| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
LETTER |
Retired Medical Inspector of the Italian State Railways
Fortaleza, CE, BRAZIL
E-mail: baschetti{at}baydenet.com.br
It is surprising that Ma and colleagues [1], in assessing the relationships between serum lipids and what they define "dietary carbohydrate factors" (i.e., total carbohydrate intake, percentage of calories from carbohydrates, glycemic index, and glycemic load), failed to consider a central factor that influences blood lipids, namely, the percentage of calories from simple carbohydrates. Had those authors taken this factor into account, their conclusions would have presumably focused on its clinical implications, by remarking that the effects of simple carbohydrates (sugars) on lipid profile can be opposite to the effects produced by complex carbohydrates (starch). Indeed, for example, in patients with high triacylglycerol levels, a diet rich in sugars increased these levels 45.2% [2], whereas they dropped by more than half in similar patients thanks to a diet in which starch represented 
90% of total carbohydrates [3].
Ma and co-workers correctly pointed out that dietary carbohydrates "have a complex relationship with serum lipids, which need to be further elucidated" [1]. Such a desirable elucidation, however, is likely to remain incomplete unless researchers will investigate the unexplored issue regarding the form in which sugars are consumed [4]. The importance of this issue should no longer be overlooked, because there are both theoretical grounds and indirect experimental evidence suggesting that sugars in concentrations exceeding 250 g/L, which would constitute the physiologic limit imposed by evolution [4], provoke unhealthy changes in serum lipids, whereas sugars respecting that evolutionary limit are innocuous [4].
As has rightly been stressed in the editorial "Medicine needs evolution", published lately in Science, it is time to recognize "evolution as a basic science for medicine" [5]. Therefore, that unexplored issue, being largely based on evolution and potentially far-reaching, should be investigated adequately by specific dietary trials [4]. Ma and colleagues might be willing to perform the first of these unprecedented investigations. The results of this pilot study will probably strengthen the growing awareness that an evolutionary approach towards human diseases is enlightening [5–8], because it can "suggest new avenues for conventional research" [9] and can "generate new questions whose answers will help improve human health" [5].
REFERENCES
Division of Preventive and Behavioral Medicine (Y.M., B.C.O., A.R.H., D.E.C., W.L., K.L.)
Division of Cardiovascular Medicine (Y.L., I.S.O.)
University of Massachusetts Medical School
Worcester, Massachusetts
Department of Biostatistics and Epidemiology (J.R.H.)
University of South Carolina
Columbia, South Carolina
E-mail: Yunsheng.Ma{at}umassmed.edu
We thank Dr. Baschetti for his recommendation. Following his suggestion, we examined the association between the consumption of types of sugars (e.g., sucrose, fructose, etc.) including starch and serum lipids using linear mixed models as in our previous analyses [1–3]. Sugar and starch intake data were computed from 24-hour dietary recalls used in our study. Results showed that daily average sucrose intake was 46.7 grams (SD = 20.6), with only 2% of subjects consuming more than 100 grams of sucrose per day, which limits the analyses using the cut-point suggested by Dr. Baschetti for sucrose intake [4]. Cross-sectionally, we found that fructose showed a marginally significant inverse association with triglycerides (p = 0.06); galactose showed a marginally significant inverse association with total cholesterol (p = 0.09), and sucrose was not associated with blood lipids. Longitudinally, starch (p = 0.04) and lactose intake (p = 0.046) were inversely associated with HDL cholesterol levels; sucrose intake was inversely associated with triglyceride levels (p = 0.01); and fructose intake was inversely associated with total cholesterol (p = 0.02). The average triglyceride level in our study was 143 mg/dl (SD = 119), with 18 subjects (3.1%) having levels over 400 mg/dl. Perhaps more associations would be found in a population with high triglyceride levels.
Our output files provide us with total quantity in grams of each type of carbohydrate per person per day, but not adjusted for the form or concentration for which they consumed. This may help to explain the apparent discrepancy in the results differing for foods containing equivalent amounts of sugar. For example, the glycemic response would be quite different for corresponding amounts of fructose if the source is from fruits or vegetables, which would tend to contain high quantities of fiber, in comparison to food additives such as high-fructose corn syrup (fructose/glucose) in soft drinks. The major limitation with classifying carbohydrates as "simple" or "complex" has been the inability of this classification scheme to predict plasma glucose and insulin responses [5]. For example, some starchy foods are digested and absorbed quickly, which means their effect on blood glucose and insulin can be very similar to the effect of ingesting sucrose or glucose; some of these foods, such as white jasmine rice, some potatoes and certain types of white bread, raise blood glucose even more than would an equivalent amount of carbohydrate from pure sucrose [6]. The glycemic index, on the other hand, is a classification scheme that takes into account the physiological response to a given amount of carbohydrate-containing food [7]. This allows people to choose foods that do not cause large postprandial spikes in blood glucose and therefore reduces demand on insulin secretion. While it does make sense that the form and concentration in which a sugar is ingested may be an important determinant of the body’s response to sugar intake (as opposed to just the actual quantity of sugar); we believe that the glycemic index covers these as well as other factors that affect the rate of absorption of sugars from foods.
It would certainly be worthwhile to further investigate the effects of the form (solid or liquid), and concentration in which sugar is ingested. We suggest that this can be done much more efficiently and effectively in a feeding trial rather than in an epidemiological study. This would ensure sufficient contrasts in exposure and would obviate the need to have subjects estimate their dietary intakes, a process associated with a variety of unavoidable errors [8–10].
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
