OUP user menu

Cot mattresses as reservoirs of potentially harmful bacteria and the sudden infant death syndrome

Richard E. Sherburn, Richard O. Jenkins
DOI: http://dx.doi.org/10.1016/j.femsim.2004.06.011 76-84 First published online: 1 September 2004

Abstract

Cot mattress materials were investigated as potential reservoirs of bacteria in relation to the sudden infant death syndrome (SIDS). The sleeping position of the infant significantly influenced bacterial population density of cot mattress polyurethane foams (p<0.0000001) and their covers (p<0.004). Staphylococcus aureus was isolated at significantly higher frequency (p<0.03) from the infant's head region of cot mattress materials. Significantly higher bacterial population densities (p<0.001) were associated with polyurethane foams from non-integral mattresses (exposed polyurethane foam), when compared to those from mattresses completely covered by polyvinyl chloride (integral type mattress). The frequency of isolation of S. aureus from polyurethane foams from non-integral mattresses was also significantly higher (p=0.03) than from foams from the integral type. The following factors were significantly associated with increased frequency of isolation of S. aureus: from the polyurethane foam, previous use of non-integral mattresses by another child (p=0.03 for all sample sites, p=0.01 for torso region); from the covers, sleeping in the prone position (p=0.003 head region, p=0.001 torso region). Prone sleeping was also significantly associated with increased bacterial population levels (p=0.01) and increased frequency of isolation of Escherichia coli (p=0.02) from the torso region of cot mattress covers. These findings could explain some recently identified risk factors for SIDS associated with type and previous use of cot mattresses. Clostridium perfringens was isolated at very low frequency and Streptococcus pyogenes was not isolated from any cot mattress materials tested.

Keywords
  • SIDS
  • Cot death
  • Cot mattresses
  • Staphylococcus aureus

1 Introduction

Bacterial infections and the presence of bacterial toxins have been reported for many sudden infant death syndrome (SIDS) cases, and a common bacterial toxins hypothesis for SIDS has been proposed [1]. At least 50% of SIDS victims have respiratory or gastrointestinal infections prior to death [2,3], and an increased risk of SIDS has been associated with urinary tract infection [4]. Toxigenic bacteria implicated in SIDS include Staphylococcus aureus[58], Clostridium perfringens[6,9], Escherichia coli[1012], Streptococcus pyogenes[13], and Bordetella pertussis[14,15]. These bacteria and/or their toxins have all been isolated from a significantly greater proportion of SIDS cases than from non-SIDS cases or implicated in epidemiological studies. A synergistic effect between bacterial toxins and known risk factors such as exposure to cigarette smoke has also been reported [16,17].

Several studies have provided evidence for a link between SIDS and cot mattress type/bedding used. An increased risk of SIDS for infants sleeping prone on natural-fiber mattresses and sheepskin bedding has been reported [18,19]. The softness of cot mattresses [20] and bedding [21] have been reported as risk factors in SIDS, with softer materials linked to increased risk. Changes in infant care practices, including wrapping a mattress in a thick impermeable cover, have been associated with a decreased risk of SIDS in New Zealand [22]. Similarly, the use of a “waterproof” wool underblanket has been reported to reduce the risk of SIDS [23]. Brooke et al. [24] reported an increased risk of SIDS associated with sleeping on older mattresses not completely covered in polyvinyl chloride (PVC). A subsequent study by these researchers concluded that routine use of an infant mattress previously used by another child was associated with an increased risk of SIDS [25].

Studies of potential links between cot mattresses and SIDS have concentrated on carbon dioxide in the cot mattress environment [26] or the microbial formation of toxic gases [27,28]. Relatively little work has been done on microbial contamination within cot mattresses. Kelley et al. [29] isolated several fungi from cot mattresses, reporting an increased frequency of potentially harmful fungi such as Aspergillus fumigatus in SIDS mattresses compared to controls. This study, however, did not investigate bacterial contamination in detail, merely reporting on the presence of Bacillus spp. in used cot mattresses.

We have investigated the possibility that toxigenic bacteria become established in cot mattress foams and/or their covers, giving rise to reservoirs of opportunistic pathogens that can colonise the upper respiratory tract of infants. In the present paper, we report overall levels of bacterial contamination and the presence of certain toxigenic bacterial species in used cot mattresses. In particular, in view of recent findings on cot mattresses and SIDS [24,25], we have compared bacterial populations of cot mattresses in relation to mattress type (completely covered by PVC or not) and previous use by another child.

2 Materials and methods

2.1 Mattress collection

Used polyurethane foam cot mattresses (30) were donated by members of Parent and Toddler groups in the Leicester (UK) area and categorised according to mattress type and previous use: integral (I) or non-integral (NI); previously used (PU) or not-previously used (NPU). Integral cot mattresses were completely covered with PVC. Previously used cot mattresses had been used by another child; not-previously used mattresses had been used by one child only. All mattresses were in regular use prior to donation and were purchased as new equipment for the infant by the donating household. Donors completed a questionnaire giving details of infant sleeping practice, home environment and infant health. Mattresses were stored at −4 °C and sampled within 24–72 h of collection.

2.2 Excision sampling of cot mattress polyurethane foam

Polyurethane foam samples (1 cm3) were taken from eight different regions of the cot mattresses, according to the usual sleeping position of the infant (head or torso), in lateral section (centre or side) and vertical section (surface, middle or base) of the mattress. The samples were labeled as: head-centre-surface (HCS); head-centre-middle (HCM); head-centre-base (HCB); head-side-surface (HSS); torso-centre-surface (TCS); torso-centre-middle (TCM); torso-centre-base (TCB); mattress-end-surface (MES, from foot of mattress). Cot mattress cover samples (1 cm2) were taken from the HCS, HCB, TCS and TCB regions.

2.3 Estimation of bacterial populations following excision sampling of polyurethane foam

Each of the mattress samples was added to 5 ml of physiological saline (0.85% w/v NaCl), incubated with rotation for 10 min (60–90 revmin−1) and agitated with a sterile glass rod for 20 s. Tests carried out by inoculating samples of cot mattress materials with known bacterial species showed that the extraction method recovered between 60% and 90% of the viable bacteria applied, depending on bacterial species. Bacterial populations extracted into the physiological saline were diluted 10-fold up to 10−6 and 0.1 ml of each dilution inoculated onto solid media. Total bacterial counts were performed on nutrient agar (Oxoid) and selective media (Oxoid) for the following species were also included: B. pertussis, charcoal agar containing defibrinated horse blood (10% v/v) and cephalexin (40 mgl−1); C. perfringens, perfringens agar base containing egg yolk emulsion (50 mll−1) and d-cycloserine (9400 mgl−1); E. coli, eosin methylene blue agar (Modified) Levine; S. aureus, Vogel–Johnson agar containing 40 ml of a 1% potassium tellurite solution; S. pyogenes, Columbia blood agar base containing defibrinated horse blood (20 mll−1), colistin sulphate (10 mgl−1) and oxolinic acid (5 mgl−1). Total bacterial counts were estimated by cultivation under both aerobic and anaerobic conditions; colonies isolated anaerobically were subcultured onto nutrient agar and incubated aerobically to differentiate facultative and strict anaerobes.

All extracts from cot mattresses were grown alongside positive and negative controls. The positive controls were B. pertussis NCTC 11089, C. perfringens NCTC 11229, E. coli NCIMB 8545, S. aureus NCTC 7447 and S. pyogenes (De Montfort University culture collection). The negative controls were S. aureus NCTC 7447 (charcoal agar, streptococcus selective medium), Clostridium sordellii NCIMB 10717 (perfringens agar), and E. coli NCIMB 8545 (Vogel–Johnson agar). The plates were incubated at 35 °C for 48 h total counts and media selective for C. perfringens, S. aureus, S. pyogenes or E. coli or for 5–10 days for media selective for B. pertussis.

Presumptive identification of isolates to the species level was confirmed by morphological, physiological and biochemical testing: colony morphology; Gram stain; S. aureus, coagulase-positive using Bacto coagulase plasma (Difco); C. perfringens, nitrate reduction, lactose fermentation, gelatin liquefaction and absence of motility; E. coli, oxidase-negative, indole production.

2.4 Parental questionnaire

A mattress donor questionnaire was used to provide data on the history of use of the cot mattresses. The questionnaire comprised two sections, one dealing with the infant the other with mattress-related factors. Data on the infant included the following: age of the last infant to use the mattress at time of the last sleep on mattress; gender; usual sleeping pattern; illnesses (runny nose, fever); usual sleeping position; smoking within the household; and feeding pattern. Data on the mattress included the following: age; period of previous use by another child; lapsed time between use of mattresses; storage of the mattress; heating and ventilation in the cot room; type of bedding (sheets) used to cover the mattress; and usual sleeping position of the infant.

2.5 Statistical analysis of data

Overall bacterial population levels (total count data, Table 1) were estimated for each mattress by calculating the mean population density for eight polyurethane foam samples or for four mattress-cover samples. The distribution of bacterial population levels (total count data) within cot mattress polyurethane foams was evaluated by comparing bacterial counts for the various regions sampled with those for the mattress-end-surface region. For cot mattress covers, the distribution of bacterial population levels was evaluated by comparing counts with those for the TCB region of the cover. Significance testing was by the t-test after log10 transformation of the data.

View this table:
Table 1

Total bacterial counts for cot mattresses donated by the public

Material sampled/cot mattressTotal bacterial counts×10−3 (aCFU cm−3; bCFU cm−2)
NumberMeanSDMedian
aPolyurethane
Mattress type
    Integral8350.5
    Non-integral211055228
Previous use
    NPU-NI61438704
    PU-NI159028111
Sampling region
    HCS3080027550
    TCS3092944837
    MES306120.5
Mattress type for HCS
    HCS-I8885
    HCS-NI21501132488
Mattress type for TCS
    TCS-I8553
    TCS-NI2115634552
Previous use for HCS
    NPU-NI-HCS6751232933
    PU-NI-HCS1540158295
Previous use for TCS
    PU-NI-TCS6668228
    PU-NI-TCS1518440157
bCovers
Mattress type
    Integral8490.4
    Non-integral21281550.8
Previous use
    NPU-NI6251081
    PU-NI21281710.7
Sampling region
    HCS30143074794
    TCS30119965140.3
    TCB30764110.2
  • Mattress codes: integral, I; non-integral, NI; used previously by another child, PU; not-previously used by another child, NPU; head-centre-surface, HCS; torso-centre-surface, TCS; mattress-end-surface, MES.

Populations of bacteria were grouped according to frequency of occurrence for individual species, as indicated in the tables of results (Tables 25). Estimates and 95% confidence intervals for the odds ratios were calculated; the frequency distribution (P-value) for data sets were calculated using the Fisher Exact test (expected frequency <5). Where appropriate, bivariate analysis was carried out to establish whether two confounding factors were either jointly or independently significant. The Fisher Exact test was used by defining one factor with four categories combined from the original dichotomous factors.

View this table:
Table 2

Influence of (a) mattress type and (b) previous use by another child on frequency of isolation of S. aureus or E. coli from cot mattress materials

Factor/organismMaterialFrequency of isolationP-value
(a) Mattress typeIntegral (n=9)Non-integral (n=21)
   S. aureusPolyur.3 (33.3)15 (71.4)
OR 1.08.5, CI 1.5–49.50.03
Cover2 (22.2)13 (61.9)
OR 1.05.7, CI 0.9–34.50.11
   E. coliPolyur.3 (33.3)10 (47.6)
OR 1.01.8, CI 0.4–9.30.69
Cover4 (44.4)8 (38.1)
OR 1.00.8, CI 0.2–3.71.00
(b) Previous use by another childNPU-NI (n=6)PU-NI (n=15)
   S. aureusPolyur.2 (33.3)13 (86.7)
OR 1.013.0, CI 1.4–124.30.03
Cover2 (33.3)11 (73.3)
OR 1.05.5, CI 0.7–42.60.15
   E. coliPolyur.2 (33.3)7 (46.7)
OR 1.01.7, CI 0.2–12.60.66
Cover3 (50.0)5 (33.3)
OR 1.00.5, CI 0.1–3.40.63
  • Polyur. = polyurethane foam; cover = cot mattress cover. Odds-ratio, OR; not previously used by another child, NPU; previously used by another child, PU; non-integral, NI. Values in parentheses are numbers of positive samples expressed as percentages. CI = 95% confidence intervals for odds ratio.

View this table:
Table 3

Distribution of E. coli and S. aureus in cot mattress materials

OrganismFrequency of isolation/odds ratio (95% confidence interval)/P-value
HCSHCMHCBHSSTCSTCMTCBMES
Polyurethane foam
S. aureus13 (43.3%)5 (16.7%)8 (26.7%)9 (30.0%)11 (36.7%)3 (10.0%)4 (13.3%)4 (13.3%)
OR(CI)5.0 (1.4, 17.8)1.3 (0.3, 5.4)2.4 (0.6, 8.9)2.8 (0.8, 10.3)3.8 (1.0, 13.6)0.7 (0.1, 3.5)1.0 (0.2, 4.4)1.0
P0.021.000.330.210.071.001.001.00
E. coli5 (16.7%)2 (6.7%)4 (13.3%)4/30 (13.3%)8/30 (26.7%)3/30 (10.0%)4/30 (13.3%)3/30 (13.3%)
OR (CI)1.8 (0.4, 8.3)0.6 (0.1, 4.2)1.4 (0.2, 8.9)1.4 (0.2, 8.9)3.3 (0.8, 2.6)1.0 (0.2, 5.4)1.4 (0.2, 8.9)1.0
P0.711.001.001.000.181.001.001.00
Cot mattress covers
S. aureus14 (46.7%)4 (13.3%)7 (23.3%)4 (13.3%)
OR (CI)5.7 (1.6, 20.3)1.0 (0.2, 4.4)2.0 (0.5, 7.6)1.0
P0.011.000.510.71
E. coli6 (20.0%)2 (6.7%)7 (23.3%)2 (6.7%)
OR (CI)3.5 (0.6, 19.0)1.0 (0.1, 7.6)4.3 (0.8, 22.5)1.0
P0.251.000.151.00
  • Head-centre-surface, HCS; head-centre-middle, HCM; head-centre-base, HCB; head-side-surface, HSS; torso-centre-surface, TCS; torso-centre-middle, TCM; torso-centre-base, TCB; mattress-end-surface, MES. HCM, HSS, TCM, and MES were not analysed for cot mattress covers.

  • Number of cot mattresses analysed (n=30).

View this table:
Table 4

Influence of (a) mattress type and (b) previous use by another child on the frequency of isolation of S. aureus or E. coli from cot mattress polyurethane foam at the head-centre and torso-centre surface regions

OrganismSampling regionFrequency of isolationP-value
(a) Mattress typeIntegral (n=9)Non-integral (n=21)
    E. coliHCS1 (11.1)4 (19.0)
OR 1.01.9, CI 0.2–19.70.86
TCS2 (22.2)6 (28.6)
OR 1.01.4, CI 0.2–8.81.00
    S. aureusHCS2 (22.2)11 (52.4)
OR 1.03.8, CI 0.6–23.10.23
TCS1 (11.1)10 (47.6)
OR 1.07.3, CI 0.8–68.90.10
(b) Previous useNPU-NI (n=6)PU-NI (n=15)
    E. coliHCS1 (16.7)3 (20.0)
OR 1.01.2, CI 0.1–15.11.00
TCS2 (33.3)4 (26.7)
OR 1.00.7, CI 0.1–5.61.00
    S. aureusHCS2 (33.3)9 (60.0)
OR 1.03.0, CI 0.4–21.90.36
TCS0 (0)10 (66.7)
OR n/an/a0.01
  • Odds-ratio, OR; not previously used by another child, NPU; previously used by another child, PU; non-integral, NI; head-centre-surface, HCS; torso-centre-surface, TCS. Values in parentheses are numbers of positive samples expressed as percentages. CI = 95% confidence intervals for odds ratios. n/a = OR ratio analysis not applicable.

View this table:
Table 5

Influence of infant and mattress-related factors on frequency of isolation of S. aureus or E. coli from cot mattress materials

Organism/region of samplingPotential risk factorFrequency of isolationOdds-ratioP-value
Polyurethane foam
    S. aureus/TCSAge of mattress: >5 years5/7 (71.4)8.5, CI 1.2–57.90.03
<5 years5/22 (22.7)
Covers
    S. aureus/HCSSleeping position: prone8/9 (88.9)24.0, CI 2.4–242.30.003
supine5/20 (25.0)
    S. aureus/TCSSleeping position: prone6/9 (66.7)38.0, CI 3.3–437.00.001
supine1/20 (5.0)
Total counts/TCSSleeping position: prone8/9 (88.9)14.9, CI 1.5–144.20.01
supine7/20 (35.0)
    E. coli/TCSSleeping position: prone5/9 (55.5)11.2, CI 1.6–80.30.02
supine2/20 (10.0)
  • Head-centre-surface, HCS; torso-centre-surface, TCS. Values in parentheses are numbers of positive samples expressed as percentages. CI = 95% confidence intervals for odds ratios.

  • Counts above the median value for the data set (341.7 CFUcm−2) were regarded as positive.

3 Results

Of the mattresses sampled, 30% (9/30) were completely covered with a waterproof barrier (PVC) and were categorised as ‘integral’ type. The remaining mattresses (70%, 21/30) were categorised as ‘non-integral’ because a waterproof barrier did not cover part, or all, of the polyurethane foam. Mattresses were also categorised according to previous use: 60% (18/30) had been used by another child (previously used mattress), while the remaining mattresses (40%, 12/30) had been used by only one child (not-previously used mattress).

S. aureus was isolated from 60% (18/30) of cot mattress polyurethane foams and 47% (14/30) of cot mattress covers. E. coli was isolated from 43% (13/30) of foams and 37% (11/30) of covers. C. perfringens was isolated from only one mattress (3%), from both the polyurethane foam and cover, while S. pyogenes was not isolated from any cot mattress materials. Upon repeat sampling of mattresses after one-week storage at −4 °C, there was no significant change (n=5, p<0.05) in either total counts or counts for S. aureus.

Colonies typical of Bordetella sp. grew on charcoal blood agar plates containing cephalexin from extracts obtained from 17% (5/30) of polyurethane foams and 13% (4/30) of the covers. Microscopic examination showed that the cells were Gram-negative coccobacilli. On subculture the organism grew slowly (visible colonies after seven days) on charcoal blood agar but did not grow on nutrient agar. Subsequent loss of isolates during subculture precluded presumptive identification to the species level.

Total bacterial count data for mattresses donated by the public are shown in Table 1. With regards to the polyurethane foam, whole mattress (all sampling sites) bacterial populations levels were significantly higher for non-integral mattresses compared with integral mattresses (p<0.001). When considering the HCS or TCS regions of the polyurethane foam, the associated bacterial populations levels were significantly higher (p<0.0000001) than for the mattress-end-surface region. Similarly, for the HCS or TCS regions of the polyurethane foam, bacterial population levels of non-integral mattresses were significantly higher (p=0.01 and 0.04 respectively) than those of integral mattresses. Previous use of a cot mattress by another child did not significantly influence bacterial population levels associated with the polyurethane foam of non-integral mattresses (p=0.17 for comparison based on whole mattress; p=0.07 for comparison based on HCS or on TCS regions).

With regards to cot mattress covers (Table 1), there was no significant difference between bacterial populations levels (whole mattress) for integral and non-integral types (p=0.11) or for previous use of non-integral cot mattresses by another child (p=0.43). Similarly, when considering only the HCS or TCS regions of mattress covers, mattress type (p=0.14 and 0.43 respectively) or previous use of non-integral mattresses by another child (p=0.23 and 0.45 respectively) did not significantly influence bacterial population levels (data not shown). Bacterial populations levels, however, were significantly higher at the HCS or TCS regions of mattress covers (p=0.003 and p<0.00001 respectively) compared to those of the TCB region.

3.1 Isolation of S. aureus and E. coli from cot mattress materials, irrespective of sample site

The frequency of isolation of S. aureus from polyurethane foams of non-integral mattresses was significantly higher (p=0.03) than that from foams of the integral type. For non-integral mattresses, S. aureus was also isolated at a significantly greater frequency (p=0.03) from the polyurethane foams of mattresses that had previously been used by another child. These findings were not evident for S. aureus isolation from cot mattress covers. Mattress type or previous-use of non-integral mattresses by another child did not significantly influence the frequency of isolation of E. coli from polyurethane foams or from cot mattress covers (Table 2).

3.2 Distribution of S. aureus and E. coli within cot mattress materials

S. aureus was isolated at a significantly greater frequency (p=0.02) from the HCS region of the polyurethane foam, when compared to the mattress-end-surface region. S. aureus was also isolated at a significantly higher frequency (p=0.01) from the HCS region of the mattress cover compared to TCB region. There were no significant differences in the frequencies of isolation of E. coli from the various regions of the polyurethane foams or of the mattress covers (Table 3).

3.3 Assessment of mattress type and previous use as factors influencing isolation of S. aureus and E. coli from the head and torso regions of cot mattress polyurethane foam

Since mean bacterial population levels were highest at the HCS and TCS regions of cot mattress polyurethane foams (Table 1), the potential influence of mattress type and previous use by another child on frequency of isolation S. aureus and E. coli from these regions was investigated (Table 4). Mattress type (integral/non-integral) did not significantly influence the frequency of isolation of either S. aureus or E. coli from these two regions of polyurethane foam. Previous use by another child significantly increased (p=0.01) the frequency of isolation of S. aureus from the TCS region of polyurethane foam from non-integral mattresses; this finding was not evident for the HCS region of the mattress. Previous use by another child did not significantly influence the frequency of isolation of E. coli from either the HCS or TCS regions of non-integral mattresses (Table 4).

With regards to cot mattress covers, mattress type and previous-use by another child did not significantly influence bacterial population density (total counts) or the frequency of isolation of S. aureus or E. coli at either the HCS or TCS regions (data not shown).

3.4 Assessment of infant and mattress related factors on frequency of isolation of S. aureus and E. coli from cot mattress materials

Data on infant and mattress related factors were obtained from parents donating cot mattresses for the study by means of a questionnaire. The factors investigated were: infant age; gender; sleeping pattern; ailments (runny nose, fever); sleeping position; smoking within the household; breast or bottle feeding over the first 16 weeks; age of mattress; period of previous use of mattress; lapsed time between use of mattresses; storage of mattress between use; double-glazing in the cot room; type of bedding (sheets) used. Infant age was the age of the last infant to use mattress at time of last sleep on mattress: range 18–173 weeks; median 83.4 weeks; inter-quartile range 56 weeks. Sleeping position was the usual sleeping position of the infant during the months/weeks prior to mattress donation (Table 5).

For infant- or mattress-related factors assessed by the questionnaire, only age of the mattress and sleeping position showed significance associations by univariate analysis in relation to isolation of S. aureus or E. coli from one or more of the various mattress sampling sites (Table 5). Use of older mattresses significantly (p=0.03) increased the frequency of isolation of S. aureus from the TCS region of polyurethane foams. Age of mattress, however, was not independently significant with regards to isolation of S. aureus from this region of polyurethane foams; bivariate analysis showed that age of mattress and previous use by another child were jointly significant (p<0.03 for independently significant).

Smoking within a household was significantly associated with growth of colonies on the medium selective for Bordetella sp. from samples from the HCS region of mattress polyurethane foams: 4/9 households with smokers; 0/20 households with no smokers (p=0.005).

For the cot mattress covers, prone sleeping significantly increased the frequency of isolation of S. aureus from both the head and TCS regions (p=0.003 and 0.001 respectively). Prone sleeping also significantly increased bacterial load (total count data) (p=0.01) at the TCS region and isolation of E. coli (p=0.02) from the TCS region.

4 Discussion

Significantly higher bacterial population levels at HCS and TCS regions of cot mattress materials, together with significantly increased frequency of isolation of S. aureus from the HCS region of the polyurethane, indicates that the bacterial flora of the cot mattress materials was influenced by the sleeping position of the infant. There is substantive evidence in the primary literature that S. aureus might have an etiological role in SIDS. The organism has been isolated more frequently from the nasal passages of SIDS infants [30], and its toxins detected more frequently in the tissues and faeces of SIDS infants than from control group infants [6,8]. S. aureus is also the species that best fits a mathematical model proposed for the common bacterial toxins hypothesis for SIDS [30]. In the present study, S. aureus was the toxigenic species isolated most often from cot mattress materials, particularly from the region where the infant's head lies. Increased frequency of isolation of this organism from cot mattress materials was significantly associated with (1) use of non-integral mattresses, (2) use of non-integral mattresses that had been used previously by another child and (3) sleeping in the prone position. Associations (1) and (2) provide plausible explanations for recently identified risk factors for SIDS: the use of mattresses not completely covered with a waterproof cover [24]; and sleeping at night on mattresses used previously by another child [25].

Sleeping in the prone position is an established risk factor for SIDS. In the prone sleeping position mucus does not drain effectively from the nasal passages. This allows proliferation of bacteria within the nasal passages [31]. This sleeping position might, therefore, promote contamination of cot mattress materials through close proximity of the nasal passages to the mattress. Similarly, the close proximity could promote infection of the nasal passages with S. aureus harboured within cot mattress materials. The present study does not distinguish cause and effect in relation to prone sleeping and increased frequency of isolation of S. aureus. The significant association of S. aureus contamination of cot mattress materials with three of the known risk factors for SIDS highlights the presence of this organism within the mattress as a potential cause for concern.

The bacterial flora of cot mattress materials would be influenced by (1) the extent of bacterial and organic nutrient contamination from the infant and from exogenous sources and (2) the capability of individual bacterial species to survive and grow in the mattress environment. The exposed polyurethane foam of non-integral mattresses is not normally subject to cleaning and is likely, therefore, to accumulate potential nutrients, such as proteins, for microbial growth over time and with mattress use. In the absence of a waterproof barrier, polyurethane foams are also directly exposed to a major source of bacterial contamination and moisture, the infant. These considerations are consistent with (1) relatively high bacterial population levels associated with polyurethane foam from non-integral mattresses and (2) relatively high frequency of isolation of S. aureus from polyurethane foam from older mattresses/those used-previously by another child. Body fluids (saliva, sweat, urine) and infant-regurgitate are likely contributors to moisture and organic matter contamination of cot mattresses. This could explain relatively high bacterial population numbers associated with HCS- and TCS regions of cot mattresses. Similarly, sleeping in the prone position is likely to promote body fluid and/or regurgitate contamination of the mattress. This might account for the observed significant associations between prone sleeping and bacterial population levels at the head and torso regions of mattress covers.

S. aureus has been shown to survive under nutrient limited conditions for several weeks on various hospital fabrics and plastics, with the longest survival times on polyester and polyethylene plastic [32]. Relatively good survival of S. aureus on polyurethane foam would favour development of reservoirs of infection for this organism in cot mattress material. We are currently investigating environmental factors influencing survival of various toxigenic bacteria, including S. aureus and E. coli, on cot mattress materials.

Growth of colonies on the medium selective for Bordetella sp. was observed with extracts from five cot mattresses collected during a whooping cough epidemic in the Leicester region. Colony morphology, microscopic morphology and slow growth on the selective medium but not on nutrient agar were consistent with Bordetella sp. Loss of the isolates on subculture precluded further identification. These observations must be interpreted with extreme caution. Clinical isolates of B. pertussis have poor survival capability outside the human host [33] and it is most likely that the cot mattress isolates were not B. pertussis.

The present research provides evidence that reservoirs of bacteria, including certain toxigenic species, exist in cot mattress materials. Such reservoirs could give rise to heavy and prolonged exposure of infants to toxigenic bacteria over a period when the immune system of infants is at its most vulnerable (two to four months of age). The findings, particularly those relating to S. aureus, could explain the lowered risk of SIDS associated with use of a waterproof cover above the mattress [22,23]. Mechanistically, the findings could also explain an increased risk of SIDS associated with sleeping on older mattresses not completely covered with PVC [24] or use of infant mattresses previously used by another child [25].

Acknowledgements

The investigation was funded by The Scottish Cot Death Trust, Glasgow, UK.

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].
View Abstract