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


Triglyceride level but not LDL particle size was an independent risk factor for MI in men

ACP J Club. 1997 Mar-Apr;126:49. doi:10.7326/ACPJC-1997-126-2-049

Related Content in this Issue
• Companion Abstract and Commentary: LDL particle size was smaller in CAD, but other lipid parameters were stronger predictors of CAD

Source Citation

Stampfer MJ, Krauss RM, Ma J, et al. A prospective study of triglyceride level, low-density lipoprotein particle diameter, and risk of myocardial infarction. JAMA. 1996 Sep 18;276:882-8.



To determine whether small-sized low-density lipoprotein (LDL) particles and elevated triglyceride levels are independent risk factors for myocardial infarction (MI) in men.


Nested case-control study of men (Physicians' Health Study).


Community study in the United States.


Physicians (n = 266) who developed an MI (fatal or nonfatal) during follow-up were studied. Exclusion criteria were a history of MI, stroke or transient ischemic attacks, cancer, renal or liver disease, peptic ulcer, or gout. Each patient was matched for age, smoking, and time from randomization with men who had not had an MI (n = 308).

Assessment of risk factors

Blood samples (85% nonfasting) were collected at baseline. Triglyceride levels and LDL particle size were the main risk factors of interest. Other risk factors were age; body mass index; total cholesterol (TC) level; high-density lipoprotein (HDL) cholesterol levels or triglyceride levels; alcohol consumption; physical activity; smoking; history of angina, diabetes, or hypertension; and family history of MI.

Main outcome measures

MI was ascertained by questionnaires (every 6 mo), telephone contact, and autopsy records reviewed by the study end point committee using World Health Organization criteria for MI. Patients with sudden death and no history of coronary disease were excluded.

Main results

An increase in the risk for MI was associated with small LDL particle size (relative risk [RR] 1.38 per 0.8-nm decrease in diameter, 95% CI 1.18 to 1.62, P < 0.001); the TC:HDL cholesterol ratio (RR 1.35, CI 1.19 to 1.53, P < 0.001); triglyceride levels (RR 1.43 per 100-mg/dL increase, CI 1.22 to 1.68, P < 0.001); and TC levels (RR 1.65 per 40-mg/dL increase, CI 1.37 to 1.99, P < 0.001). Decreased risk for MI was associated with HDL cholesterol levels (RR 0.87 per 10-mg/dL increase, CI 0.76 to 0.99, P = 0.03). When adjusted for all risk factors, including the interaction between triglycerides and TC, the risk for MI was increased with increasing triglyceride levels (RR 1.40 per 100-mg/dL increase, CI 1.10 to 1.77, P = 0.007) and increasing TC levels (RR 1.80 per 40-mg/dL increase, CI 1.44 to 2.26, P < 0.001). The association between the risk for MI and LDL particle size was no longer significant (RR 1.09 per 0.8-nm decrease, CI 0.85 to 1.40, P = 0.5).


Triglyceride and total cholesterol levels but not low-density lipoprotein particle size were independent predictors of myocardial infarction in men.

Sources of funding: National Institutes of Health; National Dairy Council; Department of Energy.

For article reprint: Dr. M.J. Stampfer, Channing Laboratory, 181 Longwood Avenue, Boston, MA 02115, USA. FAX 617-525-2008.


Many studies have shown an association between LDL cholesterol levels and CAD and MI. Several case-control studies have also shown an association between small, dense LDL particles and CAD.

The studies by Gardner and Stampfer and their colleagues show that LDL size abnormalities precede the onset of CAD but do not resolve whether small LDL particles are atherogenic or merely markers for high risk. In both studies, the risk for MI was higher among men who had a higher TC:HDL cholesterol ratio and triglyceride level.

The studies differed in some potentially important respects. Gardner and colleagues studied matched pairs of men and women, whereas Stampfer and colleagues studied men (physicians). In the Gardner study, the outcomes were broader (CAD and definite MI vs MI only) and the men had a cardiovascular mortality that was 15% of that of a comparable population of white men. Potentially important methodologic issues were nonstandardization of meal size and composition; timing of meals relative to blood samples; degradation of samples over time; and in the study by Gardner and colleagues, noninclusion of diabetes as a risk factor.

Small LDL particle size may reflect derangement in the metabolism of triglyceride-rich lipoprotein particles. Triglycerides are readily metabolized by most tissues and do not collect in macrophage cytoplasm or in the extracellular lipid pool that forms the core of the atherosclerotic plaque (LaRosa J. Personal communication). This fact may help explain why triglyceride levels were not consistent independent predictors of atherosclerosis when adjustments were made for cholesterol levels and why the role of hypertriglyceridemia as an independent CAD risk factor is still unresolved (1). Hypertriglyceridemia may indirectly promote atherosclerosis by reducing HDL cholesterol levels and increasing the more atherogenic LDL cholesterol levels (e.g., more easily oxidized, smaller, denser particles) (2) or by facilitating a more hypercoagulable state.

Triglyceride levels increase with age (more in men than women) and are more sensitive than other lipoprotein fractions to diet, exercise, and weight change. Observational studies suggest that the triglyceride level may be especially important as a risk factor for CAD in women and the elderly. However, neither of the studies advanced our understanding about these issues in women, probably because of the small number of women included. The small number may partially explain why LDL particle size and levels of triglyceride and HDL cholesterol did not predict CAD in women.

Studies have shown that lowering cholesterol levels by diet or drugs reduces the risk for CAD and may have beneficial effects on carotid arteries (3). No clinical trial designed predominantly to lower triglyceride levels has yet been done. In the Helsinki study (4), the largest decrease in cardiovascular risk was shown in patients who had the highest triglyceride levels and LDL:HDL cholesterol ratios. A study that used bezafibrate in men who had had an MI (5) showed little LDL cholesterol lowering, a 31% reduction in triglyceride levels, and coronary events and angiographic findings comparable with those in the studies that used statins to lower LDL cholesterol (6).

How should clinicians apply the results of these studies? The use of the more readily available tests, such as measurement of plasma triglyceride, TC, HDL and LDL cholesterol, and the TC:HDL cholesterol ratio, is well established. Measurement of other lipid or lipoprotein moieties, such as very-low-density lipoprotein and apolipoprotein B, may be helpful for specific patients. Until measurement of LDL particle size becomes more widely available at reasonable cost and specific treatment for hypertriglyceridemia is better established bytrials, measurement of particle size should probably remain in the research domain. Vigorous modification of CAD risk factors, including treatment of dyslipidemias, remains the cornerstone of patient management.

Marc D. Meissner, MD
Harper HospitalWayne State UniversityDetroit, Michigan, USA

Marc D. Meissner, MD
Harper HospitalWayne State University
Detroit, Michigan, USA


1. NIH Consensus Conference. JAMA. 1993;269:505-10.

2. Grundy SM, Vega GL. Arch Intern Med. 1992;152:28-34.

3. Sacks FM, Pfeffer MA, Braunwald E, eds. A Symposium: Cholesterol-Lowering Trials: New Results and Emerging Issues. Am J Cardiol. 1995;76:1C-126C.

4. Manninen V, Tenkanen L, Koskinen P, et al. Circulation. 1992;85:365-7.

5. Ericsson CG, Hamsten A, Nilsson J, et al. Lancet. 1996;347:849-53.

6. Effect of simvastatin on coronary atheroma: the Multicentre Anti-Atheroma Study (MAAS). Lancet. 1994;334:633-8.