This is the old United Nations University website. Visit the new site at

Table of ContentsNext Page


The blind men and the elephant

The blind men and the elephant

The phenomenon of different blind men describing an elephant in totally different ways depending on the part of his anatomy they touch is a parable with many applications. In the 1970s, a physician working in a country with marasmus due to severe undernutrition but with no kwashiorkor assumed that those in many African and Latin American countries who rarely saw marasmus were missing its importance. The physician believed that protein deficiency was not a serious problem if infants and young children received sufficient calories. Yet at the time clinical kwashiorkor was occurring in many countries due to a dietary deficiency of protein relative to calories. It was usually superimposed on chronic undernutrition and precipitated by episodes of infection. Most of it was appropriately identified as marasmic-kwashiorkor [1]. A few cases of apparently obese infants with the so-called sugar-baby type of kwashiorkor were described from Jamaica. Kwashiorkor has now disappeared from most of the countries in which it was formerly prevalent, but the perception that it was not a serious public health problem in many developing countries during the 1960s was erroneous. Its prevalence is documented in dozens of articles and several books of the era.

Similarly, in the early 1950s, soon after Ancel Keys first published his six-country comparison relating fat consumption to coronary heart disease, I returned to the United States from Guatemala to attend a national medical meeting and heard a presentation categorically denying such a relationship on the basis of a study in a small population of U.S. males whose percentage of dietary calories from fat ranged from about 30% to 45%. In the poor populations with which I was working in Central America, the range was more nearly from 10% to 25%, and deaths from coronary heart disease were not observed. Later the Interamerican Atherosclerosis Project, with serial autopsies of all deaths from 1962 to 1965 in 13 public hospitals in Latin America as well as New Orleans, confirmed the virtual absence of significant aortic and coronary heart disease in Latin American public hospital deaths and a high prevalence in New Orleans [2]. There was a strong association between dietary fat intake and the atherosclerotic index [3].

Similarly, the role of salt in hypertension could not be firmly established by studies in single populations. Analyses of cross-sectional surveys of various populations have generally shown a positive correlation between the intake of sodium and the level of systolic and diastolic blood pressure [4]. These have been confirmed by the Intersalt Study in 52 centres in 32 countries [5]. About 30% of the population is sensitive to salt intakes that are lower than in most industrialized-country diets [6]. In populations consuming less than 4.5 g of salt daily, an age-related rise in blood pressure is slight or absent, and the frequency of hypertension is uniformly low. As sustained intakes of salt rise progressively above 6 g daily, blood pressure rises with age and hypertension is increasingly frequent, but this is difficult to establish within the range of the salt intake of any single population.

The three preceding examples concern major disease issues, but the same risk of misleading extrapolation from studies in a single country applies to any level of observation. The paper in this issue, “The Helen Keller International food-frequency method underestimates vitamin A intake where sustained breastfeeding is common,” reports that even evaluation of a method in three different countries was apparently not sufficient to ensure its applicability to all countries [7]. The authors show that this is the case for the Helen Keller International food-frequency methodology to assess the community risk of vitamin A deficiency. The authors make sensible suggestions for modifying the way in which this simple and innovative method can be adapted for use in a broader range of countries and populations.

Nevin S. Scrimshaw


1. Scrimshaw NS, Behar M. Protein malnutrition in young children. Science 1961;133:2039-7.

2. Strong JP, McGill HC, Tejada C, Holmna RL. The natural history of atherosclerosis. Comparison of the early aortic lesions in New Orleans, Guatemala and Costa Rica. Am J Pathol 1958;344:731-44.

3. Scrimshaw NS, Guzman MA. Diet and atherosclerosis. Lab Invest 1968;18:623-8.

4. National Research Council. Diet and health: implications for reducing chronic disease risk. Washington, DC: National Academy Press, 1989.

5. Intersalt Cooperative Research Group. Intersalt: an international study of electrolyte excretion and blood pressure. Result for a 24 hour urinary sodium and potassium excretion. BMJ 1988;197:319-28.

6. Dahl LK, Heine M, Tassinari L. Effects of chronic excess salt ingestion: evidence that genetic factors play an important role in susceptibility to experimental hypertension. J Exp Med 1962; 115:1173-90.

7. Persson V, Greiner T, Islam S, Gebre-Medhin M. The Helen Keller International food-frequency method underestimates vitamin A intake where sustained breast-feeding is common. Food Nutr Bull 1998;19:342-346.

Top of Page Next Page