OUP user menu

Detection of specific antibodies in cord blood, infant and maternal saliva and breast milk to staphylococcal toxins implicated in sudden infant death syndrome (SIDS)

Linda M. Harrison, James A. Morris, Lisa A. Bishop, Robert M. Lauder, Christine A.M. Taylor, David R. Telford
DOI: http://dx.doi.org/10.1016/j.femsim.2004.06.010 94-104 First published online: 1 September 2004

Abstract

The common bacterial toxins hypothesis of sudden infant death syndrome (SIDS) is that nasopharyngeal bacterial toxins can trigger events leading to death in infants with absent/low levels of antibody that can neutralise the toxins. The aim of this study was to investigate nasopharyngeal carriage of Staphylococcus aureus and determine levels of immunity in the first year of life to toxic shock syndrome toxin (TSST-1) and staphylococcal enterotoxin C (SEC). Both toxins have been implicated in SIDS cases. Seventy-three mothers and their infants (39 males and 34 females) were enrolled onto the study. The infants had birth dates spread evenly throughout the year. In infants, S. aureus carriage decreased significantly with age (P<0.001). Between 40% and 50% of infants were colonised with S. aureus in the first three months of life and 49% of the isolates produced one or both of the staphylococcal toxins. There was a significant correlation between nasopharyngeal carriage of S. aureus in mothers and infants in the three months following the birth (P<0.001). Carriage of S. aureus in infants and their mothers was not significantly associated with levels of antibody to TSST-1 or SEC in cord blood, adult saliva or breast milk. Infants colonised by S. aureus had higher levels of salivary IgA to TSST-1 than infants who were culture negative. Analysis of cord blood samples by a quantitative ELISA detected IgG bound to TSST-1 and SEC in 95.5% and 91.8% of cases respectively. There was a marked variation in levels of maternal IgG to both TSST-1 and SEC among cord blood samples. Maternal age, birth weight, and seasonality significantly affected the levels of IgG binding to TSST-1 or SEC. Analysis of infant saliva samples detected IgA to TSST-1 and SEC in the first month after birth; 11% of samples tested positive for salivary IgA to TSST-1 and 5% for salivary IgA to SEC. By the age of two months these proportions had increased to 36% and 33% respectively. More infants who used a dummy tested positive for salivary IgA to TSST-1 compared to infants who did not use a dummy. Levels of IgA to TSST-1 and SEC detected in the breast-milk samples varied greatly among mothers. There was a trend for infants receiving breast milk with low levels of antibody to TSST-1 or SEC to have higher levels of salivary antibody to the toxins. In conclusion, passive immunity to toxins implicated in SIDS cases varies greatly among infants. Infants are able to mount an active mucosal immune response to TSST-1 and SEC in the first month of life.

Keywords
  • Sudden infant death syndrome
  • Bacterial toxins
  • Immunoglobulin A
  • Immunoglobulin G
  • Cord blood
  • Breast milk
  • Saliva

1 Introduction

Sudden infant death syndrome (SIDS) has a consistent and characteristic age distribution curve. A mathematical model based on an infectious cause for SIDS describes the curve if the causative organisms are very common, i.e., met by 50% of the infant population in any 50 day period [1]. Maternal IgG transferred across the placenta provides immunity for infants in the first weeks of independent life. The model predicts that as levels of maternal IgG fall, the risk of infection increases. Risk of infection then decreases because infants meet and acquire immunity to common organisms. The incidence of SIDS peaks between two and four months of age when circulating antibody levels are at their nadir.

The model is derived from the common bacterial toxins hypothesis which proposes that some cases of SIDS are due to the synergistic effects of absorbed nasopharyngeal bacterial toxins [1,2]. As it is assumed that the toxigenic bacteria of the normal microbial flora are common, they will be encountered by infants in the first few weeks of life and come to form part of their normal upper respiratory tract flora. This could result in infants being at risk of toxin-induced sudden death in the interval between loss of maternal immunoglobulin and acquisition of specific immunity. Previous studies suggest that Staphylococcus aureus is a prime candidate organism for a pathogenic role in SIDS. It has been identified as the most common nasopharyngeal coloniser in the first three months of life, with an isolation rate that fits the mathematical model [3,4]. SIDS cases are significantly more likely to be colonised with S. aureus than healthy age-matched infants [3,5]. A range of toxins, including superantigens, are produced by S. aureus. Staphylococcal enterotoxin (C) (SEC) and toxic shock syndrome toxin 1 (TSST-1) have been detected in the tissues of significantly more SIDS cases than controls [6,7].

The aim of this longitudinal study was to investigate passive and active immunity to the S. aureus toxins, TSST-1 and SEC in the first year of life. Levels of IgG to TSST-1 and SEC in cord blood samples and IgA to TSST-1 and SEC in breast milk and maternal and infant saliva samples were determined and related to risk factors associated with SIDS, e.g., maternal age, birth weight, gestational age and seasonality. Antibody levels in saliva were compared with nasopharyngeal carriage and toxin production by S. aureus in mothers and infants.

2 Subjects and methods

2.1 Population studied

Parents were recruited to the study at ante-natal classes. The aim was to follow the families of at least four infants (two boys and two girls) born in each calendar month for 12 successive months. The families of 73 infants (39 males and 34 females) were enrolled in the study. The infants had birth dates evenly spread throughout the year. One family dropped out of the study after eight months because they moved away from the area. The study was approved by Morecambe Bay Ethics Committee and informed consent was obtained from mothers.

2.2 Sample collection and preparation

Details of mothers recruited to the study were sent to the Maternity Unit and cord blood samples were collected by the midwives. Each sample was centrifuged at 1500g and the separated serum was stored at −20 °C until examined.

Mothers and infants were visited at home each month for 12 months following the birth. At each visit, nasopharyngeal swabs were obtained from mothers and infants. Mothers supplied samples of unstimulated saliva and breast milk (if lactating). Samples of unstimulated saliva were collected from infants using a cotton tipped swab in the first month of life and a sterile plastic pipette in subsequent months. Nasopharyngeal swabs were processed on return to the laboratory. Samples of saliva were centrifuged at 1500g and the supernatant stored at −20 °C. Samples of breast milk were stored at −20 °C.

Throughout the study parents recorded information on their smoking habits, infant sleeping position, feeding methods and dummy use.

2.3 Isolation and identification of S. aureus

Nasopharyngeal swabs were cultured on blood agar and S. aureus isolates were identified using standard methods. The isolates were stored on beads at −70 °C for further tests.

2.4 Detection of toxin production by S. aureus isolates

Toxins were prepared by growing the bacteria over a semi-permeable membrane [8]. The growth was washed from the membrane, centrifuged and the supernatant was passed through a 0.2 µm filter. Replicate samples (2 µl) were loaded onto two nitrocellulose membranes. Controls included purified toxic shock syndrome toxin (TSST-1) (Toxin Technology, Sarasota, FL, USA), purified staphylococcal enterotoxin C (SEC) (Toxin Technology) and supernatant prepared from S. aureus isolates previously tested for the production of TSST-1 and SEC (Colindale Public Health Laboratory). The membranes were blocked for 30 min in blocking buffer composed of Marvel skimmed milk powder 2% (w/v) in phosphate buffered saline (PBS). Each membrane was washed three times in washing buffer composed of PBS containing 0.05% (v/v) Tween 20, then transferred for 1 h to either horseradish peroxidase (HRP)-labelled sheep anti-TSST-1 or HRP-labelled sheep anti-SEC (Toxin Technology) diluted 1 in 300 in blocking buffer. After three washes, the membranes were transferred to 3,3′-Diaminobenzidine tetrahydrochloride (DAB) (Sigma). The substrate contained 10 mg DAB in 15 ml of Tris buffered saline (pH 7.6) and was activated immediately before use by adding 5 µl of 30% (v/v) H2O2. The colour change was stopped by rinsing the membranes in water.

2.5 Detection of IgG to staphylococcal pyrogenic toxins in cord blood

Serum samples were examined for IgG antibody to TSST-1 and SEC by an enzyme linked immunosorbent assay (ELISA) on 96 well polystyrene microtitre plates. A modification of the protocol of Al Madani et al. [9] was used. Test and control wells were coated with 100 µl of either TSST-1 or SEC at 1 µgml−1 in carbonate buffer (Sigma). Standard wells were coated with 100 µl of serial dilutions of human IgG (Sigma); concentrations ranged from 5 to 10,000 ngml−1. Plates were incubated overnight at 4 °C and washed three times with washing buffer. The plates were blocked with blocking buffer, containing 1% (w/v) bovine serum albumin (BSA) (Sigma) in PBS for 1 h at room temperature (RT). After three washes, the serum samples were diluted 1 in 50 in blocking buffer and 100 µl loaded into each test well. Blocking buffer (100 µl) was added to standard and control wells. Plates were then incubated for 1 h at 37 °C. After three washes, 100 µl of anti-human IgG, conjugated to horseradish peroxidase (HRP) (Sigma) diluted 1 in 1000 in blocking buffer were added to the test and standard wells. HRP-labelled sheep anti-TSST-1 (100 µl) or HRP-labelled sheep anti-SEC (100 µl) diluted 1 in 300 in blocking buffer was added to the appropriate control wells. The plates were incubated for 1 h at 37 °C. After three washes, 100 µl of the substrate 3,3′,5,5′-tetramethylbenzidine (TMB) (Sigma) was added to each well. The substrate contained 1 mg TMB in 10 ml of 0.05 M phosphate-citrate buffer (pH 5.0) and was activated immediately before use by adding 2 µl of 30% (v/v) H2O2. The reaction was stopped after 15 min by the addition of 50 µl 2 M H2SO4 and the absorbance was read at 450 nm with a plate reader (Organon Teknika). Samples were tested in duplicate and the results were expressed in ngml−1 derived from the linear section of the IgG standard curve in each experiment.

2.6 Detection of IgA to staphylococcal pyrogenic toxins in saliva

Test and control wells of microtitre plates were coated with 100 µl of either TSST-1 or SEC at 1 µgml−1 in carbonate buffer (Sigma). Standard wells were coated with 100 µl of serial dilutions of a standard human IgA derived from colostrum (Sigma); concentrations ranged from 5 to 10,000 ngml−1. Plates were incubated overnight at 4 °C then washed three times with washing buffer. The plates were blocked with blocking buffer for 30 min at room temperature. After three washes, 100 µl of undiluted saliva or the following dilutions made in blocking buffer were loaded into the appropriate test wells: 1 in 3; 1 in 5; or 1 in 10. Blocking buffer (100 µl) was added to standard and control wells. Plates were incubated for 2 h at RT. After three washes, 100 µl of HRP-labelled goat anti-human IgA diluted 1 in 500 in blocking buffer was added to the test and standard wells and 100 µl of HRP-labelled sheep anti-TSST-1 or HRP-labelled sheep anti-SEC, diluted 1 in 300 in blocking buffer was added to the appropriate control wells. The plates were incubated at RT for 2 h. The experiment then continued as described in Section 2.5. Samples were tested in duplicate and the results were expressed in ngml−1 derived from the linear section of the IgA standard curve in each experiment.

2.7 Detection of IgA to staphylococcal pyrogenic toxins in breast milk

A modification of the protocol of Gordon et al. [10] was used. Test and control wells of microtitre plates were coated with 100 µl of either TSST-1 or SEC at 1 µgml−1 in carbonate buffer (Sigma). Standard wells were coated with 100 µl of serial dilutions of a standard human IgA derived from colostrum; concentrations ranged from 5 to 10,000 ngml−1. Plates were incubated overnight at 4 °C then washed three times with washing buffer. The plates were blocked with blocking buffer for 30 min at room temperature. After three washes, 100 µl of undiluted breast milk, or dilutions of 1 in 5, 1 in 10 or 1 in 50 made in blocking buffer were loaded into the appropriate test wells. Blocking buffer (100 µl) was added to standard and control wells. Plates were then incubated for 2 h at RT. The experiment then continued as described in Section 2.6.

2.8 Statistical analyses

Results were analysed using SPSS software (version 11.0): McNemar's test was used to compare bacterial isolation from mother and infant at each visit. Logistic regression was used to test for seasonal and time trends. Pearson's correlation coefficient, Spearman's rho correlation coefficient, t test, Mann–Whitney U test, ANOVA and Kruskal–Wallis test were used to assess anti-toxin levels in relation to other variables. The non-parametric tests were used for data not normally distributed. These data were not normalised by log transformation because many values of nil were detected.

3 Results

3.1 Population characteristics and infant care practices

Children born to 73 mothers enrolled into the study, 39 males and 34 females. The mean gestation period was 279 days (SD±9.4, range 253–293). The average maternal age at birth was 29 years (SD±4.9, range 18–42). Birth weights ranged from 2268 to 4848 g with a mean weight of 3402 g (SD±589). There was no significant difference in the birth weight of males and females.

Forty-nine infants were being breast fed on discharge from hospital. By the age of 12 months, the number of infants still receiving breast milk was 11. Twenty-one infants (11 males and 10 females) regularly used dummies throughout the first year of life. Less than 9% of the infants were cared for by smokers, each of whom stated that they did not smoke in the same room as the infant.

All parents were advised at ante-natal classes to place their babies to sleep in the supine position. After three months of age, a small number of infants (4%) began to turn over in their cots to sleep in the prone position. By 11 months of age, 78% of the infants regularly slept prone; 8% slept supine; 11% slept on their side; and 4% showed a marked variation from night to night (Fig. 1).

Figure 1

Regular sleeping position in the first 12 months of life.

3.2 Nasopharyngeal carriage of S. aureus in infants and their mothers

The nasopharyngeal carriage rates of S. aureus in infants and mothers are shown in Fig. 2. Carriage of S. aureus changed significantly in infants. It was highest in the first three months of life then fell progressively and levelled out by 12 months of age (Logistic regression, P<0.001). Maternal carriage remained fairly constant throughout the 12 months with the highest carriage rate found in the first month of sampling.

Figure 2

Nasopharyngeal carriage of S. aureus in infants and their mothers.

There was a significant correlation between nasopharyngeal carriage of S. aureus in mothers and their infants in the first three months (McNemar test, P<0.001, P<0.001, P<0.05) but not in the subsequent months.

There was no significant association between isolation of S. aureus and seasonality.

3.3 Detection of toxin production by S. aureus isolates

S. aureus isolates (n=403) were tested for production of TSST-1 and SEC with a dot blot assay. Of the 217 isolates collected from infants, 79 (36%) produced TSST-1 and 64 (29%) produced SEC. There were 186 isolates from mothers; 72 (39%) produced TSST-1 and 49 (26%) produced SEC. Some S. aureus strains produced more than one toxin.

3.4 Detection of IgG to staphylococcal pyrogenic toxins in cord blood

Sixty-six samples of cord blood were collected and tested for the presence of IgG to TSST-1 and SEC. The mean level of antibody to TSST-1 was 5380 ngml−1 (range 0 to 15,000 ngml−1, SD±3910). Antibody levels to SEC ranged from 0 to 14,420 ngml−1 (mean 5690 ngml−1, SD±3740). There was no significant correlation between anti-TSST-1 and anti-SEC levels (r=−0.051).

Anti-TSST-1 IgG correlated negatively with maternal age at time of birth but there was a positive correlation between anti-SEC IgG and maternal age (Table 1) which was most marked for male infants (r=0.471, P=0.003).

View this table:
Table 1

Association between anti-toxic shock syndrome toxin (ATSST) and anti-staphylococcal enterotoxin (C) (ASEC) levels in cord blood (IgG) and maternal age; birth weight; gestational age; mean ATSST and ASEC levels (IgA) in breast milk samples collected in the first and second months after birth; mean ATSST and ASEC levels (IgA) in adult saliva; mean ATSST and ASEC levels (IgA) in infant saliva

AntibodyrP-value
Pearson correlations
Maternal ageATSST−0.0790.526
ASEC0.3170.009
Birth weightATSST0.1370.273
ASEC0.3050.013
Gestational ageATSST0.1140.361
ASEC0.1320.292
Breast milk (first month)ATSST0.1360.385
ASEC−0.0310.844
Breast milk (second month)ATSST0.0530.764
ASEC−0.2650.129
Spearman's rho correlations
Adult salivaATSST0.2190.077
ASEC0.3060.012
Infant salivaATSST−0.1290.303
ASEC0.1560.212

There was a positive correlation between anti-toxin levels and birth weight which was statistically significant for antibody to SEC (Table 1). This was most marked for male infants (r=0.421, P=0.01). Correlations were also positive, but not statistically significant, with gestational age for both anti-TSST-1 IgG and anti-SEC IgG (Table 1).

Table 1 summarises the associations between anti-toxin levels (IgG) in cord blood and mean anti-toxin levels (IgA) in saliva and breast milk. There were weak positive correlations between cord blood anti-TSST-1 IgG and both adult salivary anti-TSST-1 IgA and breast milk anti-TSST-1 IgA. In contrast, a weak negative correlation was found with infant salivary anti-TSST-1 IgA.

A statistically significant positive correlation was found between cord blood anti-SEC IgG and adult salivary anti-SEC IgA. Cord blood anti-SEC IgG showed a weak positive correlation with both breast milk anti-SEC IgA and infant salivary anti-SEC IgA, but the associations were not significant.

There was no significant difference in cord blood anti-TSST-1 and anti-SEC levels between male and female infants. In males, the mean level of anti-TSST-1 IgG was 5704 ngml−1 (SD±4304) and in females it was 4970 ngml−1 (SD±3362). For anti-SEC IgG the mean level in males was 5493 ngml−1 (SD±3805) and for females the mean was 5938 ngml−1 (SD±3715).

Seasonality had an influence on the levels of IgG to TSST-1 and SEC detected in the cord blood samples (Figs. 3 and 4). Anti-TSST-1 IgG varied by month of birth (F=3.341, P=0.001) as did anti-SEC IgG (F=2.048, P<0.05). Anti-TSST-1 IgG, in particular, showed the lowest levels in the winter months.

Figure 3

Seasonality of anti-TSST-1 levels in cord blood.

Figure 4

Seasonality of anti-SEC levels in cord blood.

3.5 Detection of IgA to staphylococcal pyrogenic toxins in saliva

3.5.1 Infant saliva

In the first month of life, infants produced remarkably little saliva and no quantification of anti-toxin levels could be made. Mean levels of salivary anti-TSST-1 IgA for individual infants ranged from 0 to 1318 ngml−1 (mean 108 ngml−1, SD±218). Individual mean levels of salivary anti-SEC IgA ranged from 0 to 340 ngml−1 (mean 39 ngml−1, SD±48).

Infants colonised with S. aureus had consistently higher levels of salivary IgA to TSST-1 throughout the first 12 months of life compared to infants who were culture-negative. This association was significant for saliva samples collected in the first three months using the pipette method (Mann–Whitney U test: one month, P<0.05; two months, P<0.001; three months, P<0.01). By comparison, there was no significant difference between salivary anti-SEC IgA in infants with or without nasopharyngeal carriage of S. aureus.

Higher levels of salivary anti-TSST-1 IgA were detected in infants colonised with S. aureus producing TSST-1 compared to infants from whom the bacteria were not isolated. The difference was significant in months one, two, three, four, five and eleven by the Mann–Whitney U test: month 1, P<0.01; month 2, P<0.01; month 3, P<0.05; month 4, P<0.01; month 5, P<0.01; month 11, P<0.05). Infants carrying S. aureus isolates that produced SEC also had higher levels of specific antibody in the saliva than infants with isolates that did not produce SEC. This difference was significant by the Mann–Whitney U test: month 1, P<0.01; month 2, P<0.05; month three, P<0.05; month nine, P<0.05; month eleven, P<0.05.

Table 2 summarises the associations between mean anti-toxin levels (IgA) in infant saliva and maternal age, birth weight and gestation.

View this table:
Table 2

Association between ATSST and ASEC levels in infant saliva (IgA mean) and maternal age; birth weight; gestation length

Spearman's rho correlations
AntibodyrP-value
Maternal ageATSST0.0520.663
ASEC0.0600.613
Birth weightATSST−0.1230.301
ASEC0.1290.278
Gestational ageATSST−0.1730.143
ASEC0.2160.066

There is a trend for infants born in the autumn and winter months to produce higher levels of salivary anti-TSST-1 IgA compared to those born in the spring and summer months. Salivary anti-SEC IgA did not show this trend.

More infants who used a dummy had detectable salivary IgA to TSST-1 compared to infants who did not use a dummy (significant in the first month of life P<0.05. This trend was also seen for salivary IgA to SEC.

3.5.2 Maternal saliva

All mothers had salivary IgA to TSST-1 and SEC, but antibodies were not detected in samples for individual mothers at all monthly visits. Mean levels of salivary anti-TSST-1 IgA for individual mothers ranged from 53 to 844 ngml−1 (mean 240 ngml−1, SD±146). Individual mean levels of salivary anti-SEC IgA ranged from 78 to 1759 ngml−1 (mean 346 ngml−1, SD±271).

There was no significant relationship between salivary anti-toxins and nasopharyngeal carriage of S. aureus. Adults colonised with S. aureus that produced TSST-1 generally had higher levels of salivary IgA to TSST-1 than adults colonised with an isolate that did not produce TSST-1. The difference was statistically significant only in month seven (Mann–Whitney U test P<0.05.

Table 3 summarises the associations between anti-TSST-1 and anti-SEC levels in adult saliva (IgA mean) and maternal age, birth weight, gestational age, mean anti-TSST-1, and anti-SEC levels (IgA) in infant saliva.

View this table:
Table 3

Association between ATSST and ASEC levels in adult saliva (IgA mean) and maternal age; birth weight; gestational age; mean ATSST and ASEC levels (IgA) in infant saliva

Spearman's rho correlations
AntibodyrP-value
Maternal ageATSST−0.1060.374
ASEC0.0400.737
Birth weightATSST0.0170.886
ASEC−0.0320.786
Gestational ageATSST0.1950.099
ASEC0.2440.038
Infant salivaATSST−0.1270.284
ASEC−0.0010.992

3.6 Detection of IgA to staphylococcal pyrogenic toxins in breast milk

All samples of breast milk collected from lactating mothers were positive for IgA to TSST-1 and SEC. Individual mean levels of breast milk anti-TSST-1 IgA ranged from 1126 to 5260 ngml−1 (mean 2109 ngml−1, SD±724). Individual mean levels of breast milk anti-SEC IgA ranged from 1213 to 3981 ngml−1 (mean 1945 ngml−1, SD±532).

Table 4 summarises the associations between the mean anti-toxin levels in breast milk (IgA) and maternal age, birth weight, gestational age, mean anti-toxin levels in adult saliva (IgA) and mean anti-toxin levels in infant saliva (IgA). There was a positive correlation with maternal age for anti-TSST-1 and anti-SEC in breast milk, but these associations were not statistically significant. Mean levels of both salivary anti-toxins (IgA) in infants tended to increase as mean levels of breast milk anti-toxins (IgA) decreased, but these associations were not statistically significant.

View this table:
Table 4

Association between ATSST and ASEC levels in breast milk (IgA mean) and maternal age; birth weight; gestational age; mean ATSST and ASEC levels (IgA) in adult saliva; mean ATSST and ASEC levels (IgA) in infant saliva

AntibodyrP-value
Pearson correlations
Maternal ageATSST0.1420.331
ASEC0.2160.136
Birth weightATSST−0.0670.647
ASEC0.0340.818
Gestational ageATSST0.0830.569
ASEC0.2440.091
Spearman's rho correlations
Adult salivaATSST0.0290.845
ASEC0.0260.858
Infant salivaATSST−0.2360.102
ASEC−0.1730.234

There was no significant seasonal variation in levels of IgA to TSST-1 and SEC in breast milk, but in the first three months after birth levels tended to be higher in the summer months and lower in the winter months.

4 Discussion

It has been suggested that toxins produced by common nasopharyngeal bacteria can act in synergy to cause sudden death in infants with low levels of circulating immunoglobulins [1,2]. There is considerable support for this hypothesis from theoretical, experimental and epidemiological studies [314] which show that the hypothesis can explain the key epidemiological features of SIDS, including the age distribution, the winter excess of cases and the association with prone sleeping and exposure to tobacco smoke.

S. aureus has been identified as the most common coloniser of the nasopharynx in the early months of life when SIDS cases reach a peak [3,4]. This organism has emerged as a prime candidate for a pathogenic role in SIDS. It fits the epidemiological pattern shown by the syndrome and pyrogenic toxins produced by S. aureus have been identified in the tissues of SIDS infants [6,7].

An assumption of the common bacterial toxins hypothesis is that healthy infants will mount an effective immunological response when first exposed to specific toxin antigens, or they succumb to their effects if they lack sufficient passive neutralising antibody. In this study, infants were at risk of early exposure to the pyrogenic toxins TSST-1 and SEC. Between 40% and 50% of infants were colonised with S. aureus in the first three months of life and 49% of the isolates produced one or both of the toxins.

The close association between isolation of S. aureus from the nasopharynx of mothers and their infants suggest that the mother is the primary source of infant colonisation. Previous studies have described this finding [3,15]. Peacock et al. [15] reported a concordance between the nasal strains carried by mothers and their infants. Other researchers have detected colonisation rates of 100% four days after birth and identified the nares and hands of hospital staff as a source of infant nasopharyngeal S. aureus[16]. Infants and their mothers could become colonised by the same hospital strain which might have different characteristics to strains present in the community. In a three year study of transmission of S. aureus in one maternity ward, there was a trend towards the spread of isolates producing SEC [17].

In this study, the lower carriage rate found in mothers compared to infants in the first three months of life could indicate that some infants acquire S. aureus from a maternal source other than nasopharyngeal carriage. Breast feeding might influence infant nasopharyngeal carriage of S. aureus and a 50% transmission rate has been identified between healthy mother and infant pairs [18]. Cracked, sore nipples are frequently experienced by breast feeding mothers and a strong correlation is reported between this condition and S. aureus colonisation [19]. In the first four months of life, there was a trend for more breast-fed infants than formula-fed infants to carry S. aureus in the nasopharynx, but this was not significant. The trend reached significance in the study by Peacock et al. [15] who examined 100 infant and mother pairs over a period of six months.

In infants, the decrease in carriage rates of S. aureus after three months of age coincided with a fall in the number of breast feeding mothers. At the same time, more intimate contact between infants and mothers was lost; some mothers returned to work and some infants began sleeping in their own bedrooms rather than their parent's bedroom.

Despite this early decline in colonisation, the nasopharynx remains an ecological niche for S. aureus throughout life and 20% of adults are described as persistent carriers of the organism, 60% intermittent carriers and 20% persistent non-carriers [20]. Since toxigenic strains of S. aureus are quite prevalent in the general population [21,22], adults commonly have protective antibodies to the bacterial toxins [23,24].

Analysis of cord blood samples detected placental transfer of maternal IgG to TSST-1 and SEC in 95.5% and 91.8% of cases respectively. The results indicate that the majority of adults in this study had been exposed to both staphylococcal toxins and generated an immune response. A small proportion of infants are born with no detectable passive immunity for one or other of the toxins. In addition, a marked variation in levels of maternal IgG to both TSST-1 and SEC among cord blood samples could affect the degree of protection afforded an infant. SIDS infants are reported to have lower levels of IgG to a number of bacterial toxins, including staphylococcal enterotoxin B (SEB), than an aged matched comparison group [25]. This could indicate that SIDS infants have been born with low levels of passive immunity.

Trans-placental transfer of IgG is up-regulated in the last trimester of pregnancy [26]. In this study there was a trend for levels of both anti-toxins to increase with gestational age. Essery et al. [27] found that the levels of these two anti-toxins decreased in maternal serum with weeks of gestation. The latter study used sera collected between weeks 30 and 40 of pregnancy from women with genitourinary tract infection. This study used sera from cord blood samples collected at births between 36 and 40 weeks gestation. In a full-term gestation, IgG concentration is usually higher in the cord serum than in the maternal serum [28]. It is possible that antibody levels become depleted in maternal serum through trans-placental transfer. Levels of anti-SEC increased with birth weight and maternal age, particularly in male infants. Prematurity, low birth weight, young maternal age at birth and male gender are all risk factors for SIDS [29].

The stratified design of the study allowed the effects of seasonality to be taken into account. Despite the reduction in SIDS cases seasonal variation still occurs with the greatest number of deaths reported in the winter months [30]. Levels of maternal IgG to TSST-1 and SEC in the cord blood samples showed a significant seasonal variation and infants born in the winter months had comparatively low levels of maternal IgG to TSST-1 suggesting that they might be vulnerable to the effects of the toxins at the time of year SIDS cases peak.

Levels of passive immunity could be boosted through breast feeding. There is conflicting evidence regarding the role of breast feeding in SIDS [31,32], but some epidemiological studies report a protective, independent effect [31,33]. Specific breast milk IgA antibodies have been found to common microbial colonisers of the upper respiratory tract [34]. In some cases they are reported to inhibit nasopharyngeal colonisation by pathogenic organisms [35]. Breast milk antibodies can neutralise the activity of bacterial toxins produced by gastrointestinal pathogens [36] by inhibiting binding to intestinal receptors [37,38] Studies report more cases of toxin induced gastrointestinal illness in formula-fed infants [39] and infants whose mothers produced lower titres of specific IgA in their milk [40]. S. aureus colonisation was increased in breast-fed infants both in the nasopharynx [15] and the gut [41]. Heikkila et al. [42] reported the inhibition of S. aureus by commensal flora found in breast milk.

In the present study, colonisation of the nasopharynx by S. aureus in adults or infants was not associated with increased levels of specific IgA to TSST-1 or SEC in breast milk. Gordon et al. [10], studied a smaller sample population and found that two thirds of mothers colonised with S. aureus in the nose or throat had milk samples with levels of IgA to the toxins equal to or higher than the mean. This difference might be explained by population size or by the prevalence of toxigenic S. aureus strains in each population. In this study, in the first three months of life, there was a trend for levels of breast milk IgA to TSST-1 to be higher in mothers of infants colonised with strains of S. aureus producing TSST-1.

Levels of IgA to TSST-1 and SEC detected in the breast milk samples varied greatly among mothers, data consistent with that reported by Gordon et al. [10]. Infants receiving lower levels of IgA to TSST-1 and SEC in breast milk appeared to compensate by producing higher levels of specific salivary IgA. Antibodies in breast milk neutralise bacterial toxins and inhibit antigenic stimulation at the mucosal surface.

Infants were able to mount an active mucosal immune response to TSST-1 and SEC in the first month of life; 11% of infants produced salivary IgA to TSST-1 and 5% of infants had salivary IgA to SEC. By the age of two months, these proportions had increased to 36% and 33% respectively. The data suggest that healthy infants have a competent oral mucosal immune system very early in life and are able to produce an active response when encountering a microbial antigen for the first time. This finding is consistent with other studies which report that the concentration of infant salivary IgA rises rapidly to reach a peak between four and six weeks of age [43] and production of specific IgA to bacterial antigens can be stimulated in the first month of life [44].

Some infants with no evidence of nasopharyngeal colonisation with S. aureus produced salivary IgA to the toxins. This could have been due to intermittent colonisation between sampling times. Breast milk antibodies can persist in the mouths of infants for some hours following feeding [45], but in this study some of the highest levels of infant salivary IgA to TSST-1 and SEC were detected in formula fed infants.

Regular dummy use is reported to decrease the risk of SIDS [46,47], and more infants who used a dummy had salivary IgA to TSST-1 and SEC compared to infants who did not use a dummy. Dummy users were more likely to have nasopharyngeal colonisation by S. aureus and to be formula-fed; this could account for the early immune response to the toxin antigens. The higher rates of S. aureus colonisation seen in dummy users might be due to repeated handling of the dummy by the carer replacing it when it becomes dislodged from the infant mouth.

Infants with lower levels of maternal IgG to TSST-1 associated with gestational age, birth weight and seasonality, had higher mean levels of salivary IgA to TSST-1. In contrast, mean levels of salivary IgA to both toxins in adults correlated positively with levels of cord blood IgG. This trend was significant for anti-SEC.

This study confirms that nasopharyngeal colonisation with S. aureus is common in the age range when the peak incidence of SIDS occurs. S. aureus isolates are reported to produce toxins in the temperature range 37–40 °C, with more toxin being produced at 40 °C [48]. In addition, expression of TSST-1 needs the presence of both oxygen and carbon dioxide [49]. Prone sleeping, viral infection and over heating [5052] could create a suitable environment for toxin production. Molony et al. [53] detected nasal septal temperatures greater than 37 °C in children placed in the prone position for only 30 min.

To reduce the risk of SIDS, parents are advised by health professionals to place their infants to sleep in the supine position and not to use excess bedding and clothing. In this study, no information was collected on infant temperature but infant sleeping position was recorded and the results were consistent with those of our previous study [54]. Infants were placed to sleep supine and continued to sleep this way until they were able to choose their own sleeping position. By 11 months of age, the majority of infants were sleeping prone. This information should be taken into account when conducting studies where sleeping position may be a confounding factor.

Mothers recorded 113 episodes of upper respiratory tract infection during the 12 months of the study. Our previous longitudinal study recorded precise dates of upper respiratory tract infection for infants in the first six months of life [4]. The age distribution was similar to the age distribution of SIDS in that both rose to peak at two to three months of age, but the rate of fall of upper respiratory tract infection was much less rapid than for SIDS. It is possible that viral infection marginally changed the environment of the nasopharynx so inducing staphylococcal isolates to produce small amounts of toxin.

In conclusion, this study shows that some infants are born with low or undetectable levels of antibody to bacterial toxins implicated in SIDS cases. Infants are able to mount an appropriate, protective immune response when they encounter the antigens for the first time. A combination of SIDS risk factors such as prone sleeping with a viral infection and over-heating could create the ideal nasopharyngeal environment for optimum staphylococcal toxin production and if this occurs on first exposure in an infant with low levels of protective antibody sudden death may result.

Acknowledgements

This work was supported by grants from Babes in Arms and the Peter Boizot Trust. We also acknowledge the generous support of Matthew Allen. We are grateful to the families who participated in this study, to Sandra Lively, Parentcraft Sister, for her help with recruitment and to the midwives for collection of cord blood samples.

References

  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
  30. [30].
  31. [31].
  32. [32].
  33. [33].
  34. [34].
  35. [35].
  36. [36].
  37. [37].
  38. [38].
  39. [39].
  40. [40].
  41. [41].
  42. [42].
  43. [43].
  44. [44].
  45. [45].
  46. [46].
  47. [47].
  48. [48].
  49. [49].
  50. [50].
  51. [51].
  52. [52].
  53. [53].
  54. [54].
View Abstract