Nonfoods as Dietary Supplements

R E Hughes. Cambridge World History of Food. Editor: Kenneth F Kiple & Kriemhild Conee Ornelas, Volume 2, Cambridge University Press, 2000.

Food: A substance (of natural origin) ingested to maintain life and growth.

Diet: The habitual pattern of consumption of food and drink.

Supplement: That which supplies a deficiency or fulfills a need.

The semantically inclined will, no doubt, perceive an element of inconsistency in the title of this contribution. Any food(stuff) ingested for a nutritional purpose is, it could be argued, ipso facto a dietary component. To refer to “nonfood dietary supplements” would, therefore, be meaningless.

On the other hand, foods are often defined in traditional- historical terms, and it is apparent that there are a substantial number of “nutritionally significant” substances which, although not ordinarily components of a diet, may nevertheless be ingested in special circumstances. Whether such “foreign” substances are then described as food(stuffs) or as dietary nonfood (stuffs) is very much a matter of opinion.

The issue is further clouded by a tendency to regard foods as being of natural origin, whereas certain dietary supplements, although having a clearly definable nutritional role, may nevertheless have a “nonnatural” (synthetic) origin. And whereas “true” foods are rarely challenged in terms of potential toxicity, this is not the case with supplements – as evidenced, for example, by the American report on the safety of amino acids used as dietary supplements (Anderson, Fisher, and Raiten 1993).

Again, one must distinguish between nonfoods as dietary supplements and nonfoods as dietary components. Geophagists, picaists, and drug addicts may, in certain circumstances, ingest large amounts of nonfood materials, but these fall outside the scope of this discussion. Supplementation implies that the additional material is introduced intentionally for an avowedly dietary reason and is a substance that could not, in normal circumstances, be supplied by realistic dietary manipulation.

Since the purpose of dietary supplementation is to improve the nutritional status of the subject (thus distinguishing dietary supplements from pharmacological treatments), its practice must be congruous with generally accepted nutritional thought – which, in turn, implies that the intentional use of dietary supplements is a development of fairly recent origin. Consequently, the significance attributed to many nonfood supplements has, in recent years, ebbed and flowed, thereby reflecting the kaleidoscopic nature of orthodox nutritional thought itself.

By the same token, it is equally apparent that the concept of nonfood dietary supplements is a relative one, with the lines of categorization shifting from community to community. Thus dietary fiber, a widely advocated nonfood dietary supplement in the Western European diet, would have no such status in many African communities. Equally difficult to define is the distinction between the use of a supplement in a dietary capacity and its use as a pharmacological agent. Megadoses of ascorbic acid (vitamin C) may, in this respect, be contrasted with the more moderate levels used for orthodox dietary supplementation; and the use of arginine to modify immunity and intestinal carrier systems could, it may be argued, reflect a pharmacological, rather than a nutritional, role for supplementation (Hirst 1993; Park 1993).

By the late nineteenth century, the categorization of foods in functional terms had progressed considerably. Foods were described as “body-building” (nitrogenous) or “energy-forming” (nonnitrogenous), and there was, consequently, a tendency to simplify dietary concepts and to limit precepts to recognizably bona fide members of these two groups. However, by the turn of the century, the recognition of the role of vitamins – nonfoods in a quantitative sense – and the blurring of the lines of demarcation between “energy” and “growth” foods provided a more flexible conceptual framework for the proliferation of ideas about the usefulness of nonfood supplementation.

This chapter therefore deals with substances that, without falling into the conventional categories of dietary components (protein, energy sources, vita-mins, and minerals), are nevertheless believed to enhance dietary effectiveness. They would not, in normal circumstances, be supplied by customary dietary components – either because the foods containing them are ingested in supposedly inadequate amounts (as in the case of dietary fiber in certain communities) or because the substance in question is not readily available from food(stuff) sources. This latter category is, in evolutionary terms, less likely to exist and, it could be argued, implies more of a pharmacological relationship than a nutritional one. Perhaps with this in mind, Michele Sadler, in a recent discussion of dietary supplements, has referred to them as “Functional Foods” and has defined them as lying between foods and pharmaceuticals (1993).

Three examples of “nonfood” dietary supplements are discussed below. All have achieved, at different times, some significance as supplements during the last 40 years or so, and they represent three “etiologically” different categories.


The term “bioflavonoids” has been widely used to refer to those flavonoids that are believed to have pharmacological or nutritional properties. They have no apparently essential role in nutrition and, consequently, no daily requirement can be specified. Nevertheless, for a period of some 30 years they were held to have an “adjuvant” role in maintaining good health – possibly by enhancing the activity of vitamin C.

The flavonoids are a large group of plant compounds based on a C 636 skeleton. They are of widespread distribution in the plant kingdom, being virtually ubiquitous in angiosperms and also existing in more primitive groups, such as green algae, and Bryophyta, including Hepaticae. It has been estimated that in the West, the average per capita daily intake of bioflavonoids is approximately 1 gram (g) (Middleton 1988). Higher animals have, perforce, evolved in an environment in which, of necessity, their feeding habits exposed them to a wide range of flavonoid material, ingested as “secondary” components of foodstuffs. It is, therefore, quite conceivable that this evolutionary exposure to a wide range of flavonoids has elicited physiological and biochemical responses in higher animals. Certainly some bioflavonoid features, such as their ability to chelate with metals and the antioxidant activity associated with hydroxyl groups, indicate a considerable potential for biochemical involvements.

Bioflavonoids first attracted the attention of nutritionists in the 1930s when Hungarian workers reported that certain vegetables and fruits (notably citrus) contained substances capable of enhancing the antiscorbutic properties of ascorbic acid (vitamin C) and even of partially substituting for it. It was claimed that “the age-old beneficial effect of fruit juice is partly due to its vitamin P [sic] content” (Bentsáth, Rusznyák, and Szent-Györgyi 1937: 327); it was also suggested that “citrin,” a flavonoid preparation, could prolong the survival period of scorbutic guinea pigs and that extracts of Citrus limon and Capsicum annuum could correct capillary fragility – a condition characteristic of ascorbic acid deficiency (Bentsáth, Rusznyák, and Szent-Györgyi 1936). The new factor(s) was regarded as separate from vitamin C and was designated “vitamin P” (Rusznyák and Szent-Györgyi 1936), and sometimes was named, primarily by French workers, the “C 2” factor. However, the true nature of the relationship (if any) between the bioflavonoids and vitamin C was unclear, and later work indicated that some of the earlier results were probably attributable to traces of vitamin C in the flavonoid preparations. Nevertheless, by the late 1930s the bioflavonoids had acquired a niche, albeit a minor one, in the annals of vitaminology. But considerable reservations remained about their real nutritional significance (Harris 1938: 28).

In 1949 Harold Scarborough and A. L. Bacharach published an important review article summarizing the work done between 1935 and 1949, and dismissing any suggestion that vitamin P could generally substitute for vitamin C.They centered their attention on the influence of bioflavonoids on capillary resistance and quoted with approval an earlier statement that “guinea pigs placed on a scorbutic diet supplemented with adequate amounts of ascorbic acid show a decline in capillary strength … restorable by vitamin P” (Scarborough and Bacharach 1949: 11).

The consensus of opinion in 1950 was that flavonoids were of physiological significance and that their main influence (on capillary resistance) was mediated independently of vitamin C. Nevertheless, the Joint Committee on Biochemical Nomenclature (United States) recommended in 1950 that the term “vitamin P” be replaced by the designation “bioflavonoids”; and in 1980 the (U.K.) Committee on Dietary Allowances of the Food and Nutrition Board (COMA) indicated that the bioflavonoids should be regarded as pharmacological, rather than as nutritional, agents.

Interest in the possible nutritional significance of bioflavonoids forged ahead during the 1950s (see Table VII.11.1), with a continuing emphasis upon their possible adjuvant relationship with vita-min C. It is interesting to note that almost half of the biof lavonoid papers published during the period from 1948 to 1957 dealt with the specific bioflavonoid rutin (quercetin-3-rutinoside).

Table VII.11.1. Publications during 1930-90 relating to the nutritional significance of (a) bioflavonoids and (b) carnitine Source: Based on papers abstracted in Nutrition Abstracts and Reviews and expressed per 20,000 total nutrition papers; the figures in brackets refer to the “uncorrected” total number of publications on bioflavonoids and carnitine, respectively.

Bioflavonoids Carnitine
1930-9 29 [53] 0
1940-9 51 [102] 0
1950-9 60 [172] 9 [25]
1960-9 35 [132] 9 [35]
1970-9 19 [85] 23 [104]
1980-9 7 [30] 93 [384]

Rutin was widely used at the time in model experimental systems and is still sold as a dietary supplement, sometimes in the form of extracts of buckwheat (Fagopyrum esculentum), one of its principal natural sources. Other specific compounds to receive attention by nutritionists were quercetin and hesperidin.

Work on bioflavonoids was extended in the 1950s and 1960s to include a wide range of supposed physiological and biochemical involvements. Possibly for historical reasons, workers in Eastern European countries were particularly active in this respect. K. Böhm, in his 1968 monograph, cited some 40 definable influences of flavonoids in humans – although the evidence for many of these was weak and sometimes contradictory. M. Gabor, in his 1972 monograph, dealt almost exclusively with one of these suggested areas, namely the supposed anti-inflammatory effect of flavonoids. Other, and more overtly nutritional, areas where bioflavonoids were believed to have a role were hepatic detoxication and lipid utilization (Hughes 1978).

Nonetheless, by the 1960s, interest in a nutritional role for bioflavonoids had peaked, along with their associated use as dietary supplements (see Table VII.11.1). Thereafter the main nutritional interest in bioflavonoids centered on their supposed synergistic relationship with vitamin C and their possible role as antioxidants. There was substantial evidence that bioflavonoids could “enhance” the ascorbic acid status of hypovitaminotic C guinea pigs, and much work in the 1960s and 1970s was aimed at elucidating the nature of this relationship. Discussions centered on whether bioflavonoids could actually substitute for (or synergistically assist) vitamin C in some of its roles, or whether they increased the availability of vitamin C either by protecting it from oxidative breakdown in the tissues or by enhancing its absorption from the gastrointestinal tract (Hughes and Wilson 1977).

Two aspects of flavonoid metabolism attracted some attention – their antioxidant capacity and, in some cases, an apparent mutagenicity. But studies of the nutritional significance of these features gave inconsistent and difficult-to-interpret results. For example, one study dealing with mice indicated that, whereas a dietary supplement of quercetin shortened life span, a flavonoid-rich extract of black currants (containing quercetin together with other flavonoids) extended it (Jones and Hughes 1982). Moreover, none of these studies provided any incontrovertible evidence of an essential role for bioflavonoids in nutrition, and a recent reviewer opined that their function in health and disease as natural biological response modifiers still needed to be determined (Middleton 1988). M. M. Cody, in a contribution to the Handbook of Vitamins, dealt with bioflavonoids under the heading “Substances without vitamin status” (Cody 1984: 578-85).

Nonetheless, there remains a certain amount of residual evidence favoring the use of bioflavonoids as a dietary supplement to enhance the absorption of vitamin C from the gastrointestinal tract and, possibly, its subsequent retention by the tissues (Hughes and Wilson 1977). As a consequence, to advocate the ingestion of “natural” vitamin C (as, for example, in the form of bioflavonoid-rich black currant juice) – rather than the synthetic “tablet” form of the vitamin – by persons otherwise recalcitrant to vitamin C absorption (such as the elderly) is not entirely without experimental basis.

For the decade 1980-9, however, only 30 “nutritionally orientated” bioflavonoid publications could be identified, and general interest in the erstwhile suspected nutritional role of bioflavonoids and in their consequent use in dietary supplementation would now appear to be waning.

The recent COMA report dismissed them as an “unnecessary substance” and made no recommendation for their use as a supplement.

Dietary Fiber

There were many “latent” references to dietary fiber before it was identified and characterized as a specific dietary component. Early writers on diet, such as Thomas Elyot (The Castel of Helth … [1541]), Ludovicus Nonnius (De Re Cibaria [1645]), and Thomas Moffet (Health’s Improvements [1655]), who referred to the laxative property of whole-meal bread (when compared with lower-extraction-rate breads), were in reality commenting on the fiber content.

Thomas Cogan, the Elizabethan dietitian, wrote in his The Haven of Health: “Browne bread made of the coursest of wheate floure, having in it much branne … shortly descendeth from the stomacke [and] … such as have beene used to fine bread, when they have beene costive, by eating browne bread and butter have been made soluble” (1612: 25).

Similarly, Thomas Venner, the Bath physician, wrote of whole-meal bread in his Via Recta ad Vitam Longam that

by reason of some part of the bran which is contained in it, it doth sooner descend and move the belly, for there is a kind of abstersive [laxative] faculty in the bran: wherefore for those that are healthy, and yet subject to costiveness, and also for such as would not wax grosse, it is most profitable (1638: 25).

This comment is curiously congruous with the standpoint adopted by present-day fibrophiles.

Hugh Trowell, in his extensive bibliography Dietary Fibre in Human Nutrition (1979), mentions a paper written in 1919 as the earliest to deal with definable fiber (cellulose) per se. Arthur Rendle-Short, a British surgeon, argued in the 1920s that epidemio-logical evidence suggested a relationship between lack of dietary cellulose and the incidence of appendicitis. The earliest title in Trowell’s bibliography to contain the term “fibre” was one by M.A. Bloom, published in 1930.

It was not until the 1970s, however, that a wide interest emerged in the nutritional significance of what had been regarded, until then, merely as dietary “roughage,” or waste material. Three names are usually associated with this phase in the history of fiber. One is that of T. L. Cleave, a British naval doctor, who argued (1974) that a number of pathological conditions were probably attributable to an increased intake of refined sugar and starch. Although Cleave did not emphasize the mirror image of his thesis – namely, that an increasingly refined diet was in reality a low-fiber one – he has, nevertheless, been widely, and somewhat incorrectly, hailed as a pioneer of current dietary fiber hypotheses.

A more overt and direct link between fiber lack and disease was promulgated in the 1970s by Hugh Trowell (1979; see also 1985) and by Denis Burkitt (1971), and the genesis of modern fiber studies has become almost synonymous with their names. Burkitt himself, in tracing the early history of thought on fiber, has included a fourth name, that of A. R. Walker of South Africa. The number of identifiable publications on the subject of dietary fiber rose from an annual average of 10 in the late 1960s to 125 in the early 1970s; 10 years later (in 1983), the average had risen to well over 500.Trowell’s original thesis (1960) stemmed from his observation that a number of diseases that appeared to be characteristic of affluent Western technological communities were rare or absent in the more “primitive” parts of Africa with which he was familiar. He drew particular attention to the differential incidence of diseases of the gastrointestinal tract and suggested that a high consumption of fiber-rich foods was protective against noninfective diseases of the large bowel. Similarly, and independently, Burkitt (in 1971) published his evidence that dietary fiber might be protective against colorectal cancer.

At that time, however, there was no generally accepted definition of dietary fiber, and much confusion stemmed from the use of values for “crude” fiber (the residue left after serial extraction with acid and alkali). This was a measure that excluded the bulk of the cellulose and hemicellulose material, both important components of dietary fiber. The problem of definition was, of course, inextricably linked to difficulties in the development of a method for the accurate determination of fiber (Englyst and Cummings 1990). Trowell, in 1971, defined dietary fiber as the plant cell material that was resistant to digestion by the gastrointestinal enzymes, and this became the accepted definition. More recently the term “dietary fiber” has been displaced by the resurrected form “nonstarch polysaccharide material,” which reflects current analytic procedures.

The original “dietary fiber hypothesis” was soon extended by other workers to embrace suggested relationships between dietary fiber and nongastrointestinal diseases, particularly cardiovascular disorders (for example, Judd 1985; Kritchevsky 1988). Epidemiological studies indicated a negative correlation between fiber intake and the incidence of cardiovascular disease, and experimental studies showed that certain types of dietary fiber were potent hypocholesterolemic agents. However, relationships of this type posed a problem. By definition, dietary fiber was a substance that did not leave the gastrointestinal tract. How then could one explain its supposed “extragastrointestinal” effects?

Nor was it always easy to distinguish between a direct protective effect of fiber and a displacement effect on “harmful” dietary components. Mechanisms based on a fiber-mediated inhibition of cholesterol absorption or an increased bile acid excretion were among those mooted. But much of the epidemiological work related to fruit and vegetable intake and not to intake of fiber per se. M. L. Burr and P. M. Sweetnam (1982), however, found that in Wales, a vegetarian lifestyle could be correlated with a reduced incidence of heart disease, but fiber intake could not.

The picture was further complicated by the interrelationships now known to exist between dietary fiber and the intestinal flora. This is a complex twoway relationship. Not only are certain types of fiber subject to degradation by colonic bacteria, but there is also increasing evidence that fiber itself may, in turn, modify the nature and metabolic activity of the colonic flora, with consequent modifications to the formation and absorption of a wide range of metabolites. Considerations of this type have led to suggestions that dietary fiber may modify the estrogen status in females, with consequent implications for a protective role for fiber in breast cancer and, possibly, other cancers (Hughes 1990; Adlercreutz 1991).

The current consensus would appear to favor an increased fiber intake in the Western-type diet, but an optimal intake has yet to be clearly delineated. Appeals to primitive dietary patterns are of little help in this respect, as the fiber intake of humans has fluctuated substantially during their stay on earth. Estimates based on coprolite analysis have indicated substantially higher ingestion of dietary fiber by primitive peoples than by their present-day descendants, with daily intakes of 100 to 200 grams occurring at certain periods. In addition, there have been significant changes in fiber intake at critical points in the socioeconomic development of humanity, such as during the “neolithic transition” (Eaton 1990).

Fifteen years after the birth of the fiber hypothesis, Rodney H. Taylor, in a leading article in the British Medical Journal (1984), questioned the prophylactic usefulness of high-fiber diets except for improving colonic function and alleviating constipation. Recent dietary precepts are almost equally noncommittal in their advocacy of dietary fiber. The most recent COMA recommendation (1991) echoed Taylor’s doubts. Although advocating an increase in NSP (non-starch polysaccharides) from the current average intake of 13 g/day to 18 g/day to improve bowel function, the committee was unable to find sufficient evidence for a direct mediating influence of fiber in other suggested areas, such as diabetes and cardiovascular disease. Moreover, the increased intake, it was suggested, should be attained by dietary manipulation, rather than by overt supplementation. It is possible that after a quarter of a century of often frenzied investigative activity, the dietary fiber movement is beginning to disintegrate.


Carnitine was discovered at the beginning of this century when its presence in Liebig’s meat extract – at the time a popular dietar y supplement – was reported. Twenty years later its structure was elucidated, and it was shown to be (beta-hydroxy-gamma-(trimethylamino)-butyric acid. But it attracted little biochemical or nutritional attention until the late 1950s when its role in the metabolism of fats was described. Reviewing the situation at that time, G. Fraenkel and S. Friedman wrote that “a small amount of evidence has been pieced together indicating that carnitine may be active in the metabolism of fats or their derivatives” (1957: 104).

Subsequent, in vivo studies with labeled fatty acids confirmed this conclusion, and within a few years the role of carnitine in stimulating the mitochondrial oxidation of fatty acids was generally accepted (Olson 1966). Subsequent investigations revealed the nature of the mechanism of this stimulation when it was shown that carnitine acts as a transport molecule for the movement of long-chain fatty acid molecules into the mitochondrial matrix. A number of situations in which carnitine has a derived or secondary role as a buffering agent for acyl groups have also been described (Cerretelli and Marconi 1990).

A lack of carnitine, or a reduced or defective activity of one or more of the transport enzymes, would therefore reduce the availability of fat as a source of energy. This could be significant in cases where it is believed that a substantial proportion of the energy metabolism is derived from fat – as in the newly born infant or in cardiac muscle metabolism. It would appear, however, that in the short term, substantial falls in carnitine are required before an impairment of fatty acid oxidation becomes apparent (Carroll, Carter, and Perlman 1987).

The body is able to biosynthesize carnitine, and by the 1980s the biosynthetic pathway had been elucidated. It was shown that the precursor molecules were lysine and methionine, two essential dietary amino acids. The lysine is methylated by the methionine to form the protein-bound trimethyllysine, which is then hydroxylated to form beta-hydroxy-Ntrimethyllysine. This is converted, first to gamma-butyrobetaine and then to carnitine (Rebouche and Paulson 1986).

Carnitine deficiency diseases, resulting from a defect in the biosynthetic pathway, have been described. Such a condition was first reported by A. G. Engel and C. Angelini in 1973, and since then a number of apparently different types of carnitine deficiency have been noted and discussed in the literature. Two basic types are sometimes recognized – systemic carnitine deficiency (characterized by a general reduction of carnitine in the tissues, including the liver) and muscle or myopathic deficiency, in which the reduction in carnitine occurs in the muscles. In such cases, carnitine replacement therapy is a recognized mode of treatment; supplementation must take the form of L-carnitine, the DL/D form being ineffective and, in some cases, further exacerbating the condition.

There are also a number of secondary or noncongenital syndromes that respond to treatment with carnitine, some of which are side effects of other clinical conditions (Rebouche and Paulson 1986; Smith and Dippenaar 1990). Carnitine levels are, in general, lower at birth than in adulthood, and there is evidence that some newly born infants may have a reduced capacity for carnitine biosynthesis. At birth, too, fatty acids become increasingly important as an energy source (Smith and Dippenaar 1990), and it has been suggested that all infants should receive carnitine supplements at least until the end of their first year of life (Olson, Nelson, and Rebouche 1989; Giovannini, Agostoni, and Salari 1991). A recent COMA report accepted that carnitine supplements could be necessary “for low birth weight or preterm infants” (Committee on Medical Aspects of Food Policy 1991: 135). Barbara Bowman, in a recent review, has pointed out that carnitine “appears to be a conditionally essential nutrient in malnutrition and in newborns, pregnant and lactating women, patients receiving dialysis or total parenteral nutrition, and patients with liver disease” (1992: 142).

Many of these examples of carnitine supplementation presumably stem from the correction of an abnormal feature of carnitine metabolism or by restoring a defective carnitine status to normal. Others are more overtly pharmacological than nutritional in nature. A striking example of the pharmacological use of [acetyl]carnitine was the report of a double-blind, randomized, controlled clinical trial in which the progression of Alzheimer’s disease was significantly delayed by the ingestion of 2 grams of acetylcarnitine daily for a year (Spagnoli et al. 1991).

Whether carnitine supplementation has a role in normal or “nonclinical” nutrition is more debatable. Much of the discussion in this respect has centered on the possibility of using carnitine supplements to enhance aerobic power and the capacity for physical exercise (Cerretelli and Marconi 1990). Theoretically, such supplementation could be advantageous in a situation characterized by an increased demand for “physical energy” and, more specifically, in circumstances characterized by (1) an increased use of fatty acids as an energy source, and (2) a depression in the biosynthetic capacity of the body.

A depressed biosynthetic capacity might, in turn, result from a reduced availability of the essential nutrients involved in the biosynthetic pathway. This latter consideration – the influence of other dietary factors on carnitine status – could be of some significance and is central to the concept of carnitine as a conditionally essential nutrient.

The hydroxylation reactions in the formation of carnitine from lysine and methionine require ascorbic acid (vitamin C) as a cofactor; in addition, two members of the vitamin B complex (niacin and pyridoxine) are involved, as is iron. The endogenous formation of carnitine, therefore, involves the participation of six obligatory dietary components – a requirement that places it in a “high risk” category with respect to inadequate supporting diets. A review has underlined this point:

If the exogenous supply of carnitine is temporarily cut off, and provided the subject is not suffering from protein hypo- or malnutrition (as may happen in vegetarianism) plasma carnitine concentrations do not shift … [but] however, drop sharply if the co-factors essential for carnitine synthesis are lacking (Giovannini, Agostoni, and Salari 1991: 88).

Of the essential cofactors, lysine (often the limiting amino acid in poor-quality diets and frequently present in a physiologically unavailable form [Helmut 1989]) and ascorbic acid (vitamin C) are probably the most critically important ones. There is evidence, both experimental and circumstantial, that a reduced availability of one or both of these essential nutrients results in a fall of carnitine and possibly a reduction of carnitine-mediated energy release. Thus, the intake of three compounds – preformed carnitine, lysine, and vitamin C (the three forming the “carnitine base”) – will determine the carnitine status of an individual, and one can point to a number of historically significant situations where a reduced carnitine base has resulted in the emergence of a stage consistent with what we now know to be the physiological consequences of carnitine deficiency: primarily a reduced capacity for sustained muscular exercise (Hughes 1993).

Thus, there are grounds for believing that the fatigue and lassitude, invariably present in the early stages of scurvy, could have resulted from the impairment of carnitine synthesis, which in turn would have resulted from a deficiency of vitamin C (Hughes 1981, 1993; see also Figures VII.11.1 and VII.11.2). It is tempting to speculate that the lowered aptitude for physical labor displayed by potato-eating Irish and by vegetableeating French peoples and commented on by a number of observers in the last century was, at least in part, attributable to a reduced carnitine base (Young 1780; Lewes 1859; Bennet 1877; Williams 1885). A diet that centered almost exclusively on potatoes (as was eaten in Ireland during the nineteenth century) would contain virtually no preformed carnitine. Moreover, 20 pounds of potatoes would have to be eaten to obtain the amount of lysine present in half a pound of meat.

The remarks of George Henry Lewes in 1859 are interestingly pertinent in this respect. Describing the relative capacities of the French and the English for physical work he wrote:

It is worth noting that the popular idea of one Englishman being equal to three Frenchmen, was found by contractors to be tolerably accurate, one Englishman really doing the work of two and a half men; and M. Payen remarks that the consumption of mutton in England is three times as much as that in France.… [B]y giving the Frenchmen as ample a ration of meat as that eaten by the Englishman, the difference was soon reduced to a mere nothing (1859, 1: 174).

Lewes’s reference to mutton was, in this respect, a strikingly apt one, for of all meats and fishes analyzed to date, mutton has by far the highest carnitine concentration (Smith and Dippenaar 1990).

The belief in a positive correlation between animal protein intake and capacity for physical work was a central feature of dietary thought until disproved by the biochemical reductionism of the post-Liebigean era. H. Letheby summarized some of this anecdotal evidence in his Cantor Lectures in 1868:

[T]here is always a relation between the amount of nitrogen contained [in] the food and the labor value of it. Carnivorous animals, for example, are … stronger and more capable of prolonged exertion than herbivores.… The bears of India and America, says Playfair, which feed on acorns, are mild and tractable whilst those of the polar regions, which consume flesh, are savage and untameable. The Peruvians whom Pizarro found in the country at its conquest were gentle and inoffensive in their habits, and they subsisted chiefly on vegetable food; whilst their brethren in Mexico, when found by Cortes, were a warlike and fierce race, feeding for the most part on animal diet.…The Hindoo navvies also who were employed in making the tunnel of the Bhore Ghat Railway, and who had very laborious work to perform, found it impossible to sustain their health on a vegetable diet; and being left at liberty by their caste to eat as they pleased they took the common food of the English navigators, and were then able to work as vigorously (1870: 79).

About seventy years later, Robert McCarrison (also in a series of Cantor Lectures) made almost the same point by comparing the diets and physical capabilities of different Indian races (McCarrison 1944). In general, however, by the beginning of the twentieth century, advances in our knowledge of muscle biochemistry were beginning to undermine the belief in a necessary relationship between animal protein intake and physical activity (Hutchinson 1902: 38).

It is interesting to note that the same biochemical reductionism that dismissed the supposed relationship between activity and a “strong” (animal protein) diet now prompts us to reconsider this anecdotal evidence from the standpoint of “carnitine base” status. Many of the “weak” or “poor” diets referred in this section would almost certainly be found wanting in this respect. By the same token, it has been suggested that certain significant socioeconomic changes could imply a change in carnitine base availability. Thus the “Neolithic transition,” although leading to an improved supply and availability of food, almost certainly resulted in a reduction in dietary quality and, particularly, in that of the “carnitine base” (Cohen 1990; Hughes 1993).

Such considerations are of direct interest vis-à-vis carnitine supplementation. In more general terms, they underline the importance in supplementation studies of considering each situation on its own merits. Any reduction in one or more of the components of the dietary carnitine base could indicate a need for carnitine supplementation. A sudden fall in the animal protein intake or in the availability of the lysine component or in vitamin C intake (as would appear to occur in the institutionalized elderly where the tissue concentrations of vitamin C are significantly below those believed to be functionally desirable in younger subjects) could result in a reduction in endogenously formed carnitine (Hughes 1993).

General Observations

The entire issue of dietary supplementation is clouded by adventitious circumstances and considerations. Thus it will be apparent that certain intellectual environments or passing paradigms of scientific thought can be particularly favorable to the concept of dietary supplementation. The burgeoning interest in bioflavonoids in the late 1940s and 1950s (and in fiber in the 1970s) was probably not unassociated with two acceptable features of nutritional thought at the time: (1) the belief in “subclinical” manifestations of deficiency diseases (which, in some undefined way, were believed to exist despite a normal intake of accepted nutrients); and (2) the conviction that supplementation of foodstuffs with micronutrients was nutritionally appropriate – a belief that, in cases such as the supplementation of low-extraction flour, carried the seal of government approval.

It is true that there have always been those who advocated supplementation for scientifically inadequate reasons. Such persons belong to the same category as those who, also for nonscientific reasons (such as folklore or romantic naturalism), favored brown (whole-meal) bread rather than low-extraction breads (McCance and Widdowson 1956).Where arguments for supplementation stem from external sources of this nature and are, consequently, difficult to accommodate within the current framework of scientific thought, they should always be treated with proper scientific skepticism and subjected to the appropriate scientific scrutiny. If it turns out that the advocacy of a supplement is consonant with current biochemical thought, then the arguments for its use can be that much more convincing.

The three examples outlined in this essay represent, in this respect, cases of dietary supplementation with three quite different origins. The suggested use of bioflavonoids stemmed, in essence, from a mixture of folklore and weak anecdotal evidence, supported originally by unconfirmed laboratory reports, and was enthusiastically embraced by believers in the value of natural foods as contrasted with manufactured ones. There are many other erstwhile dietary supplements that belong to this category – “vitamin B 15” (pangamic acid), “vitamin B 17” (laetrile), sea salt, and a host of herbal preparations. These supplements, for the most part, entered the nutritional field, as it were, from “outside” and for nonscientific reasons. Their alleged efficacy was, consequently, more easily disproved by accepted experimental and statistical techniques, and their use was frequently subjected to criticism and, sometimes, ridicule by scientifically orientated “establishment” nutritionists (see, for example, Bender 1985).

Fiber belongs to a somewhat different category of nonfood supplements, as its appearance as a candidate for dietary status was the consequence of epidemiological studies, although it must be admitted that in a nutritional context, it is not always easy to distinguish between “strong anecdotal evidence” (as presented, for example, by Trowell in his early studies) and statistically acceptable epidemiological evidence. Unlike the bioflavonoids, however, dietary fiber has achieved a foothold in current nutritional thought mainly because of strong correlative evidence coupled with the results of some experimental studies. Dietary intervention studies with fiber have been somewhat less successful, and accommodating the purported advantages of fiber supplementation within the current ambit of biochemical thought still poses considerable problems.

Carnitine represents a third etiologic category of potential dietary supplements. Its emergence as a factor of possible nutritional significance, in contrast to the two other examples discussed, was an “internal” event; it was not thrust upon nutritional thought, as it were, from the outside. Arguments for its acceptance in certain circumstances as a dietary supplement placed no conceptual strain on contemporary nutritional thought. In scientific terms it belongs, therefore, to a more acceptable category than the other two examples discussed. Other putative supplements whose emergence has reflected the current state of the art, rather than external and unrelated circumstances are taurine (which, alongside carnitine, receives conditional acceptance as a supplement in the current COMA report), inositol, para-amino benzoic acid, and specific amino acids. In all such cases of putative dietary supplements, the final verdict must await extensive experimental work and intervention studies.

This brief survey of the recent history of three “nonfood” dietary supplements should serve to illustrate three important and cautionary facts. First, it is useful, as far as possible, to distinguish between a pharmacological role and a nutritional role for supplements. Second, one should exercise caution before accepting claims based on anecdotal evidence or derived from statements or arguments external to (and sometimes in open contradiction to) current scientific thought. Third, and most importantly, dietary supplementation must be defined in terms of lacunae and imbalances in the existing dietary pattern, rather than in terms of absolute requirements – the concept of “conditional essentiality.”

Classical nutrition derived much of its strength (and lately, some of its weaknesses) from generalizations based on an essentially reductionist and unitary approach to dietary components. A necessary condition for its success was the virtual exclusion of any conceptually extraneous matter, such as the possible importance of nonobligatory dietary components, the significance of dietary interactions, and the importance of the changing balance between tissue demands and nutrient availability. There are signs that current nutritional thought is shedding at least some of its traditional absolutism (Hughes 1993: 40-1). Future studies will presumably be designed not so much to prove in absolute terms the usefulness of specific supplements but to define the nutritional circumstances in which such conditionally essential nutrients as carnitine and fiber would be deemed to be necessary.