Body mass index and nutrition in the aetiology of osteoporosis


Musculo-skeletal disorders are the most frequently reported long-standing illness or disability in Britain. Osteoporosis is a subheading of these disorders frequently first diagnosed following fracture. British research shows body mass index (BMI) to be the most statistically significant risk factor, showing an increase by a factor of 6.7 for hip fractures as BMI fell from 25 to under 20. Research in nine countries is introduced showing BMI is an accepted key risk factor for osteoporosis. Irregular menstruation is a risk factor for poor nutrition and later osteoporosis. Exercise on an adequate diet increases bone mineral density, but increases the risk of osteoporosis on an inadequate diet and low BMI. A minority of men and women live until their 80s with fully adequate bone density, suggesting that osteoporosis is wholly preventable. But the majority, and particularly women over 50, have a deteriorating bone density. Hormone replacement therapy is a valuable preventative for women at risk. Several studies show a correlation of low bone density and dietary intake of calories and protein, and one study of protein but not calories. Other studies show correlations with some individual vitamins and minerals. It is concluded that current recommendations for BMI need reconsideration, as the increase in osteoporosis risk as BMI falls below 25 is shared by other disorders.

Keywords: osteoporosis, body mass index, fracture, hip fracture, irregular menstruation, bone mineral density, exercise.

The Health Survey for England 1991 said [ 1]:

The most frequently reported long-standing illness or disability by both men and women, in both surveys, were musculoskeletal conditions.

The second survey referred to is the 1989 General Household Survey in which 14.7% of men and 19.6% of women over the age of 16 reported loss of quality of life from musculo-skeletal disorders. Osteoporosis is a subheading of these disorders generally first diagnosed following a fracture. Rheumatoid arthritis and osteoporosis are both associated with loss of bone substance, lower bone density and mechanical strength [ 2].

Vertebral fractures are reported to affect one-third of all women in the US by the age of 65. These may be painless, but they progress and can cause severe back pain over years or even decades. Vertebral fractures outnumber hip fractures by a factor of three: they generally happen at a younger age than hip fractures and depress quality of life for longer [31.

Osteoporosis will always be a disorder better prevented than treated, and the nutrition of those at risk will always be an important part of prevention and treatment [ 4]. A paper in the first issue of Journal of Nutritional Medicine in 1990 by Alan Gaby and Jonathan Wright quoted an American estimate of the annual cost of fractures attributed to osteoporosis of US$6.1 billion [ 5]. A British paper by Alan Dixon, chairman of the National Osteoporosis Society, estimated the annual cost of osteoporosis fractures to Britain of at least £1 billion [ 6]. The economic and social costs of osteoporosis and other musculo-skeletal disorders in lost working time and sick pay, unemployment and disability benefit, premature retirement, home help and sheltered accommodation by far exceeds the cost to the National Health Service [ 7]. In Britain, the number of fractures each year is estimated at 290 000 [ 8]. Many are not recorded.

The Report of the Working Group on the Nutrition of Elderly People 1992 of the Department of Health recommended (p. 22):

The prognostic significance of body mass index and other anthropometric measures in a British population of elderly people should be clarified.

The present paper is intended as a contribution to this clarification.

A paper by Cyrus Cooper and colleagues of the Medical Research Council Environmental Epidemiology Unit in 1988 showed various risk factors for hip fracture of 300 patients and 600 matched controls [ 9]. Of all the factors possibly amenable to preventative action, body mass index (BMI) showed the largest relative risk between high and low values and the highest statistical significance of any risk factor recorded. BMI is expressed as body weight in kilograms divided by height in metres squared (kg m-2). The data from this survey are presented in Table 1. It is seen that the risk of hip fracture increased by a factor of 6.7 as BMI fell from above 25 to below 20 kg m-2. Thinness is a major risk factor for hip fracture. Moderate or heavy alcohol consumption was also a significant risk factor affecting 9% of patients but only 2% of controls. However, alcohol makes pain easier to bear which encourages consumption. Smoking showed a statistically significant risk ratio of 1.7 between those who had ever smoked and those who had never smoked. In Table 1 the BMI was calculated for men and women separately but the data were combined to calculate the relative risk.

The increase in the risk of hip fracture as BMI falls in both men and women appears to be common to populations in the developed world. A US study in King County, Washington, of 350 women who had a hip or forearm fracture and 562 controls showed an increase in relative risk by a factor of 10 as BMI decreased from 28.1 to below 17.5 [ 10]. The relative risk for both hip and forearm fractures was doubled on average by smoking and trebled by smoking for the category of women who were very thin [ 11]. Use of hormone replacement therapy (HRT) for at least 1 year greatly reduced the effects of smoking, suggesting that smoking may depress oestrogen levels. The same King County study also showed that there was no association between hip or forearm fracture and a history of childbirth or lactation. No factor recorded in this population approached the importance of BMI as a correlate of osteoporosis.

The Osteoporotic Fractures Research Group, centred on the University of California, studied 9704 ambulatory, white women all aged 65 years or older [ 12, 13]. In a multi-variate analysis, including physical characteristics and medical history, body weight had the highest correlation with bone mass of any variable except oestrogen use. BMIs were not published. Grip strength, indicating muscle strength, was the next most important correlate of bone mass and is discussed further below.

The American Framingham Heart Study reported 191 hip fractures among 2873 women. The fracture rate increased exponentially after the age of 50. BMIs were not recorded, but the risk of hip fracture increased by a factor of 3.0 as body weight fell from the highest to the lowest fifth of the body weight distribution [ 14]. BMIs would probably have given a bigger relative risk if they had been used because BMI is a better indicator of risk for osteoporosis than body weight. It is apparently more important to be the right weight for height than to achieve some standard weight disregarding height. This report from the Framingham Heart Study used weight units called 'Metropolitan Relative Weight', quite widely used in other American papers and called 'percentage of ideal weight'. It is noteworthy that in this study the weight of the women in the quintile of highest risk of hip fracture and lowest body weight was 105% of the 'ideal' weight. The average weight of women with hip fractures was 111 +/- 2% and without fractures 121 +/- 0.2% (p < 0.0001) of 'ideal' weight as measured during the 2 years before a fracture.

A Canadian study from Waterloo, Ontario, recorded an average BMI of 22.5 among 22 women aged 45-59 with hip fractures and 25.3 among 414 healthy controls [ 15]. Forty-nine women aged 65 to 74 with hip fractures had a BMI of 21, while 229 healthy controls had a BMI average of 24.6. This is another study suggesting that it may be unwise to restrict diet over the age of 40 to a level that allows BMI to fall below 25.

Bone density is correlated inversely with the risk of fracture and directly with BMI. Kokai Kin and colleagues of Hamamatsu University, Japan, studied the effect of BMI, age and menopausal status of 1048 women and 248 men, all of whom were volunteers [ 16]. The lumbar spine density (L2-L4) of all subjects was examined.

Bone density decreased in all women aged over 40 as BMI fell below 25. The decline of bone density with age was greater the lower the BMI. The study showed a highly significant correlation of BMI and bone density in all age ranges 20-39, 40-69 and over 70 (p<0.001). The greatest rate of fall in bone strength was through the menopause. Women with BMIs under 25 lost 16.3% of bone density during the 5 years after the menopause, and the women with BMIs over 25 lost 8.0% of bone density. Losses of bone are generally cumulative so that losses in middle age increase the risk of fracture in old age.

A study from Florence Tremollieres and colleagues at the teaching hospital in Toulouse, France, of 155 post-menopausal women also reported a highly significant correlation of vertebral bone loss rate and BMI (p<0.01) [ 17]. A scatter diagram in the report of this study showed, however, that a minority of women in the age range 45-61, some with a BMI below 25, do not suffer spinal bone loss at all. Tremollieres recorded small bone density gains of up to 1.5% per year for a few women (20 out of 155) and a further number of women (20 out of 155) showed no change in bone density. The remaining majority was recorded with a bone loss ranging up to 8.0% a year and a mean annual bone loss of 1.46% for women with BMI below 25 and 0.54% for a BMI over 25. This study emphasizes that bone loss and osteoporosis are not an inevitable consequence of ageing.

Susan Harris and colleagues at Tuft's University, Boston, US, measured rates of change in bone mineral density of the spine, femoral neck and radius of 288 healthy post-menopausal women aged 41 to 71 and found a substantial number who showed no sign of reduced bone mineral density over 2-year periods [ 18]. BMI was correlated with bone mineral density at all three sites and with serum levels of oestrone and oestradiol.

A study by R. Nuti and colleagues from the Metabolic Bone Disease Centre of the University of Siena, Italy, of 885 women showed many women maintaining bone mineral density for 30 or 40 years beyond the menopause [ 19]. Body weight and bone mineral densities were highly significantly (p < 0.001) correlated for spine, trunk, arms and legs.

A study of 348 healthy elderly women in the Netherlands, all over 70 years of age, by M. E. Ooms of the University of Amsterdam found many women with wholly satisfactory bone mineral bone densities [ 20]. Some women in their eighties, over 30 years after menopause, had bone mineral densities of the femoral trochanter of 0.9 g cm-2 while other younger women had a bone mineral density as low as 0.4 g cm-2. Bone mineral densities generally declined with age but with a large minority maintaining a satisfactory bone strength into old age. The average age of these women was 80.3 and average BMI was 28.3. Perhaps the Dutch habit of cycling limits femoral bone loss.

All these studies suggest that osteoporosis is wholly preventable but that a BMI over 25, requiring appropriate nutrition, is desirable. As discussed below, adequate exercise is essential too and prevention has to begin early enough to prevent deformity.

The effects of low body weight are cumulative and are particularly damaging in the elderly. M.D. Bastow and colleagues reported in the Lancet a study of 744 elderly women with hip fractures admitted to the University Hospital, Nottingham, UK [ 21]. The summary began:

On the basis of triceps skinfold thickness and arm muscle circumference measurements, 744 elderly women with fractured neck of femur were divided into three groups well-nourished, thin, and very thin. The mortality in the three groups was 4.4%, 8% and 18% respectively.

In the study by Kin, there were 36 younger women aged 20-39 out of 162 (22%), who had irregular menstruation, and these women had a bone density 8.1% lower than the women with regular menstruation [ 16]. Women with a history of irregular menstruation before the age of 40 had on average a loss in bone density greater by 8% subsequently than the women with a history of normal menstruation, in the absence of HRT or other preventative measures. Women with a history of irregular menstrual cycles will, on average, be at increased risk of osteoporosis fractures subsequently.

The many trials of HRT have firmly established the effect of hormones, notably oestrogen, on bone density and the prevention of osteoporosis. Irregular menstruation is generally a consequence of depressed hormone levels and it is therefore to be expected that low bone density would follow irregular menstruation.

A Committee of the US National Research Council said in a report on reproductive toxicology [ 22]:

Menstrual cycles constitute the most accessible and noninvasive biologic marker of female reproductive function in humans.

A wide range of toxic substances, including smoke constituents, causes menstrual irregularity, but the value to the toxicologist is limited by the sensitivity of menstruation to nutrition. Undernutrition and deficiencies of particular nutrients depress levels of oestradiol and progesterone [ 23]. A symposium on 'The Menstrual Cycle and its Disorders and the Influences of Nutrition, Exercise and Neurotransmitters' was reviewed in the Journal of Nutritional Medicine in 1991 [ 24].

Depression in levels of oestradiol and progesterone slows down rates of cell division in follicle and embryo, and apparently also in all connective tissue including skin, cartilage and in bones. These hormones act on the cell nucleus and, in bones, on the nuclei of the osteoblasts and, in teeth, on the nuclei of the odontoblasts. Many cells of different body systems depend on an adequate supply of these hormones to promote the cell replication necessary for tissue maintenance and repair. The effects of nutrition on cell replication are partly mediated by hormones. However, all replication needed for bone growth and maintenance is not only promoted by hormones and nutrition, but also by physical exercise.

Around 15 million Americans take exercise by running, which includes what is called jogging in Britain, and there is a 50-plus Runners' Association, whose members were studied by Nancy Lane and colleagues at Stanford University, California [ 25]. Five hundred and thirty-nine runners and 422 controls were enrolled in the study. All individuals were over 50 years but under 73 years of age, and were high-school graduates. The runners had averaged 36 miles (42 km) a week for 9 years. One purpose of this study was to examine the effect of this exercise on osteoarthritis and no clinical effect was found. However, runners, both male and female, had on average a 40% higher bone mineral density than matched controls. This suggests that running for about half an hour every day prevents most of the risk of osteoporosis over the age of 50.

The effects of less strenuous exercise were recorded in a study of women aged 43-72 at the Human Nutrition Research Center on Ageing of Tuft's University, Boston [ 26]. This showed a significant correlation between miles walked per week and density of the legs and spine, and suggested that it is desirable to walk at least one mile a day regularly. The women who walked most, however, also spent most time in gardening and in sporting activity. Nevertheless, the study concluded that, "walking is a beneficial form of physical activity for maintaining skeletal integrity".

Most people over the age of 50 will never be persuaded to run for half an hour a day and many do not live where this is possible. Furthermore, for women who are too thin or on slimming diets, exercise increases the risk of bone loss by further depressing the levels of the two sex hormones oestradiol and progesterone. Exercise can cause amenorrhoea notably in thin women [ 27]. A paper by Barbara Drinkwater from the University of Washington reported a BMI average of 19.7 kg m-2 for 14 women athletes who were amenorrhoeic [ 28]. These 14 women had a mean serum oestradiol level of 38.58 pg ml-1 compared with 106.99 pg ml-1 for women with normal menstruation and had lower bone mineral density in the lumbar spine from L1 to L4 than women who had normal menstruation. Amenorrhoea caused by heavy exercise in young women on an inadequate diet causing thinness has been reported to cause stress fractures including hip fractures and scoliosis in young women, for example ballet dancers [ 29].

There are more than 100 published studies showing the benefits of exercise in strengthening bones [ 30]. Muscle strength and bone strength are associated [ 31, 32]. However, in any interpretation of numerical data associating exercise with bone density, there are confounding factors including BMI, nutritional factors and illness. Prophylactic exercise must always be adapted with care to the individual's health and circumstances.

The difference in diet of young women runners with normal menstruation or with amenorrhoea is shown in Table 2 based on data by Miriam Nelson and colleagues of Tuft's University, Boston [ 33]. Amenorrhoea, low bone density and low serum levels of oestradiol are all seen to be associated with a low intake of the main nutrients. That oestradiol and progesterone levels are associated with BMI and diet is well known from another substantial literature on women's fertility and pregnancy outcome [ 24]. Exercise only depresses hormone levels in men and women who are too thin or not eating enough. A recent paper suggests that 'exercise and diet' alone could delay the average age of osteoporotic fractures for 10 years [ 34]. The substantial number of men and women who reach their late seventies and eighties with satisfactory bone health suggests that osteoporosis may be prevented by exercise and good nutrition. Table 2 suggests the importance of diet, but only for young athletes.

A study by G. Geinoz and colleagues of Geneva University Hospital had the cooperation of 48 women, mean age 82 years, and 26 men, mean age 80 years, all hospital patients with diverse disorders [ 35]. All ingested hospital food was recorded for 3 X 3 day periods. Food available to patients was not restricted. The bone mineral densities of femoral neck and shaft, and spine were measured. Data were analyzed according to what the patients chose to eat to show differences in bone mineral densities. Women and men consuming a diet containing more than 1 g kg-1/day body weight of protein had highly significantly stronger bones than those who chose a lower protein diet below 1 g kg-1/day. The protein intake was, however, correlated with total energy intake and the energy intake, notably the fat, was also significantly correlated with bone strength (p < 0.001). The authors concluded their study:

Malnutrition can be considered as a risk factor for osteoporotic fractures.

This is a valid conclusion if malnutrition is defined as less than 1 g kg-1 of protein daily, but the way the data were analyzed introduces a bias in favour of protein as it is unclear from the data presented whether the higher bone densities were attributable to higher energy or higher protein intakes or even to higher intakes of other nutrients. The conclusion, however, is in agreement with the Boston study by Nelson that malnutrition or undernutrition causes lower bone densities.

The US National Research Council says that the suggestion that high protein intake increases the risk of osteoporosis is not supported by current evidence [36; p. 265].

Geinoz's study at Geneva University is not the only nutrition survey to show a significant positive correlation of bone density and protein intake and no study has been found that shows an association of protein and lower mineral density or osteoporosis. Thus, a 1991 Japanese study by J. M. Lacey and colleagues of 89 pre-menopausal women aged 35-40, and 89 post-menopausal women aged 55-60, also found that current protein intake was correlated with bone density both before and after the menopause [ 37]. Intake of vegetables (leafy green, yellow, orange and white) and current milk intake were also correlated with bone density. The mean protein intake per day was 70.9 +/- 19.6 g before and 72.6 +/-15.6 g after the menopause. Irregular menstruation before the menopause was significantly correlated with low bone mineral density.

A Chinese study by Ji-Fan Hu of 843 women aged 35 to 75 from five rural counties in China makes it possible to distinguish between the effects on bone density of protein and energy [ 38]. The 142 women of one pastoral county, Xianghuangqui, had a protein intake of 75 +/- 27 g while the 147 women of another county, Tuoli, which has a more arable farming economy, had a much lower average protein consumption of 51 +/-17 g/day. Energy consumption was similar and close to 1500 kcal/day in both counties. Xianghuangqui women had higher bone densities than Tuoli at all ages from 35 to 75 years. This was certainly not attributable to energy intake nor was the difference attributable to non-dairy calcium, but the higher intake of milk and dairy foods in the pastoral county could explain the whole difference in protein and calcium intakes. Milk contains, of course, a range of nutrients that could have contributed to bone health. Another Chinese county, Cangxi, with a mainly arable economy, recorded 151 women as having an exceptionally high energy intake from carbohydrate of 2455 kcal/day and a comparatively low protein intake of 57 g/day, but low bone mineral densities. This Chinese study supports the view that it is not just low energy that matters but low protein or a low intake of a range of other nutrients too, or malnutrition in its widest sense that causes low bone density.

The comparative importance of particular nutrients is the subject of much disagreement. Conflicting recommendations for daily calcium intakes reflect such disagreements. Thus, the UK Department of Health's reference nutrient intake for men and women 19 years of age and over is 700 mg/day. The British National Osteoporosis Society recommends over 1000 mg/day for adults, rising to 1500 mg/day for women over 40 through the menopause and 1200 mg/day for men and women over 60.

Calcium is essential but is not the only nutrient needed for the prevention of osteoporosis. A paper in a British symposium declares [ 39]:

The abundance of papers on calcium metabolism gives the impression the skeleton is composed entirely of chalk.

Bone is continuously replaced and remodelled. Old bone is absorbed and replaced by a protein matrix which is calcified to produce new bone. The matrix grows first with formation of a soft protein material called osteoid produced by osteoblasts. The calcification of osteoid takes about 1 month. The nutrients needed to support the replication of the osteoblasts are needed first, before the calcium necessary for calcification. The osteoblasts need all the nutrients required for cell replication and protein synthesis. Freudenheim of the University of Wisconsin has published data showing that bone density measurements are correlated with dietary intakes of B vitamins, protein and energy [ 40].

Osteoblasts have been shown to have receptors for oestrogen, which is needed to stimulate cell replication [ 41]. One effects of HRT and endogenous oestrogen is to speed up osteoid formation. Deficiencies of B vitamins, for example folic acid, depress levels of oestrogen. Also, some nutrients, notably divalent metals, are needed by hormone receptors.

Japanese studies have shown that oestrogen receptors require magnesium and that the effect of oestrogen on the cell nucleus is modulated by magnesium at physiological levels [ 42]. Magnesium is also needed together with adenosine triphosphate to provide the energy needed for osteoblast replication, protein synthesis and calcification. It is probably no coincidence that there are reports from several countries of an association of magnesium deficiency and osteoporosis. Papers from L. Steidl and colleagues of Palackeho University, Olomouc, Czech Republic, describe a highly significant association of osteoporosis with magnesium deficiency in blood serum and erythrocytes [ 43]. There was no similar deficiency of calcium or zinc in the same patients. An association of osteoporosis and magnesium deficiency has also been reported from Israel [ 44, 45] and Belgium [ 46]. A research programme at the University of Nymegen has shown the importance of magnesium in calcification of bone and prevention of calcium loss [ 47]. Magnesium deficiency is not, however, generally accepted as a cause of osteoporosis and further research is merited. Magnesium deficiency is less easy to recognize or remedy by dietary change than calcium deficiency, but may be more important in some populations.

Vitamin D deficiency is not uncommon in elderly patients particularly during winter months, and aggravates osteoporosis by retarding calcification of the osteoid. It is recommended that serum 25(OH)D levels should not be allowed to fall below 30 ng ml-1 and should preferably be above 40 ng ml-1 [ 48]. Patients spending their whole lives indoors, particularly those having reduced physical activity or long periods of bed rest, are advised to take daily supplements of from 10 to 20, but not more than 25, mu g of vitamin D [ 48].

The Report on the Nutrition of Elderly People emphasizes (p. 23):

the importance for elderly people of consuming a varied diet, including foods which are rich, on an energy basis, in vitamins and minerals.

This report also discusses protein consumption for men and women aged 50 years or over, and refers with approval to an intake close to 1 g mixed protein per kg body weight/day (p. 62).This will give higher intakes than dietary references values (DRVs or RNIs), but is quite close to the average diet that healthy people actually consume. People who are house-bound as a result of disability or in hospitals or residential homes are likely to have particularly low energy and nutrient intakes. The Report on the Nutrition of Elderly People states that

'Health professionals should be aware of the often inadequate food intake of elderly people in institutions' and stresses particularly the importance of maintaining intakes of protein, vitamins and minerals if energy intake declines.

However, constituents of diet desirable for the prevention of osteoporosis are secondary in importance in the world literature to overall dietary sufficiency. No other dietary indicator appears to have a comparable significance to BMI, which generally reflects energy balance over longer periods of time. BMI levels compatible with the prevention of osteoporosis are therefore discussed further.

Table 1, based on the Medical Research Council's work, shows that the risk of hip fracture decreases as BMI increases up to at least 25.6. Research findings in Table 3 from France, Japan and the US also show the lowest risk for fractures at BMIs over 25. There can be little doubt that the avoidance of slimming to below a BMI of about 25, and an adequate diet that maintains BMI above, but not much above, 25, are a desirable part of any programme for women approaching the menopause for the prevention of osteoporosis. In contrast, the government's strategy described in Health of the Nation [ 49] is to reduce the whole population as far as possible to BMIs below 25. Over 25 is described as 'overweight'. The government's strategy describes no lower BMI limit and makes no reference to people who are too thin.

British government advice on BMI differs from the advice of the US National Research Council, which advocates an increase of BMI with age and a desirable range of 24 to 29 over the age of 65 [36; p. 564]. An analysis of the US National Build Study data, based on 4.2 million insurance policies and 106 000 deaths, found an optimum BMI of over 25 for men and women aged 50 and over, and no significant difference between men and women [ 50].

The UK Health Education Authority (HEA) has published a coloured diagram with heights and body weights divided into areas as shown in Fig. 1 [ 51]. This diagram has been widely reproduced in many books and leaflets and distributed through the check-outs of supermarkets. Figure 1 shows that the labelled areas are separated by lines which are close approximations to BMIs. The area defined as acceptable gives a BMI range of 17.5-24.5 as shown in Table 3 and Fig. 1; 17.5 is much lower than has been found anywhere else in world literature as a recommended BMI for health. The HEA with this coloured chart are giving their stamp of approval to weight loss by adults of all ages with BMIs over 24.5.

The grounds for slimming advocated in Health of the Nation are to reduce death rates and to reduce the risk of coronary heart disease and stroke. This prompts a number of questions which a reader may reasonably ask and expect at least a brief answer.

Will a reduction of men and women's BMI to below 25 as advocated in Health of the Nation and by the HEA reduce death rates from heart disease and stroke for men and women? The largest survey to relate BMI and mortality was reported by H. Waaler in 1983 and followed 1.8 million adults, almost half the population of Norway [ 52]. There is no comparable British study. The Norwegian study showed that for men and women a BMI of 25 is close to optimum, and no reduction in death rates can be expected by reducing BMIs to 24 and 23, and below 23 death rates begin to rise and more than double at most ages before BMIs fall to 17.5, the bottom of the HEA's 'acceptable' range. Slimming not only causes osteoporosis but shortens lives if carried too far.

Is there good evidence that men and women can reduce the risk of coronary heart disease by slimming to achieve a BMI below 25? The US National Research Council's report on the 'implications for reducing chronic disease risk' by diet summarizes some of the extensive studies and concludes that the risk of coronary heart disease increases at BMIs greater than 30 [ 35]. Waaler shows that mortality from cardiovascular disease in men and women increases at both high and low levels of BMI, as do a number of smaller studies all with flat minima for the mortality curves. No studies have been found in which slimming to BMI below 25 has been shown to reduce the risk of coronary heart disease; a BMI of 25 is near the apparent optimal BMI in most studies. The cardiovascular system, like the bones, requires an optimal level of oestrogen in the blood [ 39, 53, 54], and as we have seen oestrogen levels fall on average as BMI falls below 25. There is no apparent contradiction between what is best for the cardiovascular system, for connective tissue, for cartilage and for bones.

Is there good evidence that men and women can reduce the risk of stroke by slimming to a BMI below 25? The risk of stroke also increases at the top and bottom of the BMI range. Waaler shows the lowest risk of stroke at a BMI of about 25.

The Osteoporotic Fractures Research Group in a study of 4024 ambulatory women aged 65 years or older found significant correlations between low bone mineral density and the incidence of stroke [ 55]. The strongest correlation was for intra-cerebral haemorrhages and occlusions with bone mineral density at the calcaneous and proximal radius. A study by A. Paganini-Hill and colleagues of the University of Southern California reported that mortality from stroke increased as BMI fell below 24.6. This study covered 13 986 residents of a white retirement community of median age 73 [ 56]. Thirty-five per cent of participants used oestrogen therapy which reduced the relative risk of death from stroke to 0.53 (0.31-0.91). This study suggests that the vascular system of the brain may also need an adequate level of oestrogen for good health. No research studies have been found that show slimming to BMIs below 25 reduces the risk of stroke and such slimming appears to be injurious to the cerebrovascular system of the elderly.

Younger men and women may well ask: is there perhaps some contradiction between the BMI desirable for childbearing and desirable to prevent low bone density later in life? The risk of low birth weight increases slowly as the prepregnancy BMI of the mother falls below 28 and more rapidly as BMI falls below 22. The US Collaborative Perinatal Study found that close to 90% of low birth weight babies (< 2500 g) were born to mothers with BMIs under 25, which is the average BMI of American women between the age of 25 and 34. Slimming is bad for babies as well as for their mothers.

Are there perhaps other disorders that benefit from slimming to BMIs below 25? There may be, but nothing of consequence has been found. Mortality from tuberculosis and infectious disorders, obstructive lung disease and some kinds of cancer have been shown to increase below the range of 25-29.

Better nomenclature is needed. Popular writing is beginning to challenge the advocacy of slimming by government and press [ 57]. But it does not help the cause of preventing osteoporosis or assisting sufferers to say that it is best to be 'fat' or 'overweight'. It is confusing and illogical to describe BMIs best for health and fitness using these pejorative words. New nomenclature is needed and the HEA weight diagram should be reconsidered and reworded.

No school boy or girl will wish to be labelled 'fat' or 'plump'. Such labels encourage slimming by children, beginning at the time of the growth spurt at the age of 9 or 10 in girls. If the slimming craze continues into the twenties and thirties, peak bone mass will be lower than it should be, increasing the risk in most cases for the rest of life.

TABLE 1. Distribution of relative risk of hip fracture in fifths of distribution of BMI: 300 patients with hip fracture and 600 controls

Female Male Relative risk Confidence interval

< 19.2 < 20.4 6.7 3.8-11.7
19.2-21.1 20.4-21.6 4.3 2.4- 7.4
21.1-22.9 21.6-23.2 3.3 1.9- 5.8
22.9-25.6 23.2-24.9 2.1 1.2- 3.8
> 25.6 > 24.9 1.0
Source: see Reference 9.

TABLE 2. Comparison of daily diets of 11 amenorrhoeic women runners and 17 women runners with normal menstruation
Legend for chart:

A - No Heading
B - Menstruation normal (a)
C - Amenorrhoeic Co (b)
D - (b) as per cent (a)


Energy (kcal) 2250 1730 76.9
MJ 9.4 7.2 --
Carbohydrate (g) 263 196 74.5
Fat (g) 44 34 77.3
Protein (g) 55.3 41.9 75.8
Calcium (mg) 1150 886 77.0
Bone density
L1-L4 (g cm-2) 1.196 1.099 --
Serum oestradiol
(pg dl-1 28.6 8.5 --
km run/week 64 56 --
Source: see Reference 33.

TABLE 3. BMI and the risk of osteoporosis: research findings and recommendations
Research findings BMIs kg m-2

UK Medical Research Council [9]: lowest
risk for hip fracture
Women aged over 50 over 25.6

Japan, Hamamatsu University [16]: bone
mineral density; lowest risk for spinal
Women aged 20-70 plus over 25

University of Washington State [10]: lowest
risk for hip and forearm fractures
Women aged 50-74 over 28

France, Toulouse CHU Department of
Endocrinology [17]: lowest risk for spinal
bone loss
Women aged 45-61 over 25

Norwegian weight study [52]: lowest mortality
women aged 45-70 23-27
over 70 23-31
men aged 35-65 23-27
over 70 21-29


UK Health Education Authority [51] 17.5-24.5
Health Survey for England 1991 [1] 20-25
Department of Health: Health of the Nation
1992 [49] under 25
US National Research Council [36]
Men and women aged 45-54 22-27
55-65 23-28
over 65 24-29
Sources: see References as given in Table.

GRAPH: FIG. 1. Source: Health Education Authority [ 51]. BMIs have been added.

[1] White A, Nicolaas G, Foster K, Browne F, Carey S. Health Survey for England 1991. Office of Population Censuses and Surveys, London: HMSO, 1993.

[2] Compston JE, Vedi S, Mellish RWE, Croucher P, O'Sullivan MM. Reduced hone formation in non-steroid treated patients with rheumatoid arthritis. Ann Rheum Dis 1989; 48: 483-7.

[3] Lindsay R, Cosman F. Epidemiology of osteoporosis. In: Drife JO, Studd JWW, eds. HRT and Osteoporosis. Heidelberg: Springer, 1990; 75-86.

[4] Lindsay R. Prevention and treatment of osteoporosis. Lancet 1993; 341: 801-5.

[5] Gaby AR, Wright JV. Nutrients and osteoporosis. J Nutr Med 1990; 1: 63-72.

[6] Dixon A StJ. Health of the nation and osteoporosis. Ann Rheum Dis 1992; 51: 914-8.

[7] Rogers LF. Financial considerations in osteoporosis. In: Genant HK, ed. Osteoporosis Update 1987. San Francisco: Radiology Research and Education Foundation.

[8] Grimley-Evans J. The significance of osteoporosis. In: Smith R, ed. Osteoporosis 1990: conference Royal College of Physicians of London, May 1989.

[9] Cooper C, Barker DJP, Wickham C. Physical activity, muscle strength, and calcium intake in the fracture of the proximal femur in Britain. BMJ 1988; 297: 1443-6; and personal communication 1994.

[10] Alderman BW, Weiss NS, Daling JR, Ure CL, Ballard JH. Reproductive toxicology and post- menopausal risk of hip and forearm fracture. Am J Epidemiol 1986; 124: 262-7.

[11] Williams AR, Weiss NS, Ure CL, Ballard J, Daling JR. Effect of weight, smoking and estrogens use in the risk of hip and forearm fractures in postmenopausal women. Obstet Gynecol 1992; 60: 695-9.

[12] Bauer DC, Browner WS, Cauley JA et al. Factors associated with appendicular bone mass in older women. Ann Int Med 1993; 118: 657-65.

[13] Cummings SR, Black DM, Nevitt MC et al. Bone density of various sites for prediction of hip fractures. Lancet 1993; 341: 72-5.

[14] Kiel DP, Felson DT, Anderson JJ, Wilson PWF, Moskowitz MA. Hip fracture and the use of estrogens in postmenopausal women. N Engl J Med 1987; 317:1169-74.

[15] Kreiger N, Kelsey JL, Holford TR, O'Connor T. An epidemiological study of hip fracture in postmenopausal women. Am J Epidemiol 1982; 116: 141-8.

[16] Kin K, Kushida K, Yamazaki S, Inoue T. Bone mineral density of the spine in normal Japanese subjects using dual X-ray absorptiometry: effect of obesity and menopausal status. Calcif Tissue Int 1991; 49: 101-6.

[17] Tremollieres FA, Pouilles J-M, Ribot C. Vertebral post-menopausal bone loss is reduced in overweight women. J Clin Endocrinol Metab 1993; 77: 683-6.

[18] Harris S, Dallal GE, Dawson-Hughes B. Influence of body weight on rates of changes in bone density of the spine, hip and radius in postmenopausal women. Calcif Tissue Int 1992; 50: 19-23.

[19] Nuti R, Martini G. Effects of age and menopause on bone density of entire skeleton in healthy and osteoporotic women. Osteoporosis Int 1993; 3: 59-65.

[20] Ooms ME, Lips P, van Lingen A, Valkenburg HA. Determinants of bone mineral density and risk factors for osteoporosis in healthy elderly women. J Bone Min Res 1993; 8: 669-75.

[21] Bastow MD, Rawlings J, Allison SP. Undernutrition hypothermia, and injury in elderly women with fractured femur: an injury response to altered metabolism? Lancet 1983; i: 143-6.

[22] US National Research Council. Biologic markers in reproductive toxicology. Washington, DC: National Academy Press, 1989, 171.

[23] Wynn M, Wynn A. The menstrual cycle as an indicator in prepregnancy care. J Nutr Med 1991; 2: 387-98.

[24] Pirke KM, Wuttke W, Schweiger U, eds. The Menstrual Cycle and its Disorders: Influences of Nutrition, Exercise and Neurotransmitters. Heidelberg: Springer, 1989.

[25] Lane NE, Block DA, Jones HH, Marshall WH, Wood PD, Fries JF. Long-distance running, bone density and osteoarthritis. J Am Med Assoc 1986; 255: 1147-51.

[26] Krall EA, Dawson-Hughes B. Walking is related to bone density and rates of bonemass. Am J Med 1994; 96: 20-6.

[27] Fabbri G, Petraglia F, Segre A etak Reduced spinal bone density in young women with amenorrhoea. Europ J Obstet Gynecol Reprod Biol 1991; 41: 117-22.

[28] Drinkwater BL, Nilson K, Cheshunt CH et al. Bone mineral content of amenorrheic and eumenorrheic athletes. New Engl J Med 1984; 311: 277-81.

[29] Warren MP, Brooks Gunn J, Hamilton LH, Warren LF, Hamilton WG. Scoliosis and fractures in young ballet dancers, relation to delayed menarche and secondary amenorrhoea. N Engl J Med 1986; 314: 1348-53.

[30] Gutin B, Kasper MJ. Can vigorous exercise play a role in osteoporosis prevention? A review. Osteoporosis Int 1992; 2: 55-9.

[31] Doyle F, Brown J, LaChance C. Relation between bone mass and muscle weight. Lancet 1970; i: 391-3.

[32] Ellis KJ, Cohn SH. Correlation between skeletal calcium mass and muscle mass in man. J Appl Physiol 1975; 38: 455-60.

[33] Nelson ME, Fisher EC, Catsos POet al. Diet and bone status in amenorrhoeic runners. Am J Clin Nutr 1986; 43: 910-16.

[34] Kanders B, Dempster DW, Lindsay R. Interaction of calcium nutrition and physical activity on bone mass in young women. J Bone Min Res 1988; 3: 145-9.

[35] Geinoz G, Rapin CH, Rizzoli R et al. Relationship between bone mineral density and dietary intakes in the elderly. Osteoporosis Int 1993; 3: 242-8.

[36] National Research Council. Diet and Health: Implications for Reducing Chronic Disease Risk. Washington, DC: National Academy Press, 1989.

[37] Lacey JM, Anderson JJB, Fujita T et al. Correlates of cortical bone mass among premenopausal and postmenopausal Japanese women. J Bone Min Res 1991; 6: 651-9.

[38] Hu J-F, Zhao X-H, Jia J-B, Parpia B, Colin-Campbell T. Dietary calcium and bone density among middle-aged and elderly women in China. Am J Clin Nutr 1993; 58: 219-27.

[39] Brincat M, Moniz C, Savvas M, Studd JWW. Oestrogen deficiency and connective tissues. In: Drife JO, Studd JWW, eds. HRT and osteoporosis. Heidelberg: Springer, 1990.

[40] Freudenheim L, Johnson NE, Smith EL. Relationships between usual intake and bone-mineral content of women 35-65 years of age: longitudinal and cross-sectional analysis. Am J Clin Nutr 1986; 44: 863-76.

[41] Eriksen EF, Colyard DS, Berg NJ et al. Evidence of estrogen receptors in normal human osteoblast- like cells. Science 1988; 44: 863-76.

[42] Fukai F, Murayama A. Association and dissociation of estrogen receptor with estrogen receptor-binding factors is regulated by Mg++. J Biochem 1984; 95: 1127-30.

[43] Steidel L, Ditmar R, Kubicek R. Biochemickenalezy u osteoporoz. Cas Lek Ces 1990; 129: 51-5, 147-50.

[44] Cohen L, Kitzes R. Infrared spectroscopy and magnesium content of bone mineral in osteoporotic women. Israel J Med Sci 1981; 17: 1123-5.

[45] Cohen L. Recent data on magnesium and osteoporosis. Magnesium Research 1988; 1: 85-7.

[46] Reginster JY, Strause L, Deroisy R et al. Preliminary report of decreased serum magnesium in postmenopausal osteoporosis. Magnesium 1989; 8: 106-9.

[47] Driessens FCM, Vetbeck RMA, van Dijk JWE. Is systematic treatment of osteoporosis and mandibular resorption possible? Acta Stomatol Belg 1990; 87: 108-24.

[48] Heaney RP. Prevention of osteoporotic fracture in women. In: Avioli LV, ed. The osteoporotic syndrome. New York: Wiley-Liss.

[49] Department of Health. The Health of the Nation. London: HMSO, 1992.

[50] Andres R, Elahi D, Tobin JD, Muller DC, Brant L. Impact of age on weight goals. Ann Int Med 1985; 103: 1030-3.

[51] Health Education Authority. Enjoy Healthy Eating. London: 1991.

[52] Waaler HTh. Height, weight and mortality: the Norwegian experience. Acta Med Scand 1983; suppl 679.

[53] Ross RK, Pike MC, Mack TM, Henderson BE. Oestrogen replacement therapy and cardiovascular disease. In: Drife JO, Studd JWW eds. HRT and Osteoporosis. Heidelberg: Springer, 1990.

[54] Grady D, Rubin SM, Petitti DB et al. Hormone therapy to prevent disease and prolong life in postmenopausal women. Ann Int Med 1992; 117: 1016-37.

[55] Browner WS, Pressman AR, Nevitt MC, Cauley JA, Cummings SR. Association between low bone density and stroke in elderly women. Stroke 1993; 24: 940-6.

[56] Paganini-Hill A, Ross RK, Henderson BE. Postmenopausal oestrogen treatment and stroke; a prospective study. BMJ 1988; 297: 519-22.

[57] Sanders T, Bazalgette P. You Don't Have to Diet. London: Bantam Press, 1994.



9 View Road, Highgate, London N6 4D J, UK

Share this with your friends