© 2008 American Thoracic Society doi: 10.1164/rccm.200804-546ED
Peanuts in PregnancyWhat Is a Mother to Do?
National Heart and Lung Institute In 1998, the U.K. Department of Health, uniquely among governments, recommended that expectant mothers with an atopic history should consider avoiding the consumption of peanuts during pregnancy; 2 years later, the American College of Pediatrics concurred that, although maternal dietary restrictions in pregnancy were unnecessary, peanuts were a "possible exception." In each case, the advice was designed to reduce the apparent epidemic of childhood peanut allergy in these countries, but in neither case does it seem to have been effective (1). Advice such as this may fail because it goes unheeded, or because it is erroneous. There is some evidence for the former explanation in the United Kingdom (2) and increasing concern over the latter (3); indeed, some have argued that avoidance may be counterproductive because it precludes the development of immunotolerance. It is always of concern when official policy is developed on a background of weak evidence, especially when it relates to advice for vulnerable groups. The specific etiology of childhood peanut allergy remains obscure and any understanding we have is based entirely on observational—and frequently on ecological—evidence alone. This issue of the Journal includes further observational data that may be used by protagonists in the peanuts-in-pregnancy debate. Willers and colleagues (pp. 124–131) report analyses of maternal consumption of foods in late pregnancy in relation to childhood asthma and atopy in a Dutch birth cohort (4). Their main finding was of an increased risk of asthma symptoms, but not atopy, with daily (compared with "rare") consumption of nut products, but not nuts per se. The authors are confident that "nut products" mainly comprise peanut butter in this population; nuts included both true ("tree") nuts and peanuts. However, evidence to suggest that this is a causal relationship is not compelling. First, the statistical analyses were unusual. The authors used generalized estimating equations to examine the relation between maternal diet in pregnancy and a summary measure of symptoms in the first 8 years of childhood, based on repeated annual outcome measurements. A limitation of this approach is that, given the different etiologies of wheezing phenotypes in early childhood (5), "lumping" early wheezing with later wheezing may obscure important relationships. This is illustrated by the fact that maternal fruit and vegetable intake was not associated with wheezing overall, but a protective effect of daily intake was seen on early transient wheezing. In contrast, daily nut product consumption was associated with persistent wheezing but not early transient wheezing. Puzzlingly, the authors condensed the original food frequency categories from seven to only two or three, even though this resulted in thousands of individuals in some categories. Consequently they limited their ability to detect effects and to look for trends in associations. For the analysis of nut products, "regular" (less than daily) consumption was not associated with an increased risk of asthma compared with "rare" consumption, suggesting there was no dose–response relation. While it is possible that there might be a threshold level of consumption above which nut products confer increased risk, why this should be is unclear. Second, given that this was an unexpected finding, and in view of the multiple statistical comparisons (with correlated outcomes), it is possible that the main result could have occurred by chance; P values, which would have aided interpretation in this respect, are not presented. Third, confounding by lifestyle factors and by other foods or nutrients that have not been controlled for is always a concern in nutritional epidemiology. For example, total energy intake may be important, given that peanut butter is an energy-dense food, as may consumption of whole grain foods in pregnancy and childhood; previous cross-sectional analyses in this cohort suggested a negative association between brown bread consumption and asthma in early childhood (6). Finally, there does not appear to be a plausible mechanism for the association. Given that there was no relation with nut consumption, the authors speculate that the link between wheeze and nut products reflects the high linoleic acid content of peanut butter. The Black Sharpe hypothesis is mentioned, a proposal that increased consumption of n-6 fatty acids such as linoleic acid promotes the development of atopy through an increase in prostaglandin E2 (PGE2) production (7). However, nut product consumption did not relate to the development of atopy, a strong predictor of peanut allergy, nor was it related to reported doctor-diagnosed peanut allergy (specific sensitization to peanut was not measured directly). Furthermore, the hypothesis might be challenged in relation to asthma, given that PGE2 may protect against airway inflammation and bronchoconstriction (8, 9). It would have been of interest to analyze relationships between atopic outcomes and maternal consumption of margarine, another rich source of linoleic acid, in view of a previous link with eczema in early childhood (10). So where does this leave the expectant and concerned mother—or indeed the official tasked with writing guidance for this vulnerable group? The primary debate continues to be over peanut allergy in childhood; we suggest that this report does little to advance that debate, although the absence of an increased risk of reported peanut allergy in association with heavy intake of peanut butter is reassuring. In relation to childhood asthma we can only agree with the authors: these findings should be interpreted with caution to prevent any further confusion, and need to be replicated before any further advice can be given to pregnant women. FOOTNOTES Conflict of Interest Statement: Neither author has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. REFERENCES
Related articles in AJRCCM:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
