Obesity is commonly understood to imply excess fat, but it is ordinarily classified according to excess weight. This semantic inconsistency may help to explain why the body mass index (BMI, kg/m²) – a popular marker for relative weight – performs only modestly as a predictor of medical risk [1]. Researchers have increasingly appreciated that adipose tissue has complex functions [2,3], and that these functions may vary according to anatomic region [4-8]. Some of the region-specific functions of adipose tissue are beneficial, such as storing and buffering daily fluxes of circulating lipid fuels [9]. Thus, one cannot assume that a high relative weight or global adiposity is always deleterious.

In the current era of increasing obesity, we should attempt to define and measure lipid accumulation specifically in those contexts where accumulation may represent a physiologic danger [10,11]. These contexts might be described as lipid overaccumulation [12]. At the same time we should avoid attributing culpability to components of adipose or lean tissue that are enlarged but might enhance physiologic processes or reduce the risk of disease.

This paper first describes a simple index for estimating lipid overaccumulation among adults. Next, to demonstrate the utility of the lipid overaccumulation concept, this paper tests the hypothesis that the described index is better correlated than BMI with a variety of cardiovascular risk factors. This hypothesis should not be surprising since the BMI can neither distinguish between fat and lean tissues nor identify the anatomic location or function of distinct fat depots.

The proposed index – designated the "lipid accumulation product" (LAP) – is based on a combination of two measurements that are safe and inexpensive to obtain. One is waist circumference (WC), a measure of truncal fat that includes the visceral (intra-abdominal) depot. The other is the fasting concentration of circulating triglycerides (TG), the esterified, long-chain fatty acids that circulate through blood contained stably inside lipoproteins. Both waist size and TG concentration tend to rise with age [13,14], suggesting that their values are subject to accumulation over time. Waist size and circulating TGs are each continuously associated with metabolic insulin resistance [15,16], daylong triglyceridemia [17], and cardiovascular risk [18-20].

Although previous papers have proposed that the combination of enlarged waist and elevated TGs might serve as a dichotomous risk marker [21,22], the simple index described here was developed to express a continuous risk function. If LAP is better correlated than BMI with cardiovascular risk factors, this finding would support the notion that the overaccumulation of lipid carries worse cardiovascular consequences than the less specific overaccumulation of weight.

The index described in this paper – the lipid accumulation product (LAP) – was developed in an effort to reflect the combined anatomic and physiologic changes associated with lipid overaccumulation in adults. Compared with BMI, LAP exhibited better correlations with lipid risk variables, uric acid concentration, and heart rate, but its correlation with blood pressure was roughly equivalent.

It is reasonable to speculate that the two LAP components – that is, enlarged abdominal fat depots and increased TG concentration – are each an indication that available lipid fuels have exceeded the individual's capacity to buffer and safely store this major form of acquired energy. Prior to 50 years old, the LAP appears to rise more slowly with age for women compared to men (Table 2). The women's relative delay of lipid overaccumulation is consistent with their greater amount of lower-body adipose tissue that confers increased buffering and storage capacity.

Whether an individual's excess lipid fuel appears eventually as an enlarged abdomen or as elevated circulating TG could be dictated in part by genes or by features of the individual's environmental circumstances. A special case, by way of an extreme example, might be the rare individual who is genetically disposed to extremely high TG concentrations (chylomicronemia). For the purpose of risk assessment in the general adult population, however, the alternative manifestations of lipid overaccumulation could be similarly informative. Regardless if the overaccumulation is marked by waist size, by TG concentration, or by both, the calculated value of LAP will be increased. In parallel with the LAP increments, excess lipid material will increasingly be deposited in nonadipose, "ectopic" tissues (e.g., liver, skeletal muscle, heart, blood vessels, kidneys, and pancreas) where it may adversely modify cellular metabolism, accelerate apoptosis (cell death), and interfere with cardiovascular control [10,11,31].

Ectopic lipid deposition is difficult to quantify directly, but an increased LAP value may indicate that various tissues or organs have become more vulnerable to injury from lipid overaccumulation. Lipoprotein particles with small diameters are more associated with disease risk than those with large diameters [32,33], and lipoprotein particle diameter is inversely associated with abdominal size [34]. Thus, LAP may effectively express disease risk through the hyperbolic relationship between LAP and its two component variables (Figure 2). The presence of an enlarged waist (implying a small lipoprotein particle size) allows the LAP value to increase rapidly with each unit increase in TG concentration. By contrast, the presence of a small waist (implying a large lipoprotein particle size) allows the LAP value to increase slowly with each unit increase in TG concentration. Although the NHANES III data set contains no direct measurement of particle sizes, our indirect estimate of LDL particle size (LDL cholesterol/ApoB) confirms that a small LDL particle size is better correlated with LAP than with BMI (Figure 3 and Table 4).

In contrast to an elevation in LAP value, an elevated BMI value (i.e., relative weight) is less specific in its anatomic or physiologic implications. Increased weight might represent enhancement of lean tissues, enlargement of the protective, subcutaneous adipose depots in the lower extremities [5,35-37], or systemic overload of fluid – changes that could be either salutary or simply secondary consequences of other disease processes. The commonly observed association of fluid overload with hypertension may explain the instances in which this study found blood pressure to be marginally better correlated with BMI than with LAP (Table 4 and Figure 3).

In order for LAP to gain a useful role in clinical medicine or epidemiology, at least three major questions remain to be addressed:

1. Is LAP strongly predictive of major disease outcomes? This preliminary analysis examined only intermediary outcomes (risk factors), and its data are entirely cross-sectional. However, a 20-year followup study of Swedish women reported that abdominal adiposity and elevated TG concentration were associated with increased risks of death from myocardial infarction and all causes, but that elevated BMI and cholesterol concentration were much weaker predictors [38]. In another prospective study from Scandinavia, post-menopausal women followed for 8.5 years demonstrated that the baseline combination of enlarged waist with elevated triglycerides – a dichotomous marker – was a very strong predictor of all-cause and cardiovascular mortality as well as the annual progression rate of aortic calcification [39]. More data collected prospectively in a variety of populations, including men, would help confirm that LAP – a continuous marker – has advantages over other simple indices for predicting the incidence of major diseases and mortality.

2. Is it useful to monitor LAP values as an indicator of intervention effectiveness? A recent report of intentional weight loss among overweight and obese Japanese women found that their exercise regimen and low-calorie diet were rewarded with a 37 percent reduction in TG concentration and a 27 percent reduction in truncal fat but only a 2 percent reduction in leg fat and a 12 percent reduction in BMI [40]. The authors also reported a direct correlation between changes in truncal fat and changes in fasting TG concentration or the number of heart disease risk factors, but that the changes in leg fat were inversely correlated with the changes in TG concentration or the number of heart disease risk factors. In another study, Italian men and women with diabetes were followed for two years after randomization to a physical activity counselling intervention [41]. Across six levels of aerobic energy expenditure, those who exercised more experienced significant reductions (p < 0.001) in waist circumference and circulating TG, but no reductions (p > 0.25) in either weight or BMI. Thus, the participants in both of these studies would have achieved a substantial reduction in LAP in association with improved cardiovascular risk factors, but their reduction in BMI was modest and less clearly associated with cardiovascular benefit. These observations from Asia and Europe demonstrate a potential advantage to using LAP as an intermediary variable by which to assess interventions against obesity-related risk.

3. Is LAP a practical index for adoption by clinicians or epidemiologists? Standardized waist measurements are highly reproducible [42], and they are arguably simpler and less expensive to obtain than well standardized weights and heights. The requirement of a venipuncture, however, could be a major obstacle for many potential participants or patients. The need for the fasting state could also represent a major inconvenience, although many persons accept the fasting condition when necessary for assessment of their glucose or lipid status. Of special importance in less developed economies, the laboratory cost of a single assay for TGs would be low compared with the costs of multiple assays for lipoproteins and an extensive biochemical panel.

The cross-sectional associations with LAP demonstrated in this paper should be seen primarily as a demonstration of how the concept of lipid overaccumulation may be expressed in an adult population. The utility of LAP for research or as a practical tool for use in the community will depend on the degree to which LAP can be demonstrated to enhance prediction of disease incidence. Prospective data sets that include baseline information on WC and fasting TG concentration would be well suited to evaluate LAP as a predictor of cardiovascular outcomes and mortality.

ApoA1, apolipoprotein A1

ApoB, apolipoprotein B

BMI, body mass index

HDL, high-density lipoprotein

LAP, lipid accumulation product

LDL, low density lipoprotein

NHANES III, third National Health and Nutrition Examination Survey

The author(s) declares that he has no competing interests.

HK conceived of this study, performed the calculations, and drafted the manuscript.

The pre-publication history for this paper can be accessed here: