|
|
||||||||
Original Research Communications |
| ABSTRACT |
|---|
|
|
|---|
Design: In a double-blind trial conducted in Matlab, Bangladesh, women at 13 wk postpartum were randomly assigned to receive until 9 mo postpartum 1) a single dose of 60 mg retinol as retinyl palmitate followed by daily placebos (n = 69), 2) daily doses of 7.6 mg ß-carotene (n = 72), or 3) daily placebos (n = 71). Milk samples collected at baseline and 3 mo postpartum were analyzed by enzyme-linked immunosorbent assay for secretory immunoglobulin A, lactoferrin, lysozyme, and interleukin 8; by HPLC for total retinol; and by atomic absorption spectroscopy for sodium and potassium.
Results: After mammary epithelial permeability (defined as an elevated Na:K) and baseline immune factor concentrations were controlled for, there were no significant treatment effects on immune factors at 3 mo. Increased mammary permeability was common (25% of women at baseline and 12% at 3 mo) and was associated with higher concentrations of milk immune factors. Low body vitamin A stores at baseline, as assessed by the modified-relative-dose-response test, were associated with a higher Na:K, but neither retinol nor ß-carotene supplementation affected the prevalence of increased mammary permeability.
Conclusions: Postpartum vitamin A supplementation does not increase milk concentrations of immune factors. The causes of increased mammary epithelial permeability in this population require further study.
Key Words: Vitamin A supplementation ß-carotene retinol breast milk immunity inflammation sodium RETIBETA project Bangladesh women
| INTRODUCTION |
|---|
|
|
|---|
Possible strategies for improving the vitamin A status of young infants include supplementing infants themselves, either at birth (3) or the time of immunization (2), or supplementing lactating mothers (7). Vitamin A supplementation of mothers has several advantages over supplementation of young infants. It improves the vitamin A status of the mothers, many of whom are deficient (8), as well as of the infants, with no apparent adverse side effects (7) as long as high-dose supplements are not given beyond 8 wk postpartum, to avoid teratogenic effects on a possible subsequent pregnancy (9). Another potential benefit of maternal vitamin A supplementation during lactation, which has not been investigated, is a possible increase in breast milk immune factors. These components of breast milk provide passive protection to young infants and stimulate development of the infants' own immune systems (10, 11). Vitamin A is known to be important for immune function (12) and it is conceivable that increased intakes by deficient, lactating women would have beneficial effects on immune factors in their breast milk. There may also be immunologic benefits to the woman's breast health because vitamin A supplementation was shown to decrease the incidence and severity of mastitis in mice (13). Mastitis greatly alters immune factors in the milk of cows and women; inflammatory mediators increase in particular (14). Because it appears that immune factors in milk increase infants' specific immune protection and decrease the potential for nonspecific inflammation (11), the immunologic changes associated with mastitis may affect infant as well as maternal health.
The present study investigated the effect of maternal supplementation with retinol or ß-carotene on amounts of immune factors in breast milk. Supplementation with synthetic ß-carotene capsules was included to represent a type of treatment intermediate between retinol supplementation and dietary intake of ß-carotene from fruit and vegetables and to determine whether any of the effects seen were retinol specific. The immune factors chosen for analysis were secretory immunoglobulin A (sIgA), lactoferrin, lysozyme, and interleukin 8 (IL-8), which are present in milk in high quantities (except IL-8), are fairly stable during storage, and benefit the infant through a variety of mechanisms (10, 11).
| SUBJECTS AND METHODS |
|---|
|
|
|---|
Women continued to receive health and family planning services from their regular community health workers every 14 d. In addition, project-specific visits occurred at
2 wk and 3, 6, and 9 mo postpartum. For the project-specific visits, mother-infant pairs were randomly assigned to 2 field and 2 clinic visits each, equally distributing the number of pairs from the 3 treatment groups to field and clinic visits at each follow-up point. During field visits, mothers gave a casual sample of breast milk (<5 mL collected without respect to the time since the last breast-feeding episode) but no blood sample. During clinic visits, women provided blood samples and a full sample of breast milk (ie, the complete contents of one breast were expressed with a manual breast pump after a 2-h period in which the breast was not used for feeding). Both the women and their infants were weighed by using standard methods and examined (and treated if necessary) by the study physician.
Samples
Milk samples from the baseline and 3-mo follow-up visits were used. Only women for whom samples from both time points were available were included, ie, 69, 72, and 71 women from the retinol, b-carotene, and placebo groups, respectively. Samples were collected between July and November 1994, frozen in plastic containers at 2208C at Matlab, then transported to Dhaka, where they were stored at 2708C. The samples were thawed, split, and refrozen. Frozen aliquots were transported to London on dry ice, where they were stored at 2608C until analyzed. Although sample handling conditions for these initially unplanned analyses were unorthodox, all samples were treated in the same way. Analytes were selected partly for their stability, and results, which were generally high, suggested that loss during storage was not a major problem.
Laboratory analyses
Immune factors were analyzed by using sandwich enzyme-linked immunosorbent assay (ELISA) techniques. The antibody pairs, polyclonal unless otherwise indicated, and standards used were as follows: anti-IgA (no. A0262; Dako, High Wycombe, United Kingdom) and anti-secretory component horseradish peroxidase (HRP) conjugate (no. P0166; Dako) with human milk IgA standard (no. BP148; The Binding Site, Birmingham, United Kingdom); anti-lactoferrin (no. A0186; Dako) and anti-lactoferrin-HRP conjugate (no. AHP13P; Serotec, Oxford, United Kingdom) with human milk lactoferrin as a standard (no. L0520; Sigma, Poole, United Kingdom); anti-lysozyme (no. A0099; Dako) and anti-lysozyme-HRP conjugate (Dako) with human milk lysozyme as a standard (no. L6394; Sigma); and monoclonal anti-IL-8 (no. MAB208; R+D Systems, Abington, United Kingdom) and biotinylated polyclonal anti-IL-8 (no. BAF208; R+D Systems) and avidin-HRP (no. P0347; Dako) with human IL-8 reference preparation no. 89/520 (National Institute for Biological Standards and Control, Potters Bar, United Kingdom). Bovine serum albumin as a blocking agent and bovine serum albumin in phosphate-buffered saline with 0.05% Tween 20 as a diluting agent were used for all analytes except IL-8. For the IL-8 assays, commercial blocking and diluting agents (CLB; Eurogenetics, Middlesex, United Kingdom) were used to decrease background absorbance. A quality-control milk sample from a British woman was included in duplicate on every plate, which permitted calculations of interplate CVs of 22% for sIgA (n = 20), 31% for lactoferrin (n = 20), 18% for lyso
zyme (n = 26), and 8.6% for IL-8 (n = 11). Although large, the CVs were similar to those found by others with ELISAs of milk proteins (17). The mean (±SD) recovery of IL-8 added in different amounts to the same milk sample was 141 ± 46% (n = 12). Recoveries >100% likely resulted from assay variability, which was in line with the variability of our other milk protein ELISAs but higher than the variability with the IL-8 assays in which the actual samples were run.
Serum retinol and didehydroretinol for the modified-relative-dose-response (MRDR) test and milk total retinol were measured by HPLC with retinyl acetate as an internal standard for serum and ß-apo-8'-carotenal-methyloxime for milk (15). Milk retinol concentrations were expressed per gram fat to control for the variation introduced by using 2 different sampling procedures (18). For similar reasons, aqueous milk components were expressed either per volume or per mg total protein, measured with a microtiter version of a commercial bicinchoninic acid method (Pierce Chemical Co, Rockford, IL). However, because results were not affected, only concentrations per volume are presented.
Sodium and potassium were analyzed by flame atomic absorption spectroscopy (Corning 480 flame photometer; Instrumentation Laboratory, Warrington, United Kingdom). Calibration was done with a Corning Multical and interassay CVs for a milk sample containing 55 mmol Na and 22 mmol K were 2.5% for sodium and 1.6% for potassium. Na-K ratios were used to determine the degree of mammary epithelial permeability. On the basis of the observed distribution of Na-K ratios, the following permeability categories were defined: Na:K
0.6, Na:K >0.6 to
1.0, and Na:K >1.0. Na:K rather than the sodium concentration alone was used to account for 1) the variation due to the different proportions of aqueous and fat fractions as a result of different sampling methods, and 2) the modest parallel decreases in both electrolytes that occur as a result of months of lactation (19). The highest ratio category corresponds to a sodium concentration of
18 mmol/L, which, after about the first 3 d of lactation, is considered to indicate the opening of mammary epithelial tight junctions in association with mastitis or weaning (20). Infants were not weaned before the last samples were taken at 3 mo postpartum; therefore, we consider high Na-K ratios to indicate inflammation.
Statistical analyses
The sample size selected for the main study (n = 220) was based on numbers required to detect a difference between either the retinol or the ß-carotene group and the placebo group in the proportion of infants classified as being vitamin Adeficient at 6 mo of age. This number was sufficient, at 95% significance and 80% power, to show 2030% differences between group means for sIgA, lactoferrin, and lysozyme on the basis of their concentrations in casual milk samples from Zairean women (21). Published data were not available on which to base IL-8 calculations. The software package SPSS for MS WINDOWS (version 6.1; SPSS Inc, Chicago) was used for the analyses. Data were log-transformed to normalize distributions, and geometric means and 95% CIs are given. In the analysis of variance of treatment group differences at 3 mo, the following covariates were considered: concentrations of the particular immune factor at baseline, Na:K, type of sample (casual or full), the number of days postpartum at which the sample was collected, maternal age, body mass index, parity, initial infant weight, and infant weight gain between baseline and 3 mo.
| RESULTS |
|---|
|
|
|---|
|
Because of the unexpectedly large effect of increased mammary permeability, in terms of the number of affected women (baseline: 52/212, or 25%; 3 mo: 25/212, or 12%) and the concentrations of most of the immune factors measured, we calculated geometric mean values on the basis of Na:K category rather than on the basis of treatment group and included the number of days postpartum that the sample was collected as a covariate (Table 2
). IL-8 and sIgA increased at both time points with increasing Na-K ratios; lactoferrin and lysozyme showed similar trends that were not always significant and total retinol increased with increasing Na-K ratios at baseline.
|
| DISCUSSION |
|---|
|
|
|---|
30% between groups for lactoferrin and lysozyme, but only differences of
60% for sIgA and IL-8. Others have shown only modest effects of food supplements on milk immune factors in women (22, 23). To our knowledge, there are no other data on micronutrient status and milk immune factors in humans; however, vitamin E and selenium supplementation of cows was shown to increase the number of and to improve the function of milk neutrophils, with consequent benefits for mammary gland health (24). We found a slight effect of initial vitamin A stores, as estimated from the MRDR results, on mammary epithelial permeability, as indicated by Na-K ratios. Because an elevated Na:K may indicate mammary inflammation, this result is similar to results from rodent studies in which the incidence and severity of experimental mastitis decreased with increasing dietary retinol (13). However, neither retinol nor ß-carotene supplementation affected the prevalence of elevated Na-K ratios. The major contributors to milk immune factor concentrations were mammary permeability and individual variation, as evidenced by the large effect of baseline concentrations on 3-mo concentrations of each factor. An increased Na:K was associated with an increase in all immune factors, as was shown by others (25), as well as in total retinol concentrations at baseline. Concentrations of the inflammatory cytokine IL-8 were most strongly affected, suggesting that an elevated Na:K represents true inflammation, not just a "leakiness" of the mammary epithelium due to other factors. In addition, this observation and the association of an elevated Na:K with reduced infant growth support in vivo production of IL-8, possibly by mammary epithelial cells (26), rather than ex vivo stimulation of cells during sample collection or processing.
We found a high proportion of women, especially at baseline, with elevated milk sodium concentrations characteristic of mastitis (20, 25) or breast engorgement (27). Na:K decreased as the days postpartum at which the baseline sample was collected increased, suggesting that this was partly a normal physiologic change as milk production adjusted to meet infant demand and as breast engorgement was relieved. Nevertheless, baseline concentrations of immune factors were still significantly correlated with Na:K even after the number of days postpartum at which the sample was collected was accounted for, indicating a possible residual effect of inflammation. Breast pain was not recorded during the study, but there was little obvious mastitis. We are unaware of comparable published studies of Na-K ratios in milk from a large unselected population of mothers from a developed country, but recently found similarly high Na-K ratios in milk samples from Tanzanian (28) and South African women (SM Filteau, JF Willumsen, unpublished observations, 1998). Vitamin A deficiency may have contributed slightly to the high prevalence because women with abnormal MRDR results had higher milk Na-K ratios.
Lysozyme and sIgA concentrations in the present study were comparable with those commonly found in the mature milk of Western women, whereas lactoferrin concentrations were somewhat higher (10, 11). Although the normal to high concentrations of immune factors in this population may have contributed to the lack of effect of vitamin A supplementation, there was similarly no effect in the quartile with the lowest initial immune factor concentrations. Several studies have shown a tendency toward higher milk concentrations of sIgA and lactoferrin in African than in European women (21, 22, 29), suggesting that concentrations are higher in areas with a high prevalence of ambient infection. Our results suggest that this may not only be a result of increased maternal synthesis due to general stimulation of immune responses in the mucosal immune system, but also a result of increased mammary permeability causing increased leakage of many compounds across the epithelium into the milk.
Several factors that could not be assessed within the context of this study may affect immune compounds and sodium in milk, including the amount of breast milk consumed by infants and gestational age at birth. Weaning or poor suckling by infants might result in breast engorgement, with resultant increases in milk sodium, but, in this area of rural Bangladesh, exclusive or predominant breast-feeding is virtually universal until
3 mo of age and infant formula is not used. Because mother-infant pairs were recruited at 2 wk postpartum, accurate measures of gestational age were unavailable. Because the treatment allocation was random, these confounders (ie, the amount of breast milk consumed by infants and gestational age at birth) would have been similar across groups and, at most, would have increased our variability. Because we found no important trends, the increased variability that resulted from the confounders likely had little effect on our conclusions.
Although samples were collected by using 2 different methods, a casual manual sample or a full breast sample obtained with a pump, this had no significant effect on immune factor concentrations. Similarly, the results of analyses of immune factors did not differ significantly when factors were expressed per mg total protein or per volume. These observations likely reflect the fact that the aqueous part of milk, presumably the major fraction containing these protein immune factors, is the largest component of the milk and, as has been shown by others (30), has a fairly constant composition throughout a feeding. Thus, it appears appropriate to use whatever method is easiest to collect and analyze milk for immune factor measurements.
In conclusion, neither maternal retinol nor ß-carotene supplementation of lactating Bangladeshi women significantly affected milk immune factors at 3 mo postpartum. However, even in the unsupplemented group, the concentrations of most factors were within the normal range. Milk samples with high immune factor concentrations also had high sodium concentrations, which are associated with breast inflammation. Because a relatively large proportion of women were identified with high milk sodium and IL-8 concentrationsand inflammatory stimuli in the gut of young infants can be harmful (11) and breast inflammation can be painfulthe underlying causes of this condition and its potential association with infant growth in other populations deserve further investigation. Most problems with lactation can be remedied with breast-feeding counseling (31) and it is possible that low-cost education programs on good breast-feeding practices may have important health benefits for these women and their infants.
| ACKNOWLEDGMENTS |
|---|
| FOOTNOTES |
|---|
2 Supported by the Child Health Research Appeals Trust, Institute of Child Health, United Kingdom, and by a grant from the Parkes Foundation, United Kingdom (to JJB). The RETIBETA project was supported by the US Agency for International Development, Office of Nutrition.
3 Reprints not available. Address correspondence to SM Filteau, Centre for International Child Health, Institute of Child Health, 30 Guilford Street, London WC1N 1EH, United Kingdom. E-mail: sfilteau{at}ich.ucl.ac.uk.
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
M. E. Flores-Quijano, A. Cordova, V. Contreras-Ramirez, L. Farias-Hernandez, M. Cruz Tolentino, and E. Casanueva Risk for Postpartum Depression, Breastfeeding Practices, and Mammary Gland Permeability J Hum Lact, February 1, 2008; 24(1): 50 - 57. [Abstract] [PDF] |
||||
![]() |
K. M Edmond, B. R Kirkwood, S. Amenga-Etego, S. Owusu-Agyei, and L. S Hurt Effect of early infant feeding practices on infection-specific neonatal mortality: an investigation of the causal links with observational data from rural Ghana Am. J. Clinical Nutrition, October 1, 2007; 86(4): 1126 - 1131. [Abstract] [Full Text] [PDF] |
||||
![]() |
L. J Hindle, R. Gitau, S. M Filteau, K. J Newens, D. Osrin, A. M Costello, A. M Tomkins, A. Vaidya, R. K. Mahato, B. Yadav, et al. Effect of multiple micronutrient supplementation during pregnancy on inflammatory markers in Nepalese women. Am. J. Clinical Nutrition, November 1, 2006; 84(5): 1086 - 1092. [Abstract] [Full Text] [PDF] |
||||
![]() |
V. Nussenblatt, N. Kumwenda, V. Lema, T. Quinn, M. C. Neville, R. Broadhead, T. E. Taha, and R. D. Semba Effect of Antibiotic Treatment of Subclinical Mastitis on Human Immunodeficiency Virus Type 1 RNA in Human Milk J Trop Pediatr, October 1, 2006; 52(5): 311 - 315. [Abstract] [Full Text] [PDF] |
||||
![]() |
E. Villamor and W. W. Fawzi Effects of Vitamin A Supplementation on Immune Responses and Correlation with Clinical Outcomes Clin. Microbiol. Rev., July 1, 2005; 18(3): 446 - 464. [Abstract] [Full Text] [PDF] |
||||
![]() |
C. A. Lovelady, C. P. Hunter, and C. Geigerman Effect of Exercise on Immunologic Factors in Breast Milk Pediatrics, February 1, 2003; 111(2): e148 - 152. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. L Dreyfuss and W. W Fawzi Micronutrients and vertical transmission of HIV-1 Am. J. Clinical Nutrition, June 1, 2002; 75(6): 959 - 970. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. D. Semba, N. Kumwenda, T. E. Taha, D. R. Hoover, T. C. Quinn, Y. Lan, L. Mtimavalye, R. Broadhead, P. G. Miotti, L. van der Hoeven, et al. Mastitis and Immunological Factors in Breast Milk of Human Immunodeficiency Virus-infected Women J Hum Lact, December 1, 1999; 15(4): 301 - 306. [Abstract] [PDF] |
||||
![]() |
R. D. Semba, N. Kumwenda, T. E. Taha, D. R. Hoover, Y. Lan, W. Eisinger, L. Mtimavalye, R. Broadhead, P. G. Miotti, L. Van Der Hoeven, et al. Mastitis and Immunological Factors in Breast Milk of Lactating Women in Malawi Clin. Vaccine Immunol., September 1, 1999; 6(5): 671 - 674. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |