| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
|
FJ
EXPRESS SUMMARY ARTICLE The Full-length version of this article is also available, published online February 5, 2003 as doi:10.1096/fj.02-0661fje. |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,2,
,
* Departments of Environmental Health and Neurology, Boston University Schools of Medicine and Public Health, Boston, Massachusetts, USA;
Institute of Public Health, University of Southern Denmark, Odense, Denmark;
Department of Biostatistics, Panum Institute, University of Copenhagen, Copenhagen, Denmark;
Department of Occupational and Public Health; Faroese Hospital System, Tórshavn, Faroe Islands;
¶ State Agency for Health and Occupational Safety of Schleswig-Holstein, Flintbek, Germany;
** National Center for Environmental Health, Centers for Disease Control and Prevention, Atlanta, Georgia; and
Department of Clinical Biochemistry, Odense University Hospital, Odense, Denmark
2Correspondence: Department of Environmental Health, Harvard School of Public Health, 665 Huntingdon Ave., Boston, MA 02118, USA. E-mail: pgrand{at}health.sdu.dk
SPECIFIC AIMS
Breast feeding is often associated with slowed postnatal growth. The basis for this "weanling’s dilemma" is unclear, and we therefore examined the potential effect on growth due to exposure to environmental contaminants from human milk.
PRINCIPAL FINDINGS
1. Prolonged breast feeding was associated with lower body weight and height at 18 months
We studied prospectively a Faroese birth cohort of 182 singleton children born at term. Because traditional food includes pilot whale, this fishing population has a wide range of exposures to seafood contaminants. Body weight and height were recorded at 18 and 42 months of age.
Exclusive breast feeding averaged 3.5 ± 2.0 (SD) months and showed a significant negative association with postnatal anthropometric parameters (Fig. 1
). Adjustment for confounders (birth weight, gestational age, sex, parity, maternal weight and height, gestational diabetes, maternal smoking and alcohol use, and the child’s age at examination) did not influence this finding. At 18 months, children who had been exclusively breast fed for at least 6 months weighed 0.59 kg less (95% confidence interval {CI}=0.03, 1.16 kg) and were 1.50 cm (95% CI=0.52, 2.47 cm) shorter than those not breast fed. The findings at 42 months were similar, but the associations all but vanished if body size at 18 months was included as a covariate. These results replicate those obtained in many previous studies.
|
2. Attenuated postnatal growth was associated with prenatal contaminant exposure
Concentrations of mercury in cord blood and of polychlorinated biphenyls (PCBs) in maternal milk were measured as markers of developmental exposure to the main seafood contaminants from the maternal diet. The total PCB concentration was estimated as the sum of the three main congeners CB-138, CB-153, and CB-180 multiplied by 2.0; when expressed on a lipid basis, the milk concentration also reflects prenatal exposure levels. The geometric mean concentrations (range) were 20.3 (1.9–102) µg/L and 1.52 (0.07–18.5) µg/g lipid for mercury and PCB, respectively. Because of skewed distributions, logarithmic transformations were applied to obtain a better fit of the models.
Especially the cord blood mercury concentration was negatively associated with the 18 month weight and height. Irrespective of duration of breast feeding, a doubling of the mercury concentration in cord blood was associated with a decrease in weight at 18 months by 0.19 kg (95% CI=0.03, 0.35 kg) and in height by 0.26 cm (95% CI=-0.02, 0.55 cm). Multiple regression analysis suggested that duration of breast feeding and prenatal methylmercury exposure may independently attenuate postnatal growth.
The PCB concentration was a less important predictor, and inclusion of this variable in the regression equation changed the mercury associations only negligibly. The same was true for the eicosapentaenoic acid (EPA) concentration in cord serum phospholipids, which remained a significant negative covariate after confounder adjustment.
Similar tendencies were observed at 42 months, but the prenatal contaminant exposure seemed to have no effect on the growth pattern when adjusted for body size at 18 months.
3. Calculated contaminant exposure from milk was related to attenuated postnatal growth
The effect of lactational exposure to mercury and PCB was examined using the product terms of contaminant concentrations and the duration of exclusive breast feeding. The logarithmic transformations were included as independent variables in multiple regression analyses.
Lactational exposure to mercury and PCB showed significant negative associations with weight and height at 18 months. Mercury remained significant when both contaminant parameters were included, despite the fact that a shorter biological half-life of methylmercury would render this parameter less precise.
Duration of breast feeding and lactational contaminant exposure were then entered into joint regression models (Table 1
). These results showed that the negative effect of breast feeding duration as such was completely abolished by adjustment for lactational mercury exposure. However, the results did not allow determination of the relative effect of prenatal and lactational mercury exposure. Yet mercury exposure remained statistically significant after adjustment for PCB. As before, regression results at age 42 months were similar if not adjusted for body size at 18 months.
|
4. The child’s serum PCB concentration was a predictor of lower body size
A serum sample was obtained from the child at age 54 months and the PCB concentration was measured. This parameter was used as a proxy variable for accumulated lactational exposure and was logarithmically transformed. The serum PCB showed a significant negative association with weight and height at age 42 months. However, a heavier child would also have a greater distribution volume for PCB, thus confounding the association between body size and the exposure measure. The body mass index was therefore included as a covariate in addition to the confounders previously identified. This adjustment much reduced the regression coefficient and indicated that a doubled PCB concentration was associated with a decrease in 42 month weight by 0.30 kg (95% CI=0.07, 0.55 kg) and in height by 0.63 cm (95% CI=0.13, 1.12 cm). These significant (P=0.08 in both cases) associations were again much reduced when adjusted for weight at 18 months. Adjustment for EPA did not affect any of the associations observed.
CONCLUSIONS AND SIGNIFICANCE
Breast feeding is thought to confer an advantage for growth during the first few months, but data on continued growth generally suggest that continued breast feeding has a negative effect. Although these studies were conducted in countries where human milk contaminants may have included PCB and pesticides, the association has been attributed to differences in the caloric intake, risk of malnutrition, possible confounding due to social factors, and perhaps in some cases a reverse causation. The present study replicates the association between breast feeding and attenuated postnatal growth and suggests that the list of possible causal factors should include human milk contaminants.
Although the cohort studied was limited in size, it was comparable to many previous studies. An advantage is that the Faroese background population is homogeneous and without nutritional deficiencies or serious social confounding. Moreover, the wide interval of contaminant exposures adds to the statistical power of the study.
The exposure biomarkers must be interpreted in light of the time of sample collection and the known fate of the chemicals in the body. Random imprecision of calculated lactational exposure levels would bias the regression results toward the null hypothesis. The biological half-life of methylmercury in the human body is probably 
45 days, although perhaps longer in breast-fed infants. Thus, the estimate of total lactational methylmercury exposure is likely to be imprecise. In contrast, half-lives of PCB congeners may be several years, but concentrations in body fluids will be reduced, e.g., by expansion of the lipid compartment. The strong association between weight at 42 months and the subsequent serum PCB concentration at 54 months must therefore be cautiously interpreted.
Intrauterine contaminant exposures could possibly determine postnatal growth through prenatal programming, as suggested by the well-documented, long-term consequences of intrauterine growth retardation and the observation that birth weight is an important predictor for height throughout childhood. Although environmental contaminant exposure did not affect birth weight of the cohort children, clear negative associations were seen with postnatal growth before 18 months of age. No further attenuation was apparent after this age. Adverse effects of environmental pollutants on postnatal growth have been convincingly documented in poisoning episodes, but the contaminant mixtures involved in these incidents may differ from those commonly occurring in human milk. Other studies have reported inconsistent effects of environmental PCB exposures on postnatal growth, but methylmercury and fatty acids were not taken into account. However, a negative effect of PCB on growth has been shown in experimental studies of primates and rodents. Decreased postnatal weight gain has also been reported in rats in relation to maternal methylmercury exposure.
Exposure to seafood contaminants is a worldwide concern, and levels similar to or in excess of those recorded in the Faroes have been published from other communities with high seafood or freshwater fish intakes. The present study indicates that postnatal growth may be adversely affected by these substances from a maternal seafood diet. It also suggests that the previously observed decreased growth in breast-fed infants could be due to environmental contaminants in human milk. Effects on weight and height were observed at 18 months of age and were still apparent 2 years later. Although the present study included subjects with substantially increased exposures, the safe margin for populations with lower exposures could be rather slim.
FOOTNOTES
1 To read the full text of this article, go to http://www.fasebj.org/cgi/doi/10.1096/fj.02-0661fje; to cite this article, use FASEB J. (February 5, 2003) 10.1096/fj.02-0661fje 
This article has been cited by other articles:
|
|
 |
|
  B. Ling and J. Alcorn Acute Administration of Cefepime Lowers L-Carnitine Concentrations in Early Lactation Stage Rat Milk J. Nutr., July 1, 2008; 138(7): 1317 - 1322. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Grandjean Contaminants in fish oil Am. J. Clinical Nutrition, December 1, 2005; 82(6): 1354 - 1354. [Full Text] [PDF]
|
|
|
|
 |
|
 P. Grandjean and A. A. Jensen BREASTFEEDING AND THE WEANLING'S DILEMMA Am J Public Health, July 1, 2004; 94(7): 1075 - 1075. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
