9.6 Endocrine and metabolic disorders
Oxford Textbook of Public Health
Endocrine and metabolic disorders
Basil S. Hetzel, Paul Zimmet, and Ego Seeman
Iodine deficiency disorders
The ecology of iodine deficiency
The physiology of iodine deficiency
The iodine deficiency disorders
Magnitude of the problem
Assessment of iodine status
The correction of iodine deficiency
The support of the United Nations system
The elimination of IDD at country level
The global partnership
The epidemiology of diabetes mellitus
Impact and cost of diabetes to society
Classification of diabetes mellitus—the new recommendations
New diagnostic criteria for diabetes
The epidemiology of diabetes mellitus
Complications of diabetes
Primary prevention of diabetes mellitus
Osteoporosis: a public health problem
Bone density—a surrogate for bone strength
Changes in bone density during growth and advancing age
Classification and causes of osteoporosis
Endocrine and metabolic disorders present major public health problems in many parts of the world, including mental retardation resulting from iodine deficiency, diabetic vascular disease, and fractures in the elderly resulting from osteoporosis and calcium loss. Many of these endocrine and metabolic disorders are readily amenable to prevention by public health measures. This chapter provides examples of three major endocrine and metabolic disorders of public health significance.
Iodine deficiency disorders
Basil S. Hetzel
Iodine is an essential constituent of the thyroid hormones thyroxine or 3,5,3′,5′ tetra-iodothyronine (T4), and 3,5,3′ tri-iodothyronine (T3). The major role of iodine in nutrition arises from the essential role of thyroid hormones in normal growth and development (Stanbury and Hetzel 1980; Hetzel 1989).
The relation of iodine deficiency to enlargement of the thyroid gland, or goitre, was first shown by David Marine who found that hyperplastic changes occurred regularly in the thyroid when the iodine concentration fell below 0.1 per cent (Hetzel 1989). Subsequently, in 1922, Marine and Kimball demonstrated in schoolchildren in Akron, Ohio, that endemic goitre could be both prevented and substantially reduced by administration of small amounts of iodine as potassium iodide in milk given over 10 days twice a year.
Mass prophylaxis of goitre with iodized salt was first introduced in Switzerland and in Michigan in the United States. In Switzerland, the widespread occurrence of a severe form of mental deficiency and deaf mutism (endemic retinism) was a heavy charge on public funds. However, following the introduction of iodized salt, goitre incidence fell rapidly and new cretins were no longer born. Goitre also disappeared from army recruits (Burgi et al. 1990).
A further major development was the administration of injections of iodized oil to people living in inaccessible mountain villages in Papua New Guinea (McCullagh 1963; Buttfield and Hetzel 1967). Subsequently, the prevention of cretinism and stillbirths was demonstrated by the administration of iodized oil before pregnancy in a controlled trial in the Highlands of Papua New Guinea (Pharoah et al. 1971). This work opened up the modern concept of the iodine deficiency disorders (IDD) resulting from all the effects of iodine deficiency on growth and development, particularly brain development, in an exposed population. Iodine deficiency is now recognized as the most common form of preventable mental defect (Hetzel 1983, 1989).
Although the major prevalence is in developing countries, the problem continues to be very significant in many European countries (France, Italy, Germany, Greece, Poland, Romania, Spain, and Turkey) because of the threat to brain development in the fetus and young infant (Hetzel 1989; Hetzel et al. 1990).
The ecology of iodine deficiency
There is a cycle of iodine in nature. Most of the iodine resides in the ocean. It was present during the primordial development of the Earth, but large amounts were leached from the surface soil by glaciation, snow, or rain, and were carried by wind, rivers, and floods into the sea. Iodine occurs in the deeper layers of the soil and is found in oil-well and natural gas effluents. Water from such deep wells can provide a major source for iodine. In general the older an exposed soil surface, the more likely it is to be leached of iodine (Stanbury and Hetzel 1980; Hetzel 1989).
The better known areas to be leached are the mountainous areas of the world. The most severely deficient soils are those of the European Alps, the Himalayas, the Andes, and the vast mountains of China. But iodine deficiency is likely to occur to some extent in all elevated regions subject to glaciation and higher rainfall, with run-off into rivers. More recently it has become clear that iodine deficiency also occurs in flooded river valleys such as the Ganges in India, the Mekong in Vietnam, and the river valleys of China.
Iodine occurs in soil and the sea as iodide. Iodide ions are oxidized by sunlight to elemental iodine which is volatile so that every year some 400 000 tons escape from the surface of the sea. The concentration of iodide in the sea water is about 50 to 60 µg/l; in the air it is approximately 0.7 µg/m3. The iodine in the atmosphere is returned to the soil by the rain which has concentrations in the range 1.8 to 8.5 µg/l. In this way the cycle is completed.
However, the return of the iodine is slow and small in amount compared with the original loss of iodine, and subsequent repeated flooding ensures the continuity of iodine deficiency in the soil. Hence no ‘natural correction’ can take place and iodine deficiency persists in the soil indefinitely. All crops grown in these soils will be iodine deficient. As a result human and animal populations which are totally dependent on food grown in such soil become iodine deficient. The iodine content of plants grown in iodine-deficient soils may be as low as 10 µg/kg compared with 1 mg/kg dry weight in plants in a non-iodine-deficient soil. This accounts for the occurrence of severe iodine deficiency in vast populations in Asia who are living within systems of subsistence agriculture in flooded river valleys (India, Bangladesh, Burma, Vietnam, and China).
The physiology of iodine deficiency
The healthy human adult body contains 15 to 20 mg of iodine of which about 70 to 80 per cent is in the thyroid gland. The thyroid weighs only 15 to 25 g so that it has a remarkable concentrating power.
Iodide is rapidly absorbed through the gut. The normal intake and requirement is 100 to 150 µg/day. Excess iodine is excreted by the kidney. The level of excretion correlates well with the level of intake so that it can be used to assess the level of iodine intake (see below).
The thyroid has to trap about 60 µg of iodine per day to maintain an adequate supply of T4. This is possible because of the very active iodide trapping mechanism which maintains a gradient of 100:1 between the thyroid cell and the extracellular fluid. In iodine deficiency this gradient may exceed 400:1 in order to maintain the output of T4.
Thyroid secretion is under the control of the pituitary gland through thyroid-stimulating hormone (TSH) which is a glycoprotein with a molecular weight of approximately 28 000. There are two subunits—the X subunit has virtually the same structure as other pituitary hormones, and the B subunit is specific for TSH but essentially the same across different species.
The control of TSH secretion is by a ‘feedback’ mechanism related closely to the level of T4 in the blood. As the blood T4 falls, the pituitary TSH secretion is increased to increase thyroid activity, and the output of T4 into the circulation, and so maintain the necessary level of circulating hormone. If this is not possible because of severe iodine deficiency, the level of T4 remains lowered and the level of TSH remains elevated. Both these measurements are used for the diagnosis of hypothyroidism due to iodine deficiency at various stages in life, but particularly in the neonate.
Blood thyroid hormones and TSH (the TSH of the pituitary) are determined using radio-immunoassay when known amounts of radio-actively labelled hormones are used and compete with the unknown amount of hormone in the blood in binding to a specific antibody. The technology for these determinations is still advancing and further improvements are being developed (see below).
The development of goitre
The preceding discussion of the production and regulation of thyroid hormones provides the framework for understanding the production of goitre as a result of iodine deficiency (Fig. 1). Although not the only cause, iodine deficiency is the primary cause of goitre. Other factors—’goitrogens’ such as thiocyanates can enhance the effect of iodine deficiency—are referred to as secondary factors (Ermans et al. 1980).
Fig. 1 A mother and child from a New Guinea village who are severely iodine deficient. The mother has a large goitre and the child is also affected. The larger the goitre, the more likely it is that she will have a cretin child. This can be prevented by eliminating the iodine deficiency before the onset of pregnancy.
The basic effect of iodine deficiency is to interfere with the production of thyroid hormones because iodine is an essential constituent of the T4 and T3 molecules. The lowering of output from the thyroid leads to a fall in the blood levels of T4 but some increase in T3 (the less iodinated hormone is produced preferentially in iodine deficiency).
The fall in the level of T4 leads to increase in TSH output from the pituitary with increase in uptake of iodide and increased turnover associated with hyperplasia of the cells of the thyroid follicles. The reserves of colloid containing thyroglobulin are gradually used up, so that the gland has a much more cellular appearance than normal. The size of the gland increases with the formation of a goitre. Enlargement is regarded as significant in the human when the size of the lateral lobes is greater than the terminal phalanx of the thumb of the person examined. More precise measurements can now be made using ultrasound.
Extensive reviews of the global geographic prevalence of goitre have been published (Dunn et al. 1986; Hetzel et al. 1987; Hetzel 1989; Hetzel and Pandav 1996).
The iodine deficiency disorders
The effects of iodine deficiency on growth and development of a population that can be prevented by correction of iodine deficiency are now denoted by the term IDD. These effects are evident at all stages including particularly the fetus, the neonate, and in infancy, which are the periods of rapid brain growth (Hetzel 1983). The term ‘goitre’ has been used for many years to describe the effect of iodine deficiency. Goitre is indeed the obvious and familiar feature of iodine deficiency but knowledge has greatly expanded in the last 25 years so that it is not surprising that a new term is needed (Table 1).
Table 1 The spectrum of IDD
Iodine deficiency of the fetus is the result of iodine deficiency in the mother. The condition is associated with a greater incidence of stillbirths, spontaneous abortions, and congenital abnormalities, which can be reduced by iodization. The effects are similar to those observed with maternal hypothyroidism which can be reduced by thyroid hormone replacement therapy (McMichael et al. 1980).
Controlled trials with iodized oil have revealed a significant reduction in recorded fetal and neonatal deaths in the treated group which is consistent with animal evidence indicating the effect of iodine deficiency on fetal survival (Hetzel et al. 1990).
A major effect of fetal iodine deficiency is the condition of endemic cretinism which is quite distinct from the condition of sporadic cretinism (Pharoah et al. 1980; Hetzel 1989). This condition, which occurs with an iodine intake of below 25 µg/day in contrast to a normal intake of 80 to 150 µg/day, is still widely prevalent, affecting for example up to 10 per cent of the populations living in severely iodine-deficient areas in India, Indonesia, and China. In its most common form, it is characterized by mental deficiency, deaf mutism, and spastic diplegia (Fig. 2); this is referred to as the ‘nervous’ or neurological type in contrast to the less common ‘myxoedematous’ type characterized by hypothyroidism with dwarfism.
Fig. 2 A mother with her four sons in Chengde, China: three of them (aged 31, 29, and 28 years) are cretins, born before iodized salt was introduced, but the fourth (aged 14) is normal, born after iodized salt became available.
Apart from its prevalence in Asia and Oceania, cretinism also occurs in Africa (Zaire) and the Andean region of South America (Ecuador, Peru, Bolivia, and Argentina) (Pharoah et al. 1980). In all these situations, with the exception of Zaire, neurological features are predominant. In Zaire the myxoedematous form is more common, probably because of the high intake of cassava (Ermans et al. 1980).
There is considerable variation in the clinical manifestations of neurological cretinism which include isolated deaf mutism and mental defect of varying degrees. In China the term ‘cretinoid’ is used to describe these individuals (Ma et al. 1982), which may number five to 10 times those with overt cretinism.
The apparent spontaneous disappearance of endemic cretinism in Italy and Switzerland raised considerable doubts as to the relation of iodine deficiency to the condition. However, a controlled trial in the Western Highlands of Papua New Guinea revealed that endemic cretinism could be prevented by correction of iodine deficiency with iodized oil before pregnancy (Pharoah et al. 1971; Pharoah and Connolly 1987).
The value of iodized oil injection in the prevention of endemic cretinism has been confirmed in Zaire and South America. Mass injection programmes have been carried out in New Guinea (1971–1972) and in Zaire, Indonesia, and China. Recent evaluations of these mass programmes in Indonesia and China indicate that endemic cretinism has been prevented where correction of iodine deficiency has been achieved (Hetzel 1989).
The apparent spontaneous disappearance of the condition is now attributed to increase in iodine intake due to dietary diversification as a result of social and economic development affecting more remote rural areas and the use of dietary supplements containing iodine (Burgi et al. 1990).
An increased perinatal mortality due to iodine deficiency has been shown in Zaire from the results of a controlled trial of iodized oil injections given alternately with a control injection in the latter half of pregnancy (Thilly et al. 1986). There was a substantial fall in perinatal and infant mortality with improved birth weight. Low birth weight (whatever the cause) is generally associated with a higher rate of congenital anomalies and higher risk through childhood.
Apart from the question of mortality, the importance of the state of thyroid function in the neonate relates to the fact that at birth the brain of the human infant has only reached about one-third of its full size and continues to grow rapidly until the end of the second year. The thyroid hormone, dependent on an adequate supply of iodine, is essential for normal brain development as has been confirmed by animal studies (Hetzel and Mano 1989; Hetzel et al. 1990).
Data on iodine nutrition and neonatal thyroid function in Europe have been published. These data confirm the continuing presence of severe iodine deficiency affecting neonatal thyroid function and hence a threat to early brain development (Delange et al. 1986). These data are arousing great concern about iodine deficiency, which is heightened by awareness of the hazard of nuclear radiation following the Chernobyl disaster.
There is similar evidence from neonatal observations in Zaire where rates of chemical hypothyroidism as high as 10 per cent have been found (Ermans et al. 1980), as is also found in northern India (Hetzel et al. 1987).
These observations indicate a much greater risk of mental defect in severely iodine-deficient populations than is indicated by the presence of classical cretinism. There is a continuing major problem in many European countries such as Italy, Germany, France, and Greece, while Rumania, Bulgaria, and Albania still have very severe iodine deficiency with overt cretinism (Delange et al. 1993; Hetzel and Pandav 1996).
Iodine deficiency in children is characteristically associated with goitre. The classification of goitre has been standardized by the World Health Organization (WHO) and is discussed below. The goitre rate increases with age so that it reaches a maximum with adolescence. There is a higher prevalence in girls than in boys. Observations of goitre rates in schoolchildren between the ages of 8 to 14 years provide a convenient indication of the presence of iodine deficiency in a community.
In one meta-analysis including 18 studies in which a comparison was made between iodine-deficient children and carefully selected control groups, the mean scores were found to be 13.5 IQ points apart. This meant that the iodine-deficient groups had a mean IQ which was 13.5 points lower than the non-iodine-deficient control group (Bleichrodt and Born 1994). Detailed individual studies demonstrating these defects in Italian and Spanish schoolchildren as well as those from Africa, China, Indonesia, and Papua New Guinea have been published (Delange 1994; Stanbury 1994). There is a serious problem in Europe as well as in many developing countries.
Iodine administration in the form of iodized salt, iodized bread, or iodized oil have all been demonstrated to be effective in the prevention of goitre in younger adults. Iodine administration may also reduce existing goitre in adults. This is particularly true of iodized oil administration. This obvious effect with the benefits of the correction of hypothyroidism leads to ready acceptance of the measure by people living in iodine-deficient communities.
In northern India a high degree of apathy has been noted in populations living in iodine-deficient areas. This may even affect domestic animals such as dogs. It is apparent that reduced mental function is widely prevalent in iodine-deficient communities with effects on their capacity for initiative and decision-making. This is due to the effect of hypothyroidism on brain function—this condition can be readily reversed by correction of the iodine deficiency (unlike the effects on the fetus and in infancy).
This means that iodine deficiency is a major block to the human and social development of communities living in an iodine-deficient environment. Correction of the iodine deficiency is indicated as a major contribution to development. Increase in physical and mental energy leads to improvements in work output, school learning, and general quality of life. Improved livestock productivity (chickens, cattle, and sheep) is also a major economic benefit.
Magnitude of the problem
The extent of IDD in the world has now been estimated by the WHO (1990). In an estimated at-risk population of 1 billion, in excess of 200 million have goitre and 20 million have some degree of brain damage due to the effects of iodine deficiency in pregnancy. More recently the estimate of the population at risk has been increased to 2.2 billion (WHO 1999).
One review of the problem of IDD in Europe led to an estimate of 140 million (out of a total of 850 million) still at risk of the effects of iodine deficiency, especially on brain development in the fetus and infants over the first 2 years of life. In fact most European countries still have some IDD (Delange et al. 1993).
The major regional concentration of population at risk is in Asia which contributes over 1 billion of the world total, with almost half in China (where 40 per cent of the population are at risk) (WHO 1999).
Assessment of iodine status
The assessment of iodine nutritional status is important in relation to public health programmes in which iodine supplementation is carried out. Therefore the problem is one of assessment of a population or group living in an area or region that is suspected to be iodine deficient. The major assessment criteria are as follows:
thyroid size including the rate of palpable or visible goitre classified according to accepted criteria
urine iodine excretion
the determination of the level of blood T4 or TSH.
Particular attention is now given to TSH levels in the neonate because of the importance of the level of thyroid function for early brain development.
The classification of goitre severity has now been simplified by the WHO (Table 2). There are significant differences in technique between different observers. In general, visible goitre is more readily verified than palpable goitre. Recent observations indicate that palpation of the thyroid overestimates the size of the gland as determined by ultrasonography, particularly in children. For this reason ultrasonography is now replacing palpation when suitable equipment is available (Hetzel and Pandav 1996; Delange 1997).
Table 2 Simplified classification of goitre
The determination of urine iodine excretion can be carried out on 24-h samples. However, the difficulties of such collections are usually insurmountable. For this reason determinations can be more conveniently carried out on casual samples from a group of approximately 50 or more subjects. The iodine levels are expressed as micrograms per litre and the range plotted as a histogram. The median level is used as a convenient indicator of the range. It is normally 100 µg/l.
The level of iodine excretion provides a good index of the level of iodine nutrition. The normal requirement for iodine intake is 100 to 150 µg/day increasing to 200 µg/day in pregnancy (Food and Nutrition Board 1989; Delange et al. 1993). Improvements in methodology and the availability of modern automated equipment (autoanalyser) are making the analysis of large numbers of samples feasible (WHO 1994). The remedial effects of iodization programmes can also be most conveniently monitored by determination of urine iodine excretion in 50 casual samples from schoolchildren.
The determination of the level of T4 or TSH provides an indirect measure of iodine nutritional status. The availability of radio-immunoassay methods with automated equipment has greatly assisted this approach, and TSH is now the preferred method because of better stability under tropical conditions and easier methodology. Particular attention should be given to levels of TSH in the neonate as an indication of the risk of brain damage.
In developed countries, where iodine deficiency in humans normally does not exist, all babies born are screened to ensure that they have adequate thyroid hormone levels. These screening programmes use blood from heel pricks of neonates (usually on the fourth postnatal day) spotted onto filter paper which is dried and sent to a regional laboratory. Blood levels of T4, TSH, or both are measured by immunoassay techniques. The detection rate of neonatal hypothyroidism requiring treatment in the absence of iodine deficiency is about 1 per 3500 babies screened. This rate varies little among developed countries (Burrow 1980).
Neonatal hypothyroid screening has been initiated in several less-developed iodine-deficient regions. As already noted severe biochemical hypothyroidism (T4 concentrations less than 3 µg/dl) has been reported in up to 10 per cent of neonates in northern India and Zaire (Hetzel et al. 1987). It is evident from this and from other reports that, within an iodine-deficient population, serum T4 levels are lowest at birth and lower in children than in the adult population (Delange et al. 1986). In addition, goitrogens such as cassava seem to be much more potent at reducing serum T4 levels in neonates and children than adults.
To summarize, the most critical evidence for determination of iodine nutrition status comes from measurement of urine excretion of iodine and from the measurement of blood TSH in the neonate. The results of these two determinations indicate the severity of the problem. They can also be used to assess the effectiveness of remedial measures (WHO 2001).
The correction of iodine deficiency
Since the introduction of iodized salt in Switzerland and the United States, successful programmes have been reported from a number of countries. These include Central and South America (e.g. Guatemala, Colombia), Finland, and Taiwan (Hetzel 1989). However, there has been great difficulty in sustaining these programmes in Central and South America because of political instability. More recently, countries in the former USSR have had major setbacks.
The difficulties in the production and maintenance of quality to the millions that are iodine deficient, especially in Asia, are vividly demonstrated in India, where there was a breakdown in supply. These difficulties have finally led to the adoption of universal salt iodization for India and subsequently to many other countries. This policy includes legislation that makes it illegal for non-iodized salt to be available for human or animal consumption.
In Asia, the cost of iodized salt production and distribution at present is in the range of 2 to 8 American cents per person per year (Hetzel and Pandav 1989). This must be considered cheap in relation to the social benefits described above.
However, there is still the problem of the iodine in the salt actually reaching the iodine-deficient subject. There may be a problem with distribution or preservation of the iodine content—it may be left uncovered or exposed to heat. It should be added after cooking to reduce the loss of iodine.
Potassium iodate is the preferred vehicle over potassium iodide because of its greater stability in the tropical environment (Hetzel 1989). A dose of 20 to 40 mg iodine as potassium iodate per kilogram is recommended to cover losses to ensure an adequate household level (WHO 1996). This assumes a salt intake of 10 g/day; if the level is below this, an appropriate correction can easily be made by increasing the concentration of potassium iodate.
An increase in incidence of hyperthyroidism has been described following iodized salt programmes in Europe and South America and following the use of iodized bread in Holland and Tasmania (Connolly et al. 1970; Stewart et al. 1971). A few cases have been noted following iodized oil administration in South America. The condition is easily overlooked in developing countries because of the scattered nature of the population and limited opportunities for observation. Iodine-induced hyperthyroidism occurs widely in Europe because of the persistent prevalence of iodine deficiency in the absence of effective iodization programmes. The condition is largely confined to those over 40 years of age; a smaller proportion of the population in developing countries is affected than in developed countries. Detailed observations are available from the island of Tasmania (Stewart et al. 1971; Vidor et al. 1973) and, more recently, from Zimbabwe (Todd et al. 1995). A review has recently been published.
In Zimbabwe careful investigations revealed that excessive levels of iodine in the salt was the result of faulty mixing at factory level and this had led to the occurrence of iodine-induced hyperthyroidism. This indicates that suitable monitoring procedures for salt and urine iodine are essential to prevent excessive iodine intake and iodine-induced hyperthyroidism. Monitoring is also essential to ensure that the intake is adequate to correct iodine deficiency, which has been a greater problem in Asia and Latin America.
Hyperthyroidism is accompanied by some morbidity and mortality in the older age groups. It is readily controlled with antithyroid drugs, radio-iodine, or subtotal thyroidectomy if it is available. Risk of hyperthyroidism, even with an increase to normal levels of intake, arises because an autonomous thyroid can develop independently of TSH control, which continues its high rate of secretion in spite of an increase in iodine intake. To reduce iodine-induced hyperthyroidism to a minimum, the median urine iodine level should not exceed 20 µg/dl. Adequate facilities for diagnosis and treatment are required in areas where iodine-induced hyperthyroidism has been detected.
The occurrence of iodine-induced hyperthyroidism with consequent morbidity and mortality is not regarded as a contraindication to iodization programmes (WHO 1996) in view of the enormous benefits that correction of iodine deficiency has for the whole population—particularly the reduction in child mortality, improved child learning, the improved health of women, greater economic productivity, and improved quality of life.
Furthermore, the correction of iodine deficiency prevents the formation of an autonomous thyroid and so prevents the condition of iodine-induced hyperthyroidism with consequent disappearance in subsequent generations. Hence this condition is included as an ‘iodine-deficiency disorder’ (Table 1).
Iodized oil by injection or by mouth is singularly appropriate for the isolated village communities so characteristic of mountainous endemic goitre areas. The striking regression of goitre following iodized oil administration ensures general acceptance of the measure (Fig. 3). In a suitable area, the oil (1 ml contains 480 mg iodine) should be administered to all females up to the age of 40 years and all males up to the age of 20 years. A dose of 480 mg will provide coverage for 1 year by mouth and for 2 years by injection (Benmiloud et al. 1994).
Fig. 3 Subsidence of goitre in a New Guinea woman 3 months after the injection of iodized oil. This is accompanied by a feeling of well being due to a rise in the level of the thyroid hormone in the blood. This makes the injections very popular. (Reproduced from Buttfield and Hetzel 1967.)
This is particularly important for infants receiving formula milk as an alternative to breast feeding. An increase in levels from 5 to 10 µg/dl has been recommended for full-term infants by the International Council for Control of Iodine Deficiency Disorders (ICCIDD), and a level of 20 µg/dl has been recommended for premature infants (Delange et al. 1993).
Iodized milk has been generally available in the United States, the United Kingdom, and northern Europe, Australia, and New Zealand as a result of the addition of iodophors as disinfectants by the dairy industry. This has been a major factor in the elimination of iodine deficiency in these countries. However, in most countries of southern and eastern Europe this has not occurred and the risk of iodine deficiency continues (Delange et al. 1993). Recently, the use of iodophors has been phased out with substantial drop in the level of urine iodine excretion. The possibility of recurrence of iodine deficiency in industrialized countries (e.g. the United States) now exists (Dunn 1998).
The support of the United Nations system
The United Nations system has now recognized IDD as a major international public health problem and adopted a global plan for its elimination by the year 2000 which was proposed by the ICCIDD working in close collaboration with UNICEF and WHO (Hetzel and Pandav 1996). The ICCIDD is an independent expert group of more than 400 professionals in public health, medical and nutritional science, technologists, and planners, drawn from more than 90 countries.
In 1990, the World Health Assembly and the World Summit for Children both accepted the goal of elimination of IDD as a public health problem by the year 2000. These major meetings included government representatives, including heads of state, at the World Summit for Children from 71 countries with a further 88 countries signing the Plan of Action for elimination of IDD as well as other major problems in nutrition and health.
Since 1989 a series of Joint WHO/UNICEF/ICCIDD regional meetings have been held to assist countries with their national programmes for the elimination of IDD. The impact of these meetings has been that governments now realize the importance of iodine deficiency to the future potential of their people. A dramatic example is provided by the government of the People’s Republic of China. As is well known, China has a one-child family policy which means that an avoidable hazard like iodine deficiency should be eliminated. In China iodine deficiency is a threat to 40 per cent of the population because of the highly mountainous terrain of China and flooded river valleys—in excess of 400 million people are at risk. In recognition of this massive threat to the Chinese people, the government held a National Advocacy Meeting (21–24 September 1993) in the Great Hall of the People, sponsored by the Chinese Premier, Mr Li Peng, under the Chairmanship of Madame Peng Pei-yun of the State Council. The commitment of the government to the elimination of iodine deficiency was emphasized by the Vice-Premier, Mr Zhu Rongyi, to the assembly of provincial delegations led by the provincial governors and the representatives of the international agencies (Hetzel and Pandav 1996).
In Beijing (October 1998) an international workshop was held by the Ministry of Health of China with the ICCIDD. Dramatic progress was reported as indicated by a reduction in mean goitre rate (from 20 to 10 per cent) with normal urine iodine levels. Severe iodine deficiency has persisted in Tibet because of difficulty in the implementation of salt iodization. In other provinces excess of iodine intake was noted in 10 per cent of the population. The need for continuation of monitoring with urine iodine was emphasized at the meeting.
The elimination of IDD at country level
It is now recognized that an effective national programme for the elimination of IDD requires the multisectoral approach as shown in Fig. 4. The ‘wheel’ must keep turning to maintain an effective programme.
Fig. 4 The ‘wheel’ represents the continuous ‘feedback’ process involved in the national IDD control (elimination) programme. All ‘actors’ in the programme need to understand the whole social process.
Striking progress with universal salt iodization has now occurred as indicated by the WHO/UNICEF/ICCIDD Report to the 1999 World Health Assembly. Table 3 shows the percentage of households consuming iodized salt. Where household data are not available, estimates based on production were used instead. These data show that of 5 billion people living in countries with IDD, 68 per cent now have access to iodized salt. Table 4 shows the status of development of the national programme in the 130 countries affected by IDD. It can be seen that 81 per cent of countries have an intersectoral co-ordinating body and 75 per cent have legislation in place.
Table 3 Current status of household consumption of iodized salt
Table 4 Current status of key elements of IDD control programmes
Monitoring is now recognized as a key requirement for ensuring the sustainability of IDD control programmes. Table 4 shows the number of countries monitoring both the process indicators (salt) and the outcome indicators (using goitre rates or urine iodine measurements or both) and whether laboratory facilities are available.
The major challenge is not only the achievement of effective salt iodization but its sustainability. In the past a number of countries have achieved effective salt iodization, but in the absence of monitoring, the programmes have lapsed with recurrence of IDD. To this end, the ICCIDD, the WHO, and UNICEF are now offering help to governments with partnership evaluation to assess progress towards the goal and to provide help to overcome any bottlenecks obstructing progress.
Criteria for tracking progress towards the goal of elimination of IDD have been agreed between the ICCIDD, the WHO, and UNICEF (WHO 1994). These include salt iodine (90 per cent effectively iodized) and urine iodine in the normal range (median excretion of 100 to 200 µg/l). The lower level of 100 µg/l is necessary to ensure normal brain development in the fetus and young infant. The higher level of 200 µg/l is designed to minimize the occurrence of iodine-induced hyperthyroidism.
The global partnership
Since 1990 a remarkable global partnership has come together made up of the people and countries with an IDD problem, the international agencies, particularly UNICEF, the WHO, and the ICCIDD, the salt industry from the private sector, and Kiwanis International, a World Service Club of 330 000 members throughout the world which has adopted a fundraising target of US$75 million towards the elimination of IDD by the year 2000 (Hetzel and Pandav 1996). This partnership exists to support countries and governments in their conquest of IDD. This would be the first global triumph in the elimination of a non-infectious disease.
Benmiloud, M., Chaouki, M.L., Gutekunst, R., Teichert, H-M., Wood, W.G., and Dunn, J.T. (1994). Oral iodised oil for correcting iodine deficiency: optimal dosing and outcome indicator selection. Journal of Clinical Endocrinology and Metabolism, 79, 20–4.
Bleichrodt, N. and Born, M.P. (1994). A metaanalysis of research on iodine and its relationship to cognitive development. In The damaged brain of iodine deficiency (ed. J.B. Stanbury), pp. 195–200. Cognizant Communication Corporation, New York.
Burgi, H., Supersaxo, Z., and Selz, B. (1990). Iodine deficiency diseases in Switzerland one hundred years after Theodor Kocher’s survey: a historical review with some new goitre prevalence data. Acta Endocrinologica, 123, 577–90.
Burrow, G.N. (ed.) (1980). Neonatal thyroid screening. Raven Press, New York.
Buttfield, I.H. and Hetzel, B.S. (1967). Endemic goitre in Eastern New Guinea with special reference to the use of iodised oil in prophylaxis and treatment. Bulletin of the World Health Organization, 36, 243–62.
Connolly, R.J., Vidor, and Stewart, J.C. (1970). Increase in thyrotoxicosis in endemic goitre area after iodization of bread. Lancet, i, 500–2.
Delange, F. (1994). The disorders induced by iodine deficiency. Thyroid, 4, 107–27.
Delange, F. (1997). Thyroid volume and urinary iodine in European school children: standardization of values for assessment of iodine deficiency. Bulletin of the World Health Organization, 75, 95–100.
Delange, F., Heidemann, P., Bourdoux, P., and Larsson A. (1986). Regional variations of iodine nutrition and thyroid function during the neonatal period in Europe. Biology of the Neonate, 49, 322–30.
Delange, F., Dunn, J.T., and Glinoer, D. (ed.) (1993). Iodine deficiency in Europe: a continuing concern. NATO ASI Series A: Life Sciences, Vol. 241. Plenum Press, New York.
Delange, F.D., de Benoist, B., and Alnwick, D. (1999). Risks of iodine-induced hyperthyroidism after correction of iodine deficiency by iodized salt. Thyroid, 9, 545–56.
Dunn, J.T. (1998). What’s happening to our iodine? Journal of Clinical Endocrinology and Metabolism, 83, 3398–400.
Dunn, J.T., Pretell, E.A., and Daza, C.H. (ed.) (1986). Towards the eradication of endemic goitre, cretinism, and iodine deficiency. Pan-American Health Organization, Washington, DC.
Ermans, A.M., Moulameko, N.M., Delange, F., and Alhuwalis, R. (ed.) (1980). Role of cassava in the aetiology of endemic goitre and cretinism. International Development Research Centre, Ottawa.
Food and Nutrition Board, National Academy of Sciences, National Research Council (1989). Recommended dietary allowances (10th edn). Washington, DC.
Hetzel, B.S. (1983). Iodine deficiency disorders (IDD) and their eradication. Lancet, ii, 1126–9.
Hetzel, B.S. (1989). The story of iodine deficiency: an international challenge in nutrition. Oxford University Press.
Hetzel, B.S. and Mano, M. (1989). A review of experimental studies of iodine deficiency during fetal development. Journal of Nutrition, 119, 145–51.
Hetzel, B.S. and Pandav, C.S. (ed.) (1996). SOS for a billion: the conquest of iodine deficiency disorders (2nd edn). Oxford University Press.
Hetzel, B.S., Dunn, J.T., and Stanbury, J.B. (ed.) (1987). The prevention and control of iodine deficiency disorders. Elsevier, Amsterdam.
Hetzel, B.S., Potter, B.J., and Dulberg, E.M. (1990). The iodine deficiency disorders: nature, pathogenesis and epidemiology. World Review of Nutrition Dietetics, 62, 59–119.
McCullagh, S.F. (1963). The Huon Peninsula endemic. The effectiveness of an intramuscular depot of iodised oil in the control of endemic goitre. Medical Journal of Australia, 1, 769–77.
McMichael, A.J., Potter, J.D., and Hetzel, B.S. (1980). Iodine deficiency, thyroid function, and reproductive failure. In Endemic goitre and endemic cretinism (ed. J.B. Stanbury and B.S. Hetzel), pp. 445–60. Wiley, New York.
Ma, T., Lu, T., Tan, U., et al. (1982). The present status of endemic goitre and endemic cretinism in China. Food and Nutrition Bulletin, 4, 13–19.
Pharoah, P.O.D. and Connolly, K.J. (1987). A controlled trial of iodinated oil for the prevention of cretinism: a long term follow up. International Journal of Epidemiology, 16, 68–73.
Pharoah, P.O.D., Buttfield, I.H., and Hetzel, B.S. (1971). Neurological damage to the fetus resulting from severe iodine deficiency uring pregnancy. Lancet, i, 308–10.
Pharoah, P.O.D., Delange, F., Fierro-Benitez, R., and Stanbury, J.B. (1980). Endemic cretinism. In Endemic goitre and endemic cretinism (ed. J.B. Stanbury and B.S. Hetzel), pp. 395–421. Wiley, New York.
Stanbury, J.B. (ed.) (1994). The damaged brain of iodine deficiency. Cognizant Communication Corporation, New York.
Stanbury, J.B. and Hetzel, B.S. (ed.) (1980). Endemic goitre and endemic cretinism. Wiley, New York.
Stewart, J.C., Vidor, G.I., Buttfield, I.H., and Hetzel, B.S. (1971). Epidemic thyrotoxicosis in northern Tasmania: studies of clinical features and iodine nutrition. Australian and New Zealand Journal of Medicine, 3, 203–11.
Thilly, C.H., Bourdoux, P.P., Vanderpas, J.B., et al. (1986). Neonatal and juvenile hypothyroidism in Central Africa. In Iodine deficiency disorders and congenital hypothyroidism (ed. G. Medeiros-Neto, R.M.B. Maciel, and A. Halpern), pp. 26–33. Ache, Brazil.
Todd, C.H., Allain, T., Gomo, Z.A.R., Hasler, J.A., Ndiweni, M., and Oken, E. (1995). Increase in thyrotoxicosis associated with iodine supplements in Zimbabwe. Lancet, 346, 552–3.
Vidor, G.I., Stewart, J.C., Wall, J.R., Wangel, A., and Hetzel, B.S. (1973). Pathogenesis of iodine induced thyrotoxicosis: studies in northern Tasmania. Journal of Clinical Endocrinology and Metabolism, 37, 901.
WHO (World Health Organization) (1990). Report to 43rd World Health Assembly. WHO, Geneva.
WHO (World Health Organization) (1993). Global prevalence of iodine deficiency disorders. Micronutrient Deficiency Information System (MDIS) working paper no. 1. WHO/UNICEF/ICCIDD, Geneva.
WHO (World Health Organization) (1994). Indicators for assessing iodine deficiency disorders and their control programmes. Publication WHO/NUT/94.6. WHO/UNICEF/ICCIDD, Geneva.
WHO (World Health Organization) (1996). Recommended iodine levels in salt and guidelines for monitoring their adequacy and effectiveness. Publication WHO/NUT/96.13. WHO/UNICEF/ICCIDD, Geneva.
WHO (World Health Organization) (1999). Progress towards the elimination of iodine deficiency disorders (IDD). Publication WHO/NHD/99.4. WHO/UNICEF/ICCIDD, Geneva.
WHO (World Health Organization) (2001). Assessment of iodine deficiency disorders. A guide for programme managers. WHO/UNICEF/ICCIDD, Geneva.
The epidemiology of diabetes mellitus
Diabetes mellitus affects large numbers of people in a wide range of ethnic groups and at all social and economic levels worldwide. Over 110 million people are reported to suffer from diabetes (WHO 1994) and this is almost certainly an underestimate. Projections suggest that there will be over 230 million diabetics by 2010 (Amos et al. 1997) and 300 million by 2025 (King et al. 1998), the majority with type 2 diabetes. Thus there is an urgent need for strategies to be implemented to prevent the emerging global epidemic of type 2 diabetes.
Historically, until the latter part of the nineteenth century, the main causes for morbidity and mortality worldwide resulted from epidemics of communicable diseases including typhoid, cholera, smallpox, and influenza (Hennekens and Buring 1987), diseases which are still epidemic in certain developing countries. With industrialization and modernization of societies, major improvements have occurred in housing, sanitation, water supply, and nutrition. These changes, plus the introduction of antibiotics and immunization programmes, radically changed the profile of diseases in developed countries initially, and later in developing countries (Hennekens and Buring 1987; Zimmet et al. 1990).
With improvements in public health, mortality from infectious diseases has fallen dramatically (Hennekens and Buring 1987), although recent decades have seen the emergence of devastating communicable diseases such as HIV and the Ebola virus, and the re-emergence of that old public health enemy, tuberculosis (World Health Report 1998). Paradoxically, a marked increase in the prevalence of risk factors for non-communicable diseases, such as type 2 diabetes, cardiovascular disease, hypertension, stroke, and cancer, has occurred and they have become major contributors to morbidity and mortality (Zimmet 1999).
This phenomenon is well illustrated in the Pacific and Indian Ocean islands (Zimmet et al. 1990; Zimmet 1995), the site of many of our own epidemiological studies relating to type 2 diabetes and other non-communicable disease. Rapid socio-economic development over the last 40 to 50 years has resulted in a dramatic change of lifestyle from traditional to modern. As the public health aspects of diabetes are so diverse, the main objective of this chapter is to cover aspects relevant to the new classification and diagnostic criteria for diabetes (Alberti and Zimmet 1998) and the prevention of diabetes and its devastating complications.
Impact and cost of diabetes to society
Diabetes is amongst the five leading causes of death by disease in most countries (Finch and Zimmet 1988; WHO 1994). However, mortality statistics greatly underestimate the true diabetes-related mortality, as diabetes is frequently under-reported on death certificates (Finch and Zimmet 1988), a situation that proves a handicap when it comes to creating awareness of diabetes for community public health priorities. Apart from the health impact, the economic cost of diabetes and its complications is enormous, both for health care and for loss of productivity to society (WHO 1994).
While diabetes cost the United States US$20.4 billion in 1987 and US$ 90 billion for 1994, the most recent estimate is around US$100 billion (American Diabetes Association 1998).
Classification of diabetes mellitus—the new recommendations
Diabetes is really a syndrome characterized by hyperglycaemia, but with many causes. The 1985 WHO Study Group classification included a number of clinical classes, of which the two most important were insulin-dependent diabetes mellitus or type 1 diabetes and non-insulin dependent diabetes mellitus or type 2 diabetes, as well as malnutrition-related diabetes, impaired glucose tolerance, and gestational diabetes mellitus (WHO 1985). A revision in classification was long overdue as a result of new data from genetic, epidemiological, and aetiological studies (Zimmet 1995). Considering this global epidemic, the changes proposed for the classification (Fig. 5) and diagnostic criteria for diabetes (Table 5) become very important for global comparisons of incidence and prevalence and monitoring the epidemic. Therefore the recommendations of the 1997 American Diabetes Association report on classification and criteria (American Diabetes Association 1997) followed by the 1998 WHO recommendations (Alberti and Zimmet 1998; WHO 1999) have important implications. The new classification is based on stages of glucose tolerance status with a complementary subclassification according to aetiological type.
Fig. 5 Disorders of glycaemia: aetiological types and clinical stages. (Data from WHO 1999.)
Table 5 Aetiological classification of disorders of glycaemiaa
In the new classification, hyperglycaemia, regardless of the underlying cause, can be subcategorized as follows.
Insulin required for survival (corresponds to the former insulin-dependent diabetes mellitus).
Insulin required for control, i.e. for metabolic control, not for survival (corresponds to the former insulin-treated non-insulin-dependent diabetes mellitus).
Does not require insulin, i.e. treatment by non-pharmacological methods or drugs other than insulin (corresponds to non-insulin-dependent diabetes mellitus on diet alone/or coupled with oral agents).
Impaired glucose tolerance and impaired fasting glycaemia. Impaired glucose tolerance was previously a separate class; it is now categorized as a stage in the natural history of disordered carbohydrate metabolism. Impaired glucose tolerance is coupled with ‘impaired fasting glycaemia’ (6.1–7.0 mmol/l).
New diagnostic criteria for diabetes
The diagnostic criteria for diabetes have changed on a number of occasions over recent decades (Zimmet 1999). Epidemiological data collected over the last decade led both the WHO and the American Diabetes Association to review and revise the diagnostic criteria (Table 6).
Table 6 Values for diagnosis of diabetes mellitus and other categories of glucose intolerance
As a result of these epidemiological data, the American Diabetes Association (1997) and the WHO (1999) recommended the following changes:
the fasting plasma glucose threshold is lowered from 7.8 to 7.0 mmol/l
impaired fasting glycaemia (fasting plasma glucose 6.1–6.9 mmol/l) is introduced as a new category of intermediate glucose metabolism (named impaired fasting glucose by the American Diabetes Association).
The American Diabetes Association (but not the WHO report) recommended that fasting plasma glucose rather than the oral glucose tolerance test should be the diagnostic test of choice for both clinical and epidemiological purposes. An important question is: What difference will these changes in recommended criteria make to the prevalence of diabetes and the individuals identified as having diabetes? A large number of studies have now been published comparing the old and new criteria. The overall impression is that the oral glucose tolerance test should be retained for the diagnosis of diabetes. Using fasting criteria alone, as recommended by the American Diabetes Association, misses a significant number of people with diabetes. These studies are reviewed in detail elsewhere (Shaw et al. 1999b).
The impaired glucose tolerance category includes persons whose oral glucose tolerance test is beyond the boundaries of normality by WHO criteria (see below). Impaired glucose tolerance may represent a stage in the natural history of diabetes as these persons are at higher risk than the general population for diabetes (Harris and Zimmet 1992). Subjects with impaired glucose tolerance have a heightened risk of macrovascular disease (Harris and Zimmet 1992) and, because of this and the association with other known cardiovascular disease risk factors including hypertension, dyslipidaemia, and central obesity (Alberti and Zimmet 1998), the diagnosis of impaired glucose tolerance, particularly in otherwise healthy and ambulatory individuals, may have important prognostic implications.
Thus a major question that arises with the American Diabetes Association recommendations (1997) is the value of impaired fasting glycaemia compared with impaired glucose tolerance for predicting future type 2 diabetes. Screening by the American Diabetes Association criteria for impaired fasting glycaemia alone would identify fewer people who subsequently progress to type 2 diabetes than would the oral glucose tolerance test as recommended by WHO (Shaw et al. 1999c). This latter study demonstrated the higher sensitivity of impaired glucose tolerance over impaired fasting glycaemia for predicting progression to type 2 diabetes. Thus, whilst the new category of impaired fasting glycaemia may broaden and improve our description of states of intermediate glucose metabolism, it should be seen as complementary to rather than a replacement for impaired glucose tolerance.
The American Diabetes Association hoped that their recommendations would simplify the diagnosis of diabetes. However, that has not been the outcome and their recommendations are further clouded by the fact that impaired glucose tolerance is a far better predictor of risk from cardiovascular disease mortality (Davies 1999).
Gestational diabetes mellitus is defined as diabetes first recognized during pregnancy. In the majority of cases, glucose tolerance returns to normal postpartum, but the lifetime risk for impaired glucose tolerance and type 2 diabetes is substantially increased (Alberti and Zimmet 1998). Gestational diabetes mellitus occurs in about 3 per cent of pregnancies in Western nations, but is much higher in high-prevalence communities such as the American Indians and Pacific Islanders (WHO 1994).
The epidemiology of diabetes mellitus
The explosion of interest and activity that has occurred in studies relating to the epidemiology of diabetes over the last two decades has been extraordinary. In his book, Epidemiology of Diabetes and its Vascular Lesions, the late Kelly West highlighted the potential for a future epidemic (West 1978) and, in fact, type 2 diabetes has now reached epidemic proportions in many developing nations, as well as in disadvantaged groups in developed countries (Zimmet 1992, 1999). There is every reason to expect that, over the next decade, the epidemic of type 2 diabetes will continue to escalate (Amos et al. 1997; Zimmet 1999) so that diabetes and its complications will emerge as one of the major threats to future public health resources worldwide at a huge economic and social cost, particularly in developing countries (Table 7).
Table 7 Estimates of type 1 and 2 diabetes in various regions of the world, 1995–2010 (in thousands)
Type 1 diabetes
Prevalence and incidence
Type 1 diabetes is a discrete disorder and its pathogenesis involves environmental triggers that may activate autoimmune mechanisms in genetically susceptible individuals leading to progressive loss of pancreatic islet b-cells (Atkinson and Maclaren 1994). Its frequency is low relative to type 2 diabetes, accounting for 10 to 15 per cent of cases of diabetes in European populations. Type 1 diabetes is most common in European populations and is rare in Asians, Native Americans, Pacific Islanders, and black people (Karvonen et al. 1993).
In the last 15 years there has been an explosion of publications on the incidence and prevalence of type 1 diabetes as a result of the proliferation of diabetes registers. This initiative has resulted in a vast global network of type 1 diabetes registers and a major international collaboration (WHO Diamond Project Group 1990) covering incidence, prevalence, morbidity, mortality, and molecular epidemiology.
There are far too many reports of prevalence and incidence to summarize in this chapter. However, Fig. 6 provides a view of the incidence of type 1 diabetes in selected countries. There is an almost 60-fold difference between the countries with the highest incidence (Finland) and the lowest (Korea) (Karvonen et al. 1993).
Fig. 6 Incidence of type 1 diabetes from selected countries. (Data from Karvonen et al. 1993.)
Genetic factors in the aetiology of type 1 diabetes
Type 1 diabetes is not directly inherited as a Mendelian trait but is a polygenic disorder (Harrison et al. 1999). Predisposition is mediated by different genes that interact in a complex manner with each other and with the environment. Evidence for genetic predisposition comes from twin studies that demonstrate a higher concordance rate for type 1 diabetes in monozygotic twins (25–30 per cent) than in dizygotic twins (5–10 per cent) (Skyler and Marks 1993; Harrison et al. 1999). Recent advances in genetic mechanisms are reviewed in detail elsewhere (Bier and Lernmark 1998).
Environmental factors in the aetiology of type 1 diabetes
Viruses There are seasonal trends for type 1 diabetes incidence, with the highest incidence in winter (Warram et al. 1994). Also, exposure to a number of viruses has been linked to the development of type 1 diabetes. While the Coxsackie B virus appears to be the most common association, 20 per cent of children with congenital rubella develop type 1 diabetes (Skyler and Marks 1993; Harrison et al. 1999). In this latter example, the exposure relates to in utero exposure, which is remote from clinical onset, thus raising the issue of other potential in utero influences.
Secular trends There has been a marked temporal variation in many populations studied. A rising incidence has been noticed in 35 of 68 countries studied (Karvonen et al. 1993), while a fluctuating incidence has been recorded in others (Scott et al. 1992; Michaelis et al. 1993).
Ethnic and migrant studies Studies comparing prevalence and incidence in the same ethnic group provide unique opportunities to unravel the mystery of the putative environmental factors contributing to type 1 diabetes. Type 1 diabetes is uncommon in Asian communities, and very low rates of type 1 diabetes have been confirmed in Korea and China (Karvonen et al. 1993), Thailand (Tuchinda et al. 1992), and Japan (Karvonen et al. 1993). Black Americans have an even lower incidence than Hispanics, but not as low as that recorded in Tanzanians in sub-Saharan Africa (Karvonen et al. 1993).
Finland shows the world’s highest incidence of type 1 diabetes at 35 in 100 000 (Tuomilehto et al. 1992b). This is higher than in the other Baltic States (Tuomilehto et al. 1992a), notably Estonia (Tuomilehto et al. 1991), whose populations are linguistically and ethnically very similar to that of Finland but who suffer only a third the incidence. This indicates that environmental factors have a particularly powerful influence on the appearance of type 1 diabetes, and other suggestions relating to environmental risk factors include toxins and nutritional factors. The rodenticide Vacor has also been associated with the development of type 1 diabetes (Tuomilehto et al. 1997).
Type 2 diabetes
Prevalence and incidence
Type 2 diabetes constitutes about 85 per cent of all cases of diabetes in developed countries (Zimmet 1999). The diagnosis is usually made after the age of 50 years in Europids, although type 2 diabetes is seen at younger age in high-prevalence populations such as Asian-Indians and Pacific Islanders (Zimmet 1999). In some developing countries, especially those with a high prevalence of diabetes, almost 100 per cent of persons with diabetes fall into this category. There is enormous variation in type 2 diabetes prevalence between populations (King and Rewers 1993; de Courten et al. 1997), and exceptionally high rates have been documented in populations who have changed from a traditional to a modern lifestyle, for example up to 40 per cent of adult Native Americans (de Courten et al. 1997), Native Canadians (Harris et al. 1997), Micronesians, and other Pacific Islanders, and about 20 per cent of adult Australian Aborigines, migrant Asian-Indians, and Mexican-Americans (Zimmet et al. 1990; Zimmet 1995). A comparison of age-standardized rates for type 2 diabetes in adults for a number of populations is shown in Fig. 7.
Fig. 7 Prevalence of type 2 diabetes in selected populations (30–64 years) worldwide. Age-standardized against Segi’s world population.
A steady stream of reports on a variety of populations have continued to highlight the explosion of type 2 diabetes in relation to lifestyle change and differences between different ethnic groups (Hamman 1993; King and Rewers 1993; Zimmet 1995). Populations previously free of type 2 diabetes are showing prevalences that are extraordinarily high when compared with developed countries. The reports of a spectacularly high prevalence of type 2 diabetes of over 40 per cent in adults over the age of 40 amongst the Pima Indians of Arizona (Bennett et al. 1992) and the Pacific Islanders of Nauru (Zimmet et al. 1990) have been followed by studies of related populations. The prevalence rates for a number of other different tribes of Native Americans range from 15 to 41 per cent of adults aged over 45 years, whilst amongst Alaskan Indians, who have had less contact with a Western lifestyle, only 3 per cent of those aged between 45 and 64 years have diabetes. The prevalences in the age range 30 to 64 years in the Pacific Islands of Kiribati and Western Samoa (de Courten et al. 1997) are 11 to 16 per cent, and a prevalence of 37 per cent has also been reported in the Melanesians of Papua New Guinea (Dowse et al. 1994). At the other end of the scale, very low prevalences of 1 to 2 per cent have been found in parts of Africa and China (Zimmet 1999).
The rates in Europids, both in Europe and North America, come somewhere in the middle of the extremes noted above. Only a few relevant studies have been performed in Europe, and they indicate rates between 5 per cent in the United Kingdom and 13 per cent in Spain (Shaw et al. 1999a) for those aged over 40 years. In the United States, NHANES II reported that 6 per cent of the Europid population between the ages of 30 and 64 years had type 2 diabetes (Harris et al. 1987).
A longitudinal epidemiological study in the Indian Ocean island of Mauritius has provided the best gauge of the type 2 diabetes epidemic occurring in large sections of the developing world (Zimmet 1999). As the population, currently 13 000 000, includes people of Asian-Indian, Chinese, and black (Creole) descent, and as these ethnic groups compose nearly two-thirds of the world population, the data from Mauritius provide a microcosm of the global epidemic.
Our previous population-based surveys in Mauritius (1987 and 1992) had shown a high diabetes prevalence. The 1987 prevalence of type 2 diabetes was 10 to 13 per cent in each ethnic group, rising to 20 to 30 per cent in those aged 45 to 74 years. In the most recent study in 1998 (de Courten et al. 1999), when 6294 subjects were screened for diabetes with an oral glucose tolerance test, we found a 30 per cent secular increase in diabetes prevalence in the 11 years from 1987. Diabetes now affects close to 20 per cent of the population over 30 years of age. These results, if transposed to parts of India, China, and Africa where modernization and industrialization are occurring, would result in huge increases in the number of cases of type 2 diabetes posing a great public health threat and burden.
Mauritius, apart from revealing a high diabetes prevalence in Asian-Indians and Creoles, showed the highest yet reported prevalence in Chinese (Dowse et al. 1990). These results, and evidence that the prevalence of type 2 diabetes doubled between 1984 and 1992 in Singapore’s Chinese community (Tan et al. 1999), along with high prevalence in Taiwan (Chang et al. 1998), provide an alarming indication to the size of the epidemic which could occur in the People’s Republic of China where the overall prevalence of type 2 diabetes was, until recently, believed to be less than 1 per cent (Zimmet 1999). Data are already available that suggest an almost threefold increase in prevalence in certain areas of China within the last two decades (Pan et al. 1997). This is highlighted by the data in Fig. 6 comparing the prevalence of diabetes in Chinese populations in China, Singapore, Taiwan, and Mauritius. If China were to experience just half of the current rate of diabetes in Taiwan, the number of individuals with diabetes would increase dramatically from 8 million in 1996 to over 32 million by 2010 (Zimmet 1999).
People from the Indian subcontinent are at especially high risk of type 2 diabetes. Early studies from India reported type 2 diabetes to be relatively rare, but a more recent study in an affluent Indian suburb found diabetes prevalence to be 20 per cent in men aged 45 to 74 years. In a south Indian urban community, two surveys performed 5 years apart showed a 40 per cent rise in diabetes prevalence (Shaw et al. 1999a). High diabetes rates have now also been found in Asian-Indians in communities located in Pakistan, South Africa, Fiji, and the United Kingdom (Zimmet 1992).
There could be no better rationale for developing primary prevention programmes for diabetes. However, before embarking on intervention activities, it is essential to have a better understanding of both the genetic and environmental determinants of type 2 diabetes.
Genetic factors in the aetiology of type 2 diabetes
Progress in understanding the genetic component of type 2 diabetes, particularly in defining potential candidate genes, has been slow until recently. Family and twin studies, along with the evidence for heightened genetic susceptibility in certain populations and genetic admixture studies (Zimmet 1992), provided firm evidence that the role of the genetic component was very strong. The recent achievements in molecular biology relating to type 2 diabetes have resulted from studying very well defined autosomal dominant forms in extended families with maturity-onset diabetes of the young (Hattersley 1998) and not from the large heterogeneous pool of persons with type 2 diabetes with or without the other components of the metabolic syndrome. There have been some promising results from studies of families with maturity-onset diabetes of the young where associations have been found with the glucokinase and hepatocyte nuclear factor 1a and 4a genes (Yamagata et al. 1996a, b) and with mitochondrial DNA mutations in families with type 2 diabetes and nerve deafness (McCarthy et al. 1994; Hattersley 1998).
The new WHO classification (Alberti and Zimmet 1998; WHO 1999) should bring a more rational approach to research by molecular biologists. Apart from the necessity to define diabetes correctly from an epidemiological and therapeutic perspective, the total success of the current thrust to define crucial genes for both type 1 and 2 diabetes depends on correct phenotyping (Zimmet 1999). It is not surprising that most geneticists still regard diabetes as the ‘geneticists’ nightmare’ when clinicians have been asking them to find the diabetes gene(s) while providing samples from poorly defined groups. The search for the type 2 diabetes gene(s) is already difficult enough without this handicap.
While maturity-onset diabetes of the young only accounts for approximately 1 per cent of all type 2 diabetes, the implications for discovering other genes for other forms of diabetes are important. Yet, to find gene(s) associated with any complex disorder such as type 2 diabetes, the task is more formidable. The failure to find a major putative gene for type 2 diabetes despite the advanced stage of the Human Genome Project (Collins 1999) raises many questions over the strategic approach most likely to be successful. Finding the gene(s) has enormous implications not only for screening for ‘high-risk’ individuals but for targets for potential new therapeutic compounds as well as identifying which individuals are likely to respond to different therapies (Collins 1999). Epidemiologists must continue to play a pivotal role in helping to define the best cohorts for these genetic studies.
Thrifty genes An attempt to explain the high prevalence of obesity and type 2 diabetes in the American Pima Indians, Australian Aborigines, and Pacific Islanders comes through the thrifty gene hypothesis (Neel 1962). The basis for the susceptibility to obesity and type 2 diabetes is unclear, but it could be the result of an evolutionary advantageous thrifty genotype which promoted fat deposition and storage of calories in times of plenty and provided a positive selective advantage during periods of food shortage and starvation (Neel 1962; Dowse and Zimmet 1993). This would have conferred a survival advantage during the regular famines and natural disasters that were interspersed with feast periods (Zimmet 1993), but would result in type 2 diabetes once a sedentary lifestyle and a diet with an excess of energy, simple carbohydrates, and saturated fats were adopted.
Hales and Barker (1992) have suggested that in utero factors leading to fetal malnutrition may be the cause of this chronic disease epidemic and they proposed the thrifty phenotype hypothesis. Their proposition is that the causes are entirely environmental and they discount a role for genetic factors. The intellectual exchanges between the proponents of the thrifty genotype and thrifty phenotype hypotheses have become the highlight of many international scientific meetings. The issues challenging the thrifty phenotype have been reviewed in detail by Joseph and Kramer (1996) who listed various direct and indirect pieces of evidence suggesting that the reported association of low birth weight and later type 2 diabetes may be biased rather than causal. They state that ‘selection bias, failure to define, measure, and adequately control for the confounding health consequences of social deprivation and inconsistencies in the hypotheses tested and in methods of data analysis and reporting are among the factors that weigh against a causal explanation for the associations observed’ (Joseph and Kramer 1996).
While it is clear that the association between low birth weight infants and subsequent type 2 diabetes risk is a true phenomenon, the explanation to explain it is not so simple. There are now strong arguments against an exclusive environmental role and it is likely that genes are also implicated (Zimmet 1999).
Environmental factors in the aetiology of type 2 diabetes
While genetic susceptibility to type 2 diabetes is clearly important, there is strong evidence that the disease is unmasked by environmental factors (Zimmet 1995; de Courten et al. 1997). Several environmental risk determinants (Table 8) are associated with an increased risk of type 2 diabetes including nutritional factors, physical activity, central and overall obesity, intrauterine factors, and so on. These are relevant to primary prevention as discussed below and a much more extensive reviews can be found elsewhere (Bennett et al. 1992; de Courten et al. 1997).
Table 8 Demographic, behavioural, and environmental risk determinants for type 2 diabetes
Complications of diabetes
Space limitations prevent a detailed description of the microvascular and macrovascular complications of diabetes which are reviewed in detail elsewhere (Jarrett 1992; Klein and Moss 1992; Ward 1992; WHO 1994; Hamman 1997; Tuomilehto and Rastenyté 1997). Their main relevance from a public health perspective is the relationship to human suffering and disability, and socio-economic costs through premature morbidity and mortality (Songer 1992; World Bank 1993). For example, the major microvascular complications are diabetic retinopathy, nephropathy, and neuropathy. The extent to which they occur is influenced predominantly by duration of diabetes and degree of metabolic control (WHO 1994). These complications may be apparent at the time of diabetes diagnosis, especially in type 2 diabetes (Harris et al. 1992).
Diabetes is the most common cause of adult blindness in developed countries due to retinopathy, cataracts, or glaucoma (Klein and Moss 1992; WHO 1994). Diabetic patients are 17 times more prone to kidney disease, and diabetes is now the leading known cause of endstage renal disease in the United States (WHO 1994). Diabetic neuropathy is probably the most common complication, being present in about 30 to 40 per cent of type 1 and 2 diabetes subjects (WHO 1994).
As for macrovascular disease, atherosclerosis is the most common complication of diabetes among Europids. It accounts for at least two-thirds of deaths (Zimmet and Alberti 1997), a figure that is two to three times greater than that in people without diabetes, and coronary artery disease and cerebrovascular disease are also two to three times more common in diabetics than in non-diabetics (Jarrett 1992). The WHO Multinational Study of Vascular Disease in Diabetics found marked differences in prevalence of macrovascular disease between 14 countries (WHO 1985b). The increase in atherosclerosis in diabetics compared with non-diabetics was seen in all populations, even in those where the incidence of atherosclerosis was low, but the prevalence and pattern of microvascular disease were similar between countries.
Numerous epidemiological studies have demonstrated the importance of indices of glycaemic control and the duration of diabetes in determining the prevalence and incidence of diabetic microvascular complications (Zimmet 1999). Subsequent epidemiological data supported this but it required testing in large-scale clinical trials. This happened with the landmark Diabetes Control and Complications Trial (DCCT) (1995). The DCCT demonstrated that improved glycaemic control reduces the risk of microvascular complications in type 1 diabetes. However, these findings only applied to type 1 diabetes, and whether these results could be extrapolated to patients with type 2 diabetes awaited confirmation in large trials such as the United Kingdom Prospective Diabetes Study (UKPDS 1998a). The recently published findings of the UKPDS now provide strong evidence to recommend tight glycaemic control in type 2 diabetes (UKPDS 1998a,b).
Two-thirds of type 2 diabetes patients die from cardiovascular disease (Zimmet and Alberti 1997). The clustering of type 2 diabetes, a well-documented risk determinant for cardiovascular disease, with the other risk factors including insulin resistance that constitute the metabolic syndrome is now well established (Zimmet 1995). This seems the most likely explanation for this increased cardiovascular disease mortality in type 2 diabetes. Alone, each component of the cluster conveys increased cardiovascular disease risk, but as a combination their effect is cumulative. An important and pivotal study relevant to this cardiovascular disease risk was the randomized clinical trial of cardiovascular intervention, the Scandinavian Simvastatin Survival Study (Pyorälä et al. 1997). The use of simvastatin in diabetic subjects with a previous coronary heart disease event resulted in a 55 per cent reduction in the risk of future coronary heart disease events compared with 33 per cent in non-diabetics. The message is now very clear that management should focus not only on tight blood glucose control but also on strategies for the reduction of the other important cardiovascular disease risk factors such as hypertension, dyslipidaemia, and obesity.
Primary prevention of diabetes mellitus
Diabetes is now a major global public health problem, but from a public health perspective type 2 diabetes is by far the most common form, particularly in newly industrialized nations (Zimmet 1999). Prevention of type 1 diabetes is still in an experimental stage (Bier and Lernmark 1998; Harrison et al. 1999) and so the major emphasis here will be on prevention of type 2 diabetes.
Prevention of type 1 diabetes
Although type 1 diabetes is probably the most common serious childhood disease, the very low prevalence demands a screening test of high specificity and sensitivity. Furthermore, the test would need to be inexpensive and easy to perform in order to screen a large number of individuals, which would be necessary for the prevention of type 1 diabetes as only 10 to 15 per cent of new cases come from the high-risk group, i.e. first-degree relatives (Bingley et al. 1993). Islet cell antibodies, insulin autoantibodies, glutamic acid decarboxylase antibodies, and tyrosine phosphatase antibodies are the markers which have been most commonly used to date (Zimmet et al. 1999), but there are major cost and technical limitations to islet cell antibody and insulin antibody assays.
The ability to measure a group of autoantibodies to islet cell constituents, including islet cell antibodies, insulin antibodies, glutamic acid decarboxylase antibodies, and tyrosine phosphatase antibodies, have greatly facilitated the ability to predict future type 1 diabetes (Zimmet 1999). The long preclinical phase of type 1 diabetes provides a window of opportunity for primary prevention with the possibility of a effective screening procedures, and the future availability of immune intervention (Skyler and Marks 1993; Harrison et al. 1999). A number of trials are now underway to assess the use of oral nicotinamide, prophylactic parenteral or oral insulin, oral glutamic acid decarboxylase, and intracutaneous bacille Calmette–Guérin (Atkinson and McLaren 1994; Harrison et al. 1999), all strategies which may delay or arrest the autoimmune process.
Prevention of type 2 diabetes
Compared with primary prevention of coronary artery disease, virtually no population-based studies of type 2 diabetes have been published yet apart from a major impaired glucose tolerance intervention study from China (Pan et al. 1997). The incidence of type 2 diabetes was reduced by a third in the intervention group compared with controls, and this applied equally to non-obese and obese subjects with impaired glucose (Fig. 8). Given the projected epidemic of diabetes in China (Zimmet 1999), the largest population in the world, such a cost-effective intervention provides good news.
Fig. 8 The effect of diet and exercise in preventing the progression of impaired glucose tolerance to type 2 diabetes in Da Qing, China. (Data from Pan et al. 1997.)
However, several population-based studies are currently underway (Diabetes Prevention Program 1999; Eriksson et al. 1999). As type 2 diabetes is so heterogeneous, for preventive measures to be fully effective in a community they must be based on knowledge of the risk determinants in that community (Tuomilehto et al. 1992c). There are sufficient data available to show that type 2 diabetes should be preventable (Zimmet 1999). These data come from a variety of sources, including the impact of lifestyle change on cardiovascular disease risk factors and incidence (Tuomilehto et al. 1992c), and two studies where indigenous groups at high risk of type 2 diabetes showed improvements in risk factors with traditional diet (O’Dea 1984; Shintani et al. 1991) and other lifestyle changes (Dowse et al. 1995; Uusitala et al. 1996). However, more compelling is the recent report from China where the incidence of progression from impaired glucose tolerance to type 2 diabetes was reduced by over 30 per cent through diet and exercise interventions (Pan et al. 1997).
There are two components of the implementation of primary prevention—the population and the high-risk strategies. They are generally complementary (WHO 1994). The population approach may be more appropriate in societies with particularly high genetic susceptibility (in which case the two strategies are effectively the same) (Tuomilehto et al. 1992c). The high-risk alternative has an advantage for individuals at special high risk for type 2 diabetes, for example first-degree relatives of type 2 diabetes patients, subjects with impaired glucose tolerance or previous gestational diabetes mellitus, and those with the metabolic syndrome, i.e. glucose intolerance, insulin resistance, abdominal obesity, dyslipidaemia, and hypertension (Zimmet 1995).
It is both cost and health effective to use an integrated approach for prevention activities to reduce the common risk factor levels for these non-communicable disease in the community particularly in developing nations (Zimmet 1995). Correction and prevention of obesity, exercise, avoidance of a high-fat diet, and encouraging a high-fibre diet all reduce insulin resistance and reduce the levels of some of the other risk factors for coronary artery disease (Zimmet 1995).
The components necessary for the development of a national diabetes prevention programme form a model for prevention of other non-communicable disease, a model now called the ‘Hetzel wheel’ (Fig. 9). Prevention programmes should not be commenced without a properly constituted evaluation component, and guidelines for monitoring and evaluation appear in one WHO report (WHO 1994).
Fig. 9 Control and prevention of diabetes. The components necessary for the development of a national diabetes prevention programme. The ‘Hetzel wheel’—these activities form a model for prevention of other non-communicable diseases.
To summarize, healthy lifestyle practices such as exercise, appropriate nutrition, maintenance of ideal body weight, smoking cessation, and reduced alcohol consumption targeted at whole communities are also likely to reduce the social and economic burden of type 2 diabetes and other non-communicable disease. The efficacy of primarily non-pharmacological interventions such as those used in the Chinese study of impaired glucose tolerance intervention (Pan et al. 1997) provides support for the view that interventions in impaired glucose tolerance subjects should be given a high priority. The American National Institutes of Health have risen to the challenge and funded a major multicentre impaired glucose tolerance intervention to examine the potential for prevention (Diabetes Prevention Program 1998). There may also be a role for pharmacological intervention with biguanides, a-glucosidase inhibitors, and the new therapeutic compounds of the thiozolidodione group. Both approaches are being used in the American study.
Diabetes is likely to remain a significant threat to public health in the years to come. In the absence of effective interventions for both type 1 and type 2 diabetes, the frequency will escalate worldwide with the main impact being seen in developing nations and in disadvantaged minorities in developed nations (Amos et al. 1997; King et al. 1998; Zimmet 1999). Thus prevention of diabetes and its microvascular and macrovascular complications is an essential component of future public health strategies for all nations.
Alberti, K.G.M.M. and Zimmet, P.Z., for the WHO Consultation Group (1998). Definition, diagnosis and classification of diabetes mellitus and its complications. Part 1: Diagnosis and classification of diabetes mellitus. Provisional report of a WHO consultation. Diabetic Medicine, 15, 539–53.
American Diabetes Association (1997). Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care, 20, 1183–97.
American Diabetes Association (1998). Economic consequences of diabetes mellitus in the US in 1997. Diabetes Care, 21, 296–309.
Amos, A., McCarty, D., and Zimmet, P. (1997). The rising global burden of diabetes and its complications: estimates and projections to the year 2010. Diabetic Medicine, 14 (Supplement 5), S1–85.
Atkinson, M.A. and Maclaren, N.K. (1994). The pathogenesis of insulin-dependent diabetes mellitus. New England Journal of Medicine, 331, 1428–36.
Bennett, P.H., Bogardus, C., Tuomilehto, J., and Zimmet, P. (1992). Epidemiology and natural history of NIDDM: non-obese and obese. In International textbook of diabetes mellitus (ed. K.G.M.M. Alberti, R.A. DeFronzo, H. Keen, and P. Zimmet), pp. 147–76. Wiley, Chichester.
Bier, S. and Lernmark, A. (1998). Diabetes. In The autoimmune diseases (ed. N.R. Rose and I.R. MacKay) (3rd edn), pp. 431–57. Academic Press, San Diego, CA.
Bingley, P.J., Bonifacio, E., and Gale, E.A.M. (1993). Can we really predict IDDM? Diabetes, 42, 213–20.
Chang, C.-J., Wu, J.-S., Lu, F.-H., Lee, H.-L., Yang, Y.-C., and Wen, M.-J. (1998). Fasting plasma glucose in screening for diabetes in the Taiwanese population. Diabetes Care, 21, 1856–60.
Collins, F.S. (1999). Shattuck Lecture. Medical and societal consequences of the Human Genome Project. New England Journal of Medicine, 341, 28–37.
Davies, M. (1999). New diagnostic criteria for diabetes—are they doing what they should? Lancet, 354, 610–11.
de Courten, M., Bennett, P.H., Tuomilehto, J., and Zimmet, P. (1997). Epidemiology of NIDDM in Non-Europids. In International textbook of diabetes mellitus. (ed. K.G.M.M. Alberti, P. Zimmet, and R.A. DeFronzo) (2nd edn), pp. 143–70. Wiley, Chichester.
de Courten, M., Chitson, P., Cox, H.S., Alberti, K.G.M.M., Tuomilehto, J., and Zimmet, P.Z. (1999). The rise and rise of diabetes in Mauritius 1987–1988. Diabetes, 48 (Supplement 1), A1758.
Diabetes Control and Complications Trial Research Group (1993). The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. New England Journal of Medicine, 329, 977–86.
Diabetes Prevention Program Research Group (1999). The Diabetes Prevention Program. Design and methods for a clinical trial in the prevention of type 2 diabetes. Diabetes Care, 22, 623–34.
Dowse, G. and Zimmet, P. (1993). The thrifty genotype in non-insulin-dependent diabetes. The hypothesis survives. British Medical Journal, 306, 532–3.
Dowse, G.K., Gareeboo, H., Zimmet, P.Z., et al. (1990). High prevalence of NIDDM and impaired glucose intolerance in Indian, Creole and Chinese Mauritians. Diabetes, 39, 390–6.
Dowse, G.K., Spark, R.A., Mavo, B., et al. (1994). Extraordinary prevalence of non-insulin-dependent diabetes mellitus and bimodal plasma glucose distribution in the Wanigela people of Papua New Guinea. Medical Journal of Australia, 160, 767–74.
Dowse, G.K., Gareeboo, H., Alberti, K.G.M.M., et al. (1995). Changes in population cholesterol concentrations and other cardiovascular risk factor levels after five years of the non-communicable disease intervention programme in Mauritius. British Medical Journal, 311, 1255–59.
Eriksson, J., Lindstrom, J., Valle, T., et al. (1999). Prevention of type II diabetes in subjects with impaired glucose tolerance: the Diabetes Prevention Study (DPS) in Finland. Diabetologia, 42, 793–801.
Finch, C.F. and Zimmet, P.Z. (1988). Mortality from diabetes. In The diabetes annual, Vol. 4 (ed. K.G.M.M. Alberti and L.P. Krall), pp. 1–16. Elsevier, Amsterdam.
Hales, C.N. and Barker, D.J.P. (1992). Type 2 (non-insulin-dependent) diabetes mellitus: the thrifty phenotype hypothesis. Diabetologia, 35, 595–601
Hamman, R.F. (1993). Genetic and environmental determinants of non-insulin dependent diabetes mellitus (NIDDM). Diabetes/Metabolism Reviews, 8, 287–338.
Hamman, R.F. (1997). Epidemiology of microvascular complications. In International textbook of diabetes mellitus (ed. K.G.M.M. Alberti, R.A. DeFronzo, and H. Keen) (2nd edn), pp. 1293–319. Wiley, London.
Harris, M.I. and Zimmet, P. (1992). Classification of diabetes mellitus and other categories of glucose intolerance. In The international textbook of diabetes mellitus (ed. H. Keen, R. DeFronzo, K.G.M.M. Alberti, and P. Zimmet), pp. 3–18. Wiley, Chichester.
Harris, M.I., Hadden, W.C., Knowler, W.C., and Bennett, P.H. (1987). Prevalence of diabetes and impaired glucose tolerance and plasma glucose levels in US population aged 20–74 years. Diabetes, 36, 523–34.
Harris, M.I., Klein, R., Welborn, T.A., and Knuiman, M.W. (1992). Onset of NIDDM occurs at least 4–7 years before clinical diagnosis. Diabetes Care, 15, 815–19.
Harris, S.B., Gittelsohn, J., Hanley, A., et al. (1997). The prevalence of NIDDM and associated risk factors in native Canadians. Diabetes Care, 20, 185–7.
Harrison, L.C., Kay, T.W.H., Colman, P.G., and Honeyman, M.C. (1999). Type 1 diabetes: from pathogenesis to prevention. In Diabetes in the new millennium (ed. J.R. Turtle, T. Kaneko, and S. Osato), pp. 85–100. Endocrinology and Diabetes Research Foundation, University of Sydney.
Hattersley, A.T. (1998). Maturity-onset diabetes of the young: clinical heterogeneity explained by genetic heterogeneity. Diabetic Medicine, 15, 15–24.
Hennekens, G.H. and Buring, J.E. (1987). Epidemiology in medicine. Little, Brown, Boston, MA.
Jarrett, R.J. (1992). Epidemiology of macrovascular disease and hypertension in diabetes mellitus. In International textbook of diabetes mellitus (ed. K.G.M.M. Alberti, R.A. DeFronzo, H. Keen, and P. Zimmet), pp. 1459–70. Wiley, Chichester.
Joseph, K.S. and Kramer, M.S. (1996). Review of the evidence on fetal and early childhood antecedents of adult chronic disease. Epidemiology Reviews, 18, 158–74.
Karvonen, M., Tuomilehto, J., Libman, I., and LaPorte, R. (1993). A review of the recent epidemiological data on the worldwide incidence of type 1 (insulin-dependent) diabetes mellitus. Diabetologia, 36, 883–92.
King, H. and Rewers, M. (1993). WHO Ad Hoc Diabetes Reporting Group. Global estimates for prevalence of diabetes mellitus and impaired glucose tolerance in adults. Diabetes Care, 16, 157–77.
King, H., Aubert, R.E., and Herman, W.H. (1998) Global burden of diabetes, 1995–2025. Prevalence, numerical estimates and projections. Diabetes Care, 21, 1414–31.
Klein, R. and Moss, S.E. (1992). Visual impairment and diabetes. In International textbook of diabetes mellitus (ed. K.G.M.M. Alberti, R.A. DeFronzo, H. Keen, and P. Zimmet), pp. 1373–84. Wiley, Chichester.
McCarthy, M.I., Froguel, P., and Hitman, G.A. (1994). The genetics of non-insulin-dependent diabetes mellitus: tools and aims. Diabetologia, 37, 959–68.
Michaelis, D, Jutzi, E., and Vogt, L. (1993). Epidemiology of insulin-treated diabetes mellitus in the East-German population: differences in long-term trends between incidence and prevalence rates. Diabete et Metabolisme, 19, 110–15.
Neel, J.V. (1962). Diabetes mellitus: a thrifty genotype rendered detrimental by ‘progress’? American Journal of Human Genetics, 14, 353–62.
O’Dea, K. (1984). Marked improvement in carbohydrate and lipid metabolism in diabetic Australian Aborigines after temporary reversion to traditional lifestyle. Diabetes, 33, 596–603.
Pan, X., Li, G., Hu, Y., et al. (1997). Effect of diet and exercise in preventing NIDDM in people with impaired glucose tolerance: the Da Qing IGT and Diabetes Study. Diabetes Care, 20, 537–44.
Pyorälä, K., Pedersen, T.R., Kjeksus, J., Faergerman, O., Olsson, A.G., and Thorgeirsson, G. (1997). Cholesterol lowering with simvastatin improves prognosis of diabetic patients with coronary heart disease: a subgroup analyses of the Scandinavian Simvastatin Survival Study (4S). Diabetes Care, 20, 614–20.
Scott, R.S., Brown, L.J., Darlow, B.A., Forbes, L.V., and Moore, M.P. (1992). Temporal variation in incidence of IDDM in Canterbury, New Zealand. Diabetes Care, 15, 895–9.
Shaw, J., de Courten, M., and Zimmet, P. (1999a). The epidemiology of diabetes: a world-wide problem. In Diabetes in the new millennium (ed. J.R. Turtle, T. Kaneko, and S. Osato), pp. 1–9. Endocrinology and Diabetes Research Foundation, University of Sydney.
Shaw, J., de Courten, M., Boyko, E.J., and Zimmet, P. (1999b). Impact of new diagnostic criteria for diabetes in different populations. Diabetes Care, 22, 762–6.
Shaw, J. E., Zimmet, P.Z., de Courten, M., et al. (1999c). Impaired fasting glucose or impaired glucose tolerance. What best predicts future diabetes in Mauritius? Diabetes Care, 22, 399–402.
Shintani, T., Hughes, C.K., Beckham, S., and O’Connor, H.K. (1991). Obesity and cardiovascular risk intervention through the ad libitum feeding of traditional Hawaiian diet. American Journal of Clinical Nutrition, 53, 1647S–51S.
Skyler, J.S. and Marks, J.B. (1993). Immune intervention in type 1 diabetes mellitus. Diabetes Review, 1, 15–42.
Songer, T. (1992). The economic costs of NIDDM. Diabetes/Metabolism Reviews, 8, 389–404.
Tan, C.-E., Emmanuel, S.C., Tan, B.-Y., and Jacob, E. (1999). Prevalence of diabetes and ethnic differences in cardiovascular risk factors. The 1992 Singapore National Health Survey. Diabetes Care, 22, 241–7.
Tuchinda, C., Angsusingha, K., Chaichanwalanakul, K., Likitmaskul, S., and Vannasaeng S. (1992). The epidemiology of insulin dependent diabetes mellitus (IDDM): report from Thailand. Journal of the Medical Association of Thailand, 75, 217–22.
Tuomilehto, J. and Rastenyté, D. (1997). Epidemiology of macrovascular disease and hypertension in diabetes mellitus. In International textbook of diabetes mellitus (ed. K.G.M.M. Alberti, P. Zimmet, R.A. DeFronzo, and H. Keen) (2nd edn), pp. 1559–83. Wiley, London.
Tuomilehto, J., Podar, T., Reunanen, A., et al. (1991). Comparison of incidence of IDDM in childhood between Estonia and Finland, 1980–1988. Diabetes Care, 14, 982–8.
Tuomilehto, J., Podar, T., Brigis, G., et al. (1992a). Comparison of the incidence of insulin dependent diabetes mellitus among five Baltic populations during 1983–1988. International Journal of Epidemiology, 21, 518–27.
Tuomilehto, J., Lounamaa, R., Tuomilehto-Wolf, E., et al. (1992b). Epidemiology of childhood diabetes mellitus in Finland—background of a nationwide study of type 1 (insulin dependent) diabetes mellitus. The Childhood Diabetes in Finland (DiMe) Study Group. Diabetologia, 35, 70–6.
Tuomilehto, J., Knowler, W.C., and Zimmet, P. (1992c). Primary prevention of non-insulin-dependent diabetes mellitus. Diabetes/Metabolism Reviews, 8, 339–53.
Tuomilehto, J., Tuomilehto-Wolf, E., Zimmet, P., Alberti, K.G.M.M., and Knowler, W.C. (1997). Primary prevention of diabetes mellitus. In International textbook of diabetes mellitus (ed. K.G.M.M. Alberti, P. Zimmet, R.A. DeFronzo, and H. Keen) (2nd edn), pp. 1799–1827. Wiley, Chichester.
UKPDS (UK Prospective Diabetes Study Group) (1998a). Intensive blood glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). Lancet, 352, 837–53.
UKPDS (UK Prospective Diabetes Study Group) (1998b). Tight blood pressure control and risk of macrovascular and microvascular complications in type 2 diabetes: UKPDS 38. British Medical Journal, 317, 703–13.
Uusitalo, U., Feskens, E.J.M., Tuomilehto, J., et al. (1996). Fall in total cholesterol concentration over five years in association with changes in fatty acid composition of cooking oil in Mauritius: cross sectional survey. British Medical Journal, 313, 1044–6.
Ward, J.D. (1992). Diabetic neuropathy. In International textbook of diabetes mellitus (ed. K.G.M.M. Alberti, R.A. DeFronzo, H. Keen, and P. Zimmet), pp. 1385–414. Wiley, Chichester.
Warram, J.H., Rich, S.S., and Krolewski, A.S. (1994). Epidemiology and genetics of diabetes mellitus. In Joslin’s Diabetes Mellitus (ed. C.R. Kahn and G.C. Weir) (13th edn), pp. 201–15. Lea & Febiger, Philadelphia, PA.
West, K.M. (1978). Epidemiology of diabetes and its vascular lesions. Elsevier, New York.
WHO Diamond Project Group (1990). WHO Multinational Project for Childhood Diabetes. Diabetes Care, 13, 1062–8.
WHO (World Health Organization) (1985a). Diabetes mellitus. WHO Technical Report Series, No. 727. WHO, Geneva.
WHO (World Health Organization) Multinational Study of Vascular Disease in Diabetics (MNSVDD) WHO (1985b). Prevalence of small vessel disease in diabetic patients from 14 centers. Diabetologia, 28 (Supplement), 615–40.
WHO (World Health Organization) (1994). Prevention of diabetes mellitus. WHO Technical Report Series, No. 844. WHO, Geneva.
WHO (World Health Organization) (1999). Report of a consultation. Definition, diagnosis and classification of diabetes mellitus and its complications. Part 1: Diagnosis and classification of diabetes mellitus. WHO, Geneva.
World Bank (1993). World development report 1993: investing in health. World development indicators. Oxford University Press.
World Health Report (1998). Life in the 21st century. A vision for all. WHO, Geneva.
Yamagata, K., Oda, N., Kaisaki, P.J., et al. (1996a). Mutations in the hepatocyte nuclear factor-1a gene in maturity-onset diabetes of the young (MODY 3). Nature, 384, 455–8.
Yamagata, K., Furuta, H., Oda, N., et al. (1996b) Mutations in the hepatocyte nuclear factor-4a gene in maturity-onset diabetes of the young (MODY 1). Nature, 384, 458–60.
Zimmet, P. (1992). Challenges in diabetes epidemiology—from West to the rest. Diabetes Care, 15, 232–52.
Zimmet, P. (1993). Hyperinsulinaemia—how innocent a bystander. Diabetes Care, 16, 56–70.
Zimmet, P.Z. (1995). The pathogenesis and prevention of diabetes in adults. Diabetes Care, 18, 1050–64.
Zimmet, P. (1999). Diabetes epidemiology as a trigger to diabetes research. Diabetologia, 44, 499–518.
Zimmet, P. and Alberti, K.G.M.M. (1997). The changing face of macrovascular disease in non-insulin dependent diabetes mellitus in different cultures: an epidemic in progress. Lancet, 350 (Supplement 1), S1–S4.
Zimmet, P., Dowse, G., Serjeantson, S., Finch, C., and King, H. (1990). The epidemiology and natural history of NIDDM—lessons from the South Pacific. Diabetes/ Metabolism Reviews, 6, 91–124.
Zimmet, P., Turner, R., McCarty, D., Rowley, M., and Mackay, I. (1999). Crucial points at diagnosis—NIDDM or slow IDDM. Diabetes Care, 22, B59–64.
Osteoporosis: a public health problem
Osteoporosis or bone fragility is characterized by low areal bone mineral density (BMD) and microarchitectural deterioration leading to bone fragility (Consensus Development Conference 1993). Fragility conferred by the microarchitectural abnormalities (trabecular thinning, loss of connectedness, cortical thinning, increased intracortical porosity) is unquantified, so that osteoporosis is currently synonymous with ‘low’ areal BMD.
A WHO panel defined (a) ‘normal’ areal BMD as values less reduced than 1 standard deviation (SD) below the young normal mean, (b) ‘low’ areal BMD or ‘osteopenia’ as values between –1 SD and –2.5 SD, and (c) ‘osteoporosis’ as areal BMD reduced by more than –2.5 SD below the young adult mean (Fig. 10). Severe or ‘established’ osteoporosis is areal BMD reduced by more than –2.5 SD in the presence of fractures (WHO 1994).
Fig. 10 Areal BMD at the lumbar spine in controls (open circles) and women with spine fractures (solid circles) (see text).
It is misleading to think of osteoporosis in dichotomous terms i.e. having or not having ‘osteoporosis’, having or not having ‘low’ areal BMD. Areal BMD is a continuous variable. The lower the areal BMD, the higher is the risk of fracture. The higher the blood pressure, the higher is the risk of stroke (Fig. 11). Areal BMD in patients with fractures and controls overlaps because areal BMD is a measure of risk, not certainty, of fracture (Melton et al. 1990).
Fig. 11 (a)The lower the areal BMD, the higher is the fracture risk. (b) Areal BMD in patients with fractures and controls overlaps, just as (c) systolic and (d) diastolic blood pressure overlap in patients with stroke and controls. (Modified from Hui et al. 1988.)
Osteopenia means mild osteoporosis. It is a radiological term used to alert the doctor to increased fracture risk suggested by radiolucency on radiographs before bone densitometry was available. This term serves no purpose other than suggesting areal BMD is in the lower part of the normal distribution but above the nominal quantitative definition of osteoporosis. ‘Established’ osteoporosis is also entrenched in the literature and implies that the disease is ‘established’ after the fracture. This may lead clinicians to withhold treatment unless the fracture is present (the event that treatment should prevent). The disease is bone fragility (osteoporosis, areal BMD lower than –2.5 SD) whether fracture is present or not. The result of the disease of bone fragility is fracture.
Osteoporosis is a public heath problem because of the morbidity, mortality, and financial burden imposed by fractures. The incidence of fracture rises with age and is higher in men than in women before 50 years of age because of trauma. After 50 years of age, bone fragility is responsible for the increasing fracture incidence in women and men. Lifetime fracture risk is 30 per cent in women and 15 per cent in men (Fig. 12) (Cooper and Melton 1992).
Fig. 12 The age-specific increase in hip, spine, and distal forearm fractures in men and women in Rochester, Minnesota. (Reproduced from Cooper and Melton 1993.)
There are over 325 million individuals over 65 years of age in the world. This will increase to over 1500 million by 2050. The increase will be seen particularly in Asia (Fig. 13). The lifetime risk for hip fracture is 15 per cent in women and 7 per cent in men. In 1990, 30 per cent of the 1.66 million hip fractures worldwide occurred in men. By 2025, the number of hip fractures in men will be similar to the number in women in 1990, a burden compounded by the increases in women in 2025 (Cooper et al. 1992a). The incidence of hip fracture differs more between countries than between sexes, suggesting that there may be factors other than the menopause responsible for the differing incidence of fracture in women and men.
Fig. 13 The expected increase in hip fractures in 2025 will be seen particularly in Asia (upper panels). Thirty per cent of hip fractures occur in men. By 2025, the number of hip fractures in men will be similar to the number in women in 1990 (lower panel). (Modified from Cooper et al. 1992a.)
Barrett et al. (1999) estimated the risk of a person aged 65 years suffering a fracture based on data from 583 256 men and 838 507 women. The risks of hip fracture at ages 75, 80, 85, and 90 years were as follows: 2.9, 6.4, 11.2, and 16.3 per cent in white women; 1.1, 2.3, 3.6, and 5.4 per cent in black women; 1.3, 2.6, 4.1, and 5.5 per cent in white men; 0.9, 1.4, 2.0, and 2.6 per cent in black men. White women had three- to fourfold higher risks than white men and 1.5- to fourfold higher risks than black women.
Kannus et al. (1999) reported that the number of hip fractures in patients over 50 years of age admitted to a Finnish hospital between 1970 and 1997 increased from 1857 to 7122. The incidence (per 100 000) increased from 163 to 438 (a 169 per cent rise). Median age at fracture increased from 76 to 82 years in women and from 70 to 76 years in men. After age adjustment, the incidence increased from 292 to 767 in women and from 112 to 233 in men (rises of 60 per cent and 108 per cent respectively). Age-specific incidences rose in all age groups, with greatest increases in older age groups. In contrast, Huusko et al. (1999) reported no changes in hip fracture incidence between the period 1982 to 1983 and the period 1992 to 1993. Crude incidences (per 1000 per year) in men and women in 1992 to 1993 were 1.56 and 3.23 and respective incidences were 10.24 and 16.50 in 1982 to 1983. On stratification by age and gender, respective mean annual rates per 1000 in women aged 45 to 54 years, 55 to 64 years, 65 to 74 years, 75 to 84 years, and 85+ years in the two periods were 0.5 versus 0.5, 2.7 versus 1.8, 10.4 versus 8.1, 24.0 versus 24.2, and 3.8 versus 3.9. The authors concluded that there had been a change in the age distribution, not in the age-adjusted incidence, within the last 10 years.
Hagino et al. (1999) reported increasing age-specific incidence rates for hip fractures in Japan between 1986 and 1994 (41 to 57 per 100 000 for men and 114 to 145 per 100 000 for women). These rates were similar to those in southern Europe but lower than in white populations in North America and northern Europe. Yan et al. (1999) reported 453 hip fractures in 612 170 Chinese aged 50 years or older in 1994. The age-adjusted 1-year cumulative incidence rate was 67 per 100 000 in women and 81 per 100 000 in men, lower than the rates reported in the United States in 1985 (87 and 100 per 100 000 in women and men respectively). The frequency of fracture by bicycle accident was 10 per cent in women and 28 per cent in men.
Memon et al. (1998) reported all new hip fracture cases between 1992 and 1995 in Kuwait; rates were 295 per 100 000 in women and 200 per 100 000 in men, similar to that reported in European women and in Asian women living in the United States. The rate in Kuwaiti men is almost equal to that in white American men; rates for men in other Asian countries are half or less.
The differing incidence of hip fracture may be partly due to flawed case ascertainment
Ascertainment errors may be partly responsible for the disparate results (Bacon et al. 1996). ‘Incidence’ is often calculated from discharge records. Failure to identify all cases (e.g. readmissions) may introduce errors. Small sample sizes, particularly in those over 80 years old, may result in an unstable numerator while the population forming the denominator in the calculation is often derived from dated census data. Groupings into 75+, 80+, and 85+ years are often made. Most hip fractures occur at this time and the mean age and distribution within these classifications will be changing with time. Thus the rising incidence may be an artefact of better surveillance. Retrospective analyses may result in underestimates.
Schwartz et al. (1999) calculated hip fracture incidence rates from discharge records during the period 1990 to 1992. Rates were lowest in Beijing (45.4 for men, and 39.6 for women) and highest in Reykjavik (141.3 for men, and 274.1 for women). Women had higher rates than men after age 65 years except in Beijing. Review of operating theatre or radiology logs increased the estimated number of hip fractures by 11 per cent. Miscoding increased the number by 1 per cent. When all sources of undercount and overcount were included, final estimates of hip fracture incidence ranged from 15 per cent lower to 89 per cent higher than the original rates based on discharge diagnoses.
If the differing incidences of hip fracture across time, from country to country, and between genders are correct, differences in the incidence of falls, severity of trauma, or bone fragility must be responsible. An increase in the age-specific incidence of falls may be occurring because of a higher prevalence of ill health and use of sedatives, or greater age-related bone loss in persons born more recently, producing greater bone fragility than that seen in persons born earlier.
Irrespective of whether there is a true increase in incidence of hip fracture during the last 20 years, it is clear that the number of hip fractures in the community is increasing because of the increasing number of individuals living into old age. The problem will place an increasing financial burden on the community.
The direct annual costs of hip fractures are £614 million in the United Kingdom, 3.7 billion francs in France, and $10 billion in the United States (Barrett-Connor 1995). Hospital admission accounts for 44 per cent of these costs. In Switzerland, 50 per cent of the direct costs are incurred in the first 18 days of admission. Thirty per cent of the hospital stay is due to waiting for nursing home availability. About 50 per cent of hip-fracture survivors are discharged to nursing homes. In the United States, $2 billion of the $5 billion direct annual costs for long-term care are attributable to nursing home care.
The epidemiology of spine fractures is less well documented for the the following reasons: (a) about 65 per cent of spine fractures do not reach clinical attention (Cooper et al. 1992c; Jacobsen et al. 1992); (b) the definition of ‘deformity’ or ‘fracture’ is unclear (Kanis 1994); (c) most data are prevalence not incidence, and a proportion of the fractures may be due to trauma in youth.
These problems may explain why the prevalence and incidence of vertebral fractures varies enormously. In Rochester, Minnesota, the prevalence in women was about 20 per cent (over 50-year-olds), 30 per cent (over 65-year-olds), and 50 per cent (over 85-year-olds) (Cooper et al. 1992b). Jensen et al. (1982) reported a prevalence of wedge fractures of about 20 per cent in 70-year-old women. Five per cent had crush fractures. Härmä et al. (1986) reported prevalences of about 0.5 to 3 per cent in 57 440 subjects. The incidence (per 100 person-years) in women in Rochester, Minnesota, was 0.5 in 50- to 54-year olds and 3.8 in 85- to 89-year olds; among persons aged 35 to 69 years, fracture rates were 0.7 per 1000 person-years in men, and 1.5 per 1000 person-years in women (Cooper et al. 1992b).
Vertebral fractures are a health problem in men. In Dubbo, New South Wales, the prevalence of vertebral fractures was higher in men than women, being 25 per cent versus 20 per cent, 17 per cent versus 12 per cent, and 27 per cent versus 25 per cent depending on ‘fracture’ criteria (Jones et al. 1994). In the United Kingdom, the prevalence was similar in 15 570 men and women—12 per cent or 20 per cent depending on the criteria used (O’Neill et al. 1996). In Korea, the prevalence was 12.5 per cent in men and 20 per cent in women. Elderly men had fewer crush fractures (3.7 per cent) than women (15.3 per cent) (Tsai et al. 1996). In Rotterdam, the prevalence in 724 men and 717 women was 8 per cent and 7 per cent respectively (moderate deformities), and 4 per cent and 10 per cent respectively (severe deformities) (Burger et al. 1997). In Nebraska, the prevalence in 529 men and 899 women aged 50 to 70 years was 29 per cent in men and 10 per cent in women, and 39 per cent and 45 per cent respectively in 80-year-olds (Davies et al. 1996). If vertebral fracture incidence is similar in men and women, the notion of osteoporosis as a disease of women must be discarded. Prospectively derived vertebral fracture incidence data are needed.
The incidence of forearm fractures increases after the menopause and reaches a plateau after 60 years of age. In men, the incidence is low throughout life (Fig. 12). In the Trent region in the United Kingdom, 84.2 per cent of all forearm fractures occurred in women. Rates (per 10 000 per year) were 9 in men and 42 in women (Kanis and McCloskey 1992). In Sweden 850 of the 934 forearm fractures occurred in women (Alffram et al. 1962). The fractures are rarely fatal, and usually do not require hospital admission.
Mortality from fractures
The 20 per cent excess death rate after hip fracture occurs within the first 6 to 12 months. Mortality is then no different to the age-predicted value. Mortality after hip fracture is higher in men than women, perhaps due to greater comorbidity. Survival in men with no comorbid conditions before hip fracture is no different to controls. In those with a high comorbidity index, survival is reduced in hip fracture cases and controls but it is greater in the cases. The excess mortality associated with vertebral fractures is about 20 per cent after 5 years and increases with time (Cooper and Melton 1992).
Bone density—a surrogate for bone strength
The highest incidence of fractures is 3 to 4 per 100 per year in women aged 80+ in the community reaching 10 to 12 per 100 per year in the highest-risk groups (women in nursing homes, women with very low areal BMD and multiple fractures). Thus neither the occurrence nor the absence of a fracture in an individual provides reliable information regarding causal relationships between a risk factor or treatment and fractures. That is, fractures, the true measure of fragility, cannot be used as an endpoint to study the pathogenesis, prevention, and efficacy treatment of fractures in an individual. This can only be determined in clinical trials.
Areal BMD is a surrogate of bone strength. In vivo and in vitro, the areal BMD measurement correlates with bone strength and is the best available measure of fracture risk. For a bone mass of 0.6 g/cm, fracture risk (per 1000 person-years) doubles from about 60 to 120 as age increases from 70 to 80 years (Fig. 14). At 70 years, fracture risk increases from 80 to 100 as bone mass decreases from 0.7 to 0.6 g/cm. Age and areal BMD are independent predictors of fracture risk (Hui et al. 1988). Age may incorporate factors such as falls and structural changes such as trabecular connectivity.
Fig. 14 (a)At any bone mass, fracture risk increased as age increased. (b) For any age, fracture risk increased as bone mass decreases. (Reproduced from Hui et al. 1988.)
Bone ‘mass’ and ‘density’—problems in nomenclature
Non-invasive methods measure ‘apparent’ areal BMD—a mass of mineral contained within an area or volume of tissue, not all of which is bone (Seeman 1997). Bone mass is expressed uncorrected for size (g), or an areal BMD corrected for length and width (g/cm2) but not depth. Quantitative CT measures mass as a volumetric BMD correcting for length, width, and depth (mg/cm3). This volumetric BMD measurement is also an ‘apparent’ density, although the word ‘apparent’ is omitted from the literature. Quantitative CT is erroneously claimed to be a ‘true’ density because it is volumetric. Implicit in the context of ‘true’ is the notion that this volumetric BMD is a better surrogate of bone strength than areal BMD measurements. There is no evidence for this.
‘Areal’, ‘volumetric’, and ‘apparent’ are dropped for convenience, but at the price of understanding. An increase in ‘density’ occurs through a change in structure within a length, area, or volume of tissue comprised of marrow, fat, cells, and fluid-filled spaces. In a unit of bone tissue, a doubling in the number of trabeculae produces the same increase in density as a doubling of the thickness of existing trabeculae (Fig. 15). The former may be more advantageous biomechanically before menopause, but less so after the menopause because thinner trabeculae may be more susceptible to perforation. Similarly, a doubling of cortical density may be the result of increased endocortical (medullary) or periosteal bone formation. The latter results in a bone more resistant to bending.
Fig. 15 A doubling in trabecular bone density may be due to doubling of the thickness or number of trabeculae. A doubling of cortical density may be the result of increased endocortical (medullary) or periosteal bone.
The failure to consider the structural basis of changes in ‘density’ results in serious misunderstanding (Seeman 1997). For example, BMD does not increase in growth; bone size increases. The amount of bone calcium (grams per 100 g fat-free tissue) is established in early intrauterine life and is constant from early in gestation to old age. Peak bone mass is less in women than men because women have smaller bones. Peak volumetric BMD is the same in men and women. These concepts are vital to the understanding of the structural basis of increasing bone ‘density’ during growth, bone loss during ageing, and the effects of treatment on bone strength.
Changes in bone density during growth and advancing age
Patients with fractures have reduced areal BMD relative to controls because of excessive bone loss or low peak areal BMD. Understanding of the pathogenesis of bone fragility or osteoporosis involves the study of risk and protective factors influencing the skeleton throughout life—during growth, adulthood, and old age.
Growth and peak bone density
Total body calcium increases during growth because the bone size increases. However, the growing bone gains more bone within it (as cortical thickness increases, as trabeculae are formed at the growth plate and then thicken) so that volumetric BMD is constant before puberty and then increases around puberty as trabeculae thicken further and cortical thickness increases by endocortical contraction, i.e. growth builds a larger skeleton but only a slightly more dense skeleton. Dunnill et al. (1967) showed that vertebral size increased from birth to young adulthood and is greater in men than in women. From about 2 years of age, the volume of bone within the vertebral body is constant—’density’ is constant (Fig. 16).
Fig. 16 (a)Vertebral volume increases during growth in females (solid circles) and males (open circles). (b) The volume of bone in vertebral body (bone density) is constant from early life. (Modified from Dunnill et al. 1967.)
Growth of the axial and appendicular skeleton differ in timing and regulation. At the end of the first year of life, trunk length increases at a constant velocity while growth of the legs accelerates. At puberty, growth of the trunk accelerates while growth of the legs decelerates (Tupman et al. 1962) (Fig. 17). Diseases occurring at different times of growth may result in site-specific deficits. Regions further from their peak or growing more rapidly may be more severely affected by exposure to disease than regions nearer completion of growth. Thus the effect of a risk factor on the growing skeleton depends on the developmental stage at which exposure occurs as well as the exposure ‘dose’ (Bass et al. 1999).
Fig. 17 At puberty, growth of the trunk accelerates while growth of the femur decelerates. (Modified from Tupman 1962.)
This heterogeneity in growth and mineral accrual may contribute to the wide normal range for BMD. The variance in peak lumbar spine BMD is ± 20 per cent of the mean. An understanding of the factors explaining this variance is important because the difference between BMD at the 5th and the 95th percentile is greater than the difference between BMD in patients with fractures and age-matched controls. Finding reduced lumbar spine BMD in daughters of women with spine fractures, and reduced femoral neck BMD in the daughters of women with hip fractures, is consistent with the view that a low peak BMD may be a sufficient explanation for the lower BMD found in patients with fractures (Seeman et al. 1989, 1994).
Genetics of osteoporosis
About 80 per cent of the variance in areal BMD is explained by genetic factors. Many polymorphisms of candidate genes have been proposed, but few if any reproducible data exist that identify any gene or gene product that accounts for this variance (Seeman 1999). Heritability is the proportion of the total variance (genetic plus environmental) in areal BMD attributable to genetic factors. The statement ’80 per cent of areal BMD is genetically determined leaving only 20 per cent to modify’ is flawed; heritability is not a constant proportion leaving 20 per cent ‘due to’ environmental factors that can be changed. The size of the total population variance depends on the factors chosen to describe the trait mean (such as age, gender, height, body composition). If total variance increases due to an increase in the environmental variance, without change in the genetic variance, heritability will decrease. Thus, ‘heritability’ (a proportion) and ‘genetic variance’ (an absolute) may give different impressions of the ‘strength’ of genetic factors.
Areal BMD is not a specific and unambiguous phenotype with identifiable physiological control mechanisms. It is the net result of the modelling and remodelling on the periosteal and endosteal surfaces that give the bone its mass, size, and architecture. Each surface behaves differently during growth and ageing because each is regulated differently. The ambiguity of areal BMD makes this an unsatisfactory phenotype for the identification of the genes that contribute to the regulation of skeletal growth and ageing. No gene, gene product, or gene polymorphism has been reproducibly, and therefore credibly, shown to account for a given proportion of the variance in areal BMD. The data concerning candidate markers such as polymorphisms of the vitamin D receptor, oestrogen receptor, and type 1 collagen genes are inconsistent, perhaps partly because of the questionable value of this phenotype, the use of genetic markers of uncertain biological function, flaws in study design such as small sample sizes, failure to account for confounding, lack of stratification and/or randomization prior to intervention, reliance on statistical adjustment rather than study design, and the use of post hoc analyses to infer causation. Distinct morphological structures (trabecular number, thickness, periosteal and endocortical width, cortical thickness) should be quantified by age-, gender-, and race-specific means and variances. Genetic and environmental factors should be sought that explain variance in these structures (Seeman 1999).
Bone loss—an imbalance in bone formation and resorption
After completion of skeletal growth, remodelling occurs at discrete sites throughout the skeleton (basic multicellular units) replacing old with new bone. Osteoclasts resorb a quantum of bone at the basic multicellular units forming a Howship lacuna on the surface of trabecular bone (Fig. 18) (Mosekilde et al. 1990). Osteoblasts lay down osteoid which undergoes primary and secondary mineralization.
Fig. 18 Electron micrographs showing bone resorption by an osteoclast (upper left), an intact trabeculum (upper right), and a trabeculum undergoing erosion (bottom right) with loss of connectivity (bottom left). (Modified from Mosekilde et al. 1993.)
Bone ‘loss’ is not only resorptive removal of bone, it is also failure of restoration. This imbalance between bone resorption and bone formation at the basic multicellular units is the structural requirement for bone loss. If the amount of bone removed and replaced is the same, no loss of bone occurs. Bone loss occurs if formation is reduced when resorption is normal and when resorption is increased and formation is either normal or reduced. At the menopause, bone turnover increases—there are more basic multicellular units activated and there is focal imbalance in each. Why bone formation does not ‘keep up’ is not understood.
Remodelling occurs on endosteal (endocortical, trabecular, and intracortical) and periosteal surfaces (Parfitt 1994). Trabecular bone loss is more rapid than cortical bone loss because of its greater surface area. Trabecular bone remodelling with imbalance at the basic multicellular units results in thinning, perforation, and loss of trabecular connectivity. Trabecular bone loss contributes progressively less to the overall bone loss because trabeculae disappear and the trabecular surface area for remodelling decreases.
Cortical bone loss is slower and is the net result of endocortical resorption, intracortical bone loss producing porosity, and periosteal bone formation which partly compensates for the endocortical bone resorption and results in increased bone diameter. With advancing age, cortical bone loss accounts for an increasing proportion of the overall bone loss. Continued endocortical bone resorption and increased cortical porosity increase the surface for resorption in cortical bone, ‘trabecularizing’ cortical bone. Bone loss does not decelerate in old age, it accelerates (Foldes et al. 1991; Jones et al. 1996). About 70 per cent of the total bone lost during ageing in women is cortical because 80 per cent of the skeleton is cortical bone (Sandor et al. 1992).
Gender differences in bone loss
The net amount of total body calcium lost in women is 100 to 150 g more than in men. As men have larger bones, the net loss is 30 to 40 per cent in women but only 10 to 15 per cent in men. The age-related diminution in trabecular bone of the iliac crest in women and men is similar (Fig. 19). Women lose more bone than men because cortical—not trabecular—bone loss is greater (Kalender et al. 1989). Cortical bone loss is greater because endocortical resorption is greater in women than in men, and periosteal bone formation is greater in men than in women. The percentage of trabecular surface undergoing resorption increases in women. Trabecular numbers fall due to perforation and loss of connectivity (Fig. 20). Trabecular bone loss in men occurs primarily by reduced bone formation and a fall in trabecular width.
Fig. 19. Upper panels show the similar diminution in vertebral body trabecular bone in men (left) and women (right) by quantitative CT. The diminution in vertebral cortical BMD (lower panels) is less in men (left) than in women (right). (Single energy—open circles, broken regression lines; dual energy—solid circles, continuous regression lines.) (Reproduced from Kalender et al. 1989.)
Fig. 20 Bone loss in women occurs by increased resorption, a fall in trabecular number, and loss of connectivity (left upper and lower panels). Bone loss occurs by reduced bone formation (mean wall thickness) and a fall in trabecular width in men (right upper and lower panels). (Modified from Aaron et al. 1987.)
Classification and causes of osteoporosis
Primary and secondary osteoporosis
The most common types of fractures—spine, forearm, and hip—in women and men have no identified cause. They are the morbid event of ‘primary’ or ‘idiopathic’ osteoporosis. Postmenopausal (type 1) osteoporosis refers to the occurrence of spine and forearm fractures within the first 20 to 25 years of menopause. Loss of trabecular connectivity associated with the increased bone turnover after the menopause may be responsible for the bone fragility. Senile (type 2) osteoporosis refers to the occurrence of hip fractures in men and women 75 years and older. Age-related loss of both trabecular and cortical bone contribute to bone fragility (Riggs and Melton 1983).
‘Secondary’ refers to osteoporosis occurring in the setting of an illness or a recognized risk factor for osteoporosis. Illnesses that may be associated with osteoporosis but contribute little to the public health problem of osteoporosis include acromegaly, primary hyperparathyroidism, osteogenesis imperfecta, Cushing’s disease, hypopituitarism, and multiple myeloma. Myeloma, malabsorption, primary hyperparathyroidism, and hypogonadism may have no clinical features and must be regarded with a high index of suspicion.
Risk factors for osteoporosis include age, hypogonadism, nulliparity, immobility, Caucasian or Asian race, tobacco use, alcohol abuse, sedentary living, low calcium intake, high caffeine, protein, and salt intakes, low body weight, a family or past history of fractures, small bones, and drugs (corticosteroids, anticonvulsants, cyclosporin, thyroxine, and heparin). Protective factors include black race, osteoarthritis, lactation, multiparity, high calcium intake, exercise, obesity, fluoridated drinking water, exposure to oestrogens, and thiazides.
One of the most important risk fractors for fractures independent of areal BMD or age is the presence of one or more spine fractures. Black et al. (1999) reported that the risk of sustaining further spine fractures in the presence of a prevalent spine deformity was 5.4. The risk of hip and of any non-spine fracture was 2.8 and 1.9 respectively in women with baseline deformity. The risk for new spine fractures increased as the number of baseline deformities increased, from 3.2 for one fracture to 10.6 for three or more fractures. Women with most severe deformities had a relative risk of 12.7 for a new deformity.
Risk factors for hip fracture include a maternal history of hip fracture, low body weight, tall stature, previous fractures, hyperthyroidism, inactivity, use of benzodiazepines, anticonvulsants, and caffeine, and reduced muscle strength. The hip fracture incidence (per 1000 per year) was 1.1 (95 per cent confidence intervals, 0.5, 1.6) in women with no more than two risk factors and normal BMD, and 27 (95 per cent confidence intervals, 20, 34) with five or more risk factors and low BMD (Fig. 21) (Cumming et al. 1995).
Fig. 21 The incidence of hip fractures (per 1000 women per year) was 1.1 (95 per cent confidence interval: 0.5, 1.6) in women with no more than two risk factors and normal BMD, and 27 (95 per cent confidence interval: 20, 34) in women with five or more risk factors and low BMD. (Reproduced from Cumming et al. 1995.)
Hypogonadism is an important ‘risk factor’ for osteoporosis in women and men. Delayed puberty, Kalman’s syndrome, hyperprolactinaemia, anorexia nervosa, premature ovarian failure, leanness, tobacco use, use of luteinizing hormone reducing hormone analogues, and excessive exercise are associated with reduced areal BMD in part due to oestrogen deficiency in women. Testosterone falls with age in men. Low testosterone was found in 48 per cent of men aged over 50 years and in 59 per cent of men with hip fractures. Illnesses to be considered in men with fractures associated with hypogonadism include idiopathic hypogonadism, pituitary tumours, Klinefelter’s syndrome, hyperprolactinaemia, anorexia nervosa, excessive exercise, haemochromatosis, and exogenous corticosteroids (Seeman 1995).
Reduced regional peak areal BMD establishes the relevance of age-related and sex hormone dependent bone loss. The age of onset, incidence, and type of fracture are determined by the rate and duration of cortical and trabecular bone loss. Regions dependent on trabecular bone for strength, such as the vertebral body, may not tolerate trabecular bone loss following the menopause if peak BMD is reduced. Fragility in predominantly cortical regions such as the proximal femur may emerge later, when age-related bone loss reduces cortical thickness and increases cortical porosity. In the presence of bone fragility, the type of fracture and the age of occurrence of the fracture partly depend on the age-specific pattern of trauma and falls. The mobility of middle age may predispose to spontaneous vertebral fractures, or forearm fractures due to falls defended by the outstretched hand. The poverty of forward movement in the elderly may favour an undefended fall onto the lateral hip region.
From a public health point of view, the problem is hip fractures in both women and men, and vertebral fractures in women, and probably in men. Falls in the elderly are a public health problem and certain questions still remain.
Why is there an increase in hip fracture incidence in some countries?
Why does the incidence of hip fractures vary so greatly between countries?
What are the determinants of bone fragility other than low areal BMD?
Can nutritional and lifestyle factors be modified in the community? If so, will this result in fewer fractures in the community?
Will starting treatment later in life be more cost-effective than treatment at menopause?
A public health policy regarding the use of hormone replacement therapy or any drug for all women, population-based BMD screening, or screening for hypogonadism in men, should not be advocated until concerns regarding the risks, efficacy, and compliance are resolved. Success—a reduction in fracture rates in the community—will be measured by the epidemiologist.
Aaron, J.E., Makins, N.B., and Sagreiy, K. (1987). The microanatomy of trabecular bone loss in normal aging men and women. Clinical Orthopaedics and Related Research, 215, 260–71.
Alffram, P.-A. and Bauer, G.C.H. (1962). Epidemiology of fractures of the forearm. Journal of Bone and Joint Surgery, American Volume, 44A, 105–14.
Bacon, W.E., Maggi, S., Looker, A., et al. (1996). International comparison of hip fracture rates in 1988–89. Osteoporosis International, 6, 69–75.
Barrett-Connor, E. (1995). The economic and human costs of osteoporotic fracture. American Journal of Medicine, 98 (Supplement 2A), 3S–8S.
Barrett, J.A., Baron, J.A., and Karagas, M.R. (1999) Fracture risk in the US Medicare population. Journal of Clin Epidemiology, 52, 243–9.
Bass, S., Delmas, P.D., Pearce, G., Hendricke, E., Tabensky, A., and Seeman, E. (1999). The differing tempo of growth in bone size, mass and density in girls is region-specific. Journal of Clinical Investigation, 104, 795–804.
Black, D.M., Arden, N.K., Palermo, L., Pearson, J., and Cummings, S.R., for the Study of Osteoporotic Fractures Research Group (1999). Prevalent vertebral deformities predict hip fractures and new vertebral deformities but not wrist fractures. Journal of Bone and Mineral Research, 14, 821–8.
Burger, H., van Daele, P.L.A., Grashuis, K., et al. (1997) Vertebral deformities and functional impairment in men and women. Journal of Bone and Mineral Research, 12, 152–7.
Close, J., Ellis, M., Hooper, R., Glucksman, E., Jackson, S., and Swift, C. (1999) Prevention of Falls in the Elderly Trial (PROFET): a randomised controlled trial. Lancet, 353, 93–7.
Consensus Development Conference (1993). Diagnosis, prophylaxis, and treament of osteoporosis. American Journal of Medicine, 94, 646–50.
Cooper, C. and Melton, L.J., III (1992). Epidemiology of osteoporosis. Trends in Endocrinology and Metabolism, 3, 224–9.
Cooper, C., Campion, G., and Melton, L.J., III (1992a). Hip fractures in the elderly: a world-wide projection. Osteoporosis International, 2, 285–89.
Cooper, C., Atkinson, E.J., O’Fallon, W.M., and Melton, L.J., III (1992b). Incidence of clinically diagnosed vertebral fractures. A population-based study in Rochester, Minnesota, 1985–1989. Journal of Bone and Mineral Research, 7, 221–7.
Cooper, C., Atkinson, E.J., Kotowicz, M., et al. (1992c). Secular trends in the incidence of postmenopausal osteoporosis. Calcified Tissue International, 51, 100–4.
Cumming, S.R., Nevitt M.C., Browner W.S., et al., for the Study of Osteoporotic Fractures Research Group (1995). Risk factors for hip fracture in white women. New England Journal of Medicine, 332, 767–73.
Davies, K.M., Stegman, M.R., Heaney, R.P., and Recker, R.R. (1996). Prevalence and severity of vertebral fracture: the Saunders County Bone Quality Study. Osteoporosis International, 2, 160–5.
Dunnill, M.S., Anderson, J.A., and Whitehead, R. (1967). Quantitative histological studies on age changes in bone. Journal of Pathology and Bacteriology, 4, 275–91.
Foldes, J., Parfitt, A.M., Shin, M-S., Rao, D.S., and Kleerekoper, M. (1991). Structural and geometric changes in iliac bone: relationship to normal aging and osteoporosis. Journal of Bone and Mineral Research, 6, 759–66.
Hagino, H., Yamamoto, K., Ohshiro, H., Nakamura, T., Kishimoto, H., and Nose, T. (1999). Changing incidence of hip, distal radius, and proximal humerus fractures in Tottori Prefecture, Japan. Bone, 24, 265–70.
Härmä, M., Heliövaara, M., Aromaa, A., and Knekt, P. (1986). Thoracic spine compression fractures in Finland.Clinical Orthopaedics and Related Research, 205, 188–94.
Hui, S.L., Slemenda, C.W., and Johnston, C.C., Jr (1988). Age and bone mass as predictors of fracture in a prospective study. Journal of Clinical Investigation, 81, 1804–9.
Huusko, T.M., Karppi, P., Avikainen, V., Kautiainen, H., and Sulkava, R. (1999). The changing picture of hip fractures: dramatic change in age distribution and no change in age-adjusted incidence within 10 years in Central Finland. Bone, 24, 257–9.
Jacobsen, S.J., Cooper, C., Gottlieb, M.S., et al. (1992). Hospitalization for vertebral fracture among the aged: a national population-based study, 1986–1989. Epidemiology, 3, 515–18.
Jensen, G.F., Christianson, C., Boese, N.J., et al. (1982). Epidemiology of post-menopausal spinal and long bone fractures: a unifying approach to post-menopausal osteoporosis. Clinical Orthopaedics and Related Research, 166, 75–81.
Jones, G., Nguyen, T., Sambrook, P., Kelly, P.J., and Eisman, J.A. (1994). Progessive loss of bone in the femoral neck in elderly people: longitudinal findings from the Dubbo Osteoporosis Epidemiology Study. British Medical Journal, 309, 691–5.
Jones, G., Nguyen, T., Sambrook, P.N., Kelly, P.J., Gilbert, C., and Eisman, J.A. (1996). Symptomatic fracture incidence in elderly men and women: the Dubbo Osteoporosis Epidemiology Study (DOES). Osteoporosis International, 4, 277–82.
Kalender, W.A., Felsenberg, D., Louis, O., et al. (1989). Reference values for trabecular and cortical vertebral bone density in single and dual-energy quantitative computed tomography. European Journal of Radiology, 9, 75–80.
Kanis, J.A. (1994). What constitutes evidence for drug efficacy in osteoporosis? Drugs and Aging, 3, 391–9.
Kanis, J.A. and McCloskey, E.V. (1992). Epidemiology of vertebral osteoporosis. Bone, 13, S1–S10.
Kannus, P., Niemi, S., Parkkari, J., Palvanen, M., Vuori, I., and Jorvinen, M. (1999). Hip fractures in Finland between 1970 and 1997 and predictions for the future. Lancet, 353, 802–5.
Maggi, S., Kelsey, J.L., Litvak, J., and Heyse, S.P. (1991). Incidence of hip fractures in the elderly: a cross-national analysis. Osteoporosis International, 1, 232–41.
Melton, L.J., III, Eddy, D.M., and Johnston, C.C., Jr (1990). Screening for osteoporosis. Annals of Internal Medicine, 112, 516–28.
Melton, L.J,. Lane, A.W., Cooper, C., et al. (1993). Prevalence and incidence of vertebral deformities. Osteoporosis International, 3, 113–19.
Memon, A., Pospula, W.M., Tantawy, A.Y., Abdul-Ghafar, S., Suresh, A., and Al-Rowaih, A. (1998). Incidence of hip fracture in Kuwait. International Journal of Epidemiology, 27, 860–5.
Mosekilde, L. (1990). Consequences of the remodelling process for vertebral trabecular bone structure: a scanning electron microscopy study (uncoupling of unloaded structures). Bone and Mineral, 10, 13–35.
O’Neill, T.W., Felsenberg, D., Varlow, J., Cooper, C., Kanis, J.A., and Silman, A.J. (1996). The prevalence of vertebral deformity in European men and women: the European Vertebral Osteoporosis Study. Journal of Bone and Mineral Research, 11, 1010–18.
Parfitt, A.M. (1994). Osteonal and hemi-osteonal remodelling: the spatial and temporal framework for signal traffic in adult human bone. Journal of Cellular Biochemistry, 55, 273–86.
Riggs, B.L. and Melton, L.J., III (1983). Evidence for two distinct syndromes of involutional osteoporosis. American Journal of Medicine, 75, 899–901.
Sandor, T., Felsenberg, D., Kalender, W.A., Clain, A., and Brown, E. (1992). Compact and trabecular components of the spine using quantitative computed tomography. Calcified Tissue International, 50, 502–6.
Schwartz, A.V., Kelsey, J.L., Maggi, S., et al. (1999) International variation in the incidence of hip fractures: cross-national project on osteoporosis for the World Health Organization program for research on aging.Osteoporosis International, 9, 242–53.
Seeman, E. (1994a). Relative contributions of low peak bone mass and bone loss to reduced bone density in osteoporosis. Osteoporosis International, 4, 15–25.
Seeman, E. (1994b). Osteoporosis in men. American Journal of Medicine, 95, 22–8.
Seeman, E. (1995). The dilemma of osteoporosis in men. American Journal of Medicine, 98 (Supplement 1A), 75S–87S.
Seeman E. (1997). From density to structure: growing up and growing old on the surfaces of bone. Journal of Bone and Mineral Research, 12, 1–13.
Seeman, E. (1999). The genetic determination of the population variants in bone mineral density. In The aging skeleton (ed. C. Rosen), pp. 77–94, Academic Press, New York.
Seeman, E., Hopper, J.L., Bach, L.A., Cooper, M., McKay, J., and Jerums, G. (1989). Reduced bone mass in daughters of women with osteoporosis. New England Journal of Medicine, 320, 554–8.
Seeman, E., Hopper, J.L., Tsalamandris, C., and Formica, C. (1994). Bone density in daughters of women with hip fractures. Journal of Bone and Mineral Research, 9, 739–43.
Tsai, K.-S., Twu, S.-J., Chieng, P.-U., Yang, R.-S., and Lee, T.-K. (1996). Prevalence of vertebral fractures in Chinese men and women in urban Taiwanese communities. Calcified Tissue International, 59, 249–53.
Tupman, G.S. (1962). A study of bone growth in normal children and its relationship to skeletal maturation. Journal of Bone and Joint Surgery, British Volume, 44B, 42–67.
WHO (World Health Organization) (1994). Assessment of fracture risk and its application to screening for postmenopausal osteoporosis, pp. 1–129. Technical Report Series 843. WHO, Geneva.
Yan, L., Zhou, B., Prentice, A., Wang, X., and Golden, M.H.N. (1999). Epidemiological study of hip fracture in Shenyang, People’s Republic of China. Bone, 24, 151–5.