The recent article by Gulson et al. (1) should provide reassurance to most women that their breast milk is free of significant lead contamination. Their data found that breast milk contained lower levels of lead than infant formula or infant food. Although small amounts of lead may be found in all human tissues, the milk of the vast majority of women does not present a lead hazard to their babies. Unfortunately, the authors formed unwarranted conclusions from their data, and the title of the press release, "International Study Finds Mothers' Lifetime Lead Exposures May Put Breast-fed Newborns at Risk," (2) was misleading. As a result, women reading this press release may decide not to breast-feed, thus depriving their babies of the most healthful food available to them and placing their infants at increased risk for a variety of infectious diseases. This press release is in conflict with the decade-old public health goals to increase both the percentage of mothers who breast-feed and duration of breast-feeding in the United States (3). Why did this happen? What lessons can we learn from the chain of events that led to the release of such misinformation? How can we avoid exposing the public to misinformation that can have dangerous public health consequences?
Gulson et al. (1) examined a small, nonrepresentative sample of 15 eastern European women emigrants and their Australian-born babies (n = 16) and compared them with 6 second-generation Australian women and their 8 babies. None of the women were exposed to lead except through background levels in the diet. The study reported no difference in either blood lead or breast milk lead concentrations of European emigrants versus Australian mothers. All had low blood lead values [geometric mean (GM) = 2.02 µg/dl; range 0.91-3.61 µg/dl]. They also had very low levels of lead in their breast milk (GM = 0.7 µg/kg; range 0.09-2.09 µg/kg), lower than lead levels in infant formula (GM = 1.8 µg/kg; range 0.36-4.3 µg/kg) or infant foods (GM = 4.1 µg/kg; range 1.4-27 µg/kg). The authors reported that maternal blood isotope ratios and breast milk lead concentrations predicted infant blood isotope ratios, although the level of statistical significance for this analysis (p = 0.09) was, at best, marginal and the coefficient for infants' blood lead concentration was negative. Gulson et al. (1) also estimated the percentage of each child's blood lead attributable to breast milk (36-80%) and/or infant formula (24-68%). The authors did not estimate the uncertainty surrounding these percentages, they did not provide data on infants' blood lead levels, and they did not compare infants' blood levels related to breast milk and formula. We believe the data of Gulson et al. (1) suggest that compared with other infants, breast-fed infants are not at increased risk from lead poisoning.
The authors' analysis of predictors of lead isotopic ratios is not meaningful in establishing risk for lead poisoning. While the isotopic ratio may be useful in establishing the source of lead, it is the mother's body burden of lead and total blood lead level that influence breast milk concentrations of lead and, ultimately, the infant's exposure to lead from breast milk.
Gulson et al. (1) speculate on which women are highly exposed to lead and then suggest a public health practice recommendation of screening women for lead body burdens. While it may be useful to recommend screening for highly exposed women, we suspect that only a small fraction of women are highly exposed. Screening recommendations should accurately target exposed women, such as those employed in lead-exposed jobs where blood lead levels are routinely measured and medical removal already occurs (4). In addition, action levels at which breast-feeding is contraindicated must be set before targeted screening could be made meaningful. No action level was provided by Gulson et al. A useful action level must weigh both the risks and benefits from breast milk. The Health Resources and Services Administration (HRSA) has published a blood level action level of 40 µg/dl or above as a contraindication for breast-feeding (5). None of the 2,925 women 15-44 years of age who participated in Phase 2 of the Third National Health and Nutrition Examination Survey had a concentration that high, and only two women had levels even half that high (6). The Health and Human Services Federal Advisory Committee on the Prevention of Childhood Lead Poisoning has not considered a recommendation about screening of women to prevent lead poisoning in children, but it is the appropriate group to do so if such a recommendation were necessary.
We became aware of the study by Gulson et al. (1) when the Centers for Diseases Control and Prevention (CDC) received media inquiries about the NIH press release describing the article. We were surprised that the press office would endorse a title that could discourage women from breast-feeding, in spite of well-established public health goals to the contrary. We became alarmed after we examined the article and contrasted the actual data in the report with the authors' conclusions and the press release. Healthy People 2000 set the nation's goal for the proportion of mothers breast-feeding their infants at 75% during the early postpartum period, 50% through 6 months, and 25% through the first year of life (3). Because these goals have not been achieved, they have been directly adopted into Healthy People 2010. Informing the public about important scientific findings is valuable, but it is crucial that the media message be accurate. Just as editors and press officials must verify technical information with technical experts other than the authors, so must they verify health policy guidance with public health experts.
As noted in the draft of Healthy People 2010 (7),
Breast milk is widely acknowledged as the most complete form of nutrition for infants, with a range of benefits for infants' health, growth, immunity, and development.
Breast-feeding benefits the infant and the mother. The American Academy of Pediatrics considers breast-feeding to be the ideal method of feeding and nurturing infants (8). The study by Gulson et al. (1) included a very small number of infant-mother pairs, and any conclusions drawn from it should be made cautiously. Their data provided no evidence that breast-feeding puts children at risk for lead poisoning. We believe that the authors' recommendations to screen certain women for elevated blood lead levels to prevent childhood lead poisoning through breast-feeding are premature. Such recommendations, if made, should be developed by an appropriate group who would consider both the benefits and possible risks from breast-feeding. After all, babies must be fed something, and all the evidence, including the report by Gulson et al. (1), support the axiom "breast is best."
Thomas Sinks
Richard J. Jackson
National Center for Environmental Health
Centers for Disease Control
and Prevention
Atlanta, Georgia
References and Notes
1. Gulson BL, Jameson CW, Mahaffey KR, Mizon KJ, Patison N, Law AJ, Korsch MJ, Salter MA. Relationships of lead in breast milk to lead in blood, urine, and diet of the infant and mother. Environ Health Perspect 106:667-674 (1998).
2. International Study Finds Mothers' Lifetime Lead Exposures May Put Breast-fed Newborns at Risk. Press Release. NIEHS PR 17-98. Research Triangle Park, NC:National Institute for Environmental Health Sciences, 1998.
3. Healthy People 2000. National Health Promotion and Disease Prevention Objectives. PHS 91-50213. Washington, DC:US Government Printing Office, 1991.
4. Occupational Safety and Health Standards. Asbestos. 29 CFR§1910.1025 (1995).
5. Lawrence RA. A Review of the Medical Benefits and Contraindications to Breastfeeding in the United States (Maternal and Child Technical Information Bulletin). Arlington, VA:National Center for Education in Maternal and Child Health, Health Resources and Services Administration, 1997.
6. National Center for Health Statistics. Phase 2, Third National Health and Nutrition Examination Survey Data. 1991-1993. Hyattsville, MD:US Department of Health and Human Services. Machine readable public-use data tapes, 1996.
7. Healthy People 2010 Objectives: Draft for Public Comment. Washington, DC:Office of Disease Prevention and Health Promotion, 1998.
8. American Academy of Pediatrics Work Group on Breastfeeding. Breastfeeding and the use of human milk. Pediatrics 100(6):1035-1039 (1997).
Sinks and Jackson take strong issue with a recent paper in EHP by Gulson et al. (1) on the biokinetics of lead transfer from maternal bone and other body lead stores to nursing infants via human breast milk. I wrote an invited Research Highlights paper (2) that accompanied and discussed the larger context and implications of Gulson et al. While Sinks and Jackson did not include my paper in their letter, their criticisms apply to a number of issues addressed in my article.
Sinks and Jackson would have us believe Gulson et al. (1) were guilty of a public health heresy when the latter noted that there may be potential risks for nursing infants if lead intakes via breast milk are elevated because of elevated long-term maternal lead exposures. My quantitative lead exposure risk calculations for nursing infants in Table 1 make it clear that the potential impact of maternal lead burdens for nursing infants across a range of maternal blood lead (BPb) values is not a trivial matter.
The clearly indicated goals of Gulson et al. (1) were to examine and quantitatively characterize 1) in vivo lead movement in nursing mothers, specifically bone lead resorption during lactation and nursing, and 2) the toxicokinetic interplay between endogenous (bone) and exogenous (diet) lead in the bodies of these mothers as they relate to transport of maternal lead to breast milk and then to nursing infants. Their findings document that bone lead releases can contribute significantly to breast milk lead and ultimately to infant lead intake in terms of lead source fractional input. The breast milk study was the latest in a published peer-reviewed series by Gulson et al. that used the method of stable lead isotopic ratio analysis to quantify the contribution of bone lead to BPb and lead in other metabolic compartments, and the temporal character of such inputs.
The findings revealed by Gulson et al., when examined with the many studies of breast milk lead levels in lactating and nursing women, indicate that the toxicokinetic parameters governing lead transfer at low concentrations of BPb to breast milk apply to other cases where there were or are high maternal lead exposures. This especially applies to the ratio of breast milk lead to BPb concentrations, which appears to increase at higher maternal BPb levels.
Neither the Gulson et al. paper nor my perspective article engaged in undue speculation about risks to the early infant from lead exposures arising from quite elevated breast milk lead concentrations. A comparative analysis of the many studies documenting the quantitative ratios between maternal BPb and associated breast milk lead levels readily shows that 1) as BPb increases, not only does the amount of milk lead increase but the fractional distribution may also increase; and 2) at high maternal lead exposures sufficient to produce high maternal BPb levels, mothers will have breast milk lead that may be problematic for their infants' lead exposures.
Sinks and Jackson take Gulson et al. to task for suggesting screening of nursing mothers, particularly those suspected of past or present high lead exposures. This is a peculiar criticism. That lead will enter breast milk from maternal body lead stores in proportion to the lead exposures in nursing mothers is far from new information. New data of Gulson et al. that show a significant fraction of maternal lead released into breast milk would be derived from very high bone lead levels following maternal chronic high lead exposures merely add a transgenerational dimension to established phenomena. They permit one to conclude that such maternal exposures should be monitored. This is the only way to identify the extent of lead releases to breast milk and ultimately to infants. The suggestion is hardly inappropriate.
Sinks and Jackson cite a 1997 report from the government's Health Resources and Services Administration's (HRSA's) National Center for Education in Maternal and Child Health on the medical benefits and contraindications for infant breast milk feeding. The report was authored by an authority on the topic, Ruth A. Lawrence (3). Lawrence's report says clearly that in the case of likely elevated lead exposures, it is advisable to not only screen children (to presumably include nursing infants) but to screen mothers as well.
Sinks and Jackson state that Lawrence's report (3) identified a maternal BPb level of 40 µg/dl as the upper limit of safe in terms of the amount of lead entering breast milk and amounts of lead ingested by nursing infants. However, the only citation in Lawrence's report used as the basis for the ceiling figure of 40 µg/dl is a summary 1994 article in the Centers for Disease Control and Prevention's (CDC's) Morbidity and Mortality Weekly Report (MMWR) on preliminary Phase 1 National Health and Nutrition Examination Survey (NHANES) III data (4). This MMWR article, however, contains no discussion of lead in breast milk and provides no evidence for selection of any particular infant breast-feeding safety limit in terms of maternal BPb. In particular, it cannot be used to justify 40 µg/dl as the upper limit of a safe maternal BPb for nursing mothers. It is also clear that Lawrence (3) is not comfortable with the use of a maternal BPb as high as 40 µg/dl. As noted, Lawrence recommended testing nursing infants for lead exposure even if maternal BPb concentrations are below 40 µg/dl and to do environmental lead assessments if maternal BPb is above 10 µg/dl.
Lawrence (3) also implies that a breast milk lead level is acceptable if such levels do not materially add to infant lead burdens. Specifically, not only would no net accumulation occur if infant lead intake is less than 5 µg/day, but infant lead burdens acquired in utero would begin to show net excretion (negative lead balance) at such low lead intakes.
Use of the high and obsolete maternal BPb value of 40 µg/dl as the upper limit for producing safe milk lead content, even if it were somehow still relevant, raises the obvious question of what breast milk lead level would be associated with this maternal BPb guideline. Lawrence (3) states that breast milk lead content can typically range up to 15-20% of the maternal BPb level. This is a range consistent with a number of breast milk lead studies in which BPb levels were elevated. Selection of 15% as the ratio results in a breast milk level of 60 µg/l (6 µg/dl) for a maternal BPb of 40 µg/dl.
I am not aware of any published toxicokinetic analysis or any other credible quantitative risk assessment of breast milk lead intakes by nursing infants that would validate a maternal BPb level of 40 µg/dl as the upper limit of safe with regard to resulting lead levels in breast milk and infant BPb levels. I have carried out an analysis of the 40 µg/dl BPb value, as part of a series of selected maternal BPb levels, in terms of resulting infant BPb levels. The results are in Table 1.
Table 1 presents modeled infant exposures using the EPA's Integrated Exposure-Uptake Biokinetic (IEUBK) computer model for this purpose. This well-validated model is in routine use by risk assessors (5). The modeling results provide both geometric mean (GM) BPb concentrations and the percentages of these nursing infants, 0-6 months of age, who would exceed the CDC action level of 10 µg/dl and exceed the medical intervention, Class III CDC risk level of 20 µg/dl. Infant BPb mean levels in Table 1 are model-estimated from postnatal milk lead intakes plus body lead at birth from prenatal maternal exposures.
Table 1 shows that the only "safe" level, in terms of the fraction of infants with >20 µg/dl, would be maternal BPb of 15 µg/dl if no more than 5% of infants are to exceed the 20-µg/dl figure. If the risk management restriction is no more than 1% to exceed this value, a maternal BPb value of 10 µg/dl, as seen in Table 1, should not be exceeded. If no more than 5% of breast-feeding infants are to exceed the 10 µg/dl action level, then a maternal BPb <10 µg/dl is required.
A sensitivity analysis using various IEUBK modeling runs shows that the infant body lead burden at birth, from in utero accumulation via maternal lead exposures, is mainly expressed through estimated BPb values in the first 6 months of infant life, as compared to the second 6 months of infant life or as compared to exposure integrated over the entire first year of infant life. This is to be expected, given the relatively high biokinetic mobility of lead in the very young. However, it is precisely in the first 6 months of infant life that breast-feeding is done. Therefore, both breast milk lead and prior infant body lead burden are significant sources of lead in breast-feeding infants of mothers with elevated lead exposures. In essence, the only maternal BPb level that is in fact "safe" in terms of CDC Class III elevated infant BPb figures also approximates the CDC infant BPb action level of 10 µg/dl. In terms of the child action level of 10 µg/dl, a maternal BPb <10 µg/dl appears prudent.
The 1991 CDC statement on childhood lead poisoning (6) identified a BPb level of 10 µg/dl as being the body lead threshold associated with the earliest toxic effects in infants and toddlers. The CDC document also accepted the risk assessment premise that there is no known threshold for lead's subtle toxicity.
Sinks and Jackson argue that the most recent NHANES III, 1991-1994, indicates that there are no women in the United States who are likely to be nursing their infants and who have BPb values anywhere close to the 40 µg/dl Sinks and Jackson claimed as permissible for nursing mothers. They cite some actual numbers noted in the NHANES III data tapes (7). Such prevalence data are aggregated cluster sample depictions at a single time point of the U.S. population lead exposure picture, stratified by national socioeconomic and demographic strata. One cannot legitimately disaggregate such national depictions or "snapshots" to generate comparisons for individual community prevalences or to use actual BPb values contained in any particular statistical sampling cell in the aggregation process. Such limits are discussed in, among other things, the Executive Summary of the 1988 report to the U.S. Congress on childhood lead poisoning (8) by the Agency for Toxic Substances and Disease Registry.
Sinks and Jackson offer the simplistic and incorrect argument that if women in the United States were nursing their infants and had elevated BPbs from workplace exposures, they would be readily and reliably detected by Occupational Safety and Health Administration (OSHA)-required exposure and medical surveillance. Many small operations are either exempt from OSHA requirements because of size or are rarely if ever inspected because of severe resource constraints on federal or various state OSHA agencies. Such assessments similarly would not detect the women who potentially have elevated BPb values owing to environmental, not occupational, lead exposures. A much better approach, regardless of sources of high maternal exposure, would be the approach endorsed by Lawrence (3): maternal/infant screening.
The letter by Sinks and Jackson may be seen by some as another example of an overall CDC retreat from lead as a persisting child health issue. Needleman (9) noted an overall backpedaling in efforts and decline in momentum to finally address the lead issue in a meaningful way by the federal government. All this raises a legitimate question among scientists and clinicians interested in lead: Is lead still considered to be a child health risk issue at the CDC or elsewhere in the federal government?
The actual content of new research should be understood before wholesale attacks on such research are launched. This is especially the case where complex research designs and equally complex results are at work. In those cases where breast milk does not contain worrisome lead concentrations from high maternal lead exposures, per Table 1, I agree with the Sinks and Jackson comment that "breast is best."
Paul Mushak
PB Associates
Durham, North Carolina
References and Notes
1. Gulson BL, Jameson CW, Mahaffey KR, Mizon KJ, Patison N, Law AJ, Korsch MJ, Salter MA. Relationships of lead in breast milk to lead in blood, urine, and diet of the infant and mother. Environ Health Perspect 106:667-674 (1998).
2. Mushak P. New findings on sources and biokinetics of lead in human breast milk: mother's bone lead can target both nursing infant and fetus. Environ Health Perspect 106:629-631 (1998).
3. Lawrence RA. A Review of the Medical Benefits and Contraindications to Breastfeeding in the United States. Maternal and Child Health Technical Information Bulletin. Arlington, VA:National Center for Education in Maternal and Child Health, Health Resources and Services Administration, 1997.
4. Blood lead levels-United States. 1988-1991. Morb Mortal Wkly Rep 43:545-548 (1994).
5. U.S. EPA. Guidance Manual for the Integrated Exposure Uptake Biokinetic Model for Lead in Children. EPA/540/R-93/081. Washington, DC:Office of Emergency and Remedial Response, 1994.
6. CDC. Preventing Lead Poisoning in Young Children: A Statement by the Centers for Disease Control. Atlanta, GA:Centers for Disease Control and Prevention, 1991.
7. National Center for Health Statistics. Phase 2, Third National Health and Nutrition Examination Survey Data. 1991-1993. Hyattsville, MD:US Department of Health and Human Services. Machine readable public-use data tapes, 1996.
8. ATSDR. The Nature and Extent of Lead Poisoning in Children in the United States: A Report to Congress. Atlanta, GA: Agency for Toxic Substances and Disease Registry, 1988.
9. Needleman HL. Childhood lead poisoning: the promise and abandonment of primary prevention. Am J Public Health (in press).
We appreciate Sinks' and Jackson's interest in our article on lead in breast milk and would like to reinforce and clarify a couple of points. Paul Mushak has responded comprehensively to their letter, and we are in complete agreement with his response.
We agree wholeheartedly that "breast is best," and our low concentrations of lead in breast milk confirm this. Our abstract (1) was quite emphatic that
Breast-fed infants are only at risk if the mother is exposed to high concentrations of contaminants either from endogenous sources such as the skeleton or exogenous sources.
Sinks and Jackson are dismissive of the use of lead isotopic ratios as not being "meaningful in establishing risk for lead poisoning." Perhaps this is true in the strict sense of risk assessment, but lead isotopic ratios are the only realistic method of determining the source of a mother's lead burden.
Sinks and Jackson also appear dismissive of the recommendation of screening women for lead body burden via blood lead levels. We stress that this would be only for those women that were exposed in the past or are currently exposed to lead. We would certainly extend any screening to women employed in lead-exposed jobs as they cite. However, the overwhelmingly vast majority of women (and their partners and infants) at risk are those exposed during house renovations involving leaded paint, from dust from ceiling and wall cavities, from "take-home" dust from their jobs, from hobbies, etc.; this can affect all socioeconomic levels. In fact, the most disturbing aspect of lead poisoning from do-it-yourself activities is that most of these people are unaware of the dangers.
No action level was given in our paper because there are no accepted guidelines for blood lead levels and breast-feeding apart from the Lawrence document (2) noted by Sinks and Jackson, the relevance of which has been questioned by Mushak in his letter. In the absence of guidelines, we recommend that if the mother-to-be is concerned that she may have been heavily exposed to lead from any source at any time, she request a blood lead measurement either before conception or at least during the first trimester. If the blood lead level is greater than two times the CDC "level of concern" (i.e., >20 µg/dl), we recommend that she have her breast milk tested by a reputable laboratory.
In addition, we suggest that mothers-to-be maintain healthy diets and consume the NIH recommended daily intake of calcium of 1,100-1,200 mg Ca/day during pregnancy and up to 1,400 mg Ca/day during breast-feeding (3). This will not only potentially lessen the mobilization of endogenous lead from the maternal skeleton as shown in our recent studies (4,5) but also lessen the uptake of exogenous lead from the gastrointestinal tract.
Brian L. Gulson (on behalf of the authors)
Graduate School of the Environment
Macquarie University
Sydney, Australia
References and Notes
1. Gulson BL, Jameson CW, Mahaffey KR, Mizon KJ, Patison N, Law AJ, Korsch MJ, Salter MA. Relationships of lead in breast milk to lead in blood, urine, and diet of the infant and mother. Environ Health Perspect 106:667-674 (1998).
2. Lawrence RA. A Review of the Medical Benefits and Contraindications to Breastfeeding in the United States. Maternal and Child Health Technical Information Bulletin. Arlington, VA:National Center for Education in Maternal and Child Health, Health Resources and Services Administration, 1997.
3. NIH Consensus Conference. Optimal calcium intake. NIH Consensus Development Panel on Optimal Calcium Intake. JAMA 272:1942-1948 (1992).
4. Gulson BL, Jameson CW, Mahaffey KR, Mizon KJ, Korsch MJ, Vimpani G. Pregnancy increases mobilization of lead from maternal skeleton. J Lab Clin Med 130:51-62 (1997).
5. Gulson BL, Mahaffey KR, Jameson CW, Mizon KJ, Korsch MJ, Cameron MA, Eisman JA. Mobilization of lead from the skeleton during the postnatal period is larger than during pregnancy. J Lab Clin Med 131:324-329 (1998).
Last Updated: January 22, 1999
