Current issues of ACP Journal Club are published in Annals of Internal Medicine


Serum sialic acid predicts cardiovascular mortality

ACP J Club. 1991 May-June;114:91. doi:10.7326/ACPJC-1991-114-3-091

Source Citation

Lindberg G, Eklund GA, Gullberg B, Råstam L. Serum sialic acid concentration and cardiovascular mortality. BMJ. 1991Jan 10;302:143-6.



To determine whether serum sialic acid level predicts either short- or long-term mortality from cardiovascular disease.


Cohort of men and women followed for 20 years.


A geographically defined part of the county of Värmland, Sweden.


Residents were contacted in 1964 and 1965 as part of a global health survey. They were between 45 and 74 years of age during any of the 20 years of follow-up. Complete data were available for over 98% of participants, who constituted 78% (26 693) of men and 81% (27 692) of women of the total population in their age range. Follow-up consisted of 329 543 person-years for men and 355 665 person-years for women.

Assessment of risk factors

At screening, blood samples were taken from nonfasting participants, and serum sialic acid levels were measured by automatic analyzer. Quartiles of sialic acid levels were determined for each 5-year age stratum. Serum cholesterol levels were measured, and blood pressure, height, and weight were recorded.

Main outcome measure

All deaths occurring in participants who were aged 45 to 74 years, and listed in the Swedish mortality register as cardiovascular deaths, were counted for analysis.

Main results

Mean sialic acid levels were 68.8 mg/dL (SD 8.0) for men, and 69.2 mg/dL (SD 8.0) for women. 21% of men and 12% of women died during follow-up, with 54% and 41%, respectively, of these deaths being cardiovascular-related. The relative risk for cardiovascular and noncardiovascular death increased with increasing serum sialic acid concentration, to 2.38 and 1.50, respectively, for men, and to 2.62 and 1.89, respectively, for women, for the highest relative to the lowest sialic acid quartile. Relative risk for cardiovascular mortality was higher than for noncardiovascular mortality when corelated with quartile of sialic acid concentration (P < 0.001 for men and P < 0.05 for women). For both cardiovascular and noncardiovascular death, an increasing trend was seen for all quartiles for both men and women (P < 0.001 in all cases). Elevated relative risk with increased serum sialic acid level remained significant in multivariate analysis considering total serum cholesterol level, diastolic blood pressure, and body mass index {no data reported}.


Serum sialic acid level predicted cardiovascular mortality and, to a lesser degree, noncardiovascular mortality.

Source of funding: Värmland County Council.

Address for article reprint: Dr. G. Lindberg, Centre for Public Health Research, S-65182 Karlstad, Sweden.


Lindberg and colleagues report on a fascinating observation that serum sialic acid level predicts cardiovascular (and noncardiovascular) mortality. The study's classical design of measuring a factor and relating it to subsequent mortality was greatly strengthened by its large community-based sample. However, this study raises several concerns. First, it is unclear whether any hypothesis about sialic acid had been set out by the investigators 20 years ago or whether this was part of the series of protein measurements that was done without a particular hypothesis. Second, because the measurement of sialic acid was done in the 1960s, another sialated protein may have been measured. Third, many more risk factors for coronary disease are known now than were known 20 years ago; for example, the only lipid measured in that study was total cholesterol. Thus, the relation of other lipids and lipoproteins to sialic acid could not be determined. Also, no data on cigarette smoking were provided. Fourth, none of the data from multivariate analysis were provided to support the claim that sialic acid is an independent predictor.

The investigators allude to 2 additional problems. 1 is the consideration of other possible operative factors. Fibrinogen may be a confounder because it is an acute-phase reactant like sialic acid and has been linked to cardiovascular disease. Regarding this possibility, the investigators could only speculate about a biologic basis for their observation.

Before relegating this paper to "one-upmanship" at rounds (i.e., for asking the junior resident whether the sialic acid level was measured in the patient), further investigation is clearly necessary to examine the factors that alter serum sialic levels and to understand the biology linking it to the development of atherosclerosis.

Simon W. Rabkin, MD
University HospitalVancouver, British Columbia, Canada