Objective. Rebreathing of exhaled air is one proposed mechanism for the increased risk for sudden infant death syndrome among prone sleeping infants. We evaluated how carbon dioxide (CO2) dispersal was affected by a conventional crib mattress and 5 products recently marketed to prevent prone rebreathing.
Setting. Infant pulmonary laboratory.
Equipment. An infant mannequin with its nares connected via tubing to an 100-mL reservoir filled with 5% CO2. The sleep surfaces studied included: firm mattress covered by a sheet, Bumpa Bed, Breathe Easy, Kid Safe/Baby Air, Halo Sleep System, and Sleep Guardian. The mannequin was positioned prone face-down or near-face-down. The sleep surfaces were studied with the covering sheet taut, covering sheet wrinkled, and with the mannequin arm positioned up, near the face.
Measurements. We measured the fall in percentage end-tidal CO2 as the reservoir was ventilated with the piston pump. The half-time for CO2 dispersal (t1/2) is an index of the ability to cause or prevent rebreathing.
Results. Compared with the face-to-side control, 5 of 6 surfaces allowed a significant increase in t1/2 in all 3 prone scenarios. The firm mattress and 4 of the 5 surfaces designed to prevent rebreathing consistently allowed t1/2 above thresholds for the onset of CO2 retention and lethal rebreathing in an animal model (J Appl Physiol. 1995;78:740).
Conclusions. With very few exceptions, infants should be placed supine for sleep. For infants placed prone or rolling to the prone position, significant rebreathing of exhaled air would be likely on all surfaces studied, except one.
- sudden infant death syndrome
- carbon dioxide rebreathing
- commercial sleep products
- consumer product safety
- SIDS =
- sudden infant death syndrome •
- CO2 =
- carbon dioxide •
- t1/2 =
- half-time for CO2dispersal
It is recommended that healthy term infants be placed to sleep supine on firm bedding surfaces.1 This infant care guideline is based on the consistent finding that sudden infant death syndrome (SIDS) occurs less commonly among infants placed supine for sleep.2–7 Exceptions to this guideline may include infants who are preterm, those with craniofacial anomalies predisposing to upper airway obstruction, or those with severe gastroesophageal reflux disease.1
Despite the Back to Sleep campaign, it seems that there is continued reluctance among many caregivers to place infants supine for sleep.8 ,9 Reasons for this reluctance include concerns about choking and aspiration, sleep disruption, or the adverse developmental effects attributed to supine sleep.10Continued use of the prone sleep position may also reflect a lack of awareness of the Back to Sleep message.8 ,9 It is likely that many of these factors contribute to rates of prone sleeping that continue to exceed 20% in the United States.8
Several mechanisms have been proposed to explain the increased risk for SIDS among prone-sleeping infants. First, infants are at greater risk for thermal stress when prone.11–13 Second, they seem to arouse from sleep less predictably.14–16 Third, prone sleeping infants may rebreathe their own exhaled air. Rebreathing is suggested by the observation that prone victims are often discovered face straight-down with their noses and mouths covered by underlying soft bedding that traps exhaled carbon dioxide (CO2).17–23
When infants sleep prone, physiologic studies have shown that rebreathing can also occur with subtle changes in posture or in the sleep environment. For example, for a prone infant on a conventional firm crib mattress, positioning an arm alongside the face, or wrinkling the underlying sheet causes significant CO2rebreathing.18 24–26 The lethal potential of these subtle exposures is less obvious than those caused by soft bedding.18 ,23 ,27 Nevertheless, a threat could be posed to the prone infant with blunted arousal or ventilatory responses to hypercarbia or hypoxemia.28–31
A number of commercial products claim to modify SIDS risk by reducing the potential for CO2 rebreathing within the infant sleep environment. However, there are few published reports evaluating their efficacy. In this study, we used a mechanical model of infant respiration32 to evaluate the effect of these products on the dispersal of CO2 in 3 different prone sleeping scenarios: 1) face-straight-down with sheet taut, 2) face-straight-down with sheet wrinkled, and 3) face-near-straight-down with the arm positioned alongside the nose and mouth. Our intent was to define whether these commercial sleep products enhanced, limited, or had no effect on the dispersal of CO2 compared with a control.
Sleep Surfaces Studied
Five sleep surfaces that advertised to reduce rebreathing of CO2 were tested. Consumer Reportsrecommended a conventional firm, foam-block infant mattress covered by a tight-fitting sheet; this was also studied.1 ,33
Four of the 6 products may be classified as passive devices because they rely on passive diffusion of exhaled air to reduce rebreathing. Two of the 6 sleep systems are active devices that use fans or pumps to hasten gas dispersion.
The passive devices (and their marketing claims) included:
Baby Air (KidSafe, Tarzana, CA): a nylon mesh netting stretched over a rigid wooden frame covered by a second mesh sheet. The frame of the Baby Air is to be placed on top of another crib mattress. Its advertising claims that it will “reduce the risk of SIDS” and that the “Baby Air breathable surface allows infants to sleep safely in any position, even face-down.”
Breathe Easy (US Family Products, Springfield, MO): a waffle pattern foam pad covered with a perforated, fitted sheet and placed on top of a crib mattress. Its advertising claims that it will “reduce carbon dioxide in your baby's crib and that parents can … minimize lethal carbon dioxide accumulation hazard with our Breathe Easy sleeping surface.”
Bumpa Bed (Baby Jogger, Yakima, WA): a combination crib bumper pad and mattress made of breathable foam, covered by a polyester/cotton sheet. Its advertising claims that it is made of “… breathable foam, which we feel may have an impact on reducing the risk of SIDS.”
Conventional crib mattress: a recommended, firm, foam-block mattress with a nylon reinforced, vinyl covering.33
The passive devices were tested using the sheets or other covering materials supplied by the manufacturers. The conventional foam-block crib mattress was studied using 1 of 2 sheets: 1) 100% cotton flannel fitted sheet, and 2) 50% cotton/50% polyester fitted sheet.
The active devices (and their marketing claims) included:
Halo Sleep System Crib Mattress (Halo Sleep Systems, Plymouth, MN): an innerspring mattress with a perforated sleep surface and a small fan in the sidewall that moves air slowly through the perforations. It was tested with its combination sheet/mattress pad. Its advertising claims that the “… ventilation system of the Halo crib mattress prevents the accumulation of carbon dioxide …”
Sleep Guardian (Sleep Guardian, Charleston, SC): a foam pad with a cloth cover used on top of a crib mattress. The pad covers less than one half of the crib mattress surface. An air pump injects air through tubing into the pad. Its advertising claims that it “… prevents this carbon dioxide from pooling and thus becoming dangerous.”
Tests for Rebreathing
A mechanical model was used to quantify the rebreathing potential of each sleep surface.32 The model used an infant mannequin, with its head weighted to approximate the density of human tissue. The nares were connected via tubing to a 100-mL reservoir and piston pump. Tubing dead space was 15 mL. A Y-connector in the tubing was connected in-line to an infrared CO2analyzer (Biochem LifeSpan 100, BCI International, Waukesha, WI).
The 100-mL reservoir was filled with 5% CO2 in air. After a 15-mL dead-space system flush, the piston pump delivered 30-mL breaths at a rate equal to 15 breaths per minute. The rate and depth of breathing was chosen to allow for the recording of a smooth washout of CO2 from the sleep microenvironment. Results obtained in this manner have correlated well with animal model data.20 The time from the initial percent end-tidal CO2 measurement to the point corresponding to half the initial concentration represents the washout half-time for CO2 dispersal (t1/2).
Windows, doors, and air vents were closed minimizing cross ventilation within the test environment.34
A control CO2 dispersal t1/2 was established with the mannequin positioned prone on an uncovered hospital mattress surface with its face turned to the side. After the control study, the mannequin was placed on the sleep surfaces and positioned to simulate sleep scenarios that have been shown to cause rebreathing in the prone infant: 1) mannequin prone with face-straight-down and underlying cover taut; 2) mannequin prone with face-straight-down and underlying cover wrinkled; and 3) mannequin prone with face-near-straight-down and 1 arm positioned next to the face.
Ten repetitions of the CO2 washout were performed with each product under each of the 3 sleep scenarios.
Statistics and Data Analysis
To determine whether the sleep systems would significantly limit CO2 dispersal, the results were compared with the control t1/2 and with t1/2 thresholds associated with the onset of rebreathing (t1/2 > 21.1 seconds) and lethal rebreathing (t1/2 > 24.0 seconds) in a rabbit model.32 Statistical comparisons were completed using analysis of variance with post-hoc Dunnett's test for multiple comparisons against a single control.35Differences were considered significant if P was <.05.
Washout t1/2 values (mean ± standard deviation) for each system and simulation are shown in Table 1. For each of the 3 simulations, differences were noted among the t1/2 values for the 6 commercial sleep products and the control (analysis of variance,P < .0001). When compared with the face-to-side control, all sleep systems imposed significant delays in washout t1/2 during the 3 sleep simulations except for the Halo Sleep System.
Figures 1 to 3 depict washout t1/2 (mean values) for each sleep system in the 3 simulations studied. In the first sleep simulation (Fig 1), an infant mannequin was placed with its face down on each of the 7 sleep products. The sheet or cover supplied by the manufacturer of each product was kept taut. For each system, the bar represents the mean t1/2 of 10 washout repetitions and the line of stars above it represents the longest t1/2 among the repetitions. The top solid line at 24.0 seconds represents the washout t1/2 at which lethality occurred in animal studies.32 The middle line at 21.l seconds is the t1/2 at which there was a rise in Paco 2 in animal studies.32 The lowest line at 13.4 seconds is the control, face-to-side washout t½. The t1/2 for 5 of the systems exceeded the threshold associated with a rise in Paco 2, and 3 had t1/2s above the lethal threshold. There was no difference between the results for the Halo Sleep System and the control study.
In the second sleep simulation (Fig 2), the mannequin was placed face-down on each of the 6 sleep products and the sheet or cover supplied by the manufacturer of each product was wrinkled or puckered beneath the mannequin face. Four systems tested with a wrinkled cover yielded t1/2s that exceeded the lethal threshold. Studies using the firm mattress covered by either a polyester/cotton or a flannel sheet yielded t1/2s exceeding the threshold for increases in Paco 2. There was no difference between the Halo Sleep System and the control.
In the third sleep simulation (Fig 3), the mannequin was placed face-near-straight-down and turned slightly toward an arm that was positioned next to its face. If possible, the cover or sheet supplied by the manufacturer was slightly wrinkled, as might be expected when an infant pulls an arm toward its face. Six of the systems had 1 or more repetitions where the t1/2exceeded the lethal threshold. The Halo Sleep System t1/2s were not different from the control.
There were 2 intuitive design styles among our sample of products marketed to reduce rebreathing risk. The first group claimed to prevent the accumulation of CO2 by the passive conductance of exhaled air through mesh netting, foam channels, or foam undersurfaces. The second group of products used fans or pumps to actively assist with the dispersal of exhaled air.
We chose a mechanical model of face-down rebreathing to evaluate the effect of these commercial sleep products on CO2dispersal. This model and related models have proven useful in other studies of the sleep microenvironment.32 ,36 ,37 In previous studies, bedding associated with face-down sudden death was shown to have the longest CO2 washout half-time.20 In addition, good correlations between washout t1/2 and elevations in arterial Pco 2 have been reported.32 These earlier studies, in which the mechanical model was validated by comparison to results from living rabbits, established t1/2s above which the rabbits died. Although a sleep surface design intended to prevent the accumulation of CO2 may be desirable, we believe that designs which permit lethal rebreathing, based on comparisons to empirically derived thresholds, should be rejected.
Our study demonstrated that most products in this sample do not hasten CO2 dispersal. In fact, most of the products tested permitted CO2 dispersal half-times that exceeded thresholds associated with elevations of arterial Pco 2 and with lethal rebreathing. Only 1 product was able to maintain washout t1/2s below these thresholds in all 3 simulations of face-down sleep. This crib mattress includes a perforated synthetic rubber surface, a firm innerspring coil, and a side-wall fan that circulates air. Our study suggests that products intended to reduce rebreathing by infants sleeping prone and face-down should incorporate design features that actively disperse exhaled air.
These commercial products were studied under 3 different simulations of prone sleep. In the prone face-straight-down simulations, the sleep surfaces were tested separately with the surface covering first taut and then wrinkled. In a third scenario, the face was positioned near-straight-down, so that the nares did not contact the underlying bedding. However, an arm was placed alongside the face in contact with the airway. We found that for nearly all surfaces, wrinkling the surface covering prolonged the washout t1/2. Prolongations of dispersal t1/2s were also observed when the arm was positioned alongside the head in the face-near-straight-down studies. These findings confirm that subtle changes in the sleep microenvironment imposed by wrinkling the sleep surface or positioning the arm in proximity to the airway will aggravate the rebreathing experienced by prone sleeping infants.26
Death in a rebreathing microenvironment is likely to be multifactorial. For example, failure to arouse from sleep and change position may play a role in the substantial proportion of infant deaths where the victim is discovered prone with its face straight-down in underlying bedding.21 Although studies of infants' responses to exogenous CO2 show that some breathe harder and some arouse, this response is quite unpredictable particularly among infants at higher risk for SIDS.30 ,31 Products that intend to reduce SIDS risk by reducing rebreathing might benefit those infants with diminished responses to hypercarbia and hypoxia, by mitigating the exposure to these stressors.
Current guidelines specify that infants should be placed supine on firm sleep surfaces.1 We also evaluated a recommended33 firm foam-block mattress. Compared with the control, prolongation of the CO2 dispersal t1/2s were observed on the conventional firm mattress, with either covering sheet, in each of the 3 prone sleep scenarios. This finding suggests that even firm mattresses could pose a rebreathing threat when vulnerable infants sleep prone. This finding may be of relevance to recent studies showing that unaccustomed prone sleepers, ie, infants who typically sleep supine but are inadvertently placed or roll prone have an increased risk of SIDS.38 ,39These studies have shown that from 43% to 71% of SIDS victims, unaccustomed to prone sleep, were discovered in the face-straight-down position. Rebreathing stresses may contribute to the elevated SIDS risk observed among this group of infants unaccustomed to prone sleep.
To date, no commercial sleep product has been shown by case–control study to reduce SIDS risk. Nevertheless, it is possible that a product that actively hastens CO2 dispersal may benefit special groups of infants who are placed or who spontaneously role prone for sleep and are thus exposed to the risk of rebreathing exhaled air. These may include infants with craniofacial anomalies predisposing to upper airway obstruction, preterm infants on supplemental oxygen,40 ,41 infants with severe gastroesophageal reflux disease, and term infants who are unaccustomed to prone sleep. Additional study evaluating the potential benefit of such products to these special infant groups is warranted.
With a single exception, we have shown that commercial products marketed to prevent rebreathing by prone infants do not perform as advertised. Our study findings underscore the need for the standardized evaluation of sleep products using similar mechanical models before they are marketed to consumers. Furthermore, the development of uniform performance rating standards for sleep products may be useful to health care providers and consumers as an aid to the evaluation of the safety of these products.
This work was supported by Research Grant 47720 from the Children's Hospital Foundation (Minneapolis).
- Received July 9, 1999.
- Accepted October 29, 1999.
Reprint requests to (P.L.C.) Children's Hospitals and Clinics—Minneapolis, 2525 Chicago Ave S, Minneapolis, MN 55404. E-mail:
This work was presented at the Society for Pediatric Research-Academic Pediatric Society Meeting; May 4, 1999; San Francisco, CA; and at the SIDS Alliance National Meeting; April 8–11, 1999; Atlanta, GA.
- Kattwinkel J,
- Brooks J,
- Keenan ME,
- Malloy M,
- Willinger M
- Wigfield RE,
- Fleming PJ,
- Berry PJ,
- Rudd PT,
- Golding J
- ↵Dwyer T, Ponsonby AL, Blizzard L, Newman NM, Cochrane JA. The contribution of changes in the prevalence of prone sleeping position to the decline of sudden infant death syndrome in Tasmania. JAMA. 195;273:783–789
- Mitchell EA,
- Brunt JM,
- Everard C
- ↵Fllick LH, Kemp JS, Cibulka N, et al. Senior caregivers influence on infant sleep position among low income African American mothers. Presented at: SIDS Alliance National Conference; April 8–11, 1999; Atlanta GA
- Ponsonby AL,
- Dwyer T,
- Gibbons LE,
- Cochrane JA,
- Jones ME,
- McCall MJ
- Fleming PJ,
- Gilbert R,
- Azaz Y,
- Berry PJ,
- Rudd PT,
- Stewart A
- Kahn A,
- Groswasser J,
- Sottiauz M
- Lijowska AS,
- Reed NW,
- Mertins-Chiodini BA,
- Thach BT
- ↵Schmid WR. Potential rebreathing danger of presumed safe sleep surfaces and a new approach to design. Pediatr Pulmonol. 1996;425. Abstract
- Van der Hal AL,
- Rodriguez AM,
- Sargent CW,
- Platzker ACG,
- Keens TG
- Kemp JS,
- Thach BT
- ↵How to buy a crib mattress. Consumer Reports. May 1993:287
- Ryan EL
- ↵Glantz SA. Primer of Biostatistics. 4th ed. New York, NY: McGraw-Hill; 1997:99–103
- Bolton DPG,
- Taylor BJ,
- Campbell AJ,
- Galland BC,
- Cresswell CA
- Carleton JN,
- Donoghue AM,
- Porter WK
- Oyen N,
- Markestad T,
- Skjaeven R,
- et al.
- Copyright © 2000 American Academy of Pediatrics