Sleep Biol.

Rhythms
DOI 10.1007/s41105-016-0063-9

ORIGINAL ARTICLE

Melatonin-rich milk fortified with alpha s1 casein tryptic

hydrolysate improves primary insomnia: a randomized placebo
controlled trial
Angela Campbell1 • Alister Neill1

Received: 15 February 2016 / Accepted: 24 May 2016

Ó Japanese Society of Sleep Research 2016

Abstract To determine the effect of a powdered skim milk pharmacological preparations but smaller than those
product iNdream3; (Synlait Milk Ltd) on sleep in subjects reported for commonly prescribed pharmaceutical hypnotic
with insomnia. A 9 week double blind randomized placebo products.
controlled crossover trial was conducted in adults with Clinical trials registry number ACTRN12612000478819
primary insomnia (mean age 40 ± 14 years, 14/19 female,
insomnia severity index [15,). Subjects were randomized Keywords Insomnia
Melatonin
Actigraphy
Skim milk
to Product A: iNdream3 ‘NightMilk’ and Product B: powder
‘DayMilk’ for 3 weeks each with a 1 week washout. Total
sleep time (TST), sleep efficiency (SE), sleep onset latency
(SOL) and sleep quality were measured both subjectively Introduction
(sleep diary and questionnaires) and objectively (actigra-
phy), over 1 week during baseline and active treatment Insomnia is defined as difficulty initiating or maintaining
arms. At the end of each arm subjects underwent home sleep, waking too early or sleep that is chronically non-
polysomnography (PSG). Consuming NightMilk compared restorative [1] occurring despite adequate opportunity for
with DayMilk improved sleep symptom questionnaire sleep, for at least 1 month and causing significant daytime
scores: Pittsburg Sleep Quality Index efficiency (?5.2 vs distress or impairment. Co-morbid insomnia is used when
-0.4, p = 0.039), daytime dysfunction (-0.58 vs no the sleep complaint is due to another medical disorder,
change, p = 0.019), sleep related disturbance (-4.3 vs mental disorder, substance abuse or sleep disorder.
-1.5, p = 0.011) and SOL from diary, (-8 min vs Insomnia is clearly very common with prevalence esti-
?5 min, p = 0.045). Percent stage N3 sleep-PSG (?8 min mates that vary according to the definition used. Interna-
vs -9 min, p = 0.004) and SE– by actigraphy (?2.3 vs tionally around 30 % of adults report symptoms of
?0.5 %, p = 0.004) increased. The number of awaken- insomnia, 15–20 % have symptoms that cause significant
ings–actigraphy improved (-1.8/night vs ?1.5/night, impairment and 5–10 % of people have insomnia according
p = 0.003). TST was not different between treatments. In to standardized clinical criteria [2]. A large New Zealand
subjects with primary insomnia consuming NightMilk community survey suggested that 13 % of people aged
produced small but significant improvements in insomnia 20–59 are affected by an insomnia symptoms and excessive
symptoms, sleep efficiency, depth of sleep but did not daytime sleepiness [3, 4]. Māori are affected dispropor-
increase TST. The size of the effects seen was similar in tionately (prevalence in the study population: Māori 19.1 %,
magnitude to other less well designed evaluations of non- non-Māori 8.9 %). The risk of reporting a chronic sleep
problem (lasting longer than 6 months) increased with
increasing socioeconomic deprivation and age.
& Angela Campbell Insomnia sufferers experience a range of daytime
[email protected]
symptoms including daytime fatigue, impaired memory,
1
WellSleep, Department of Medicine, University of Otago poor concentration and irritability [5]. Although, often
Wellington, PO Box 7343, Wellington, New Zealand dismissed as a nuisance symptom, in clinically diagnosed

123
 Sleep Biol. Rhythms

insomnia, it is consistently associated with increased health The aim of this study is to determine whether NightMilk
care costs, occupational and social dysfunction and (with naturally enhanced levels of melatonin and an alpha
impaired quality of life [6]. S1 casein tryptic hydrolysate, iNdream3TM) improves the
Currently the most popular medical treatment for insomnia sleep of insomniacs beyond the effects of DayMilk (low
is the prescription of benzodiazepine agonists and sedating melatonin and no bioactive ingredients) alone.
antidepressants. These are available as prescription producing
a rapid improvement in sleep, but carry a significant risk of Materials and methods
side effects, tolerance and dependence, so only short term
therapy is advised. Of further concern is that sedatives use Recruitment
increases the mortality risk [7] so there is a strong clinical
need for safe, evidence-based alternative therapy. Potential participants were recruited from the community
Cognitive Behavioral Therapy for Insomnia (CBT-I), and underwent a 3 stage selection process: (1) Question-
seeks to change sleep related behaviors and thoughts, thereby naires, (2) Sleep Physician review, (3) Baseline
targeting the factors that cause insomnia. CBT-I is more polysomnography. The study was undertaken at WellSleep,
effective than hypnotic medication in the long term [8] but is University of Otago Sleep Investigation Center.
resource intensive and is not widely available outside spe-
cialized research settings. Newer online versions are making Inclusion criteria
this treatment more accessible and have been evaluated with
good results [9]. Patient drop out has been a concern given the The inclusion criteria required that participants:
sleep restriction component of treatment often produces an
initial worsening of daytime symptoms before improvement 1. Meet standard research diagnostic criteria for
is seen after 3–4 weeks into the treatment [10]. insomnia.
Melatonin is a hormone secreted from the pineal gland (a) One or more of the following insomnia
in the brain and that has an important role in the regulation symptoms for at least the last 2 months.
of the sleep-wake cycle. It is synthesized and released only
during the period of darkness from sundown to sunrise. (i) Difficulty initiating sleep
Once in the blood stream it acts to lower core body tem- (ii) Difficulty maintaining sleep
perature, which reduces arousal and increases sleep (iii) Waking up too early, or
propensity. Studies of oral melatonin at various doses show (iv) Sleep that is chronically non-restora-
that, it is effective in shifting the timing of sleep (e.g., jet tive or poor in quality
lag, sleep phase disorders) and in elderly individuals with (b) The above sleep difficulty occurs despite
insomnia it improves subjective sleep quality [11]. Mela- adequate opportunity and circumstances for
tonin has a superior safety profile compared with conven- sleep (not caused by sleep restriction).
tional sedatives and also innocent of any addictive effects. (c) In addition to a sleep complaint participants
Cow’s milk has long been considered sleep promoting in must experience at least one of the following
the western tradition, however, recent studies show that forms of daytime impairment related to the
100 g or 500 g of commercial milk has no effect on sleep nighttime difficulty.
or next day alertness in the elderly [12, 13]. Milk contains
small amounts of melatonin which exhibit a marked daily (i) Fatigue/malaise
rhythm, concentrations increase tenfold in milk produced at (ii) Attention, concentration or mem-
night [14]. This phenomenon, may be a mechanism by ory impairment
which hormones can pass through the milk and into the (iii) Social/vocational dysfunction
infant, improving nocturnal sleep [15]. The possible benefit (iv) Mood disturbance/irritability
to sleep from melatonin rich night milk was studied in (v) Daytime sleepiness
elderly institutionalized patients. Melatonin rich night (vi) Motivation/energy/initiative
milk, but not regular milk, produced a significant increase reduction
in morning and evening care-giver reported activity likely (vii) Proneness for errors/accidents at
related to improved sleep [13]. The second possibly active work or while driving
component of the Nightmilk, alpha S1 casein hydrolysate (viii) Tension headaches and or GI
has been the subject of research in milk and reports an symptoms in response to sleep loss
anxiolytic effect in rats [16] and subjective improvement in (ix) Concerns or worries about sleep.
some insomnia symptoms [17]. 2. Were over 24 years of age.

123
Sleep Biol. Rhythms

Exclusion criteria 1. EEG recorded at C3/A1 and C4/A2 (all electrodes

standard gold-plated Grass cup electrodes, Grass
Participants were excluded if they: instruments)
2. EOG placed at the left and right outer canthi (ROC/
1. Suffered from any other major medical or psycho-
A1; LOC/A2) with the right placed 1–1.5 cm higher
logical problem that may affect sleep, including but
than the left to measure vertical as well as horizontal
not limited to;
eye movements.
(a) Chronic Medical conditions such as. 3. EMG placed over the buccinators muscle, approxi-
mately 1–2 cm apart and slightly below the corners
(i) Cardiovascular disease, respiratory
of the mouth.
problems, gastrointestinal problems,
4. Reference electrodes placed on the forehead at sites
pain conditions, neurologic conditions,
CZ and Fp1.
metabolic disorders, sleep disorders,
5. Body position
psychiatric disorders (or a history of).
6. Oro-nasal airflow
(b) Have significant life style or personal factors 7. ECG/heart rate
that likely significantly impact their sleep. 8. Abdominal and thoracic RIP
(c) Take substances affecting sleep e.g., excessive 9. Pulse oximetry
caffeine or cigarettes, alcohol before bed, 10. Leg movements
illicit drug use.
All studies were scored by an experienced sleep tech-
(d) Use medications that alter sleep (and unable/
nologist using Profusion PSG3 (Compumedics Ltd). Sleep
unwilling to stop) i.e., sedatives, anti-
stages were scored according AASM 2007 standard criteria
depressants
[19] using the alternative hypopnea definition for sleep
2. Allergic to milk or lactose intolerant disordered breathing. Polysomnography at the end of each
3. Warfarin use randomized period and baseline 2 recorded sleep related
signals only.
Questionnaires
Study design
Potential subjects completed the PROMISÒ Sleep Related
Randomized double blind placebo controlled crossover
Disturbance (SRD) and Sleep related Impairment (SRI),
trial over 9 weeks. The products were sealed in identical
Pittsburgh Sleep Quality Index (PSQI), the Insomnia
sachets and simply labeled A or B. Participants were ran-
Severity Index (ISI) and our own pre-study questionnaire.
domized to which milk they received first using prospec-
To proceed to the next stage subjects needed to score an ISI
tive urn randomization. Randomization was undertaken by
of 15 (or at least showed moderate sleep disturbance with
a researcher following the completion of the 3 stage
moderate daytime dysfunction) and show no signs of the
enrolment process. Each participant acted as their own
exclusion criteria.
control Table 1.

Appointment with sleep physician Power analysis

Sleep and underlying medical history were taken during a Number of subjects was determined based on data from
consultation with a Sleep Physician ensuring likelihood of Buysse et al [20]. This study used actigraphy to determine
meeting inclusion criteria was high and specifically night to night variability of sleep efficiency in subjects with
ensuring a low likelihood of sleep disorders other than chronic insomnia. The current study was powered to detect
insomnia. a 6 % increase in SE with NightMilk using actigraphy
(n = 19).
Baseline polysomnography
Questionnaires
Studies were set up in the home by an experienced sleep
physiologist using the Somte PSG sleep system (Compumedics Questionnaires (PSQI, ISI, ESS, FOSQ (general produc-
Ltd, Melbourne Australia), according to the accepted method- tivity, activity, vigilance, social, total), PROMISÒ Sleep
ology of Campbell et al [18]. The montage included; Related Disturbance and PROMIS Sleep Related

123
 Sleep Biol. Rhythms

Table 1 Summary of the nine-week study protocol

Baseline Week Week Week One week Baseline Week Week Week
week 1 2 3 washout week 6 7 8

Sleep diary (whole week) x x x x

Actigraphy (whole week) x x x x
Home PSG (only for last day x x x x
of week)
Questionnaires x x x x
Product (if randomized to A A A A B B B
first)
PSG Polysomnography

Impairment) were completed on the morning following the amount was significantly altered due to external influences.
home PSG, at the end of the recording week. This was most commonly due to staying out late partying
or being significantly woken up often by a partner or child.
Milk drink These nights were excluded prior to analysis and unbinding
(n = 9 patients, total of 23/532 nights excluded (actigraphy
Subjects were instructed to take the milk drink 30 min and SD), 1 patient PSG data).
prior to bedtime each night for 3 weeks. The sachets
labeled A and B were identical. Both subjects and Statistical analysis
researchers were blinded to product details until after
analysis of data. Product A contained a reduced lactose To examine differences in the effects of night and day milk
skim milk powder produced by separating milk produced student t tests (paired, two-tailed) were performed for each
from cows at night from milk produced at other times of outcome measure both within and between the two milk
the day, producing a readily soluble powder containing types. Because the data for SOL latency and REM latency
melatonin (85.5 pg/ml) and modified with an alpha s1-ca- were not normally distributed Mann–Whitney U two-
sein tryptic hydrolysate. Product B was reduced lactose sample rank-sum tests were used instead of the t tests to
skim milk powder with a lower level of melatonin (8.8 pg/ assess differences. t tests were also performed between the
mg). Melatonin levels were confirmed using radioim- two baseline weeks to ensure washout was effective. The
munoassay techniques. Each sachet contained 30 g of baseline data was not different between weeks for any
product (Nightmilk sachet contained 2.565 ng of mela- variable. Given the number of outcome measures a sign
tonin) and subjects were instructed to make the drink up test was undertaken for subjective outcomes and objective
with 250 ml of water. outcomes for DayMilk arm results versus NightMilk arm
results. The purpose of the sign test was to combine the
Outcome measures information from all the tests to see whether the results
were randomly positive or negative for NightMilk over
The study assessed both objective and subjective measures DayMilk. All analyzes were completed prior to un-blinding
of sleep. The main outcome measure, for which the study of study group assignment.
was statistically powered, was sleep efficiency by actigra-
phy. SOL was measured subjectively over each week by
the sleep diary and questionnaires, however, also measured Results
objectively during the same week by actigraphy and then
by PSG on a single night at the end of each arm (Sunday– Recruitment and drop-outs
Thursday). Sleep quality was scored subjectively on sleep
diary for each night over one week from 1-very poor to The majority of participants who were excluded or
5-very good and the total week scored were compared. declined, did so from the first interview which was done
Actigraphy (Actiwatch2) was analyzed using 1 min over the phone or via email (Fig. 1). Some were lost after
epochs set to medium wake threshold selection (Philips the questionnaires were sent out. A few were excluded after
Actiware 6.0.8). an assessment by a sleep physician, who identified them as
Actigraphy and sleep diary measures are all averaged having excluding factors. The 9 that were excluded after
from valid nights of sleep during each week of recording. A baseline were mostly due to exhibiting sleep disordered
night of sleep was deemed invalid if the sleep timing or breathing. Of the 3 that dropped out after randomization;

123
Sleep Biol. Rhythms

Expressed interest via phone, email Table 2 Subject demographics

or in person N=213
Subject demographics N = 19
Sex 14F 5 M
Excluded or declined after Age (mean ± SD, years) 40 ± 14
initial interview N=123
Ethnicity 17 NZE 2 other
Insomnia type 9 Mixed 7 onset
Sent screening questionnaires 3 maintenance
N=90
ISI score (mean ± SD range) 17.3 ± 3.4, 14–23
ESS score (mean ± SD, range) 4.8 ± 3.8, 1–13
Met exclusion criteria or didn’t reply PSQI (mean ± SD range) 10.8 ± 2.4, 7–15
N=46
Insomnia type determined from questions 1–3 of the ISI
NZE New Zealand European, ISI insomnia severity index, ESS
Underwent physician review Epworth sleepiness score, PSQI Pittsburgh sleep quality index
N=-44

Discarded by physician (ISI) was used to assess severity for which 11 were mod-
N=13
Other reasons for insomnia/medication resulting in exclusion erate (ISI [ 15), 5 mild (ISI \ 15) and 3 severe (ISI [ 21).
Only 1 patient had excessive daytime sleepiness as repor-
Met all criteria to start baseline ted by their Epworth Sleepiness score (ESS [ 10/24). All
N=31
scored over 5 on the PSQI indicative of poor sleep.
As per the entry criteria patients were nonsmoking and
Discarded during initial baseline most drank alcohol very moderately. For those that did
N=9
Sleep disorder identified on PSG n=6 Declined n=3 drink a lot on occasion those nights were removed from the
weekly averages and they were asked not to drink on the
Randomized home PSG nights. Most drank caffeinated drinks daily but
N=22 none excessively. They were asked to not drink caffeine or
alcoholic beverages on the night of home PSG study. None
Drop outs N=3 all prior to starting protocol of the participants were on medications that would alter
One moved out of town
One had a non-related illness resulting in hospitalization sleep.
One felt she couldn’t continue without sleeping pills

Home polysomnography data

Successfully completed N=19

There was largely no change in the PSG variables across

Fig. 1 Recruitment and patient drop out the 4 nights of testing (Table 3). However, NightMilk
significantly increased stage N3 sleep time (?8 min)
one was transferred to another city for work, one grew ill compared to DayMilk (-9 min) which produced signifi-
with gallstones and another felt she couldn’t last the study cantly less stage 3 compared to baseline (p = 0.004).
without taking her sleeping pills. Subjects had significantly less stage N1 sleep compared to
baseline (-7 min) but was not significant against the pla-
Patient characteristics cebo (p = 0.13). TST did not change for either group.

The majority of subjects were female and of European or Actigraphy data

NZ European/Pakeha ethnicity (Table 2). Ages ranged
from 27 to 73 years with a mean age of 40 (±14) years. Actigraphy data shows NightMilk produced a significant
Most reported problems both getting and staying asleep increase in mean SE (?2.3 %) with no change with Day-
and all were at least moderate on one of these (using the Milk (Table 4). SOL was also significantly improved with
ISI). The insomnia type was assigned to them based on the NightMilk compared to DayMilk. NightMilk also produced
subjective and objective data; most people had a mix of a significant decrease in awakenings, this along with a
sleep maintenance and sleep onset problems (47 %) or just decrease (not significant in WASO) have combined to
problems with sleep onset (37 %). Few had sleep mainte- improve sleep efficiency despite no change in TST. No
nance problems only (16 %). The Insomnia Severity Score differences in any of the other measures.

123
 Sleep Biol. Rhythms

Table 3 Home PSG data

PSG measure Baseline DayMilk DayMilk Baseline NightMilk NightMilk Estimated effect size p value

SOL (min) 9.3 (8.8) 8.3 (14.9) 7.5 (7.9) 6.5 (15.3) 0.01 0.86
SE (%) 85.2 (4.4) 83.9 (9.9) 86.3 (9.0) 87.9 (5.7) 0.3 0.20
TST (min) 436.9 (40) 414.5 (66) 437.9 (47) 436.8 (48) 0.46 0.07
Arousal index/hour 10.5 (3.9) 11.3 (6.2) 12.3 (4.8) 11.5 (4.5) 0.45 0.075
Stage N1 sleep (min) 41.1 (14.7) 38.5 (17.2) 44.7 (12.9) 37.6 (12.5)* 0.58 0.13
Stage N2 sleep (min) 218.6 (30) 215.6 (50) 230.7 (39) 226.4 (45) 0.18 0.46
Stage N3 sleep (min) 79.0 (21.7) 70.3 (23.2)* 68.7 (24.9) 76.4 (22.2) 0.85 0.004**
REM sleep (min) 98.2 (15.4) 90.1 (24.3) 93.8 (22.2) 96.4 (18.3) 0.13 0.097
Values are mean (SD) for all variables apart from SOL and REM latency are median n = 16 for all variables
SOL sleep onset latency, SE sleep efficiency, TST total sleep time, REM rapid eye movement
* Significantly different from baseline
** Significantly different DayMilk vs NightMilk

Table 4 Actigraphy variables

Actigraphy measure Baseline DayMilk DayMilk Baseline NightMilk NightMilk Estimated effect size p value

SOL (min) 12.8 (10.3) 9.4 (17.2) 14.2 (14.2) 11.3 (10.6)* 0.8 0.02**
SE (%) 82.7 (4.8) 82.9 (4.9) 81.5 (4.8) 83.8 (3.8)* 0.45 0.036**
WASO (min) 58.6 (17.9) 56.7 (18.5) 61.3 (21.3) 56.2 (11.1) 0.31 0.29
Awakenings (#) 29.2 (8.8) 30.7 (8.9) 30.8 (8.0) 29.2 (7.3) 0.5 0.039**
TST (min) 427.9 (4.5) 430.6 (34) 423.4 (38) 426.7 (53) 0.41 0.48
Values are mean (SD) except for SOL which is median n = 18 for all variables
SOL Sleep onset latency SE Sleep efficiency WASO Wake after sleep onset TST Total sleep time
* Significant difference from baseline
** Significant difference DayMilk vs NightMilk

Sleep diary data only showed improvement in 1 (Sleep Related

Impairment).
NightMilk significantly reduced median SOL (-8) com- Along with NightMilk producing subjectively better
pared to DayMilk (?5), though neither was significantly sleep, when subjects were asked which product most
different from baseline from subjective sleep diary mea- improved their sleep (A or B), 8 chose A (NightMilk) and 6
sures (Table 5). There was also an improvement from chose B (DayMilk) (with 5 unsure).
baseline for sleep quality and TST but which wasn’t sig-
nificant compared to DayMilk.
Discussion
Questionnaire data
In this randomized placebo controlled cross-over trial of
There was a significant improvement in questionnaire adults with primary insomnia NightMilk improved
derived scores of sleep related disturbance in the Night- insomnia symptoms, sleep efficiency and sleep architecture
Milk arm compared to baseline and compared to DayMilk (increased % slow wave sleep). Total sleep time measured
(Table 6). The total PSQI improved from baseline with objectively by actigraphy or PSG did not significantly
NightMilk but the change wasn’t significant compared to change but this may have been subject to a ceiling effect.
DayMilk. However, examining the sub scores of the PSQI
reveals a significant improvement in daytime dysfunction Polysomnography and actigraphy
and sleep efficiency between NightMilk and DayMilk.
NightMilk produced small but significant improvements NightMilk consumption produced small but statistically
from baseline in 4 out of the 6 questionnaires used (ISI, significant improvements in actigraphy derived mean sleep
Sleep Related Disturbance, PSQI, FOSQ) while DayMilk efficiency and in number of awakenings compared to

123
Sleep Biol. Rhythms

Table 5 Sleep diary data

Sleep diary measure Baseline DayMilk DayMilk Baseline NightMilk NightMilk Estimated effect size P value

SOL (min) 40 (31.5) 45.0 (26.0) 37.0 (27.0) 29.0 (26.0) 0.14 0.045**
WASO (min) 53.9 (47.9) 48.9 (48.8) 51.7 (39.1) 42.7 (41.4) 0.22 0.33
Sleep quality? 11.9 (3.7) 12.4 (3.1) 12.0 (3.2) 13.6 (3.9)* 0.3 0.11
TST (min) 375.0 (76) 390.9 (68) 385.3 (50) 401.7 (54)* 0.01 0.49
SE (%) 77.9 (12.8) 80.3 (12.3) 80.0 (9.7) 84.4 (8.4) 0.26 0.39
Values are mean (SD) except for SOL which is median n = 19
SOL sleep onset latency, WASO wake after sleep onset
* Significant difference from baseline
** Significant difference DayMilk vs NightMilk
?
Sleep quality added over 1 week—subjective from a range of 1-very poor to 5-very good

Table 6 Questionnaire data

Questionnaire Baseline DayMilk DayMilk Baseline NightMilk NightMilk Estimated effect size p value

ISI 14.3 (4.3) 13.3 (3.9) 14.5 (5.2) 12.1 (4.5)* 0.33 0.098
PROMISÒ SRI 21.2 (5.0) 19.7 (4.7)* 21.2 (5.4) 19.5 (6.1) 0.1 0.049
PROMISÒ SRD 29.1 (5.3) 28.6 (4.9) 28.8 (6.0) 24.5 (5.0)* 0.59 0.011**
PSQI-Total 9.7 (2.9) 9.2 (2.5) 9.4 (2.9) 7.8 (3.3)* 0.4 0.10
PSQI-Daytime dysfunction 1.79 (0.8) 1.79 (1.0) 1.79 (0.9) 1.21 (1.0)* 0.7 0.019**
PSQI-Efficiency 69.1 (11.8) 68.7 (13.3) 67.7 (14.0) 72.9 (15.9)* 0.7 0.039**
PSQI-Sleep latency 4.16 (1.6) 3.74 (1.4) 3.79 (1.5) 3.32 (1.6)* 0 0.45
PSQI-Sleep quality 1.84 (0.6) 1.63 (0.5) 1.79 (0.42) 1.37 (0.5)* 0.4 0.15
Epworth sleepiness score 4.37 (3.9) 4.21 (5.2) 4.5 (3.9) 4.47 (3.2) 0.03 0.38
Mean (SD) n = 19 for all variables
ISI insomnia severity Index, SRI sleep related Impairment, SRD sleep related disturbance, PSQI Pittsburgh sleep quality index
* Significant difference from baseline
** Significant difference DayMilk vs NightMilk

DayMilk. Sleep stage analysis showed significant deeper 1 (Sleep Related Impairment). NightMilk when compared
sleep with an increase in the proportion of Stage N3 sleep with DayMilk significantly improved sleep related distur-
and relatively less time in stage N1 on nights after taking bance, daytime dysfunction and sleep efficiency. The
NightMilk. By contrast, DayMilk was associated with subjective improvement in SE of 5.2 % with no change
decreased TST during the PSG nights and significantly less with DayMilk was in agreement with the actigraphy
stage N3 sleep. These results are consistent with NightMilk measure.
having mild sleep enhancing effect when compared to Subjects reported subjectively faster SOLs (sleep diary)
DayMilk. with NightMilk, but both objective measures (actigraphy and
PSG) suggested no change in SOL. NightMilk appears to be
Sleep diary and questionnaire increasing the depth of sleep and acting to counteract the
perception of delayed sleep onset in subjects with insomnia.
On average the insomnia sufferers perceived an improve- An alteration in the perception of sleep is scientifically
ment in sleep onset latency with NightMilk. Sleep diary plausible with some insomnia sufferers reporting quite
measured sleep onset latency reduced from 37 min to marked sleep reduction despite having relatively normal
29 min after taking NightMilk. This was reflected in PSQI objectively measured total sleep time, a phenomena called
derived sleep latency but this was not significant when sleep state misperception [21]. This is seen in our own data
compared with the effect of DayMilk. NightMilk produced where the average perceived SOL is 48 min compared to
small but significant improvements from baseline in 4 out 8–12 min when measured for both actigraphy and PSG.
of the 6 questionnaires used (ISI, Sleep related disturbance, It should be acknowledged that slight objective
PSQI, FOSQ) while DayMilk only showed improvement in improvements in SOL might be difficult to measure having

123
 Sleep Biol. Rhythms

to respond to various rule based definitions of the sleeping insomnia was indeed their primary sleep complaint and we
state. In addition because actigraphy uses bodily movement followed the AASM research criteria for insomnia guide-
to estimate SOL periods of lying still, still awake, will be lines [30] wherever possible.
misclassified as sleep [22]. The PSG nights occurred only While the study numbers were small, the study was
once at the end of the 3rd week and involved some dis- powered on objective measures. The subjects were pri-
ruption to the usual routine, this data may have been more mary insomniacs and whether the results can be gener-
varied as a result and less representative than the sleep alized to those with secondary insomnia (such as another
diary which was averaged over whole weeks. It has been medical condition) is unknown at this stage. It is also
suggested that subjective and objective measures of sleep likely that both subjective complaints and objectively
onset match up best when your objective measure is PSG measured changes are subject to a ceiling effect
time from lights out to 10 min of consolidated stage 2 sleep becoming more difficult to document in subjects with
[23]. It would be interesting to further analyze our data longer actual sleep times. Despite a careful screening
using this definition of SOL. process aiming to identify insomnia sufferers it was
When matched against other sleep aid randomized surprising that the objectively measured total sleep time
controlled trials these improvements are comparable to was on average 430 min, which is deemed normal by the
findings for Tart cherry juice [24], melatonin [25] and PSQI. This finding reduced the ability of the study
Circadin [11] (prolonged release melatonin). However, the (ceiling effect) to show large changes in some of the
improvements of NightMilk are inferior in the short-term sleep measures.
to that shown by prescription hypnotics such as benzodi- This study has measured sleep variables in multiple
azepines, non-benzodiazepines and sedating antidepres- ways—for example SOL by PSG, actigraphy and sleep,
sants [26]. A review of psychological and behavioral outcomes vary and are likely to reflect the sensitivity of
therapies (CBTi) for chronic insomnia report moderate to each measuring technique. Sign testing for subjective
large effect sizes for these therapies in terms of improving outcomes produce a p value of 0.022 when comparing
sleep measures [27]. outcomes during the DayMilk arm versus NightMilk—fa-
In this study we found that DayMilk produced no voring NightMilk Objective outcome measures of DayMilk
improvement from baseline in any of the 33 sleep param- versus NightMilk give p = 0.006 in favor of NightMilk.
eters measured, except for a tiny improvement in the Given some milk proteins, for example those from the
PROMIS sleep related disturbance questionnaire. This hydrolysis of alpha S1 casein, have been shown to be
finding is in agreement with previous research using nor- associated with improved sleep and anxiety outcomes it is
mal milk as a control [13]. It is interesting to note that in possible such proteins have an enhancing effect acting in
contrast to using milk to get to sleep, a simple high gly- synergy with melatonin. Previous research with casein
cemic index (GI) meal is shown to produce a rather large tryptic hydrolysate (CTH) [17] using the PSQI found
improvement in SOL, at least in healthy individuals [28]. improvements in subjectively rated sleep quality and sleep
An evaluation of alternative or natural remedies of efficiency in workers with insomnia complaints and
insomnia concluded that rigorous scientific evidence of their decreased stress-related symptoms and perceived sleep
beneficial effect were absent for the vast majority with only problems in women [31].
valerian root and antihistamines showing a mild but con- Some other studies assessing sleep aids have made
vincing hypnotic effect. For countless others such as Cha- efforts to control subject diets, the tart cherry juice study
momile, St. Johns wart, magnesium, tryptophan, the studies asked subjects to avoid foods known to contain or influence
are limited by small numbers, lack of placebo, no statistical melatonin [24]. In addition they were required to record
analysis and sparse use of objective sleep measures [29]. their diet for the baseline week and to replicate it during the
By contrast the evaluation of NightMilk has been sub- supplementation week as closely as possible. This would
jected to a rigorous analysis using appropriate design. The appear logical as the timing and content of a nighttime
double blind nature of the study has ensured no investigator meal has been shown to significantly influence sleep onset
bias entered into the results, and the placebo (DayMilk) latency [28]. Diet is just one external influence on sleep,
was indistinguishable from NightMilk which is not always there are an infinite amount of variables that you could
possible with sleep aids. Randomization was undertaken endeavor to control for research purposes (such as physical
with a well described protocol and subject completion rate activity, mental stimulation, light levels, bed partners). In
was high. The use of both objective and subjective data has the current study we felt it was not relevant to try and
ensured that confidence can be taken from the results and control diet for a product taken prior to sleep where the
we can easily compare with other studies. The subjects in likely active ingredients were not commonly found in other
the study were well screened to ensure that primary food products.

123
Sleep Biol. Rhythms

Conclusion objectively-impaired vigilance: implications for the clinical

management of insomnia disorder. Sleep. 2014;37:229–37.
11. Lemoine P, Nir T, Laudon M, Zisapel N. Prolonged-release
NightMilk resulted in some small but significant melatonin improves sleep quality and morning alertness in
improvements in both objective and subjective sleep in insomnia patients aged 55 years and older and has no withdrawal
group of primary insomniacs. Eight of the 19 patients felt, effects. J Sleep Res. 2007;16:372–80.
12. Yamamura S, Morishima H, Kumano-go T, Suganuma N, Mat-
it improved their sleep and that they would purchase a
sumoto H, Adachi H, Sigedo Y, Mikani A, Kai T, Masuyama A,
NightMilk product as an aid for sleep. Takano T, Sugita Y, Takeda M. The effect of Lactobacillus
helveticus fermented milk on sleep and health perception in
Acknowledgements Mathew Johnson for assistance with patient elderly subjects. Eur J Clin Nutr. 2007;63:100–5.
recruitment, data entry and analysis. Dalice Sim for statistical advice. 13. Valtonen M, Niskanen L, Kangas AP, Koskinen T. Effect of
melatonin-rich night-time milk on sleep and activity in elderly
Compliance with ethical standards institutionalized subjects. Nordic J Psychiatr. 2005;59:217–21.
14. Peuhkuri K, Sihvola N, Korpela RA. Dietary factors and fluctu-
Funding Funding was provided through an unrestricted research ating levels of melatonin. Food and Nutr Res 56:17252. 2012.
grant from Synlait Milk Ltd, Ingredia SA, and NZ Trade and Enter- http://dx.doi.org/10.3402/fnr.v56i0.17252.
prise, with support from the Ministry for Primary Industries and 15. Cohen Engler A, Hadash A, Shehadeh N, Pillar G. Breastfeeding
DairyNZ through the Primary Growth Partnership. may improve nocturnal sleep and reduce infantile colic: potential
role of breast milk melatonin. Eur J Pediatr. 2012;171:729–32.
Conflict of interest No author has any conflict of interest to declare. 16. Violle N, Messaoudi M, Lefranc-Millot C, Desor D, Nejdi A,
No author received any direct financial payment or goods in kind Demagny B. Ethological comparison of the effects of a bovine a
from Synlait Milk Ltd. S1-casein tryptic hydrolysate and diazepam on the behaviour of
rats in two models of anxiety. Pharmacol, Biochem Behav.
Ethical approval All procedures performed in studies involving 2006;84:517–23.
human participants were in accordance with the ethical standards of 17. de Saint-Hilaire Z, Messaoudi M, Desor D, Kobayashi T. Effects
the national research committee. of a bovine alpha S1-Casein tryptic hydrolysate (CTH) on sleep
disorder in Japanese general population. Open Sleep J.
Informed consent Informed consent was obtained from all individ- 2009;2:26–32.
ual participants included in the study. 18. Campbell A, Neill AM. Home polysomnography for the diag-
nosis of suspected obstructive sleep apnea. J Sleep Res.
2011;20:207–13.
19. Iber C, Ancoli-Israel S, Chesson A, Quan S, for the American
References Academy of Sleep Medicine. The AASM manual for the scoring
of sleep and associated events: rules, terminology and technical
1. American Academy of Sleep Medicine. The international clas- specifications, 1st ed. Illinois: American Academy of Sleep
sification of sleep disorders. Diagnostic and coding manual. 2nd Medicine; 2007.
ed. Darien: American Academy of Sleep Medicine; 2005. 20. Buysse DJ, Cheng Y, Germain A, Moul DE, Franzen PL, Fletcher
2. Ohayon MM. Epidemiology of insomnia: what we know and M, Monk TH. Night-to-night sleep variability in older adults with
what we still need to learn. Sleep Med Rev. 2002;6:97–111. and without chronic insomnia. Sleep Med Rev. 2010;11:56–64.
3. Paine SJ, Gander PH, Harris R, Reid P. Who reports insomnia? 21. Edinger JD, Krystal AD. Subtyping primary insomnia: is sleep
Relationships with age, sex, ethnicity, and socioeconomic state misperception a distinct clinical entity? Sleep Med Rev.
deprivation. Sleep. 2004;27:1163–9. 2003;7:203–14.
4. Paine SJ, Gander PH, Harris RB, Reid P. Prevalence and con- 22. Lichstein KL, Stone KC, Donaldson J, Nau SD, Soeffing JP,
sequences of insomnia in New Zealand: disparities between Murray D, Lester KW, Aguillard RN. Actigraphy validation with
Maori and non-Maori. Aust NZ J Pub Health. 2005;29:22–8. insomnia. Sleep. 2006;29:232–9.
5. Roth T, Ancoli-Israel S. Daytime consequences and correlates of 23. Hauri PJ, Olmstead E. What is the moment of sleep onset for
insomnia in the United States: results of the 1991 National Sleep insomniacs? Sleep. 1983;6:10–5.
Foundation Survey. II. Sleep. 1999;31:177–84. 24. Howatson G, Bell PG, Tallent J, Middleton B, McHugh MP, Ellis
6. O’Keeffe KM, Gander PH, Scott WG, Scott HM. Insomnia J. Effect of tart cherry juice (Prunus cerasus) on melatonin levels
treatment in New Zealand. NZ Med J. 2012;125:46–59. and enhanced sleep quality. Eur J Nutr. 2012;51:909–16.
7. Kripke DF, Langer RD, Kline LE. Hypnotics’ association with 25. Buscemi N, Vandermeer B, Hooton N, Pandya R, Tjosvold L,
mortality or cancer: a matched cohort study. BMJ Open. Hartling L, Baker G, Klassen TP, Vohra S. The efficacy and
2012;2:e000850. doi:10.1136/bmjopen-2012-000850. safety of exogenous melatonin for primary sleep disorders. J Gen
8. Sivertsen B, Omvik S, Pallesen S, Bjorvatn B, Havik OE, Kvale Intern Med. 2005;20:1151–8.
G, Nielsen GH, Nordhus IH. Cognitive behavioral therapy vs 26. Buscemi N, Vandermeer B, Friesen C, Bialy L, Tubman M,
zopiclone for treatment of chronic primary insomnia in older Ospina M, Klassen TP, Witmans M. The efficacy and safety of
adults: a randomized controlled trial. JAMA. 2006;295:2851–8. drug treatments for chronic insomnia in adults: a meta-analysis of
9. Espie C, Kyle S, Williams C, Ong JC, Douglas NJ, Hames P, Brown RCTs. J Gen Intern Med. 2007;22:1335–50.
JSL. A randomized, placebo-controlled trial of online cognitive 27. Riemann D, Perlis M. The treatments of chronic insomnia: a
behavioral therapy for chronic insomnia disorder delivered via an review of benzodiazepine receptor agonists and psychological
automated media-rich web application. Sleep. 2012;35:769–81. and bahvioural therapies. Sleep Med Rev. 2009;13:205–14.
10. Kyle S, Miller CB, Rogers Z, Siriwardena AN, MacMahon KM, 28. Afaghi A, O’Connor H, Chow CM. High-glycemic-index car-
Espie CA. Sleep restriction therapy for insomnia is associated bohydrate meals shorten sleep onset. Am J Clin Nutr.
with reduced total sleep time, increased daytime somnolence, and 2007;85:426–30.

123
 Sleep Biol. Rhythms

29. Meolie A, Rosen C, Kristo D, Kohman M, Gooneratne N, Derivation of research diagnostic criteria for insomnia: report of
Aguillard RN, Fayle R, Troell R, Townsend D, Claman D, Hoban an American Academy of Sleep Medicine Work Group. Sleep.
T, Mahowald M, Clinical Practice Review Committee, American 2004;27:1567–96.
Academy of Sleep Medicine. Oral nonprescription treatment for 31. Kim JH, Desor D, Kim YT, Yoon WJ, Kim KS, Jun JS, Pyun KH,
insomnia: an evaluation of products with limited evidence. J Clin Shim I. Efficacy of a s1-casein hydrolysate on stress-related
Sleep Med. 2005;1:173–87. symptoms in women. Eur J Clin Nutr. 2006;61:536–41.
30. Edinger JD, Bonnet MH, Bootzin RR, Doghramji K, Dorsey CM,
Espie CA, Jamieson AO, McCall WV, Morin CM, Stepanski EJ.

123