigation of

potential sources of heterogeneity.

Supported by program grants from the British Heart Foundation

(to Profs.Danesh and Lowe)and the Medical Research Council (to

Prof.Pepys)and by the and Beverly Sackler Research

Award in the Medical Sciences (to Prof.Danesh).Dr.Hirschfield was

supported by a Medical Research Council Clinical Training Fellow-

ship.

We are indebted to Prof.Simon and Dr. Whitlock

for helpful comments,to Kelsey Juzwishin for epidemiologic sup-

port,to the laboratory staff of the Icelandic Heart Association,to

Drs.M. and D.Goodier of the Clinical Chemistry Depart-

ment at the Royal Free Hospital,to Fiona Key and Craig at the

University Department of Medicine at the Glasgow Royal Infirmary,

and to Roche Diagnostics for donating the C-reactive protein assay

kits.