Asthma and Allergic Diseases: Is There a Downside to Cleanliness and Can We Exploit It?{

Asthma and allergic diseases: is there a downside to cleanliness and can we exploit it?{

J Crane*

Department of Medicine, Wellington School of Medicine and Health Sciences, Wellington, New Zealand

When hay fever was first described in the early 19th century it was an uncommon disorder. Since then asthma, allergic rhinitis and eczema, have become common conditions particularly in industrialised western economies. International prevalence studies reveal wide variation in the prevalence of asthma, and allergic disease, but confirm this view although the studies also show these diseases to be by no means rare in most countries. The reasons remain unclear but the ‘‘hygiene hypthesis’’ postulating an inverse relationship between hygiene, in its broadest sense, and allergic disease, fits some of the epidemiology of these diseases and has an associated immunological hypthesis to support it. Recent studies suggest that many hygiene related factors may influence immune development in favour of an allergic phenotype. Antibiotics in the first of life have been associated with increased risks of allergic disease in later childgood, and farming exposures to protection. Interest in hte role of bowel flora in modifying immune development has been suggested as an explanation for the risk associated with antibiotic exposure and has led to studies exploring the effects of modifying bowel flora with probiotics both to influence established allergic disease and to prevent it. The challenge is to continue to reap the enormous benefits that accrued from modifying infectious disease by both public and individual health strategies but at the same time convincing the immune system that it has been exposed to them. European Journal of Clinical Nutrition (2002) 56, Suppl 4, S39 – S43. doi:10.1038/sj.ejcn.1601660

Descriptors: asthma; atopy; antibiotics; endotoxin

Introduction 1819, John Bostock gave the first detailed description of the The last two decades has seen an unprecedented increase in signs and symptoms of hay fever, describing his own disease, the reporting of asthma symptoms and in the diagnosis of in a paper read to the medical society in London (Bostock, allergic diseases in many countries around the world amongst 1819). Nine years later, in 1828, he read a second paper to both adults and children. A recent review of this pheno- the same society, in which he documented the disease in a menon described these changes as an epidemic (Holgate, further 18 cases that had been brought to his attention 1999). The cause, or causes, of this increase in asthma and (Bostock, 1828). At the beginning of the twenty first century, allergic disease prevalence is unknown. A number of hypo- some 170 years after these initial descriptions, the prevalence theses have been suggested, and have recently been the of hay fever in many developed countries is self reported at subject of considerable debate (von Mutius et al, 2001). between 20 and 30%. While there are many reasons why a While the symptoms of asthma have been described since relatively minor disorder such as hay fever might have gone antiquity (Cohen, 1992) and the allergic basis suggested as unnoticed amidst the everyday mortality and morbidity early as the sixteenth century (van Helmont, 1662), until the from major infectious diseases of the early nineteenth early part of the eighteenth century such diseases appear to century, it seems unlikely that a physician such as Bostock, have been more of curiosity than a significant health pro- already on the lookout for the signs and symptoms of hay blem. This pattern of disease development is typified by the fever, would have taken 9 y to recruit 18 cases if the development of hay fever or seasonal allergic rhinitis. In prevalence was in any way approaching that found today. These historical perspectives suggest that environmental *Correspondence: J Crane, Department of Medicine, Wellington School of change in its broadest sense is the major determinant of Medicine and Health Science, PO Box 7343, Wellington, New Zealand. these allergic diseases. E-mail: [email protected] However, their study has not proved straightforward. One {Based on a presentation to the symposium: Fermented food, fermentation and intestinal flora, National Institute of Nutrition, Mexico reason is that there are no universally agreed definitions, and City, Mexico, 17 – 18 May 2001. studies of their prevalence relies largely on the individual’s Asthma and allergic diseases J Crane S40 description of symptoms with the attendant problems of ferences. A striking feature of the study was that English- interpretation, increasing community awareness, combined speaking countries, in particular the UK, Australia, New with a reduced willingness to tolerate symptoms (Magnus & Zealand, and to a lesser extent the United States and Jaakkola, 1997). Similarly the criteria applied to these symp- Canada, had much higher rates than equally developed toms by health professionals have been changing. Some countries in continental Europe. Nevertheless there are authors have suggested that the evidence for this increase important differences in prevalence related to socio- in prevalence is far from conclusive and will require studies economic gradient. Broadly, with some important excep- that incorporate objective markers associated with allergic tions, these trends for asthma were mirrored by allergic disease before we can be certain that the prevalence has rhinitis and allergic dermatitis. indeed risen (Magnus & Jaakkola, 1997). It seems likely that at least a significant proportion of the increased prevalence can be attributed to an increased awareness of the symptoms The hygiene hypothesis of allergic diseases by a better informed public, particularly The rise in the prevalence of atopic diseases has been linked in developed countries. The symptoms that are used to to the decline in infectious disease. Gerrard et al (1976) define asthma, allergic rhinitis and eczema are non-specific. examined serum IgE amongst 275 individuals from 58 Thus wheezing and coughing and nasal discharge are symp- Metis families, descendants of Cree Indians in Northern toms that may arise in relation to colds or influenza. In the Saskatchewan. They compared total IgE and the prevalence case of eczema or allergic dermatitis, skin rashes particularly of atopic disease with 819 individuals from 176 white in infancy and early childhood from a wide variety of causes families living in the same region. The prevalence of are very common. Each of the symptoms may arise entirely eczema was 7-fold higher and asthma almost 3-fold higher in the absence of any allergic phenomenon and yet be in the white community compared with the Metis commu- indistinguishable from the same symptoms caused by nity. In contrast, geometric mean total IgE, was 3-fold higher allergy. Despite these obvious caveats, the prevalence at in the Metis community compared with the white commu- least since the nineteenth century almost certainly has nity. In their concluding statement these authors spelled out increased and a better understanding of the factors asso- what was to be termed the hygiene hypothesis a decade later. ciated with this increase might lead to preventative measures that could reduce the incidence. The prevalence of asthma, eczema and urticaria was greater in the white than in the Metis community and contrasted with the increased prevalence of helminth Geographical variation in the prevalence of infestation as well as of other untreated viral and asthma and allergic disease bacterial diseases in the Metis community. It is suggested International comparisons of asthma prevalence have only that atopic disease is the price paid by some members of recently been undertaken (Leung & Ho, 1994; Robertson et al, the white community for their relative freedom from 1993; Burr et al, 1994). Burr et al (1994) compared symptoms diseases due to viruses, bacteria and helminths. and exercise fall in peak flow, as a marker of airway hyperre- sponsiveness, in 12-y-old children in Wales, New Zealand, In noting this inverse relationship between infection and South Africa and Sweden. Asthma prevalence and a fall in atopic disease they raised, as others had previously in rural PEFR with exercise were reasonably similar in New Zealand and urban Gambia (Godfrey, 1975) and elsewhere, the para- and Wales but less prevalent in South Africa and Sweden. dox, that in communities exposed to more infection and Asher et al (1998) in ISAAC used simple written and video helminth infestation in particular, total IgE was elevated but questionnaires in the first large-scale international study of atopic disease infrequent. asthma and allergic disease prevalence in children. Almost In 1989 Strachan noted a strong inverse relationship 258 000 6 – 7-y-olds, from 91 centres in 38 countries, and between birth order and hay fever in cohort data in the UK 464 000 13 – 14-y-olds from 155 centres in 56 countries took (Strachan, 1989). He first articulated and coined the phrase part. There was a 15-fold variation in asthma symptoms ‘hygiene hypothesis’ to explain these associations, suggest- between countries, with economically developed countries ing that a greater exposure to infection in early childhood tending to have higher rates of asthma symptoms than less- might arise from contact with siblings. Subsequently a developed countries. This socio-economic gradient was number of studies have identified inverse prevalence rela- apparent within Europe, with Western European countries tionships between atopy or atopic disease and specific infec- tending to have higher prevalence than Eastern European tions (Matricardi et al, 1997; Shaheen et al, 1996; Shirakawa countries. Very high rates were found in some, but not all, et al, 1997). In Italy, with high rates of hepatitis A infection, South American countries, often higher than in Spain or a clear positive relationship has been established between Portugal. In contrast much lower prevalence rates were number of siblings and positive hepatitis A serology found in developing countries such as in India and mainland (Pasquini et al, 1984), providing a direct link between China. The socio-economic gradient apparent in the ISAAC family size, the risk of infection and decreased risk of study does not however appear to explain all of the dif- atopic disease. The epidemiological associations were soon

European Journal of Clinical Nutrition Asthma and allergic diseases J Crane S41 followed by a complementary immunological hypothesis, in Table 1 Antibiotic use ever, by having current wheeze, a which reduced infection and microbial exposure in early life history of wheeze, asthma, asthma and current wheeze, hayfever and eczema would delay the switch from Th2 to Th1 dominant immune response in infancy (Martinez & Holt, 1999). It is in this Adjusted postnatal switch from Th2 to Th1 patterns that infection and odds ratio 95% CI microbial exposure would be expected to operate, microbial Wheeze last 12 months 1.08 0.54 – 2.16 exposure being a potent Th1 stimulus. Wheeze ever 1.86 1.06 – 3.26 The utility of the hygiene hypothesis is that it provides a Asthma ever 2.74 1.10 – 6.85 Asthma ever þ current wheeze 1.22 0.45 – 3.29 clear link between the immunology and the epidemiology of Hayfever ever 1.99 0.93 – 4.26 atopic disease. It might help to explain the long-term time Eczema ever 1.23 0.71 – 2.13 trends, and the sudden appearance of hay fever in the mid nineteenth century. The profound reductions in infectious Table adapted from Wickens et al (1996). disease through improved sanitation and housing, reduced crowding and, later, antibiotic treatment and immunization likely to report the symptoms of asthma hay fever or eczema in the developed economies through the late nineteenth and (Wickens et al, 1999). This remained significant for wheezing twentieth centuries. The same argument would provide ever and a diagnosis of asthma after adjustment for a variety probably the best single explanation for the positive asso- of potential confounding factors (Table 1). We were able to ciations between atopic disease and Western lifestyle show that the risk was greater for antibiotic use in the first (Woolcock & Peat, 1997). When this increased susceptibi- year of life compared with antibiotic use after the first year lity to allergens is combined with increased exposure to (Table 2). We were also able to demonstrate a dose – response common environmental allergens, and to a greater array of effect with children receiving three or four courses of anti- potential allergens, the hypothesis becomes even more biotics in the first year of life having an adjusted odds ratio compelling as an explanation for the long-term rise in the for asthma of 4.0, compared with an odds ratio of 2.3 prevalence of atopic disease. Certainly there is no doubt that children receiving one to two courses in the first year hygiene in its broadest sense has improved beyond recogni- (Table 2). Forooqi & Hopkin (1998) showed an increased tion since the early nineteenth century when hay fever was risk for asthma in a large general practice cohort for children first described and the end of the twentieth century when receiving antibiotics in the first year of life. They were also most of the major infectious diseases, at least in developed able to show that this effect was independent of the number countries have largely disappeared and atopic diseases are at of GP visits, was greater with broad spectrum antibiotics, and epidemic proportions. was evident regardless of the indication for the antibiotic If the hygiene hypothesis stands the tests of time and treatment. Alm et al (1999), in a case – control study of inquiry then it comes with interesting avenues for preven- Rudolf Steiner families in Sweden, showed that in children tion. For example, a number of studies have shown birth adopting an anthroposophic lifestyle, which included less order and family size effects for atopic diseases. As men- antibiotic use, less childhood immunization, and a diet tioned, these are based on the premise that the first-born much higher in fermented vegetables, had significantly less child is exposed to considerably less infection than sub- atopy, as measured by skin prick testing or RAST, than sequent children who are exposed to a variety of infectious matched children living in the same region. In this study agents by their older siblings. This has been extended to the there was an objective measure of atopic status rather than a study of creches, with the suggestion that children who reported history of illness or symptoms. They were not, attend creches in early childhood have more wheezing however, able to tease out the separate effects of individual associated with an increased number of respiratory tract aspects of the anthroposophic lifestyle. The most recent infections, but subsequently are less likely to develop study comes from a case – control study in Belgium in asthma (Ball et al, 2000). Three further developments based on the hygiene hypothesis have been considered, early Table 2 Antibiotic use ever by having current wheeze, a history of antibiotic exposure, early exposure to farms and farm wheeze, asthma, asthma and current wheeze, hayfever and eczema animals, and the role of endotoxin in shaping the allergic phenotype. Adjusted odds ratio 95% CI

Antibiotics used under 1 y Antibiotics Used first year of life 4.05 1.55 – 10.59 Only used after first year of life 1.64 0.60 – 4.46 There have now been four studies in which exposure to Not used at all 1.00 antibiotics early infancy and childhood have been associated Number of courses under 1 y with a reduced prevalence of asthma and other allergic None during the first year of life 1.00 diseases in later childhood. In a study of children attending 1 – 2 courses 2.27 1.14 – 4.51 3 þ courses 4.02 1.57 – 10.31 Rudolf Steiner schools in New Zealand we found that children who had not been exposed to antibiotics were less Table adapted from Wickens et al (1996).

European Journal of Clinical Nutrition Asthma and allergic diseases J Crane S42 which an increased risk of asthma and atopic disease was negative bacteria is capable of causing immune deviation found for antibiotic use in early childhood, but this was only towards a Th1 profile (Lapa e Silva et al, 2000). Such exposures significant for children where there was a family history of would be predicted to reduce allergic sensitization by redu- atopic disease (Droste et al, 2000). cing continued Th2 responses to environmental allergens. These studies need to be interpreted cautiously. There While endotoxin exposure in early life is now being suggested remains the possibility that the effects are caused by reverse as a possible augmentor of Th1 responses, it should not be causation and that the antibiotic treatment was adminis- forgotten that its role is paradoxical. In many occupational tered for symptoms that were due to allergic disease{. The settings, and even possibly in the domestic environment, studies require confirmation, ideally in randomized controlled endotoxin can induce acute airway obstruction, in asthmatics trials of antibiotic avoidance in early childhood. Never- and non-asthmatics (Liu & Redmon, 2001). theless they present a consistent picture in different communities. It is not known how this effect might be mediated. Antibiotics are known to modulate the immune Bowel flora and probiotics system with many of them having an immune depressant The associations between exposure to microbes or microbial effect (Pulverer et al, 1993). It is conceivable that they are products and sensitization to allergens, like the studies of suppressing or eradicating particular infections that are antibiotics, has led to renewed interest in bowel flora, parti- themselves important for reducing a Th2 immune pattern. cularly stimulation of the developing immune system during However as a high proportion of antibiotics are prescribed, at the first year of life. Studies of bowel flora amongst children least in childhood and in developed countries, for viral in Sweden and Estonia have shown that allergic children are infections this would seem less likely. Interest has centred less likely to be colonized with lactobacilli and higher counts on the possible role of antibiotics, and in particular broad of aerobic organisms than their non-allergic counterparts spectrum antibiotics, in altering bowel flora. While it has (Bjorksten et al, 2000). Isolauri et al (2000) have taken this been known for a long time that antibiotics alter bowel flora one step further and used probiotic organisms in randomized this does not appear to have been studied in great detail. The controlled trials both to treat allergic dermatitis in young diarrhoea associated with antibiotic use is well described and children and more recently to prevent allergic dermatitis in occurs often as a result of changes in bowel flora. The advent high-risk children (Kalliomaki et al, 2001). These studies are of DNA probing for specific bacterial species will make the small clinical trials, but in the case of prevention, a 50% study of these effects much easier. reduction in allergic dermatitis at 2 y was observed. Mothers were given probiotics during the latter stage of the preg- nancy, and some infants were also given probiotic supple- Farms and farm animals ments from birth. All these findings are very preliminary Another off-shoot of the hygiene hypothesis that has been but suggest the possibility of modifying immune responses investigated recently concerns a possible role for farm expo- in early life. Probiotic supplementation is particularly sure during childhood. A number of studies predominantly appealing as it could be introduced widely as a public from Central Europe, in particular Switzerland and Germany, health primary prevention strategy. have suggested that early farm exposures, especially farm There is accumulating evidence that in situations where animal exposures, may result in a reduction in the sub- children are likely to be exposed to large numbers and a wide sequent development of atopic diseases. School-age children variety of microbes their immune systems are modified, exposed to farm animals in the first year of life experience resulting in a lower frequency of the atopic phenotype. approximately one-third the risk of asthma of their non- This effect may at least in part explain why allergic diseases exposed counterparts (Braun-Fahrlander et al, 2001). are significantly more common in the developed world. While these results come from the new hygiene paradigm, Similarly there has been an enormous improvement in Blackley (1873) in fact first noted them in the 1870s, when he hygiene and cleanliness both personal and public since hay was studying both the epidemiology and causes of hay fever fever was first described 170 y ago, and the hygiene hypo- and asthma. He noted that hay fever was very uncommon thesis provides an intriguing explanation. It is unlikely, amongst farm workers regularly exposed to grass pollens. Two however, that it represents the whole story. The prevalence separate possibilities might explain these findings. On the of asthma, for example, has increased markedly all over the one hand, higher levels of exposure to common environ- world since the 1960s. This has occurred both in developed mental allergens may induce a degree of protection against and in developing countries. The last two decades have sensitization. A recent study amongst schoolchildren has seen significant increases in asthma prevalence in the UK, shown that higher levels of exposure to cat allergens signifi- Australia and the US and in many other countries where it cantly reduce the risk of sensitization, associated with the would be difficult to demonstrate any major change in production of a modified Th2 response and elevated levels of hygiene or microbial exposure. The ISAAC study showed IgG4 (Platts-Mills et al, 2001). An alternative explanation is the prevalence of asthma symptoms to be 3-fold higher in that farm and particularly farm animal exposures result in the UK compared with Scandinavia or Germany, for exam- increased exposure to endotoxin. Endotoxin from Gram- ple. It is difficult to conceive that hygiene factors could be so

European Journal of Clinical Nutrition Asthma and allergic diseases J Crane S43 different between these European countries as to lead to such Cohen SG (1992): Asthma in antiquity: the Ebers Papyrus. [Classical article.] Allergy Proc. 13(3), 147 – 154. a large variation in prevalence. While the hygiene hypo- Droste J, Wieringa M, Weyler J, Nelen V, Vermiere P & Van Bever H thesis may explain an increase in atopic disease over the last (2000): Does the use of antibiotics in early childhood increase century, it is unlikely to explain changes in the prevalence of the risk of asthma and allergic disease? Clin. Exp. Allergy 30, asthma over the last decade or two. Nevertheless it provides 1547 – 1553. Farooqi I & Hopkin J (1998): Early childhood infection and atopic an interesting and promising avenue for modifying the disorder. Thorax 53, 927 – 932. development of allergic disease. Gerrard J, Geddes C, Reggin P, Gerrard C & Horne S (1976): Serum IgE Amidst the growing enthusiasm for the hygiene hypo- levels in white and Metis communities in Saskatchewan. thesis and its possible extension to explain the increasing Ann. Allergy 37,91– 100. Godfrey RC (1975): Asthma and IgE levels in rural and urban prevalence of Th1-mediated disease we should not forget the communities of The Gambia. Clin. Allergy 5(2), 201 – 207. enormous benefits that increased hygiene, both public and Holgate S (1999): The epidemiology of allergy and asthma. Nature private, has bestowed on developed countries in particular, (Suppl) 402,b2– 4. nor should we lose sight of the fact that for the vast majority Isolauri E, Fau-Arvola T, Arvola T, Fau-Sutas Y, Sutas Y, Fau-Moilanen E, Moilanen E, Fau-Salminen S & Salminen S of the world’s population, their public hygiene provisions (2000): Probiotics in the management of atopic eczema. and their burden of infectious disease is probably worse Clin. Exp. Allergy 30(11), 1604 – 1610. than for Charles Blackley in Britain in the mid-nineteenth Kalliomaki M, Fau-Salminen S, Salminen S, Fau-Arvilommi H, century. Arvilommi H, Fau-Kero P, Kero P, Fau-Koskinen P, Koskinen P, Fau-Isolauri E & Isolauri E (2001): Probiotics in primary prevention of atopic disease: a randomised. Lancet 357, 1076 – 1079. Acknowledgements Lapa e Silva JR, Lefort J & Vargaftig B (2000): Endotoxins, asthma, and allergic immune responses. Toxicology 152,31– 35. Julian Crane is supported, in part, by a senior research Leung R & Ho P (1994): Asthma, allergy and atopy in three south east fellowship from the Health Research Council of New Asian populations. Thorax 49, 1205 – 1210. Zealand. Liu A & Redmon AJ (2001): Endotoxin: friend or foe? Allergy Asthma Proc. 22,337– 340. Magnus P & Jaakkola J (1997): Secular trends in the occurrence of Addendum ({) asthma among children and young adults: critical appraisal of Since this presentation Celedon et al have published a pro- repeated cross sectional surveys. Br. Med. J. 314, 1795 – 1799. spective study of the relationship between antibiotic expo- Martinez F & Holt P (1999): Role of microbial burden in aetiology of Lancet sure and allergic disease upto age 5 years which shows no allergy and asthma. 354 (Suppl), S12 – S15. Matricardi PM, Rosmini F, Ferrigno L, Nisini R, Rapicetta M, association. Chionne P et al (1997): Cross sectional retrospective study of Celedon JC, Litonjua AA, Ryan L, Weiss ST, Gold DR. Lack prevalence of atopy among Italian military students with anti- of Association between Antibiotic Use in the First Year of Life bodies against hepatitis A virus. Br. Med. J. 314, 999 – 1003. Pasquini P, Kahn H, Pileggi D, Pana A, Terzi J & D’Arca T (1984): and Asthma, Allergic Rhinitis, or Eczema at Age 5 Years. Am J Prevalence of hepatitis A antibodies in Italy. Int. J. Epidemiol. 13(1), Respir Crit Care Med 2002; 166: 72 – 5. 83 – 86. Platts-Mills T, Vaughan J, Squillance S et al (2001): Sensitisation, asthma and a modified Th2 response in children exposed to cat References allergen: a population based cross sectional study. Lancet 357, Alm J, Swartz J, Lilja G, Schneyius A & Pershagen G (1999): Atopy in 752 – 756. children of families with an anthroposophic lifestyle. Lancet 353, Pulverer G, Ko H & Beuth J (1993): Immunomodulating effects 1485 – 1488. of antibiotics influencing digestive flora. Pathol. Biol. Paris 41, Asher M, Anderson H, Stewart A & Crane J (1998): Worldwide 753 – 758. variations in the prevalence of asthma symptoms: The Inter- Robertson CF, Bishop J, Sennhauser FH & Mallol J (1993): Interna- national Study of Asthma and Allergies in Childhood (ISAAC). tional comparison of asthma prevalence in children: Australia, Eur. Resp. J. 12, 315 – 335. Switzerland, Chile. Pediatr. Pulmonol. 16(4), 219 – 226. Ball T, Castro-Rodriguez J, Griffith K, Holberg C, Martinez F & Wright A Shaheen SO, Aaby P, Hall AJ, Barker DJ, Heyes CB, Shiell AW et al (2000): Siblings, day-care attendance, and the risk of asthma and (1996): Cell mediated immunity after measles in Guinea-Bissau: wheezing during childhood. New Engl. J. Med. 343, 538 – 543. historical cohort study. Br. Med. J. 313, 969 – 974. Bjorksten B, Naaber P, Sepp E & Mikelsaar M (2000): The intestinal Shirakawa T, Enomoto T, Shimazu S & Hopkin JM (1997): The inverse microflora in allergic Estonian and Swedish 2-year-old children. association between tuberculin responses and atopic disorder. Clin. Exp. Allergy 29, 342 – 346. Science 275,77– 79. Blackley C (1873): Experimental Researches on the Causes and Nature of Strachan DP (1989): Hay fever, hygiene, and household size. Br. Med. Catarrhus Aestivus. London: Balliere, Tindall & Cox. J. 299, 1259 – 1260. Bostock J (1819): Case of a periodical affection of the eyes and chest. van Helmont J (1662): Oriatrike, or a Physic Refined. London: Med. Chir. Trans. x, 161 – 162. Lodowich LLoyd. Bostock J (1828): On the catarrhus aestivus or summer catarrh. von Mutius E, Platts-Mills T, Woodfolk J & Sporik R (2001): Med. Chir. Trans. xiv, 437 – 446. The increase in asthma can=cannot be ascribed to cleanliness. Braun-Fahrlander C, Eder W, Schreuer M et al (2001): Exposure to (Editorial.) Am. J. Respir. Crit. Care Med. 164, 1106 – 1109. farming environment during the first year of life protects against Wickens K, Pearce N, Crane J & Beasley R (1999): Antibiotic use in the development of asthma and allergy. Am. J. Respir. Crit. Care early childhood and the development of asthma. Clin. Exp. Allergy Med. 163(A157). 29(6), 766 – 771. Burr M, Limb E, Andrae S, Barry D & Nagel F (1994): Childhood Woolcock AJ & Peat JK (1997): Evidence for the increase in asthma asthma in four countries: A comparative survey. Int. J. Epidemiol. worldwide. Ciba Found. Symp. 206, 122 – 134; discussion 134 – 139; 23(2), 341 – 347. 157 – 159.

European Journal of Clinical Nutrition