Harbottle unpublished results). The possible risks of intestinal blood loss ..... Lewis are, of course, right to point out that intervention against a risk factor is not ...
3 Dellinger EP. Severe necrotizing soft-tissue infectionsmultiple disease entities requiring a common approach. 7AMA 1981,246:1717-21. 4 Tehrani MA, Ledingham IMcA. Necrotizing fasciitis. Postgrad Medj 1977;53:237-42. 5 Janevicius RV, Sang-Erk Hann, Batt MD. Necrotizing fasciitis. Surg Gvnecol Obstet 1982;154:97-102. 6 Stamenkovic I, Lew PD. Early recognition of potentially fatal necrotizing fasciutis-the use of frozen-section biopsy. N Engl JMed 1984;310:1689-93. 7 Cruickshank JG, Hart RJC, George M, Feest TG. Fatal streptococcal septicaemia. BAf7 1981;282:1944-5. 8 Cone LA, Woodard DR, Schlievert PM, Tomory GS. Clinical and bacteriologic observations of a toxic shock-like syndrome due to Streptococcus pyogenes. N Engl J Med 1987;317: 146-9. 9 Strasberg SM, Silver MS. Hemolytic streptococcus gangrenean uncommon but frequently fatal infection in the antibiotic era. Am3'Surg 1968;115:763-8. 10 Stevens DL, Tanner HM, Winship J, et al. Severe group A streptococcal infections associated with a toxic shock-like syndrome and scarlet fever toxin A. N Engl J Med 1989; 321:1-7. 11 Kim YB, Watson DW. A purified group A streptococcal pyrogenic exotoxin -physicochemical and biological properties including the enhancement of susceptibility to endotoxin lethal shock. JExpMed 1970;131:611-28. 12 Wannamaker LW. Streptococcal toxins. Rev Infect Dis 1983;5 (suppl 4):S723-32.
Milk for babies and children SIR,-Professor Brian Wharton's advice for feeding infants cannot be applicable to the babies and children in the Third World. ' Moreover, in my opinion, it is also inappropriate for the best nutrition of babies and children in developed countries. Professor Wharton ignores the unique position of human milk for babies and infants. Mothers throughout the world, irrespective of whether they come from poor or rich nations, should breast feed their infants for 'two years or more. Exclusive breast feeding should be practised for four to six months. Then the baby should be weaned on to locally available, home made, additional food. Breast feeding alone provides for infants optimum nuturition up to 4 to 6 months of age, immunological protection, and bonding between mother and infant, it is also economical and provides a contraceptive effect for the mother as well as safeguarding against breast and ovarian cancer." Commercial interests rather than nutritional ones have dominated the production of various kinds of milk powders throughout the world; in Bangladesh, for example, about 500 million takas ($15m) worth of milk powders are imported a year. About 50 different brands are available in the country from all over the world, including about 10 varieties of infant formula (Consumer Association of Bangladesh, unpublished data).2 About six brands ofpowder milk are soya based and popularly sold for treating diarrhoea. One European brand even boasts of being suitable for treating acute diarrhoea. There is aggressive marketing of the powdered tinned milk throughout the country both in towns and rural areas. As Bangladesh is poor and the literacy rate is 32% (22% among women) mothers' perception is that any powdered milk is a breast milk substitute. Due to ignorance and poverty these powdered milks, particularly the infant formulas, are prepared dilute and in second hand, dirty bottles. As a result children who receive this mixture get diarrhoea and become malnourished. Secondhand bottles are easily available and the teat is cheap, but mothers cannot clean them because there is no safe water and fuel is scarce. In Brazil it was found that infants who were not breast fed at all had 14 2 and 3-6 times greater risk of death from diarrhoea and respiratory infections respectively compared with infants who were breast fed and receive no milk supplements." Infants who received both breast milk and formula or cows' milk were 4-2 and 1 6 times more likely to die from diarrhoea and respiratory infections than those who were exclusively breast fed.7 The BMJ is read not only in Britain, Europe, and North America but by the medical profession
1392
throughout the world. Western based education still has a deep root in the medical colleges and among the professionals of the Third World, and this kind of article will certainly influence many doctors in the developing countries when they give feeding advice to mothers for their infants. Although this article is directed at developed countries, it could be disastrous for many children and families in the Third World. MQ-K TALUKDER Institute of Postgraduate Medicine and Research, Dhaka- 1 200, Bangladesh 1 Wharton B. Milk for babies and children. BMJ 1990;301:774-5. (6 October.) 2 United States Agency for International Development. Breast feeding for child survival strategy. Washington, DC: USAID, 1990. (Report 20523.) 3 Layde PM, Webster LA, Bangham AL, et al. The independent associations of parity, age at first full-term pregnancy, and duration of breast feeding with the risk of breast cancer. J Clin
learns to chew, to feed himself by hand, and to increase his experience of tastes and textures. There is little to be gained by postponing this developmental milestone. Why are follow on milks being recommended? The cost difference between follow on formula and cows' milk is not negligible. Calculating the difference in weekly household expenditure, and using the weight and energy data from Fomon's hypothetical infants' I estimate the additional weekly cost of one follow on formula compared with cows' milk to be £1.88 and £1.43 at 6 and 10 months respectively. Because of their distinct formulation follow on milks may also escape the legal strictures of the World Health Organisation's international code of marketing with regard to advertising of breast milk substitutes. If the spectre of protein malnutrition slips its bony toe in the door in Britain, how much more vulnerable will the mothers of the Third World be to persuasion?
Epidemiol 1989;42:%3-73. 4 Yuan JM, Yu MC, Ross RK, et al. Risk factors for breast cancer in Chinese women in Shanghai. Cancer Res 1988;48:1949-53. 5 McTiernan A, Thomas DB. Evidence for a protective effect of lactation on risk of breast cancer in young women. Am J Epidemiol 1986;124:353-8. 6 Byers T, Grahams S, Rzepka T, Marshal J. Lactation and breast cancer: evidence for a negative association in premenopausal women. Amy Epidemiol 1985;121:664-74. 7 Victoria CG, Vaughan JP, Lombardi C, et al Evidence for protection by breast feeding against infant deaths from infectious diseases in Brazil. Lancet 1987;i:319-21.
SIR,-Professor Brian Wharton proposed that follow on milks offer advantages for older infants and toddlers. ' Little evidence exists to support this claim, and the implicit suggestion that a high protein milk might compensate for impoverished weaning diets is to be deplored. Available evidence indicates that the protein energy requirements of an infant will be met by breast or formula milk provided that sufficient volume is taken to meet energy needs. If the safe level of protein intake falls from 1-86 g protein/kg body weight to 1-48 g/kg and the energy requirements fall from 485 kJ/kg to 423 kJ/kg between 6 and 12 months of age2 protein needs during this period will be met by an energy adequate diet in which the percentage of protein energy falls from 6-4% to 5 9%. The percentage of protein energy in breast milk is 7-4%; formula milks range from 8-6% to 11-3% and follow on milks from 11-9% to 17 8% (calculated from Professor Wharton's data). Fomon et al calculated that the protein needs of hypothetical 6 and 10 month old infants will be met by a mixture of formula milk and weaners and question whether the follow on milks available in the United States (also high in protein) are "designed to meet the special needs of the older infant."3 Even if we accept the unlikely postulate of a weaning diet that is virtually free of protein, the addition of nonprotein dietary energy will improve the rate of nitrogen retention "until the response reaches a plateau which represents the limitations imposed by the dietary protein level."2 The regular weaning diet is high in protein3 and in Western countries usually offers a higher energy density than milk. Iron deficiency, especially in Asian babies is a more important nutritional problem in Britain than that of protein energy.45 The weaning diet of 100 Asian toddlers, which included a large contribution from formula milk and was adequate in protein and energy, failed to meet the recommended dietary intake for iron (M B Duggan and L Harbottle unpublished results). The possible risks of intestinal blood loss due to early feeding with cows' milk should also be seen in context. Infants in the study cited by Professor Wharton had been fed on pasteurised cows' milk that had not been heat treated, and the authors noted that in Finnish studies heat treatment of cows' milk prevented intestinal blood loss.6 The bioavailability of the iron in formula milks is uncertain, and they are often given by bottle. During weaning the infant
MAUREEN B DUGGAN
Centre for Human Nutrition, Sheffield S10 2TN 1 Wharton B. Milk for babies and children. BMJ 1990;301:774-5. (6 October.) 2 World Health Organisation energy and protein requirements. WHO Tech Rep Ser 1985;724. 3 Fomon S, Sanders KD, Ziegler EE. Formulas for older infants. J Pediatr 1990;116:690-6. 4 Erhardt P. Iron deficiency in young Bradford children from different ethnic groups. BMJf 1986;292:90-3. 5 Grindulis H, Scott PH, Belton NR, Wharton BA. Combined deficiency of iron and vitamin D in Asian toddlers. Arch Dis Child 1986;61:843-8. 6 Zeigler EE, Fomon SJ, Nelson SE, et al. Cow milk feeding in infancy: further observations on blood loss from the gastrointestinal tract.J Pediatr 1990;116:11-8.
AUTHOR'S REPLY,-Having cut my nutritional teeth in a developing country (Uganda), I share many of Dr Talukder's concerns. I did, however, state that the editorial was looking at the question of "what advice should be given to mothers living in developed countries." Dr Talukder emphasises the value of breast feeding; I quoted the study showing that even in Scotland babies who had been breast fed for 13 weeks or more had fewer gastrointestinal upsets and fewer admissions to hospital. Many aspects of medical care, preventive and curative, differ because ofthe genetic characteristics of the population, the age and the development of the individual, and the environment in which he or she lives: microbiological, cultural, socioeconomic, etc. What is good for Europe is not necessarily good elsewhere-nor did I suggest that it was. I am a great supporter of breast feeding, but bottle fed babies have rights too. Dr Duggan is mainly concerned with follow on milks, particularly their higher protein *content when compared with infant formula and their higher cost when compared with whole cows' milk. Cows' milk contains even more protein than a follow on milk. Presumably she therefore agrees with my advice to mothers who are bottle feeding to continue with an infant formula, but if a mother insists on changing, as some do, then rather than cows' milk she would favour follow on milk because it has a lower protein content. Her opinion is not clear because in the later comparisons she compares the cost of a follow on formula with cows' milk for a 6 and 10 month old child-is she really advocating using whole cows' milk despite her concern about protein intakes at this age quite apart from its other disadvantages? Pasteurised cows' milk is by definition heat treated. Dr Duggan has taken most exception to the possible use of a follow on formula between 12 and 24 months. I agree that this is controversial. What does she recommend for the liquid part of the diet at this age? Skimmed and semiskimmed milks have a high protein to energy ratio. Whole cows' milk certainly is acceptable, and I mentioned it first, but follow on milks are worth considering
BMJ VOLUME 301
15 DECEMBER 1990
because they contain added iron and vitamin D and have limited saturated fat and sodium contents, without limiting energy content. Regarding cost we can quibble over figures, but let's accept an extra cost ofE l.43 a week of using a follow on milk (or infant formula) instead of cows' milk. This is the equivalent each day of one to two disposable nappies, or a newspaper, or two cigarettes, or 28 ml beer. Nevertheless, for some families even this extra cost may be the proverbial straw-but the answer to this is to change other aspects of public policy, not to continue to recommend substandard child feeding customs that are being abandoned by many other countries in the developed world. BRIAN WHARTON
Yorkhil Hospitals. Glasgow G3 8SJ
Sexual expression in paraplegia SIR, -Dr J M Kellett emphasises the potential for people with paraplegia to have satisfying sexual relationships despite difficulties with potency, sexual responsiveness, and immobility. But for patients with neuropathic bladder sexual expression may be marred by incontinence or voiding on intercourse. Learning how to self catheterise may solve these problems. Provided that patients are motivated and have a bladder that can retain an adequate volume of urine they can become dry with self catheterisation.2 Regular bladder drainage also protects the upper tracts.3 Disability is not necessarily a bar to self catheterisation. Patients can learn the technique despite paraplegia, poor manual dexterity, and lack of perineal sensation.4 By enabling patients to regain bladder control and freeing them from bulky external appliances, self catheterisation does much to enhance their self esteem and sexuality.5 Complications are few and long term results excellent.3 Dr Kellett recommends practical advice about sexual technique and aids. When appropriate this might include discussion of the benefits of clean intermittent self catheterisation. PIPPA OAKESHOTT
London SW8 2UD
26:405-12.
The mean predicts the number of deviants SIR,-Professor Geoffrey Rose and Mr Simon Day have shown that in a population the proportion of deviant values of such characteristics as obesity and blood pressure is strongly related to the mean value and conclude that we should attend to the population as a whole and not just its deviants.' For reasons related to natural biological variability a person may be unusually small and heavy, have a high blood pressure, be of low birth weight, and so on, and there is no reason to expect that excess risk of disease should attach to such people as even if we are all absolutely "normal" someone will still be the smallest, heaviest, most hypertensive, and so on. Such people are neither abnormal nor deviant of course:
15 DECEMBER 1990
blood pressure. Until we have evidence that intervention in normotensive patients is helpful we should leave well alone. And what of the great cholesterol debate? Whether to shift the population's behaviour or just the top centile, will depend on evidence from controlled trials. Sometimes, of course, the "sick population" approach is unarguable-for example, when a whole Third World region is trapped in famine we should attack the causes, but the approach would not preclude us from feeding a starving person. L S LEWIS
Newport, Dyfed SA42 OTJ
1 Rose G, Day S. The population mean predicts the numbers of deviant individuals. BMJ 1990;301:1031-4. (3 November.)
AUTHOR'S REPLY,-Mr Jon Nicholl and Dr L S Lewis are, of course, right to point out that intervention against a risk factor is not warranted unless it is believed that the relation is causal and at least partially reversible. These conditions are met in the examples we studied (hypertension, obesity, heavy drinking, and high sodium intake), but any other examples must each be judged individually, including the question of heterogeneity in the make up of the population. The laboratory custom of reporting a "range of normal" in purely statistical terms (±2 standard deviations) often causes confusion, being quite different from clinical or biological significance. A particular value may be uncommon yet benign, or common and yet sinister. Statistical measures of deviance have no biological meaning: they only identify what is unusual. Clinical action must be guided by absolute measures. GEOFFREY ROSE London School of Hygiene and Tropical Medicine, London WC1E 7HT
JON NICHOLL Medical Care Research Unit, Sheffield S10 2RX 1 Rose G, Day S. The population mean predicts the number of deviant individuals. BMJ 1990;301:1031-4. (3 November.) 2 Wilcox AJ, Russell IT. Birthweight and perinatal mortality. I. On the frequency distribution of birthweight. Intj Epidemiol
1983;12:314-8. 3 Skjaerven R, Wilcox AJ, Russell D. Birthweight and perinatal mortality of second births conditional on weight of the first.
Inty Epidemiol 1988;172830-8.
1 Kellett JM. Sexual expression in paraplegia. BMJ 1990;301: 1007-8. (3 November.) 2 Hunt GM, Whitaker RH, Doyle PT. Intermittent self catheterisation in adults. BMJ 1984;289:467-8. 3 Diokno AC, Sonda P, Hollander JB, Lapides J. Fate of patients started on clean intermittent self catheterisation therapy 10 years ago. J Urol 1983; 129:1120-1. 4 Hunt GM. Recent advances in intermittent catheterisation. Z Kinderchir 1989;44(suppl 1): 50. 5 Hill V, Davies W. A swing to intermittent clean self catheterisation as a preferred mode of management of the neuropathic bladder for the dextrous spinal cord patient Paraplegia 1988;
BMJ VOLUME 301
On the other hand, some people may have unusual values because they have a disease and thereby are at risk of other disease. The reasons for their differences from the mean may be pathological rather than biological. Thus the distribution of values in a population may be made up of a distribution of values in normal people reflecting natural biological variation and an overlapping distribution of values in a few diseased people. This is widely thought to be the case, for example, for birth weight,23 for which it is supposed that no excess risk of postnatal mortality or morbidity attaches to low birth weight babies who were expected to be small for biological reasons such as having a small mother. The risk may, however, be great in babies who are small for pathological reasons, such as disease brought about by poor prenatal nutrition or antenatal smoking. Altering the mean in ways that only shift the distribution of "normal" values will have no effect on the prevalence of disease related deviance and hence of disease, even though the deviant tail will alter as your authors have shown. In some cases the main biological determinants of the characteristic may be known, and an "expected" value can then be devised, and deviance from this expected value may allow individual values that are due to disease to be identified. This is, for example, the case for forced expiratory volume, which is usually expressed as a percentage of the predicted value based on age, height, and sex. For most characteristics, however, it is not possible to distinguish between values that were determined, in part, by disease and other identical values that were the result of natural biological variation. In these cases, as the tail of a distribution will contain a larger proportion of people whose values are deviant for pathological rather than biological reasons and who may need and benefit from treatment, it is surely correct to pay attention to the tail of the distribution.
SIR,-Professor Geoffrey Rose and Mr Simon Day may favour arguments for managing "sick populations" rather than "deviant individuals" but they rest their case for debate on a specious appeal to statistics.' If a social or biological characteristic is distributed in anything remotely like a bell shape then the proportion or number of individuals that exceed any arbitrary absolute position (systolic blood pressure 3 140 mm Hg, body mass index >30 kg/m2, alcohol intake ¢300 ml/week, sodium intake ¢250 mmol/day) will of course be predicted by the position of the whole curve (that is, its mean). But deviance from the mean in statistical terms is described by such relative measures as centiles or standard deviation and not by absolute cut off points. We cannot say whether skewing or sliding is at work in the 52 study populations, but the graphs (A-D) look, unsurprisingly, like the normal cumulative probability curve rather than like straight lines. We need to discuss the issue of norms, deviance, and absolutes, not as metaphysic but according to reason and causality. In what sense are moderate drinkers collectively responsible for the heavy drinkers? Before we focus our attention on deviants we need to be sure not only that they are in some sense at risk but that we have an effective treatment. Thus we do not direct medical effort against high intelligence, but we do direct drugs against high
SIR,-The premises on which Professor Geoffrey Rose and Mr Simon Day build their argument that the normal majority must change is fallacious, their reasoning faulty, and their evidence irrelevant. ' The only common denominator of their data seems to be the trivial observation that shifting population means also shifts the tails of distribution. As these observations are based on interpopulation comparisons of surrogate measures for health, no conclusion can be drawn as to the health of individual populations. There is no evidence that lowering populations means for weight, blood pressure, alcohol consumption, or 24 hour urinary sodium excretion makes people live longer or happier lives. The people who differ from arbitrarily assigned values are called "deviant individuals" or simply "deviants." In this way the standard deviations of statistics become reified as moral lapses of individuals. Yet, in the case of alcohol intake, teetotallers, who are equally "deviant" statistically do not earn this label. It is a short step from labelling people as deviants to moral exhortation and the language of blame: "The population thus carries a collective responsibility for its own health and wellbeing, including that of its deviants" and "it is no longer possible to regard normal (majority) behaviour as of no wider consequence." Note the imperceptible shift by which weight, blood pressure, or urinary sodium excretion become behaviour and responsibility. Professor Rose and Mr Day would have us believe that studying the determinants of "average blood pressure and weight, alcohol intake, average population 'mood,' intellectual performance, aggression, etc" could tell us something about "the causes of hypertension, obesity, alcoholism, depression, violence, and so on." This is analogous to claiming that the causes of poverty can be discovered by analysing the determinants of average income, or that the causes of gigantism or
1393
