Carbohydrates and diabetes
The role of dietary factors in the aetiology of non-insulin dependent diabetes (NIDDM) has been the subject of many epidemiological studies. It is important to emphasize that there are major difficulties in assessing nutritional aetiologies of any chronic disease. In addition to the problems inherent in various epidemiological approaches (ecological studies, case control and cohort studies), the instruments for measuring intake of food and nutrients (diet records, 24 hour recalls and food frequency questionnaires) all have limitations. Feeding studies in animals and humans may help to confirm or refute observations made in epidemiological studies.
Carbohydrates and dietary fibre in the aetiology of diabetes
The suggestion that refined carbohydrates, and sugars in particular, might be involved in the aetiology of NIDDM dates back to the writings of early Indian physicians. Over 40 studies have examined the role of sugars in the aetiology of NIDDM, with about half suggesting a positive association and a comparable number suggesting no association. Some have even suggested an inverse association between diabetes incidence and sucrose intake (111). Further evidence to suggest that sucrose is not an important contributing factor in the aetiology of NIDDM comes from carefully controlled studies in people with NIDDM (112). Isoenergetic substitution of moderate amounts of sucrose in the diets of individuals participating in randomized cross-over experiments do not result in deterioration in glycemic control. There is similarly little evidence that readily digested starchy foods increase the risk of developing NIDDM.
On the other hand, there is rather more support for the suggestion that foods rich in slowly digested or resistant starch or high in soluble dietary fibre might be protective. Countries with high intakes of these foods have low rates of diabetes and Trowell drew attention to the fact that the reduced mortality rates for diabetes during and after the Second World War paralleled the increased intake of dietary fibre during that period (113). These observations on their own provide no more evidence for a protective role for these foods than do comparable studies suggesting a causal role of sucrose. However, there is some corroborative evidence for a protective role of dietary fibre (non-starch polysaccharide) and slowly absorbed or resistant starch and the foods which are rich in these nutrients.
One study has shown that consumption of legumes rich in soluble dietary fibre was inversely associated with risk of glucose intolerance (114). Experimental studies provide further confirmation. In controlled experiments, diets high in soluble fibre-rich foods (115) or foods with a low glycemic index are associated with improved diurnal blood glucose profiles as well as long term overall improvement in glycemic control as evidenced by reduced levels of glycated haemoglobin (116). Some other studies provide indirect support for this hypothesis. Diabetes risk appears to be lower in vegetarians than in those who are not vegetarians (117). The diet of vegetarians is characterised by a high intake of dietary fibre, but differs in other ways from that of non-vegetarians. In addition to not eating meat and animal products, vegetarians also have less saturated fat, more polyunsaturated fat and a diet which differs in micronutrient composition when compared with non-vegetarians.
Not all studies have been able to confirm the protective effect of dietary fibre. For example, in two studies, no associations were found between intakes of carbohydrate or fibre and the risk of diabetes. Despite these negative findings (118) the overall weight of evidence suggests that individuals who consume diets rich in soluble dietary fibre or which have a low glycemic index are likely to be at reduced risk of developing diabetes. Conclusive evidence of a causal association is unlikely to be forthcoming.
Carbohydrates in the treatment of diabetes
Diabetic dietary recommendations - general principles
Before the discovery of insulin in the 1920s, radical restriction of dietary carbohydrate was the cornerstone of diabetes treatment. However, even after widespread availability of insulin, people with insulin-dependent diabetes mellitus (IDDM) were recommended to consume diets in which carbohydrate provided less that 40% total energy. It is difficult to establish the reason for this advice since there is no experimental data proving the benefits of carbohydrate restriction for people who are adequately insulinised. Carbohydrate restriction will lead to improvement in glycemic control in overweight NIDDM patients if such restriction is accompanied by weight loss but these observations provide no scientific justification for the value of restricting carbohydrate-containing foods. Dietary recommendations for the management of diabetes have been made in most countries, but those of the American Diabetes Association and European Association for the Study of Diabetes have been particularly widely quoted (119,120). The two sets of recommendations are in broad agreement. (See Table 10 for carbohydrate dietary recommendations for diabetes).
Carbohydrate and dietary fibre intake for diabetes
In the 1970s a series of studies from various research groups showed that a high carbohydrate diet (up to 60% total energy from carbohydrate) was associated with improved glycemic control and reduced levels of LDL cholesterol when compared with a low carbohydrate diet (40% total energy). These findings, together with the observation that people with diabetes in societies which traditionally consume a high carbohydrate diet have low rates of ischaemic heart disease (IHD), led some official diabetes organizations to recommend a change from the traditional low carbohydrate to a high carbohydrate diet.
Sugars
Avoidance of 'simple sugars', especially sucrose has undoubtedly been the most widely recommended component of the diabetic dietary recommendations. This is based on the incorrect assumption, however, that sugars will aggravate hyperglycemia to a greater extent than other carbohydrates. Indeed there is evidence to the contrary. It has long been known that blood glucose levels following a sucrose or fructose load are lower than after a comparable oral glucose load or starchy foods containing a similar amount of carbohydrate (121). This observation led to longer term studies in which diets including sucrose within mixed meals were compared with sucrose-free diets. The result of such studies confirmed that moderate intakes of sucrose (up to about 50g daily) can be incorporated into the diets of people with diabetes provided the sucrose is consumed as part of a meal and does not displace fibre-rich carbohydrate (112). These findings apply to those with both insulin-dependent and non-insulin-dependent diabetes.
TABLE 10 Some dietary recommendations for diabetes
Components of dietary energy |
- Saturated fatty acids, <10% of total energy |
- w -6 polyunsaturated fatty acids, <10% of total energy |
|
- Protein, 10-20% of total energy |
|
- Carbohydrate and cis-monounsaturated fatty acids, for the remainder |
|
Carbohydrate issues |
- Low glycemic index foods and those rich in soluble fibre recommended |
- Vegetables, fruits, pulses and cereal-derived foods preferred |
|
- Sucrose, <10% total energy acceptable in certain circumstances |
|
- Timing of intake essential for those on insulin |
|
Special 'diabetic' and 'dietetic' foods |
- Non-alcoholic beverages sweetened with non-nutritive sweeteners are useful |
- Other special foods not encouraged |
|
- No particular need of fructose and other 'special' nutritive sweeteners over sucrose |
General advice regarding carbohydrate intake
While some foods have predictable and consistent glycemic indices, others will vary from country to country or indeed sometimes widely within a single country. Locally determined information is therefore essential. However, with regard to carbohydrate-containing foods one may in summary say that the staple cereals, breads, pasta, pulses, vegetables and fruit are generally appropriate sources of carbohydrate for people with diabetes. Such foods should form a major component of all meals and snacks. They should not adversely influence blood glucose levels in the short term, and may help to achieve optimum glycemic control in the long term. In addition, they are usually good sources of a wide range of essential macronutrients. Fibre-depleted starchy foods, foods rich in simple sugars, sucrose and other added sugars need not necessarily be totally excluded but should generally be restricted.
Carbohydrates and cardiovascular disease
History and epidemiology of carbohydrates and cardiovascular disease (CVD)
The hunter-gatherer diet had in excess of 50% of its energy from carbohydrate (122) and cardiovascular disease as we know it today, with underlying atherosclerotic vascular disease, was probably non-existent. Not all of this carbohydrate was digestible. As much as 50% of the energy intake for indigenous Australians (Aborigines or Kooris) living on the Victorian Plains was derived from fructo-oligosaccharides of the inulin type, from a tuberous plant known as Murnong (123,124). This was indigestible and fermented in the colon. Prior to European settlement in Australia, atherosclerotic vascular disease was unknown, yet life expectancies at birth were over 60 years. The question is how much of this cardiovascular protection was attributable to nutrition. Furthermore, how much of this was related to the amount and type of carbohydrate intake?
Possible effects of dietary carbohydrate on cardiovascular disease
There are a variety of ways in which a high carbohydrate diet might be protective of cardiovascular disease risk:
1. Maintenance of insulin sensitivity - especially in the basal state. High carbohydrate diets tend to lower basal (fasting) glucose and insulin over several days (125). In turn, this decreases risk factors (hyperglycemia and hyperinsulinaemia) for cardiovascular disease.2. Fermentable carbohydrate in the colon produces absorbable SCFA (short chain fatty acids), with potential regulation of hepatic gluconeogenesis and insulin handling (126). Effects on lipoprotein metabolism are also in evidence (127).
3. Providing companion dietary compounds (micronutrients and phytochemicals) which tend to be protective of the cardiovascular system (128).
4. Displacement of nutritionally disadvantageous components of the diet (e.g. saturated animal fat.
5. Increasing satiety and decreasing the energy density of the diet, making obesity less likely (129-131).
Optimising intake of carbohydrates for CVD protection
Food sources
The food source of carbohydrate may have a bearing on the particular aspect of cardiovascular risk being addressed. Examples of the potential changes in risk with increased intake of particular food carbohydrate sources are shown in Table 11 (arrows show direction of change). Interestingly, whatever the specific effects on cardiovascular risk factors, the sum total of effect on cardiovascular events of increased plant food intake is favourable.
Patterns of eating
If carbohydrate-containing foods are low in fat and consumed as snacks throughout the day, cardiovascular risk is lower (132). More recent reviews of feeding patterns and BMI (body mass index) raise questions about the adequacy of food intake methodology to capture snacking information reliably (133).
Staple foods
One of the major cultural, economic and ecosystem issues relating to food which confronts public health policy-makers is how much carbohydrate should be provided by a staple crop, and how much should come from various other food sources. Populations with a carbohydrate staple such as potatoes, maize, rice or wheat tend to have low CVD rates (134).
Food security, however, would indicate that over-dependence on one crop is not desirable. There is a strong case for food variety to improve cardiovascular risk profiles (135-137).
TABLE 11 Effect of food carbohydrates on cardiovascular risk factors
|
Body fatness |
Lipoproteins, triglycerides |
Blood pressure |
Glycemic status |
Thrombosis |
Antioxidant status |
CV events |
Cereal |
?¯ |
® |
? |
¯ |
? |
? |
¯ |
Fruits (not fruit juice) |
?¯ |
® |
|
® ¯ |
? |
¯ |
|
Vegetable |
¯ |
¯ |
¯ |
® ¯ |
|
¯ |
¯ |
Legumes |
?¯ |
¯ |
|
¯ |
? |
? |
¯ |
Nuts |
® |
¯ |
? |
|
? |
¯ |
¯ |
Combination * |
¯ |
¯ |
¯ |
¯ |
¯ |
¯ |
¯ |
* The combined effects of these foods is best judged by indices of total plant food intake or variety
Carbohydrates and cancer
Introduction
Interest in the role of carbohydrate in the aetiology of human cancer has been fuelled in recent decades by the debate about the role of dietary fibre in colorectal carcinogenesis. Carbohydrates include sugars and oligosaccharides, starches and non-starch polysaccharides (NSP). In epidemiological studies it is often difficult to distinguish between the effect of the sugars and starches and the role of total energy intake and the associated overweight. An additional problem has been that consumption of a diet rich in, for example, root vegetables, is often associated with low income and poor variety in the diet. Thus there is a range of confounding factors which often make interpretation of epidemiological data difficult.
Carbohydrate, unlike protein or some fats, does not yield potent carcinogens during cooking or storage, and indeed most of the discussion about the role of carbohydrate in human carcinogenesis has concerned cancer prevention in the large bowel. However, there is a clear and indisputable relationship between being overweight and cancers at a number of body sites (138) and, of course, a relationship between consumption of excess energy and being overweight. In those situations where carbohydrate intake has been positively correlated with cancer risk it has been usual to assume that the carbohydrate intake is simply a surrogate measure of excess energy intake resulting in being overweight. There has been much more interest in the possible protective role of carbohydrate, in particular against colorectal cancer. Starting from the premise that colorectal cancer is caused by a luminal carcinogen or promoter formed by bacterial action on some benign substrate (139), it was postulated that dietary fibre protects by:
a. Modifying the colonic bacterial flora to one less likely to produce toxic metabolites;b. Being itself fermented to yield an environment in the colon less conducive to bacterial production of carcinogens/promoters;
c. Causing stool bulking, thereby decreasing the concentration of luminal carcinogens or promoters; and
d. Speeding the rate of transit of the colonic contents, allowing less time for carcinogens or promoters to act.
Diet has a profound effect on the flora of the caecum (140). Increased dietary fibre results in a non-specific increase in most components of the gut bacterial flora because the enzymes responsible for the breakdown of the macromolecules are largely extracellular (141) and so the released products are available as nutrients to the whole flora. It is not clear, therefore, how this general increase in bacterial population density would decrease the rate of carcinogen production. More recently, non-digestible oligosaccharides like inulin and its hydrolysate oligofructose, have been shown to selectively stimulate the growth of colonic bifidobacteria, opening the way to selectively and significantly modify the composition of the colonic microbiota (142). The major products of bacterial fermentation of carbohydrate are the short chain fatty acids (SCFA), and these acidify the caecum (143). Most of the bacterial enzymes responsible for the production of carcinogens/mutagens have pH optima of 7 or above (138) and so acidification of the caecal lumen would decrease the toxicity of luminal contents to the gut mucosa (144). There is very strong support that dietary fibre protects against colorectal cancer by causing stool bulking, thus decreasing luminal carcinogens or promoters. The greatest effect is seen with wheat bran, and this has been documented by many groups in many countries.
Sugars and oligosaccharides
Sucrose
Since sucrose intake is a significant contributor to total energy intake, it is a standard feature of diet surveys and so there has been a lot of information gathered on its relation to human cancer risk. In general this has shown remarkably little evidence for such a relationship (145). Since sucrose is readily digested, it is a contributor to total energy intake and so might be expected to be associated with those cancers associated with high energy intake, such as colon cancer and the hormone-related cancers, but such an association has not been observed (146).
Lactose
The vast majority of northern Europeans and their descendants worldwide, retain their small bowel lactase throughout their life. Dietary lactose would be well-digested in their small bowel and so would be expected to behave epidemiologically like sucrose. However, the majority of the adult population of the rest of the world is lactose intolerant, having lost their small intestinal lactase during childhood (147). The populations where a high percentage of adults are lactose-tolerant are the only ones in which large bowel cancer is common.
In lactose-intolerant persons, the lactose reaches the large bowel and is fermented to short chain fatty acids (SCFA), thereby contributing to caecal acidification and laxation, and having an effect similar to that of dietary fibre. If dietary fibre is protective against large bowel carcinogenesis then lactose would be expected to also be protective of lactose-intolerant populations (148).
Oligosaccharides
There has been considerable interest in the potential protective effects of fructo-oligosaccharides which are present in some plant foods and are also available commercially. These oligosaccharides are thought to modify the caecal bacterial flora (increasing the numbers of lactobacilli and bifidobacteria, and decreasing the numbers of bacteroides and clostridia) and to acidify the caecum (as a result of rapid fermentation to short chain fatty acids), thereby causing a decreased rate of production of luminal carcinogens/promoters. There is at present, however, only preliminary evidence available which does not permit final evaluation of this interesting hypothesis.
Starches
There have been some studies (149) on the relation between starch intake and colorectal, gastric and breast cancers. Starchy foods in general are associated with an increased risk of these cancers but the correlations may be secondary to other factors. One study of breast cancer indicated that starch was a risk factor in a northern Italian population (150) but considered that starch might only be a marker for total energy intake (which has long been known to be a risk factor for breast cancer). A large study of 23 populations found that starchy foods such as potatoes were inversely correlated with risk of cancers of the colon, breast, prostate and endometrium (146).
Dietary fibre
Burkitt proposed a hypothesis that 'dietary fibre' is a major contributor to protection against colorectal carcinogenesis (139). For the purposes of that hypothesis dietary fibre was defined as the dietary carbohydrate that reaches the large bowel. In the 1970s there were no good assays for dietary fibre and so epidemiologists used intakes of fibre-rich foods as a surrogate measure and obtained results in population studies that showed a very strong protection.
These results were treated as hypothesis-supporting and were followed by a host of case-control studies, which tended to give less convincing support. A major problem with all case-control studies of colorectal cancer is the lack of specificity of the symptoms, which leads to a lag between the onset of symptoms and the actual diagnosis of cancer. Since the presence of symptoms during this lag is likely to cause an insidious change in the diet, it is necessary to use diet recall methods to determine the pre-symptoms diet. Such methods are notoriously inaccurate. This problem was obviated in a prospective study of U.S. nurses, which observed a protective effect of fibre (151).
A further problem with these studies, however, was the method used to determine dietary fibre intake. On the assumption that no starch reaches the colon, methods have been developed to assay non-starch polysaccharides (NSP) and this has been used as a measure of dietary fibre. However, from measures of daily stool weight, we can calculate the amount of carbohydrate that must reach the colon simply to feed the bacteria that make up more than 25% of faeces (152). From this it has been concluded that at least 60-70g carbohydrate reaches the colon per day (153), as compared to the 12-15 g NSP per day used in case-control studies. A proportion of dietary starch is undigested during small bowel transit and so reaches the colon in addition to the NSP (152). Indeed there is probably far more starch than NSP reaching the colon.
In examining whole grain intake in 15 case-control studies, a strong protective effect was observed that was not seen when refined grain was studied (154). A review of 58 case control or cohort studies found a similar strong protective effect for cereals and cereal fibre, as shown in Table 12.
TABLE 12 Summary of a review of 58 studies of diet and colorectal cancer risk
Number of papers |
Effect on colorectal cancer risk |
|||
Protection |
No effect |
Promotion |
||
Cereal intake |
36 |
23 |
9 |
4 |
Cereal fibre |
16 |
13 |
3 |
0 |
These data strongly suggest that cereals, particularly wheat bran, protect against colorectal carcinogenesis. In support of this conclusion, one study (155) demonstrated that diet intervention to increase intake of wheat bran and decrease fat, although it did not affect the rate of colorectal adenoma formation, it prevented their growth to more than 1 cm in diameter. This is the key step in colorectal carcinogenesis (148). Further, a U.S. cohort study (where most of the cereal would have been wheat) showed a strong dose-response effect in the protection by cereal intake (156).
Summary
There is little evidence of any significant correlation between intake of mono-, di- and oligosaccharides and cancer at any site that could not be explained by total energy intake.
There is stronger evidence of a positive correlation between intake of starch or refined carbohydrate and risk of cancers of the colon and breast. This may be because starch intake is a surrogate for total energy intake. There is no good hypothesis to explain why the products of starch digestion (which would be absorbed and delivered to the colon by the vascular route) should be organotropic carcinogens for the colon and breast. Indeed if butyrate has its anticarcinogenic effects (157,158) in vivo as well as in vitro then a protective effect would have been more likely.
There is still dispute about the protective effect of dietary fibre against colorectal cancer. Since the assay of dietary fibre appears to be very inaccurate it is remarkable that any effect has been seen at all. When better surrogate measures of dietary fibre are used there is much stronger evidence of protection. The data suggest a protective effect for whole grain cereals, particularly wheat (which is much richer in fibre and which has the greatest stool bulking effect).
Carbohydrates and dental caries
Introduction
Dental caries is one of the most widespread oral diseases. Its causation has been associated with food carbohydrates since the beginning of scientific approaches to dental problems. It is still true that eating food containing carbohydrates is a risk factor. Research, especially epidemiology, during the last 25 years, however, has markedly modified current views.
Dental caries - principle of development of lesions
In 1962, three principal factors were defined, all of which are required for the development of carious lesions (177):
1. Microorganisms in the mouth with the potential to form acid from carbohydrates.2. Substrates locally available in the mouth as a source of energy for the metabolism of the acidogenic oral bacteria.
3. Host properties. The human being is the host for oral microorganisms and the relevant host properties (besides presence of teeth to be colonized by bacteria) are susceptibility to chemo-bacterial attack and protective resistance/defence, as well as the potential for regeneration.
Microorganisms
It is important to realize that microorganisms, such as bacteria, are normal inhabitants of the mouth. Oral bacteria are harmless in thin layers, and are cariogenic only in thick, organized and undisturbed "plaque" on the teeth. This had been shown by measurement of acid formation on the teeth after rinsing with sugar solutions (178). Immediate drops in pH were found when thick plaque was present, whereas after removal of the plaque no acidity could be detected on the cleaned surface. The same phenomenon was reported from a number of laboratories where pH telemetry is used to measure the acidogenic potential of carbohydrate-containing foodstuffs: if the experimenter does not let plaque grow for at least 2 days, no acid formation can be observed (179).
Substrates
In the aetiology of caries, the sources of substrates for cariogenic bacteria are mainly foodstuffs and drinks containing sugars and other carbohydrates. This utilization of some of the food by bacteria is a local side effect in the mouth during food passage, in contrast to the systemic effect of carbohydrates as a source of energy for the host.
It is the result of the side effect, namely acid formation in the bacterial plaque on the teeth, which causes demineralization. The carbohydrates consumed exert no direct damaging effect on the teeth. During sleep and when no food is available, the acidogenic plaque bacteria can metabolize and survive on a minimum supply of substrate derived from carbohydrate sidechains of salivary mucins. At these low substrate concentrations, no cariogenic amounts of acid are formed.
Host factors/properties
Although extremely hard and therefore quite resistant to wear, the enamel covering the tooth crowns is slightly soluble in acids. This fact is the essence of the risks threatening dental health, because eating of food with a certain degree of acidity, as well as drinking of acidic fruit juices and other acidic beverages occurs daily at meals and between meals. Moreover, since a normal diet is rich in carbohydrates, additional acids can be formed if bacterial plaque is present on the teeth.
One way of minimizing these risks from the side of the host is to minimize the frequency of eating and drinking. Other possibilities are to adopt a good oral hygiene habit, to try to increase resistance of teeth against acid, and to improve the repair mechanisms which are provided naturally by secretion of saliva. Saliva contains the buffering bicarbonate system which at least in part can neutralize damaging acids. Moreover, saliva is a solution of calcium and phosphate which at neutrality is oversaturated with respect to enamel apatite. This results in automatic remineralization when, after an acid attack, the pH returns to normal. Such a "normal" remineralization process is very slow, however, and if demineralization is stronger and lasts longer than the time for remineralization, then carious destruction of teeth occurs.
At this point the action of fluoride ions comes in. The presence of fluoride ions protects tooth enamel mineral from becoming demineralized. This protection is only partial, but fluoride ions have a second important capacity - they speed up and improve remineralization and repair. This process requires lifelong frequent tooth-fluoride contacts to maintain dental health. Fluoride sources may be drinking water or tea, or use of a fluoride-containing toothpaste twice a day.
Dietary sugars and caries prevalence
An association between intake of sugars and dental caries was first studied experimentally in the early 1950s on inmates of the Vipeholm asylum in Sweden (180). The Vipeholm Study was the first to reveal the distinction between the effects of amount of sugars eaten, versus the frequency of sugar intake. The experiment showed that restriction of sugar intake to four main meals daily did not significantly increase the caries activity, even if large amounts of sugar (300 g/day) were given. On the other hand, when 8 or 24 between-meal sugar-containing snacks were given daily, caries incidence rose dramatically.
Sugars do not give rise to production of dangerous amounts of acid in the oral cavity when plaque is absent or only present in thin layers. Therefore, it is feasible to separate modern epidemiological research from early findings accumulated in the pre-dental hygiene era. This is the more important because during the same time span in which oral hygiene practice developed, the sale of toothpastes containing fluoride in caries-inhibiting concentrations rose up to more than 95% in many countries.
It is obvious that while dietary sugars are a determinant in the development of caries, they are not the most important factor in the aetiology of the disease. Studies done with groups of children have indicated that frequent consumption of candy did not seem to be a significant determinant of caries, but instead, the oral hygiene status appeared to be the more important caries risk factor (181,182). The Netherlands is one of the developed countries where caries prevalence within the last 25 years has decreased rapidly, although sugar consumption is still more than 90% of what it was in 1965. In Sweden, Norway and New Zealand, sugar consumption between 1982 and 1985 increased, but nevertheless, regular epidemiological monitoring of caries data showed that the caries prevalence in children continued to decrease.
Summary
Although a relationship between sugars and dental caries is accepted by all clinicians and researchers working in the dental field, the degree of emphasis on the importance of this factor in prevention and control of the disease varies. The information we have available today, based on studies into the situation in the 1990s, should allow for a more scientific and rational approach to the role of fermentable carbohydrates in dental caries. If one intended to advise in principle against consumption of all cariogenic food, it would be irrational not to advise, for instance, against consumption of milk and fruit, but only against consumption of sucrose. The recommendation of a varied diet, oral hygiene and fluoride use seems to be the better alternative.
The most important observation emerging from recent epidemiologic studies is that more and more populations are characterized by a decreasing caries prevalence in the young generation, mostly independent from intake of sugars and other carbohydrates. A basic "personal prevention package" of oral hygiene habits - cleaning with a toothbrush and using fluoride toothpaste - is probably sufficient to keep 75 per cent of adolescents caries free. In short, dental health problems do not require any dietary recommendations in addition to, or other than, those required for maintenance of general health.