JOGNN

RESEARCH

Effect of Cobedding Twins on
Coregulation, Infant State, and Twin
Safety
Kathryn M. Hayward, C. Celeste Johnston, Marsha L. Campbell-Yeo, Sheri L. Price, Shauna L. Houk,
Robin K. Whyte, Susan D. White, and Kim E. Caddell

Correspondence
Kathryn Hayward
School of Nurisng,
Dalhousie University, 5869
University Avenue, PO Box
15000, Haliax, NS, Canada,
B3H 3J5.
kathryn.hayward@dal.ca
Keywords
cobedding
coregulation
preterm
twin
multiples
infant sleep
neonatal
RCT

ABSTRACT
Objective: To evaluate the efficacy of cobedding on twin coregulation and twin safety.
Design: Randomized controlled trial (RCT).
Setting: Two university affiliated Level III neonatal intensive care units (NICUs).
Participants: One hundred and seventeen sets (N = 234) of stable preterm twins (<37 weeks gestational age at birth)
admitted to the NICU.
Methods: Sets of twins were randomly assigned to be cared for in a single cot (cobedded) or in separate cots (standard
care). State response was obtained from videotaped and physiologic data measured and recorded for three, 3-hour
sessions over a one-week study period. Tapes were coded for infant state by an assessor blind to the purpose of the
study.
Results: Twins who were cobedded spent more time in the same state (p < .01), less time in opposite states (p < .01),
were more often in quiet sleep (p < .01) and cried less (p < .01) than twins who were cared for in separate cots. There
was no difference in physiological parameters between groups (p = .85). There was no difference in patient safety
between groups (incidence of sepsis, p = .95), incidence of caregiver error (p = .31), and incidence of apnea (p = .70).
Conclusions: Cobedding promotes self-regulation and sleep and decreases crying without apparent increased risk.

JOGNN, 44, 193-202; 2015. DOI: 10.1111/1552-6909.12557
Accepted December 2014

Kathryn M. Hayward, MN,
RN, IBCLC, is an assistant
professor in the School of
Nursing, Dalhousie
University, Halifax, Nova
Scotia, Canada.
C. Celeste Johnston is an
Emeritus professor at the
Ingram School of Nursing,
McGill University,
Montreal, Quebec and
scientist in the Department
of Newborn Medicine,
IWK Health, Halifax, Nova
Scotia, Canada.

obedding is defined as caring for two or
more infants in the same cot/incubator and
is considered a developmental initiative to minimize adverse neurodevelopmental sequelae of
preterm birth (Als et al., 1994; Lutes & Altimier,
2001). However, there is a paucity of research to
support or refute the benefits of cobedding. The
practice of cobedding in nurseries is based on
anecdotal benefits to twins and to parents (Lutes
& Altimier, 2001; Nyquist & Lutes, 1998). Randomized controlled trials (RCTs), particularly multicentered trials, are required to determine the safety of
the practice of cobedding and the benefits to the
twins and the mothers.

C

(Continued)

The authors report no conflict of interest or relevant
financial relationships.

http://jognn.awhonn.org

As a result of their shared environment and mutual
tactile communication in utero, twin fetuses exhibit
a synchrony in sleep and awake states and a high
incidence of coincidental movements (Gallagher,
Costigan, & Johnson, 1992; Klaus & Klaus, 2000;
Nyquist & Lutes, 1998; Touch, Epstein, Pohl, &
Greenspan, 2002) and heart rate accelerations

C

(Gallagher et al., 1992; Sherer, Nawrocki, Peco,
Metlay, & Woods, 1990). At birth, twins are often
separated as individual health needs are met. Als
(1986) hypothesized that infants actively communicate how they perceive and cope with their
environment. She referred to this communication
as the synactive theory of development. Nyquist
and Lutes (1998) used this theory to provide an
explanatory model for understanding how cobedding may assist preterm twins in coping with the
extrauterine environment. Suggestions for cobedding are based on the premise that extrauterine
adaptation of twin neonates is enhanced by
close physical contact with the other twin, rather
than the sudden deprivation of such stimuli (Als,
1986; Byers, Yovaish, Lowman, & Francis, 2003;
Lutes & Altimier, 2001; Nyquist & Lutes, 1998;
Tomashek & Wallman, 2007). Swaddling multiples
in the same blanket and boundary provides
them with the opportunity to coregulate and to
continue to progress in their unique interactive
development.

2015 AWHONN, the Association of Women’s Health, Obstetric and Neonatal Nurses

193

 RESEARCH

Effect of Cobedding Twins

Questions remain concerning the efficacy of cobedding, the
short- and long-term benefits of cobedding, and the
circumstances under which cobedding should be practiced.

Review of the Literature

Marsha L. Campbell-Yeo,
PhD, NNP, RN, is an
assistant professor in the
School of Nursing and
Departments of Pediatrics,
Psychology and
Neuroscience, Dalhousie
University, Halifax, Nova
Scotia, Canada.
Sheri L. Price, PhD, MN,
RN, is an assistant
professor in the School of
Nursing, Dalhousie
University and an affiliate
scientist in the Women’s
Health Program at the IWK
Health Centre, Halifax,
Nova Scotia, Canada.
Shauna L. Houk, MN, RN,
is a representative of the
Parents of Multiple Births
Association and an assistant
professor in the School of
Nursing, Dalhousie
University, Halifax, Nova
Scotia, Canada.
Robin K. Whyte, BSc, MB,
BS, FRCP(C), is a professor
of medicine, Department of
Pediatrics, Dalhousie
University, Halifax, Nova
Scotia, Canada.
Susan D. White, BN, RN, is
a neonatal educator in the
Perinatal Program
Newfoundland and
Labrador, St Johns,
Newfoundland, Canada.
Kim E. Caddell, BN, RN, is
a neonatal nurse and
research coordinator at the
IWK Health Centre in
Halifax, Nova Scotia,
Canada.

194

Classic research on infant cues and state modulation provides evidence of the newborn infant’s
ability to not only respond to his or her external
environment, but also to initiate communication
through the use of cues (Barnard, 1999; Klaus
& Klaus, 2000; Sumner & Spietz, 1998). Investigators exploring the psychological and social
effects of cosleeping (mothers sharing beds
with their infants) found positive consequences
on child development (Lewis & Janda, 1998;
McKenna, 1997; McKenna & Mosko, 1994).
Cobedding theoretically enhances twin coregulation, improves respiratory status, decreases
oxygen requirements, increases weight gain, and
facilitates mutuality in their circadian rhythms and
sleep/awake patterns (Byers et al., 2003; Lutes &
Altimier, 2001; Nyquist & Lutes, 1998; Tomashek
& Wallman, 2007; Touch et al., 2002). In addition,
cobedding has been reported to enhance synchrony in caregiving and feeding of twins, which
in turn may facilitate mutuality in their circadian
rhythms and sleep/awake patterns (Byers et al.,
2003; Hayward, 2003; Lutes & Altimier, 2001;
Nyquist & Lutes, 1998; Touch et al., 2002).
Synchronization or coregulation is the way newborn twins support each other in the physiologic
transition to postnatal life and in the achievement
of stable sleep/awake states through activities mediated by physical contact (Als, 1986).
Through synonymous clustering of twin care, the
potential for synchronization in the twins’ activities
is optimized (Lutes & Altimier, 2001). Infant state
is the manner in which infants respond to their environments in an organized and predictable way;
this affects how parents provide care. There are
six commonly recognized infant states: two sleep
states, three awake states, and one transitional
state (Barnard, 1999). At one end of the sleep
wake cycle is quiet sleep, which is characterized
by a lack of body activity, smooth and regular
repirations, lack of facial or eye movement, and
occassional bursts of sucking movements. The
infant is generally unresponsive. This is nonrapid
eye movement (REM) sleep. Active sleep reflects
REM sleep. The infant has rapid eye movement,
irregular respirations, is more responsive to

JOGNN, 44, 193-202; 2015. DOI: 10.1111/1552-6909.12557

external stimuli, may smile and make brief fussy
or crying sounds. At the other end of the sleep
wake cycle is crying that is characterized by
irregular respirations, facial grimaces, cries,
color changes, and a variable sensitivity to
stimuli. Achieving stable sleep awake states is
a major developmental task of infants, reflecting
maturation of the central nervous system, and is
evidenced by increasing periods of quiet sleep,
decreasing periods of active sleep, and stable
transitions between the states (Als, 1986).
The National Association of Neonatal Nurses
(NANN, 2011) reported a paucity of scientific data
to support or reject the practice of cobedding
twins and higher order multiples. Questions
remain concerning the efficacy of cobedding,
the short and longterm benefits of cobedding,
and the circumstances under which cobedding
should be practiced (Lai, Foong, Foong, & Tan,
2012; NANN, 2011). Despite the potential benefits, there are few scientifically rigorous studies of
cobedding.
The purpose of this multisite RCT was to evaluate the efficacy of cobedding on twin coregulation. Data related to safety were also collected.
Twin coregulation was defined as the percentage
of time that twins coregulate their sleep/wake cycles and physiological parameters. Safety was
measured as rates of infection and of caregiver
errors. Further safety variables included abnormalities in heart rate, respiratory rate, oxygen saturation levels, and temperature stability as indicated by the incidence of heart rate decelerations,
apnea, fever, and low oxygen saturation. The aim
of this study was to assess the effect of cobedding on preterm twin coregulation and safety. The
specific coregulatory measures of quiet sleep and
crying states were further evaluated for duration of
time spent in each state.

Hypotheses
1. Preterm twins who are cobedded have enhanced coregulatory behavior without associated adverse effects compared to twins
who are not cobedded (standard care).
2. Preterm twins who are cobedded spend
more time in quiet sleep than twins who are
not cobedded (standard care).
3. There is no difference in infection rates, caregiver error, heart rate, temperature, or oxygen saturation between cobedding twins and
those receiving standard NICU care.

http://jognn.awhonn.org

 RESEARCH

Hayward, K. M. et al.

Assessed for
eligibility
(n =320 sets)

Met eligibility
requirements
(n=177 sets)
Refused to Participate (n = 34sets)
Other Reasons (n = 26 sets)

Randomized
(n = 117 sets)

Randomized to Cobed
Allocated to intervention (n = 63 sets)
Received allocated intervention (n = 62 sets)
Withdrew from the study (n = 1 set)

Randomized to Standard Care
Allocated to intervention (n= 54 sets)
Received allocated intervention (n=53 sets)
Withdrew from the study (n = 1 set)

Analysed:
Hypothesis 1 (n = 35)
Hypothesis 2 (n = 53 sets)
Hypothesis 3 (n = 53 sets)

Analysed:
Hypothesis 1 (n = 47 sets)
Hypothesis 2 (n = 62 sets)
Hypothesis 3 (n = 62 sets)

Figure 1. Flow diagram of subject enrollment, treatment allocation, and inclusion in analysis.

Methods

Sample

Design

Based on a priori data, it was determined that a
sample of 152 sets of twins was required to reach
a 50% change in corregulatory behavior, power of
80%, and a confidence interval of 95%. Population
effect was set at 0.5 to detect a moderate effect
(0.5) between groups (Hayward et al., 2007).

The primary outcome for this study was infant
coregulatory behavior. Secondary outcomes
focused on time spent in quiet sleep and
incidence of adverse events. A randomized
controlled trial was employed to examine the
effect of cobedding (caring for twins in the same
incubator/cot) compared to twins cared for in
a traditional neonatal environment (separate
incubators/cots) on twin coregulation, specifically outcome measures of infant sleep cycles,
state, and physiological variables (heart rate,
oxygen saturation, and temperature). Secondary
outcmes addressed the incidence of adverse
events, measuring the specifice outcomes: rate
of infection, caregiver error, and incidence of
abnormalities in physiological parameters (heart
rate drops, apnea, fever, and decreased oxygen
saturation) between cobedding twins and those
receiving traditional NICU care. The study was
conducted in two tertiary-level university affiliated
NICUs.

JOGNN 2015; Vol. 44, Issue 2

Preterm twins born before 37 weeks gestational
age, who were considered medically stable,
whose parents were able to understand written
and verbal English and who provided consent
were eligible for enrollment. Twins were excluded
if at the time of entry into the study either twin was
less than 1000 grams body weight, was receiving
ventilator support or phototherapy, had confirmed
or suspected sepsis, had chest tubes or umbilical
catheters insitu, or had known congenital anomalies. Randomization was stratified by weight (less
than or greater than 2000 grams) and by study site
(see Figure 1). Concealed committed allocation
was determined by an offsite computer-generated
program.

195

 RESEARCH

Effect of Cobedding Twins

Ethical approval was granted by the Institutional
Review Board governing the NICUs in each of the
two tertiary care health centres. Parents of twins
who met the inclusion criteria were approached
and provided written and verbal information on
the study. Written consent was then obtained from
each parent.

Procedures
Standard care. Twins randomized to receive standard care were placed in separate incubators or
based on infant weight placed in cots. Infants
weighing more than 2000 grams were moved to
a cot after 24 hours. The temperature of the incubator was set based on the weight of the larger
twin. All twins receiving standard care had cardiopulmonary monitoring during the week of data
collection and if medically indicated. As a safety
precaution, each twin was assigned a color and
their equipment was color coded to match the
individual twin.
Cobedding. Twins randomized to the cobedding
group were placed side-by-side, diaper clad, in
an incubator or were partially clothed in an open
cot. All twins assigned to the cobedding group
had cardiopulmonary monitoring maintained until
24 hours prior to discharge. Similar to the standard
care group, twins were assigned a color and a
specific side of the incubator/cot with all of their
equipment color coded to match.
Measures. Infant state and sleep cycle data were
collected using video recording. Using two cameras, individual twins were videotaped for 3 hours
on admission to the study to collect baseline data
prior to group allocation. Following group assignment, the videotaping procedure was repeated on
Days 3 and 7. Primary outcome data were collected for a period of one week following enrollment. Secondary outcome data were collected
until discontinuation of cobedding, generally
24 hours prior to discharge, at which time the monitors were discontinued and twins separated.
Physiologic data on heart rate were collected by
using a data acquisition system with a sampling
rate of 100 Hz averaged on a beat-to-beat basis. Arterial oxygen saturation was collected via
a pulse oximeter placed on a hand or foot of
the twins and connected to the data acquisition
system.

196

dard thermometers. Adverse events included
number of heart rate decelerations <85 bpm for
>20 seconds, apnea, fever, low oxygen saturation,
and incidence of infection. Incidence of infection
was measured by signs and symptoms of sepsis resulting in septic workups (analysis of blood
or body fluid and/or X-ray), incidents of treatment
with antibiotics, or infection recorded by positive
blood cultures. Incidence of medication and treatment error was defined as the incidence of one
twin receiving an intervention or diagnostic test
intended for the other twin. Treatment errors included medication error, non-invasive treatment
errors (X-ray or ultrasound), and invasive treatment errors (blood work). Tracking of errors was
through review of quarterly NICU reports of adverse occurrence reports. Study protocol adherence was monitored and recorded daily by the
site research coordinator.
Outcome variables. The sleep states for each twin
were compared for synchrony of activity (coregulatory behavior), length of time in quiet sleep state,
and length of time of in crying state and was coded
in 5-minute intervals over three separate 3-hour
periods of time (baseline 0, Day 3, and Day 7).
Synchrony of activity was measured as the number of times the infants were coded in the same
state over the 3-hour period of time. The videotapes were scored by offsite coders’ blind to the
purpose of the study and the study hypothesis.
Coders were trained in coding infant state according to the Nurse Child Assessment Satellite Tool’s
(NCAST) six sleep/awake states (quiet sleep, active sleep, drowsy, quiet alert, active alert, and crying) (Sumner & Spietz, 1998). Given the inability
to blind the cobedding intervention, coders were
trained separately to an ICC of .85 and were assigned to code either the cobedding or standard
care videos. Following the initial training, every
10th videotape of each twin was coded by a second coder for inter-rater reliability ensuring an ICC
above .85 was maintained throughout the study.
Secondary outcomes included adverse events, infection rates, and caregiver error.

Statistical Analysis

Twin skin and axillary temperatures were measured by skin temperature probes and unit stan-

Following descriptive statistics, each hypothesis
was tested separately (see Figure 1). Variables
were checked for normality of distribution. Square
root transformation was performed to normalize
the data and allow for parametric analysis. Hypothesis 1 was tested using repeated measures
(each 3-hour block of time) multivariate factorial
analysis of variance (RM-MANOVA) with group

JOGNN, 44, 193-202; 2015. DOI: 10.1111/1552-6909.12557

http://jognn.awhonn.org

 RESEARCH

Hayward, K. M. et al.

assignment as the independent between-subjects
variable and observation periods (Days 0, 3, 7) as
the within-subjects variable. Dependent variables
were percentage of incidents twins spent in the
same states and opposite states. Hypothesis
2 and 3 each each took into consideration the
unique nature of studying twins. Twin data, although collected individualy, are not independent.
Given that they were randomized as small groups
(twin sets), their nonindependence needed to be
taken into consideration in the analysis phase.
(Murray, Varnell, & Blitstein, 2004). Hypothesis
2 was tested using a nested RM-MANOVA with
a nested group assignment as the independent
between-subjects variable, observation periods
(Days 0, 3, 7) as the within-subjects variable, and
time in quiet sleep and crying as the dependent
variables. The measures related to Hypothesis 3
were tested using generalized estimating equations (GEE) for either continuous or categorical
measures. The demographic data for each twin
set was tested using an independent sample t
test and chi-squared as appropriate.

Twins who were cobedded were in quiet sleep more often and
cried less than twins who were cared for in separate cots.

and did not want to change or prolong course in
NICU. Cobedding twins did not differ significanty
from standard care twins in measured baseline
and demographic variables (Table 1). Two sets of
twins were withdrawn from the study prior to study
completion.

Coregulatory Behavior
The RM-MANOVA showed an overall betweensubject (standard vs. cobed) significance,
F(2) = 8.114, p < .01. Cobedded twins showed
greater time in the same behavioral state than
twins who received standard care, F(1) = 8.32,
p < .01 (see Table 2). Coregulatory behavior
changed differently over time for twins who
were cobed compared to twins who received
standard care, F(4) = 3.93, p < .01. There
was a significant change across time, between
groups when comparing the variable coregulatory
behavior. Over time, there was a greater increase
in the percent of time cobed twins spent in the
same states than twins who received standard
care, F(1) = 7.05, p < .01. There was a greater
increase in the percentage of time twins receiving
standard care spent in opposite states than twins
who were cobed, F(1) = 11.12, p < .01. When
comparing the univariate between group interaction looking at coregulatory behavior (same
state and opposite states) collapsed across
time, there was a significant difference between
groups, same state: F(1.97) = 4.92, p < .01,
opposite states, F(1.63) = 8.46, p < .01 (see
Table 2).

Results
Twins were recruited over a 5-year period between May 2006 and March 2011. Of the 320 sets
screened for recrutiment, 177 were eligible and
117 sets of twins and their parents were enrolled
(see Figure 1). One set of twins was withdrawn
from the study after receiving randomization
to standard care with parents stating a wish to
cobed. Reasons for refusal to participate were
plans to transfer to home hospital, parents wanting to be randomized to cobed, not interested,
only one parent in agreement, too stressed, did
not want to delay time with cardiac monitoring,

Table 1: Infant Demographics
Cobed
n (%)

Standard
M

SD

df

p

LL

UL

126

1655.7

421.4

106

1711.2

459.0

230

1.0

0.34

−58.5

169.5

GA entered

126

34.2

1.6

106

34.1

1.9

230

−0.5

0.65

−0.7

0.4

126

31.8

2.4

106

32.1

2.4

229

0.8

0.43

−0.4

0.9

GA at birth

SD

na (%)

Birth weight

b

M

95% CI
t

Monozygotic

26 (20.6%)

28 (26.9%)

1

0.26

Males

64 (50.8%)

47 (44.3%)

1

0.33

Females

62 (49.2%)

59 (55.7%)

1

0.33

Note. CI = confidence interval; LL = lower limit; UL = upper limit; GA = gestational age.
a
Data missing on two participants as participants withdrew from study after randomization.
b
Weight in grams.

JOGNN 2015; Vol. 44, Issue 2

197

 RESEARCH

Effect of Cobedding Twins

Behavioral States

quiet sleep, F(79) = 2.99, p < .01 (see Figure 2).
When comparing the univariate between group
interaction looking at individual states collapsed
across time, twins who were cobed were in quiet
sleep more, F(94) = 4.5, p < .01, and cried less,
F(87) = 1.5, p < .05, than twins receiving standard
care (see Figure 3).

NCAST identifies six states of consciousness: two
sleep states (quiet alert and active alert), one
transitional state (drowsy) and three awake states
(quiet alert, active alert and crying) (Sumner &
Spietz, 1994). All six NCAST infant states were examined for differences in duration of time spent
in each state (see Table 3). The multivariate test
for a between subjects main effect in infant states
was significant, F(474) = 2.9, p < .01. There was
also a significant multivariate main effect for time,
F(12) = 14.5, p < .01. Infant states changed differently over time for twins who cobed compared to
twins who received standard care, F(948) = 1.6,
p < .01. When comparing twins over time, twins
who were cobed increased the percent of time
spent in quiet sleep where as twins receiving
standard care decreased the percent of time in

There were no significant differences betweengroup in incidence of sepsis (p = .95, confidence
interval [CI] [0.10, 0.11]), incidence of caregiver
error (p = .31, CI [–0.02, 0.05]), and incidence
of apnea (p = .70, CI [1.37, 2.04]) (see Table 4).
In addition, there were no significant differences
in heart rate (cobed mean = 151 bpm, standard
mean = 152 bpm, p = .26, CI [–0.5, 1.89]), oxygen saturation (cobed mean = 95.1%, standard
mean = 95.6%, p = .62, CI [0.52, 1.27]), and
temperature (cobed mean = 37.0 C, standard
mean = 36.9 C, p = .37, CI [0.77, 0.29]).(See Table 4.) Only 26.9% of cobed infants experienced

There was no difference in infection rates, caregiver error, heart
rate, temperature, and oxygen saturation between cobedded
twins and those receiving standard NICU care.

Safety and Adverse Events

Table 2: Comparison of Percentage of Time Twins Spent in the Same State as Opposed to
Opposite States
95% CI
Measure

Group

Study day

M

SE

LL

UL

p

0

48.2

2.67

42.90

53.50

3c

46.64

2.67

38.33

48.97

7d

38.58

2.89

32.81

44.33

b

48.42

2.31

43.81

53.00

c

55.44

2.31

50.86

60.03

50.00

2.50

45.03

54.97

0b

9.44

1.75

5.96

12.93

c

11.98

1.40

9.19

14.78

d

10.16

1.18

7.81

12.51

b

0

11.29

1.51

8.28

4.29

3c

3.84

1.21

1.43

6.25

d

3.31

1.02

1.28

5.34

<0.01

a

Same state

Standard

Cobed

b

0

3
7

d

<0.01

Opposite statee
Standard

3
7
Cobed

7

Note. LL = lower limit; UL = upper limit.
Cobed n = 36 sets of twins; standard n = 47 sets of twins.
a
The percentage of time twin infants were in the same state.
b
Day 0 (prior to randomization).
c
Day 3 of data collection.
d
Day 7 of data collection.
e
The percentage of time twin infants were in opposite states (sleep vs. awake).

198

JOGNN, 44, 193-202; 2015. DOI: 10.1111/1552-6909.12557

http://jognn.awhonn.org

 RESEARCH

Hayward, K. M. et al.

Table 3: Mean Score and Standard Deviations for Infant State Between Nested
Groups Across Time
Ma (SD)
Measure

Cobed group

Standard group

Day 0b

17.0 (16.9)

17.7 (16.1)

Day 3c

24.1 (23.5)

17.9 (14.9)

Day7d

25.3 (23.9)

16.3 (13.5)

<0.01

Quiet sleep

<0.01

Active sleep
b

60.3 (16.5)

61.0 (17.5)

Day 3c

56.8 (24.1)

52.8 (17.1)

Day 7d

52.7 (23.6)

54.4 (15.3)

Day 0

<0.01

Drowsy
b

13.3 (11.5)

12.2 (10.6)

c

11.2 (11.4)

12.8 (11.6)

Day 7d

10.6 (10.6)

13.2 (10.4)

Day 0
Day 3

<0.01

Quiet alert
b

0.9 (2.4)

1.2 (3.1)

c

Day 3

1.5 (3.3)

1.6 (3.7)

Day 7d

1.6 (3.6)

1.2 (3.0)

Day 0

p

<0.01

Active alert
b

3.6 (5.2)

3.3 (4.9)

c

0.7 (2.7)

4.5 (5.7)

d

1.3 (3.3)

4.8 (5.4)

Day 0
Day 3
Day 7

<0.01

Crying
Day 0b

2.4 (4.4)

2.6 (4.3)

c

0.8 (2.5)

2.8 (4.8)

d

0.5 (1.6)

2.2 (3.5)

Day 3
Day 7

Note. Cobed group n = Day 0(119), Day 3(112), Day 7(98). Standard group n = Day 0(101), Day 3(91), Day 7(68).
a
M reflects the percentage of times the infants were counted in
each state.
b
Day 0 (prior to randomization).
c
Day 3 of data collection.
d
Day 7 of data collection.

an apnic episode compared to 30.2% of infants
receiving standard care.

Discussion
To our knowledge, this is the first clinical trial to
examine the effect of cobedding on twin coregulation and safety. We found that over time (7
days), there was a greater increase in the amount

JOGNN 2015; Vol. 44, Issue 2

of time cobedded twins spent in the same states
when compared to noncobedding twins. Nyquist
& Lutes (1998) and Lutes & Altimier (2001) also
reported coregulatory behavior and similarities in
sleep patterns in twins who were cobed. This similarity in sleep patterns is possibly related to twins
waking each other with their movement resulting
in similar wake/sleep cycles.
An important new finding of this study was that duration of quiet sleep increased over the 7 days of
data collection for cobedded twins, whereas twins
receiving standard care decreased the amount
of time in quiet sleep. Our findings are consistent
with others that have suggested that cobedding
duration may be linked with a higher degree of
physiological stability and regulatory behaviors.
Byers et al. (2003) reported significantly lower
activity heart rate levels in cobedding infants after
cobedding for 48 hours that they contributed
to a decrease in stress levels associated with
increasing exposure to cobedding. In a small
cohort study examining the responses of twins
to intermittent periods of cobedding, twin infants
with prior cobedding experience of no fewer than
24 hours were compared to infants that had never
been cobedded. Infants’ responsive behaviors
were then observed and recorded in an infant observation guide, and videotaped in three bedding
arrangements on five occasions for duration of
30 minutes in each bedding condition (two during
cobedding, two when separated, and one with
infant individually wrapped in a warm blanket). Although not statistically significant, Stainton, Jozza,
and Fethney (2005) reported a trend toward lowered stress levels and higher self regulatory
behavior in cobedding twins and a larger effect
with increasing duration of cobedding. Similarly,
two recent reports examining the pain reactivity,
infant regulation, and stress response of preterm
twins who were cobedding compared to noncobedding twins, while undergoing a heel lance
in the NICU demonstrated that cobedding twins
exhibited a faster physiological recovery (time it
took for heart rate and oxygen saturation levels to
return to baseline) (Campbell-Yeo et al., 2012) and
a more tightly regulated salivary cortisol response
when compared to a standard noncobedding
condition (Campbell-Yeo et al., 2014). In keeping
with others examining the effect of close maternal
contact provided through skin-to-skin contact
with mothers, it may be possible that the close
presence of the other twin enhances maturation
of self regulatory behavior and modulates stress
response (Feldman, Rosenthal, & Edelman,
2014).

199

 RESEARCH

Effect of Cobedding Twins

Figure 2. Estimated marginal means of time in quiet sleep.

In this study we were the first investigators to
rigorously examine the effect of cobedding on
coregulation and sleep over time. Given the
added work of caring for multiples and importance of strategies to promote twin sleep patterns
and parents ability to rest, these are important

findings for parents of multiples. Several unanswered questions remain regarding the possibility
that cobedding may promote more mature self
regulatory pathways in at risk preterm infants.
Further studies are needed to elucidate the sustained benefits of cobedding over greater periods

Figure 3. Estimated marginal means of time crying.

200

JOGNN, 44, 193-202; 2015. DOI: 10.1111/1552-6909.12557

http://jognn.awhonn.org

 RESEARCH

Hayward, K. M. et al.

Table 4: General Estimating Equation for
Infant Safety
95% CI
Measure

p
a

Temperature

M (%)

SD

0.37

0.77

Cobed

36.9

3.0

Standard

36.9

0.1

HR drop <85b

−0.5

0.26

Cobed
b

Standard

Sat. drop <80%

c

Standard
e

151

11.6

152

10.1

0.62

Cobed

95.1 (26.9d )

2.8

95.6 (30.2d )

2.6

0.70

Total apnea
Cobed

3.6

3.2

Standard

3.2

3.5

Sepsise

0.95
d

0.4

d

0.4

0.1 (5.6 )

Cobed

0.1 (5.7 )

Standard
e

Days antibiotics

0.41
0.8

3.7

Standard

1.0

4.6

Cobed
Standard

UL
0.29

1.89

0.52

1.27

1.37

2.04

0.10

0.11

5.76 14.05

Cobed

Errorse

LL

Despite the strengths of this study, there were
some limitations. These include the inability to
blind care providers to the intervention and the
decrease in power related to our decrease in
sample size. Sample size was based on the
primary hypothesis focusing on synchrony of
infant state; the study was underpowered for
addressing infant safety outcomes. Additional
studies are needed to confirm our findings and
explore the effects of cobedding on maturation
of self- regulation, sleep patterns, and safety over
sustained periods in hospital and home settings.
The National Association of Neonatal Nurses
(2011) and others alert health care providers to
the paucity of evidence supporting the safety and
benefits of cobedding in the home environment
(Campbell-Yeo et al., 2009, 2012, 2013; Touch
et al., 2002). As parents often mimic hospital
sleep practices it is important that future research
address the potential risks of cobedding in the
home environment (Campbell-Yeo et al., 2009,
2012, 2013; Touch et al., 2002). Both quantitative
and qualitative studies are needed to explore
parents’ decisions regarding choice of sleep
environment and how this choice affects the
family.

Conclusion
−0.02

0.31
0.02 (1.6d )
d

0 (0 )

0.05

0.1
0

Note. CI = confidence interval; LL = lower limit; UL = upper limit;
HR = heart rate; Sat. = oxygen saturation.
a
Celsius.
b
Beats/minute.
c
Saturation percentage.
d
Total percentage.
e
Count of events.

Twins in this study who were cobedded spent
more time in the same state, less time in opposite
states, were in quiet sleep more often, and cried
less than twins who were cared for in separate
cots. There was no difference in patient safety between groups. These results will add to the growing body of knowledge related to cobedding and
will contribute to the development of evidencebased recommendations for developmental care
of twins in hospital. Results may also be used to
support future research studying cobedding in the
home environment.

of time as well as infants born full term and older
infants.

Acknowledgment
We found that no adverse effects regarding increased risk related to temperature instability, respiratory compromise, infection, or
misidentification, were evident for twins undergoing cobedding compared to twins receiving standard care. Our findings add to the current body
of similarly reported evidence related to the safety
of cobedding of preterm twins in the NICU setting
(Campbell-Yeo et al., 2012; Chin, Hope, & Christos, 2006; Hayward et al., 2007; Lai et al., 2012;
LaMar & Dowling, 2006; NANN, 2011; Stainton
et al., 2005; Touch et al., 2002).

JOGNN 2015; Vol. 44, Issue 2

Funded by Nova Scotia Health Research Foundation and Canadian Nurses Foundation. The
authors thank Drs. Krista Ritchie and Colleen
O’Connell for statistical support.

REFERENCES
Als, H. (1986). A synactive model of neonatal behavioural organization:
framework for the assessment of neurobehavioural development
in the premature infant and support of infants and parent in the
neonatal intensive care environment. Physical & Occupational
Therapy in Pediatrics, 6, 3–53.

201

 RESEARCH

Effect of Cobedding Twins

Als, H., Lawhon, G., Duffy, F., McAnulty, G. B., Gibes-Grossman, R., &

Lai, N., Foong, S., Foong, W., & Tan, K. (2012). Cobedding in neonatal

Blickman, J. G. (1994). Individualized developmental care for the

nursery for promoting growth and neurodevelopment in stable

very low-birth-weight preterm infant medical and neurofunctional

preterm twins. Cochrane Database of Systematic Reviews, 12,

effects. Journal of the American Medical Association, 272, 853–
858.
Barnard, K. E. (1999). Beginning rhythms: The emerging process of

twins cared for in co bedding in the neonatal intensive care

sleep-wake behaviors and self-regulation. Seattle, WA: NCAST

unit. Journal of Obstetric, Gynecologic, & Neonatal Nursing, 35,

Publications, University of Washington, School of Nursing.

193–198. doi:10.1111/J.1552-6909.2006.00025.x

Byers, J., Yovaish, W., Lowman, L., & Francis, J (2003). Co-bedding

Lewis, R., & Janda, H. (1998). The relationship between adult sex-

versus single-bedding premature multiple-gestation infants in

ual adjustment and childhood experience regarding exposure

incubators. Journal of Obstetric, Gynecologic, & Neonatal Nurs-

to nudity, sleeping in the parental bed, and parental attitudes

ing, 32, 340–347.

toward sexuality. Archives of Sexual Behavior, 17, 349–362.

Campbell-Yeo, M., Johnston, C., Joseph, K., Feeley, N., Chambers, C.,
& Barrington, K. (2009). Co-bedding as a comfort measure for
twins undergoing painful procedures (CComForT trial). British
Medical Journal Pediatrics, 9, 76. doi:10.1186/1471-2431-976
Campbell-Yeo, M., Johnston, C., Joseph, K., Feeley, M., Chambers, C.,

doi:10.1007/BF01541812
Lutes, L., & Altimier, L. (2001). Co-bedding multiples. Newborn Infant
Nursing Review, 2(4), 242–246.
McKenna, J. J. (1997). Co-sleeping (bed sharing) among infants and
toddlers. Developmental and Behavioral Paediatrics, 18(6), 408–
412.

& Barrington, K. (2012). Cobedding and recovery time after heel

McKenna, J., & Mosko, S. (1994). Sleep and arousal, synchrony and

lance in preterm twins: Results of a randomized trial. Pediatrics,

independence among mothers and infants sleeping apart and

13(3), 500–506. doi:10.1542/peds.2012-0010

together (same bed): An experiment in evolutionary medicine.

Campbell-Yeo, M., Johnston, C., Joseph, K., Feeley, N., Chambers, C., & Barington, K. (2014). Co-bedding between preterm

Acta Paediatra Supple, 397, 94–102. doi:10.1111/j.16512227.1994.tb1371.x

twins attenuates stress response following heel lance: Results

Murray, D., Varnell, S., & Blitstein, J. (2004). Design and analysis of

of a randomized trial. Clinical Journal of Pain, 30(7), 598–

group-randomized trials: A review of recent methodological de-

604.

velopments. American Journal of Public Health, 94(3), 423–432.

Chin, S., Hope, L., & Christos, P. (2006). Evaluating the effects of

doi:http://dx.doi.org/10.2105/AJPH.94.3.423

cobedding on weight gain and physiologic regulation in preterm

National Association of Neonatal Nurses. (2011). Cobedding twins and

twins in the NICU. Advances in Neonatal Care, 6(3), 132–149.

higher order multiples. Advances in Neonatal Care, 9(6), 61–67.

doi:10.1016/j.adnc.2006.02.008

doi:10.1097/ANC.0b013e31823b53bd

Feldman, R., Rosenthal, Z., & Edelman, A. (2014). Maternal-preterm

Nyquist, K., & Lutes, L. (1998). Co-bedding twins: A developmentally

skin-to-skin contact enhances child physiologic organization

supportive care strategy. Journal of Obstetric, Gynecologic, &

and cognitive control across the first 10 years of life. Biological Psychiatry, 75, 56–64. doi:10.1016/j.biopsych.2013.08.
012
Gallagher, M. W., Costigan, K., & Johnson, T. (1992). Fetal heart rate accelerations, fetal movement, and fetal behavior patterns in twin

Neonatal Nursing, 27(4), 450–456.
Sherer, D. M., Nawrocki, M. N., Peco, N. E., Metlay, L. A., & Woods, J.
R. (1990). The occurrence of simultaneous fetal heart rate accelerations in twins during nonstress testing. Obstetrics & Gynaecology, 76(5 part 1), 817–821.

gestations. American Journal of Obstetrics and Gynecology,

Stainton, C. Jozza, E., & Fethney, J. (2005). Responses to co-bedding

167(4 part 1), 1140–1144. doi:10.1016/S0002-9378(12)80057-

of low birth weight twins in neonatal intensive care. Neonatal

3

Paediatric and Child Health Nursing, 8, 4–12.

Hayward, K. (2003). Cobedding of twins: A natural extension of the so-

Sumner, G., & Spietz, A. (1998). NCAST: Caregiver/parent-child inter-

cialization process? American Journal of Maternal Child Nursing,

action teaching manual. Seattle, WA: NCAST Publications, Uni-

28, 260–263.

versity of Washington, School of Nursing.

Hayward, K., Campbell-Yeo, M., Price, S., Morrison, C., Whyte, R.,

Tomashek, K., & Wallman, C. (2007). Cobedding twins and higher-

Cake, H., & Vine, J. (2007). Co-bedding twins: How pilot study

order multiples in a hospital. Pediatrics, 120(6), 1359–1366.

findings guided improvements in planning a larger multicentre
trial. Nursing Research, 56(2), 137–143.
Klaus, M., & Klaus, P. (2000). Your amazing newborn. Boston, MA:
Perseus Books.

202

1–10. doi:10.1002/14651858.CD008313.pub2
LaMar, K., & Dowling, D. (2006). Incidence of infection for preterm

JOGNN, 44, 193-202; 2015. DOI: 10.1111/1552-6909.12557

doi:10.1542/peds.2006-3096
Touch, S., Epstein, M., Pohl, C., & Greenspan, J. (2002). The impact
of co-bedding on sleep patterns in preterm twins. Clinical Pediatrics, 41, 425–431. doi:10.1177/000992280204100608

http://jognn.awhonn.org