Effect of firmness of mattress on chronic non-specific low-back pain: Randomised, double-blind, controlled, multicentre trial

Article (PDF Available)inThe Lancet 362(9396):1599-604 · December 2003with122 Reads
DOI: 10.1016/S0140-6736(03)14792-7 · Source: PubMed
Abstract
A firm mattress is commonly believed to be beneficial for low-back pain, although evidence supporting this recommendation is lacking. We assessed the effect of different firmnesses of mattresses on the clinical course of patients with chronic non-specific low-back pain. In a randomised, double-blind, controlled, multicentre trial, we assessed 313 adults who had chronic non-specific low-back pain, but no referred pain, who complained of backache while lying in bed and on rising. Mattress firmness is rated on a scale developed by the European Committee for Standardisation. The H(s) scale starts at 1.0 (firmest) and stops at 10.0 (softest). We randomly assigned participants firm mattresses (H(s)=2.3) or medium-firm mattresses (H(s)=5.6). We did clinical assessments at baseline and at 90 days. Primary endpoints were improvements in pain while lying in bed, pain on rising, and disability. At 90 days, patients with medium-firm mattresses had better outcomes for pain in bed (odds ratio 2.36 [95% CI 1.13-4.93]), pain on rising (1.93 [0.97-3.86]), and disability (2.10 [1.24-3.56]) than did patients with firm mattresses. Throughout the study period, patients with medium-firm mattresses also had less daytime low-back pain (p=0.059), pain while lying in bed (p=0.064), and pain on rising (p=0.008) than did patients with firm mattresses. A mattress of medium firmness improves pain and disability among patients with chronic non-specific low-back pain.
For personal use. Only reproduce with permission from The Lancet publishing Group.
ARTICLES
THE LANCET • Vol 362 • November 15, 2003 • www.thelancet.com 1599
Summary
Background A firm mattress is commonly believed to be
beneficial for low-back pain, although evidence supporting
this recommendation is lacking. We assessed the effect of
different firmnesses of mattresses on the clinical course of
patients with chronic non-specific low-back pain.
Methods In a randomised, double-blind, controlled,
multicentre trial, we assessed 313 adults who had chronic
non-specific low-back pain, but no referred pain, who
complained of backache while lying in bed and on rising.
Mattress firmness is rated on a scale developed by the
European Committee for Standardisation. The H
s
scale starts
at 1·0 (firmest) and stops at 10·0 (softest). We randomly
assigned participants firm mattresses (H
s
=2·3) or medium-
firm mattresses (H
s
=5·6). We did clinical assessments at
baseline and at 90 days. Primary endpoints were
improvements in pain while lying in bed, pain on rising, and
disability.
Findings At 90 days, patients with medium-firm mattresses
had better outcomes for pain in bed (odds ratio 2·36 [95% CI
1·13–4·93]), pain on rising (1·93 [0·97–3·86]), and disability
(2·10 [1·24–3·56]) than did patients with firm mattresses.
Throughout the study period, patients with medium-firm
mattresses also had less daytime low-back pain (p=0·059),
pain while lying in bed (p=0·064), and pain on rising
(p=0·008) than did patients with firm mattresses.
Interpretation A mattress of medium firmness improves pain
and disability among patients with chronic non-specific low-
back pain.
Lancet 2003; 362: 1599–604
See Commentary page 1594
Introduction
Non-specific low-back pain is defined as pain between the
costal margins and the inferior gluteal folds that is
generally accompanied by painful limitation of motion, is
affected by physical activities and posture, and might be
associated with referred pain.
1
The diagnosis implies that
the syndrome is not related to underlying disorders, such
as fractures, spondylitis, direct trauma, or systemic
processes. Although the pain is frequently believed to be
the result of degenerative disc syndrome, protrusion or
hernia of intervertebral discs, facet-joint degeneration, or
other disorders associated with position or movement of
the spine, such as scoliosis, vertebral instability, or
spondylolisthesis, in 85% of patients no organic cause can
be established.
1
Several biomechanical factors raise the risk of low-back
pain.
2,3
In healthy people, characteristics of mattresses
may trigger pain, especially in the morning.
4
People who
have chronic low-back pain seem to be more sensitive to
the firmness of mattresses than healthy people.
5
In
patients undergoing percutaneous transluminal coronary
angioplasty, the combination of an alternating air mattress
and an exercise programme reduces the frequency and
severity of low-back pain during the 48 h after the
procedure,
6
and in a population-based study, feeling low-
back pain in bed or on rising was the factor most strongly
associated with low-back pain.
7
In daily practice, physicians are frequently requested to
counsel on the characteristics of beds and mattresses to
lessen back pain. In a survey of orthopaedic surgeons,
95% believed that mattresses played a part in the
management of low-back pain, with 76% recommending
a firm mattress.
8
However, evidence supporting this
advice is lacking. The effect of mattress characteristics on
low-back pain has been analysed in a limited number of
studies,
6,9–12
but the results are weakened by shortcomings
in the methods. Few treatments are effective for chronic
low-back pain, a disorder that causes most social costs in
developed countries.
13,14
We assessed the effect of the firmness of mattresses on
the clinical course of chronic low-back pain and disability.
We postulated that the effect of the mattress would be
more noticeable while lying in bed or on rising.
Methods
Study population
We recruited adults who had participated in a previous
population-based study on the prevalence and risk factors
for common low-back pain in adolescents and their
parents.
7
Adults were eligible if they reported low-back
pain while lying in bed or on rising. A research assistant
visited people at home to assess inclusion and exclusion
criteria before inviting them to participate.
Inclusion criteria were: age 18 years or older, presence
of chronic low-back pain for 3 months or more without
referred pain, presence of pain while lying in bed or on
Effect of firmness of mattress on chronic non-specific low-back
pain: randomised, double-blind, controlled, multicentre trial
Francisco M Kovacs, Víctor Abraira, Andrés Peña, José Gerardo Martín-Rodríguez, Manuel Sánchez-Vera, Enrique Ferrer,
Domingo Ruano, Pedro Guillén, Mario Gestoso, Alfonso Muriel, Javier Zamora, María Teresa Gil del Real, Nicole Mufraggi
Scientific Department, Kovacs Foundation, Palma de Mallorca,
Spain (F M Kovacs
PhD, M Gestoso MD, M T Gil del Real MPH,
N Mufraggi
MD); Unit of Clinical Biostatistics (V Abraira PhD,
A Muriel MSc, J Zamora
PhD), and Service of Physical Rehabilitation
(A Peña
MD), Hospital Ramón y Cajal, Madrid; Service of
Neurosurgery, Hospital de la Zarzuela, Madrid
(J G Martín-Rodríguez
MD); Service of Trauma, Clínica de la Luz,
Madrid (M Sánchez-Vera
MD); Service of Neurosurgery, Hospital
Clínic, University of Barcelona, Barcelona (E Ferrer
PhD);
Department of Anatomy, University of Barcelona, Barcelona
(D Ruano
PhD); and Service of Trauma, Clínica CEMTRO, Madrid
(P Guillén
PhD)
Correspondence to: Dr Francisco M Kovacs, Scientific Department,
Kovacs Foundation, Paseo Mallorca 36, E-07012 Palma de Mallorca,
Spain
(e-mail: kovacs@kovacs.org)
Articles
For personal use. Only reproduce with permission from The Lancet publishing Group.
rising, and voluntary agreement to particpate. Exclusion
criteria were: habitual prostration, signs of possible
systemic disease,
15
a diagnosis of inflammatory disease or
cancer, a diagnosis or clinical suspicion of fibromyalgia
(defined as pain spread throughout large muscle masses
with unjustified fatigue or non-restful sleep), pregnancy,
habitually sleeping in a different bed 2 or more nights per
week, taking anti-inflammatory medication with a 24 h
effect at any time of the day, and taking hypnotic,
analgesic, anti-inflammatory, or relaxant medication for
any reason from 1700 h to the time at which pain on rising
was assessed.
For participants who fulfilled the inclusion criteria and
who shared beds (eg, couples), only one person using the
bed was allowed to participate. Daytime medications for
low-back pain were not withdrawn. New mattresses were
installed for free but patients were not further
remunerated. We told participants that the study objective
was to assess the effect of the mattress on low-back pain,
but not that two different kinds of mattresses were going
to be compared. Participants were informed that they
could withdraw from the study at any time and that if the
new mattress made their pain worse we would supply a
new replacement of their choice free of charge. All
patients gave written informed consent to participate. The
study protocol was approved by the ethics committees of
the participating institutions.
Intervention
Randomisation was done in a central office, according to a
table of random permutations
16
before interventions were
assigned. One of the researchers (MG) randomly selected
the starting point for reading the table of permutations.
The staff of the central office, which was independent
from the research staff involved in recruiting the patients,
wrote correlative numbers on the front of opaque
envelopes. A numeric code from the table of permutations
was copied in the inside of the envelope (the number on
the front corresponded to the order of that numeric code
in the table). Envelopes were then sealed and
interventions were assigned to numbers in the table. Once
a patient had been included in the study, the research
assistant informed the person responsible for randomising
patients, who wrote the participant’s name on the
envelope showing on its front the number corresponding
to the order in which the patient had been included in the
study. The person in charge opened the envelope and
assigned the participant to one group or the other,
depending on the number shown inside the envelope.
After baseline assessments, the mattresses of all
participants were substituted for new spring mattresses of
the same size. The firmness of mattresses (H
s
) was rated
according to the European Committee for
Standardization scale.
17
The scale starts at 1·0 (firmest)
and stops at 10·0 (softest). The firm mattresses we used
were H
s
2·3 the medium-firm mattresses H
s
5·6. We
selected these mattresses because they cost similar
amounts (average price: 450 [US$522] for firm and
445 [US$516] for medium firm, respectively), and their
firmness represented the extreme and medium values of
those available in the market.
Mattresses were distinguishable only by fictitious names
that were unrelated to firmness and were similar to
commercially available models. The mattresses were
installed in the participants’ homes under identical
conditions by the same workers, who were unaware of
which type of mattresses they were installing. Existing
mattress support bases were substituted with a firm base if
the original base supported less than 50% of the mattress
surface (ie, wooden or plastic slats). Only the person who
did the randomisation knew which mattress had been
installed, but that person had no access to data obtained
throughout the trial.
We assessed patients at baseline and at 90 days. Each
patient was assessed at home with validated self-
assessment instruments and by a research assistant. At
baseline, patients self-assessed intensity of pain while lying
in bed, the intensity of pain on rising, and the degree of
disability. Pain was assessed with a visual analogue scale
(VAS), which ranges from zero (least) to ten (most
intense pain).
18
We asked patients to rate their low-back
pain as soon as they woke up (pain while lying in bed) and
within 30 min of rising (pain on rising). Disability was
assessed by a previously validated Spanish version of the
Roland Morris questionnaire,
19
consisting of 24 items
related to activities of daily living. Scores range from zero
(no disability) to 24 (maximum disability). We asked
patients to fill out the the VAS and Roland Morris
questionnaire with no-one else present.
At baseline assessment, the research assistants recorded
the variables that might affect the clinical course of low-
back pain and those related to exposure to the mattress:
age, sex, weight, height, socioeconomic status,
occupation, exposure to repeated flexion-extension
movements with load or whole-body vibration at work,
smoking habit, sports activity, duration of low-back pain,
low-back pain in bed, and low-back pain on rising, use
of medication for low-back pain (listing separately
analgesics, anti-inflammatory agents, muscle relaxants,
and other agents), most common sleeping position,
sharing of the bed, partner’s low-back pain in bed,
number of minutes in bed daily (weekdays and
weekends); age, length, width, and thickness of the
original mattress, type of support base; thickness of the
board placed over the support base if any; and subjective
feeling on firmness of the original mattress. Finally, the
research assistant fully explained to patients how to fill out
the scales for intensity of pain and disability.
At 90 days, every participant assessed their intensity of
low-back pain while lying in bed and on rising, and
disability. The research assistants asssessed the following
variables: characteristics of occupation, exposure to
repeated flexion-extension movements with load and to
whole-body vibrations at work, smoking, body position
when sleeping, sharing of the bed, subjective feeling on
firmness of the new mattress, whether their partner had
pain in bed, use of medication for low-back pain, daily
low-back pain, low-back pain while lying in bed and on
rising throughout the study (follow-up period), whether
he or she experienced more pain in bed when more time
spent in it.
Primary outcomes were the intensity of pain while lying
in bed and on rising, and the degree of disability.
Secondary outcomes included low-back pain, low-back
pain in bed or on rising throughout the study period, and
more intense pain in bed when lying down for an
extended time. We recorded complaints of pain while
lying in bed from the participants’ partners as a side-
effect.
Statistical analysis
We established the size of the study population at
125 patients per group, according to Lemeshow’s tables,
20
assuming a difference in the proportion of patients
improving in each group of at least 20% and that the
prevalence of improvement in one of the groups would be
50%. A type I error of 0·05 and a type II error of 0·10
were accepted. We increased the study population to
ARTICLES
1600 THE LANCET • Vol 362 • November 15, 2003 • www.thelancet.com
For personal use. Only reproduce with permission from The Lancet publishing Group.
313 participants to compensate for an anticipated 20%
loss of patients during follow-up.
Analyses were done by a team of statisticians who were
unaware of mattress assignment. The scores on the last
measurement (day 90) were subtracted from those
obtained at baseline; positive values indicated
improvement—improvement increasing with increasing
value—and negative values corresponded to worsening.
Analyses were done on SPSS (version 10.0) and PRESTA
(version 2.21)
21
for collinearity diagnosis. We present data
on continuous variables as medians and ranges, and crude
analyses were done with the Mann-Whitney U test
because distribution of data departed from normality. We
analysed categorical variables with the
2
test. We did
analyses by intention to treat and per protocol. To assume
the most conservative approach, we decided that for the
intention-to-treat analysis the poorest observed results
would be assigned to losses in the group showing the best
evolution, and vice versa.
22
We used multiple logistic regression models to assess
the association between the independent variable group
and improvement of pain in bed, pain on rising, and
disability, after adjustment for possible confounding
factors.
23
Improvement was defined as a positive change,
whatever the magnitude, between baseline and follow-up
assessment. Variables with imbalance between the groups
at baseline were included in the models, as well as those
that could exert confounding on the effect of the mattress
on low-back pain. Therefore, the three maximum models
included age, sex, height, and weight (combined as the fat
coefficient),
24
height-to-length bed ratio, characteristics of
job, exposure to repeated flexion-extension movements
with load, exposure to whole-body vibration at work,
sports (we classified sports as not relevant if they were
done once or less per week, and relevant if they were done
twice or more per week or at competition level), smoking
(yes or no), history of low-back pain (years), sharing of
bed (yes or no), age of original mattress (years), subjective
perception of the firmness of the new mattress, posture
while sleeping at baseline and at 90 days (correct=supine
with knees bent, lying on the side in the fetal position, or
three quarters—ie, any position with knees bent between
fetal and prone; incorrect=other positions), VAS in bed at
baseline, VAS on rising at baseline, Roland Morris
ARTICLES
THE LANCET • Vol 362 • November 15, 2003 • www.thelancet.com
1601
365 preselected
313 randomised
52 excluded
45 did not meet
inclusion criteria
7 refused to
participate
158 assigned
firm
mattresses
158 in intention-
to-treat
analysis
158 in per-protocol
analysis
155 in intention-
to-treat
analysis
152 in per-protocol
analysis
155 assigned
medium-firm
mattresses
3 lost to
follow up
Trial profile
Firmness of mattress
Firm Medium firm
(n=158) (n=155)
Characteristic
Sex
Male 42 (26·6%) 42 (27·1%)
Female 116 (73·4%) 113 (72·9%)
Median (range) age (years) 44·0 (18·078·0) 45·1 (19·082·0)
Median (range) weight (kg) 69 (45122) 65 (45104)
Median (range) height (cm) 164 (146190) 164 (145190)
Educational level
No studies or primary 69 (43·6%) 71 (45·8%)
education
Secondary education or 89 (56·3%) 84 (54·2%)
higher
Smoking habit
Never or ex-smoker 113 (71·5%) 112 (72·3%)
Current smoker 45 (28·5%) 43 (27·7%)
In employment 88 (55·7%) 91 (58·7%)
Characteristics of occupation
Sedentary or ambulatory without 106 (67·1%) 105 (67·7%)
strain
Ambulatory with strain or 52 (32·9%) 53 (32·2%)
non-ambulatory with strain
Exposures at work to
Flex-extension movements with load 42 (26·6%) 40 (25·8%)
Whole-body vibrations 8 (5·1%) 6 (3·9%)
Sport activity (>twice per week or 86 (54·4%) 77 (49·7%)
competition level)*
Median (range) time in bed (min) 480 (300771) 480 (300626)
Most common position during sleep
Supine knees bent 9 (5·7%) 6 (3·9%)
Supine knees straight 35 (22·2%) 41 (26·5%)
Prone 16 (10·1%) 10 (6·5%)
Fetal 86 (54·4%) 87 (56·1%)
Three-quarters 7 (4·4%) 4 (2·6%)
Other 5 (3·2%) 7 (4·5%)
Shared bed 111 (70·3%) 111 (71·6%)
Characteristics of mattress
Median (range) age (years) 7 (233) 8 (234)
Median (range) length (cm) 190 (180210) 182 (180200)
Median (range) width (cm) 135 (90180) 135 (80190)
Median (range) thickness (cm) 16 (825) 16 (722)
Median (range) thickness of board (cm) 2 (0·38) 2 (15)
Type of base
English mesh 16 (10·3%) 13 (8·6%)
Square-link mesh 11 (7·1%) 7 (4·6%)
Box spring 54 (34·6%) 64 (42·1%)
Board with >50% support 38 (24·4%) 30 (19·7%)
Firm base 35 (22·4%) 30 (19·7%)
Other 2 (1·3%) 8 (5·3%)
Subjective perception of the
firmness of mattress
Very soft 1 (0·6%) 5 (3·2%)
Soft 17 (10·8%) 21 (13·5%)
Neither soft nor firm 67 (42·7%) 72 (46·5%)
Firm 59 (37·6%) 51 (32·9%)
Very firm 13 (8·3%) 6 (3·9%)
Back pain of partner in bed 48 (43·2%) 40 (36·0%)
Median (range) duration of low-back 10 (142) 9 (053)
pain (years)
Median (range) duration of low-back 7 (142) 6 (053)
pain while lying in bed (years)
Median (range) duration of low-back 7 (142) 5·5 (053)
pain on rising (years)
Median (range) pain in bed on VAS 5 (010) 5 (09)
Median (range) pain on rising on VAS 7 (010) 7 (010)
Median (range) pain-related disability 9 (020) 8 (021)
on RMQ
Taking medication for low-back pain 40 (25·3%) 32 (20·6%)
RMQ=Roland Morris questionnaire. *Football, swimming, volleyball, judo,
basketball, athletics, sailing, tennis, gymnastics, aerobics, indoor football,
paddle-tennis, squash, and handball.
Table 1: Baseline characteristics
For personal use. Only reproduce with permission from The Lancet publishing Group.
questionnaire score at baseline, mean daily time in bed
([min]; working days=mean time for 5–7 days; days off
work=mean time for 2–7 days), type of base (firm=firm
base or boards supporting 50% of the mattress’ surface,
not firm=English mesh, square link mesh, box spring, or
boards supporting <50% of the mattress’ surface), and
medication for low-back pain (yes or no). For each
regression model, we took improvement as the dependent
variable. The collinearity of the maximum models was
assessed with the criteria proposed by Belsley.
25
We
deemed variables to be confounding if the estimate of the
coefficient of the variable group changed by more than
10% when that variable was removed from the maximum
model. A backward strategy was used.
Results
365 patients were selected for the study. 52 were excluded
for the following reasons: use of medication under the
conditions of the exclusion criteria (27), fibromyalgia
(eight), refusal to take part in the study (seven), presence
of a systemic inflammatory disease (six), sleeping away
from home more than 2 days per week (three), and
habitual prostration (one). Therefore, 313 patients
(84 men and 229 women) were included in the study
and randomised. 158 participants were assigned firm
mattresses and 155 mattresses of medium firmness
(figure). Three patients from the firm mattress group were
lost to follow-up because of a change in their place of
residence.
In the intention-to-treat analysis, baseline data of
patients in the two groups were similar (table 1). Most
patients reported a long history of low-back pain. Pain
intensity on VAS was graded as moderate to severe
(median [range] score while lying in bed: 5 [0–10] firm,
5 [0–9] medium-firm mattresses; and on rising, both
groups 7 [0–10]), and the degree of disability on
Roland Morris questionnaire was notable (9 [0–20] firm,
8 [0–21] medium-firm mattresses, table 1).
Mean time between randomisation and installation of
new mattresses was 10 days (SD 6·5). All participants had
bed bases supporting more than 50% of the mattress
surface and, therefore, no bed-base changes were made.
At day 90, participants in both groups had experienced
improvements compared with baseline in the intensity of
pain while lying in bed (mean intensity improvement 70%
firm and 80% medium-firm mattresses), intensity of pain
on rising (each 57%), and disability (30% and 50%).
Although there were differences in favour of medium-firm
mattresses in all variables at day 90, only those in the
degree of improvement of disability and pain on rising
throughout follow-up were significant (p=0·008, table 2).
Differences in the degree of improvement of pain on rising
(p=0·053), having had low-back pain throughout follow-
up (p=0·059), and pain while lying in bed throughout
follow-up (p=0·064) were close to significance (table 2).
The firmness of the new mattress was generally perceived
accurately by participants; 77·2% of patients in the firm
mattress group perceived their mattress to be firm or very
firm, compared with 43·5% in the medium-firm mattress
group who though their mattress was firm (p<0·0001,
table 2). When they entered the study, 63 of the 111
individuals who shared beds with participants in the firm
mattress group and 71 of the 111 who shared beds with
those from the medium-firm mattress group had pain in
bed (p=0·27). At the end of the study, 19 from the firm
mattress group and 14 from the medium-firm mattress
group had pain (p=0·35).
In most patients, the change of mattress was associated
with an improvement of pain while lying in bed (firm
mattress 77·8% vs 82·6% medium-firm mattress, odds
ratio 1·35 [95% CI 0·77–2·36], p=0·29) and on rising
(80·4% vs 85·8%, odds ratio 1·48 [0·81–6·68], p=0·201).
Pain-related disability improved in both groups, although
in a significantly higher proportion of patients in the
ARTICLES
1602 THE LANCET • Vol 362 • November 15, 2003 • www.thelancet.com
Firmness of mattress p
Firm mattress Medium firm
(n=158) (n=155)
Characteristic
Characteristics of occupation 0·399
Sedentary or ambulatory 108 (68·4%) 97 (63·8%)
without strain
Ambulatory with strain or 50 (31·6%) 55 (36·2%)
non-ambulatory with strain
Exposures at work
Flex-extension movements 38 (24·1%) 43 (28·6%) 0·396
with load
Whole-body vibrations 4 (2·5%) 4 (2·6%) 0·956
Most common position during 0·379
sleep throughout follow-up
Supine knees bent 4 (2·5%) 2 (1·3%)
Supine 41 (25·9%) 31 (20·5%)
Prone 12 (7·6%) 18 (5·3%)
Fetal 93 (58·9%) 104 (68·9%)
Three-quarters 6 (3·8%) 6 (4·0%)
Other 2 (1·3%) 0
Shared bed 109 (69·0%) 106 (69·7%) 0·866
Subjective perception of the <0·0001
firmness of mattress
Very soft 0 1 (0·6%)
Soft 0 5 (3·2%)
Neither soft nor firm 36 (22·8%) 81 (52·6%)
Firm 102 (64·5%) 64 (41·6%)
Very firm 20 (12·7%) 3 (1·9%)
Collapsed categories <0·0001
Very soft/soft/neither soft 36 (22·8%) 87 (56·4%)
nor firm
Firm/very firm 122 (77·2%) 67 (43·5%)
Back pain of the partner in bed 0·407
No 90 (82·6%) 91 (86·7%)
Yes 19 (17·4%) 14 (13·3%)
Back pain of the partner in bed* 0·192
No 26 (64·4%) 28 (77·8%)
Yes 16 (35·6%) 8 (22·2%)
Back pain of the partner in bed 0·386
No 57 (95·0%) 61 (91·0%)
Yes 3 (5·0%) 6 (9·0%)
Taking medication for
low-back pain 0·983
No 135 (85·4%) 130 (85·5%)
Yes 23 (14·6%) 22 (14·5%)
Had low-back pain throughout 0·059
follow-up
No 36 (22·8%) 48 (32·4%)
Yes 122 (77·2%) 100 (67·6%)
Had low-back pain in bed 0·064
throughout follow-up
No 44 (27·8%) 57 (37·7%)
Yes 114 (72·2%) 94 (62·3%)
Had low-back pain on rising 0·008
throughout follow-up
No 36 (22·8%) 55 (36·7%)
Yes 122 (77·2%) 95 (63·3%)
More intense pain in bed with 0·087
more time in it throughout
follow-up
No 73 (51·8%) 89 (61·8%)
Yes 68 (48·2%) 55 (38·2%)
Degree of improvement
Median (range) pain while 3·35 4·0 0·276
lying in bed on VAS (10 to 10) (5·0 to 9·20)
Median (range) pain on rising 4·0 (7 to 9) 4·0 (3 to 10) 0·053
on VAS
Median (range) disability
on RMQ 3·0 (10 to 19) 4·0 (14 to 19) 0·008
*Only partners who had pain while lying in bed on entering study. Only partners
free from pain while lying in bed on entering study.
Table 2: Results of assessment at 90 days
For personal use. Only reproduce with permission from The Lancet publishing Group.
medium-firm mattress than in the firm mattress group
(81·9% vs 68·3%, 2·10 [1·24–3.56], p=0·005). After
installing the new mattresses, worsening was observed in
some patients for pain in bed (firm mattresses 17·1%,
medium-firm mattresses 9·0%), pain on rising (firm
mattresses 12·7%, medium-firm mattresses 6·5%), and
disability (firm mattresses 24·1%, medium-firm
mattresses 9·0%). However, none of the patients
requested a change of the mattress during the course, or
after completion of, the study.
In the multivariate analysis, to avoid collinearity-
related difficulties, we eliminated the variable height-to-
length bed ratio, and centred the variables fat coefficient
and minutes spent in bed by subtracting their means.
23
Intensity of basal pain while lying in bed, pain on rising,
and perceived firmness of the new mattress were
confounding variables for intensity of pain in bed.
Intensity of basal pain on rising and perceived firmness of
the new mattress were confounding variables for
improvement of pain on rising. After adjustment for these
variables, the final model showed that patients who
received the medium firmness mattresses were around
twice as likely to improve than were patients with firm
mattresses for low-back pain while lying in bed, low-back
pain on rising, and disability (table 3).
For the per-protocol analysis, we did not include data
for the three patients lost to follow-up. This exclusion did
not change the direction of results, although it increased
the differences between the groups, and intensity of pain
on rising became significant in the crude and multivariate
analyses (data not shown).
Discussion
In patients with chronic low-back pain, mattress
conditions affect the degree of pain-related disability and
the intensity of pain while lying in bed and on rising. The
substitution of old mattresses with firm and medium-firm
new ones was associated with more frequent
discontinuation of drug treatment and relevant
improvements in pain and disability than other strategies
for chronic low back pain.
13
The design of this study does not permit us to quantify
the Hawthorne and placebo effects on the observed
improvement. However, the double-blind design of the
trial allows us to conclude that beyond such effects, the
use of a mattress of medium firmness improves the clinical
course of low-back pain in a higher proportion of patients
than the use of a firm mattress. In fact, although the mean
differences between groups for some quantitative variables
were small, all the differences favoured the use of a
mattress of medium firmness.
The underlying mechanisms explaining the results
of this trial are probably related to the duration of
exposure to the mattress, which represents roughly a
third of a person’s life, and to the effect of its firmness on
pressure distribution and muscular function when lying
in bed.
Side-effects were rare. The installation of a firm or
medium-firm mattress was associated with a reduction in
the number of partners with low-back pain while lying in
bed. Pain in bed, however, occurred in nine of the
134 partners who did not have it previously.
We focused on low-back pain while lying in bed and on
rising, and on the potential effect on back pain throughout
the day, which was explored by a simple question related
to its presence during the study period. Because of the low
sensitivity of this method, the differences in favour of the
medium-firm mattress were only close to significance. In
addition, there were clinical and significant differences in
favour of a medium-firm mattress in disability.
Accordingly, the beneficial effect of the characteristics of
the mattress seems to carry on beyond the time when the
patient is lying in bed or immediately after rising. This
possibility should be further explored in future trials.
Although patients were unaware of the type of mattress
they were receiving, they generally perceived correctly the
firmness of their mattress. The general belief that “harder
means better” reduced the positive effect of the medium-
firm mattress in the crude analysis, which in turn is
consistent with the importance of cognitive and
psychosocial factors in the course of chronic low-back
pain.
26–28
This finding, together with the degree of
improvement in the two groups, suggesting a contribution
from the Hawthorne and placebo effects, strongly shows
the need for using adequate methods (masked randomised
controlled trial) in studies aimed at investigating mattress-
related backache effects. Most previous controlled studies
on the effects of the mattress were not blinded adequately
for mattress characteristics,
4,6,9–11
which may have
undermined the validity of their results. One previous
study used a double-blind design, but the follow-up period
was only 1 night, and it focused on the quality of sleep and
not on the evolution of low-back pain.
5
The medium-firm mattress was associated with an
improvement in disability related to low-back pain. This
effect is particularly important since, although some
cognitive and psychosocial interventions slightly improve
the degree of disability, very few medical or physical
interventions have achieved this objective.
13,29,30
Several of
the items in the Roland Morris questionnaire might be
affected by stiffness in bed or on rising, which might
account for such a result. The results of this study also
suggests that, although psychosocial factors have an effect
on disability,
26–28
some biomechanical factors that were not
previously considered may also have an effect and should
be taken into consideration for future studies.
The external validity of our results might be limited by
several factors. This study was done in patients with
chronic low-back pain and no referred pain, and the effect
of the firmness of the mattress on referred pain as well as
on an acute exacerbation of backache is unknown.
Psychosocial factors affect the degree of disability
associated with low-back pain and might be specific to
each setting, which might alter slightly the generalisability
of findings on this variable. Finally, we used mattresses of
two different firmnesses but they were spring mattresses.
Although the firmness scale is independent of the
composition of the mattress, the results might differ with
other kinds of mattresses. For ethical reasons, we offered
patients any mattress for free after 90 days in case their
low-back pain had worsened with the one installed during
the study. This action limited follow-up to that period,
and further studies should assess long-term effects of the
mattress.
Our findings stress that recommendations for daily
living, such as what kind of mattress to use, may have a
ARTICLES
THE LANCET • Vol 362 • November 15, 2003 • www.thelancet.com
1603
Odds ratio (95% CI) p
Variable
Improvement of pain while lying in bed on VAS* 2·36 (1·134·93) 0·023
Improvement of pain on rising (VAS) 1·93 (0·973·86) 0·061
Improvement of pain-related disability (RMQ) 2·10 (1·243·56) 0·006
RMQ=Roland Morris questionnaire. *Adjusted by perceived firmness of new
mattress, pain on rising at baseline, pain while lying in bed at baseline.
Adjusted by perceived firmness of new mattress and VAS on rising at
baseline.
Table 3: Results of multiple logistic regression model for
intention-to-treat analysis
For personal use. Only reproduce with permission from The Lancet publishing Group.
relevant effect on the clinical course of low-back pain. The
effects should be assessed with sound methods similar to
those used for other medical treatments.
Contributors
F M Kovacs and M Gestoso designed the study, recruited patients,
coordinated study execution, wrote the report, scientifically reviewed the
paper, and approved the final draft. V Abraira reviewed the study design,
planned and did the statistical analysis, reviewed the scientific content,
and approved the final draft. A Muriel and J Zamora did the statistical
analysis and approved the final draft. A Peña, J G Martín-Rodríguez,
Manuel Sánchez-Vera, E Ferrer, D Ruano, P Guillén, M T Gil del Real,
and N Mufraggi reviewed the design of the study, supervised coordination
of the study, reviewed the report drafts, and approved the final version.
Conflict of interest statement
None declared.
Acknowledgments
This study was done and financed by the Kovacs Foundation, a non-
profit Spanish research institution with its own funding resources that
promotes evidence-based health care in clinical practice. FLEX, a Spanish
bedding company, provided and installed the mattresses used in the trial
without charge; nobody from that company participated in study design or
in the collection, analysis, and interpretation of data. We thank Marta
Pulido for editing the report and for editorial assistance.
References
1 Deyo RA, Weinstein JN. Low back pain. N Engl J Med 2001; 344:
363–70.
2 Pope MH, Wilder DG, Krag MH. Biomechanics of the lumbar spine:
basic principles. In: Frymoyer JW, ed. The adult spine: principles and
practice. New York, NY: Raven Press, 1991.
3 Lings S, Leboeuf-Y de C. Whole body vibration and low back pain: a
systematic, critical review of the epidemiological literature 1992–1999.
Int Arch Occup Environ Health 2000; 73: 290–97.
4 Koul PA, Bhat MH, Lone AA, Koul AN, Wahid A. The foam
mattress-back syndrome. J Assoc Physicians India 2000; 48: 901–02.
5 Enck P, Walten T, Traue JC. Associations between back pain, quality
of sleep and quality of mattresses: double-blind pilot study with hotel
guests. Schmerz 1999; 13: 205–07.
6 Scriver V, Crowe J, Wilkinson A, Meadowcroft C. A randomized
controlled trial of the effectiveness of exercise and/or alternating air
mattress in the control of back pain after percutaneous transluminal
coronary angioplasty. Heart Lung 1994; 23: 308–16.
7 Kovacs FM, Gestoso M, Gil del Real MT, López J, Mufraggi N,
Méndez JI. Risk factors for nonspecific low back pain in
schoolchildren and their parents: a population-based study. Pain 2003;
103: 159–268.
8 Levy H, Hutton WC. Mattresses and sleep for patients with low back
pain: a survey of orthopaedic surgeons. J South Orthop Assoc 1996; 5:
185–87.
9 Garfin SR, Pye SA. Bed design and effect on chronic low back pain.
a limited controlled trial. Pain 1981; 10: 87–91.
10 Hagino C, Erfanian P. Before-after study to determine the
effectiveness of an adjustable wood frame-foam and wool mattress bed
system (The Natura Mattress System) in reducing chronic back pain
in adults. J Can Chiropract Assoc 1997; 41: 16–26.
11 Monsein M, Corbin TP, Culliton PD, Merz D, Shuck EA. Short-term
outcomes of chronic back pain patients on an airbed vs innerspring
mattresses. MedGenMed 2000; 2: E36.
12 Jacobson BH, Gemmel HA, Hayes BM, Altena TS. Effectiveness
of a selected bedding system on quality of sleep, low back pain,
shoulder pain, and spine stiffness. J Manipulative Physiol Ther 2002;
25: 88–92.
13 Van Tulder MW, Koes BW, Bouter LM. Conservative treatment of
chronic low back pain: a meta-analysis of 80 randomized trials of 14
interventions. In: Van Tulder MW, Koes BW, Bouter LM, eds. Low
back pain in primary care. Amsterdam: Institute for Research in
Extramural Medicine, 1996.
14 Andersson GBJ. The epidemiology of spinal disorders. In:
Frymoyer JW, ed. The adult spine: principles and practice, 2nd edn.
New York, NY: Raven Press, 1997.
15 Agency for Health Care Policy and Research. Clinical practice
guideline number 14: acute low back pain problems in adults.
Rockville, MD: US Department of Health and Human Services, 1994.
16 Moses LE, Oakford RY. Tables of random permutations. Standford,
CA: Standford University Press, 1963.
17 European Committee for Standarization. Domestic Furniture. Beds
and mattresses: test method for the determination of functional
characteristics. CEN/TC207/WG1/TG4. Brussels: Comité Européen
de Normalisation, 1996.
18 Huskisson EC. Measurement of pain. Lancet 1974; 2: 1127–31.
19 Kovacs FM, Llobera J, Gil del Real MT, et al. Validation of the
Spanish version of the Roland-Morris questionnaire. Spine 2002; 27:
538–42.
20 Lemeshow S, Hosmer DW, Klar J, Lwanga SK. Adequacy of sample
size in health studies. Chichester, UK: John Wiley, 1990.
21 Abraira V, Zaplana J. PRESTA, un paquete de procesamientos
estadísticos. Proceedings de la Conferencia Iberoamericana de
Bioingeniería. 100. Asturias, Spain: Gijón, 1984.
22 Pocock SJ. Clinical trials: a practical approach. New York: John Wiley,
1983.
23 Kleinbaum DG, Kupper LL, Muller KE. Applied regression analysis
and other multivariable methods, 2nd edn. Boston, MA: PWS-KENT,
1988.
24 Pronnett S. Le marathon, equilibre energetique, endurance et
alimentation du coureur sur route. Paris: Vigot, 1983.
25 Belsley DA. Conditioning diagnostics: collinearity and weak data in
regression. New York, NY: John Wiley, 1991.
26 Schmidt AJ. Cognitive factors in the performance level of chronic low
back pain patients. J Psychosom Res 1985; 29: 183–87.
27 Sandstrom J, Esbjornsson E. Return to work after rehabilitation: the
significance of the patient’s own prediction. Scand J Rehabil Med 1986;
18: 29–33.
28 Waddell G, Somerville D, Henderson I, Newton M, Main CJ. A fear
avoidance beliefs questionnaire (FABQ) and the role of fear avoidance
beliefs in chronic low back pain and disability. Pain 1993; 52: 157–68.
29 Kovacs FM, Abraira V, Pozo F, et al. Local and remote sustained
trigger point therapy for exacerbations of chronic low back pain: a
randomized, double-blind, controlled multicenter trial. Spine 1997:
22: 786–96.
30 Kovacs FM, Llobera J, Abraira V, et al. Effectiveness and cost-
effectiveness analysis of neuroreflexotherapy for subacute and chronic
low back pain in routine general practice: a cluster randomized,
controlled trial. Spine 2002; 27: 1149–59.
ARTICLES
1604 THE LANCET • Vol 362 • November 15, 2003 • www.thelancet.com
    • "This was not an RCT, compared sleeping at work to sleeping at home, with no control for the effect of shift pattern, and did not specify and control the amount of hours each resident slept. The only high quality RCT in this area (Kovacs et al. 2003a) enrolled 313 adults from the community who reported persistent non-specific low-back pain. Using double blinding, participants were randomly assigned to use mattresses of two different levels of firmness (medium-firm and hard). "
    [Show abstract] [Hide abstract] ABSTRACT: Introduction Although back pain (BP) is a very common cause for sickness absence, most people stay at work during BP episodes. Existing knowledge on the factors influencing the decision to stay at work despite pain is limited. The aim of this study was to explore challenges for coping with BP at work and decisive factors for work attendance among workers with high physical work demands. Methods Three focus groups (n = 20) were conducted using an explorative inductive method. Participants were public-employed manual workers with high physical work demands. All had personal BP experience. Thematic analysis was used for interpretation. Results were matched with the Flags system framework to guide future recommendations. Results Workers with BP were challenged by poor physical work conditions and a lack of supervisor support/trust (i.e. lack of adjustment latitude). Organization of workers into teams created close co-worker relationships, which positively affected BP coping. Workers responded to BP by applying helpful individual adjustments to reduce or prevent pain. Traditional ergonomics was considered inconvenient, but nonetheless ideal. When pain was not decisive, the decision to call in sick was mainly governed by workplace factors (i.e. sick absence policies, job strain, and close co-workers relationships) and to a less degree by personal factors. Conclusion Factors influencing BP coping at work and the decision to report sick was mainly governed by factors concerning general working conditions. Creating a flexible and inclusive working environment guided by the senior management and overall work environment regulations seems favourable.
    Full-text · Article · Mar 2015
    • "In addition to its effect on sleep, bed comfort is often linked to low back pain, a leading cause of temporary disability and sick leave (Ehrlich, 2003). Although the causes of non-specific back complaints are not well understood, it has been shown that low quality bed systems may contribute significantly to the onset or persistence of low back pain (Kovacs et al., 2003). Consequently, it comes to no surprise that ergonomic aspects of bed design have evolved substantially during the last decade. "
    [Show abstract] [Hide abstract] ABSTRACT: This study aims at evaluating spinal alignment during sleep by combining personalized human models with mattress indentation measurements. A generic surface model has been developed that can be personalized based on anthropometric parameters derived from silhouette extraction. Shape assessment of the personalized surface models, performed by comparison with 3-D surface scans of the trunk, showed a mean unsigned distance of 9.77 mm between modeled and scanned surface meshes. The surface model is combined with an inner skeleton model, allowing the model to simulate distinct sleep postures. An automatic fitting algorithm sets the appropriate degrees of freedom to position the model on the measured indentation according to the adopted sleep posture. Validation on lateral sleep positions showed good intraclass correlations (0.73–0.88) between estimated and measured angular spinal deformations, indicating that silhouette-derived body shape models provide a valuable tool for the unobtrusive assessment of spinal alignment during sleep.Relevance to industry: A common drawback of the available techniques to assess spinal deformation on bedding systems is that they interfere with the actual sleep process. The current study presents a novel method based on silhouette-derived body shape models in order to estimate spine shape during sleep unobtrusively.
    Full-text · Article · Sep 2012
    • "Thereagainst , mattresses with high stiffness lead to shoulder pain, and cause the decrease in sleep quality and improper distribution of body loads131415. Kovacs et al. observed that patients with low back pain feel less comfortable on a mattress with high stiffness than on a mattress with average stiffness [16]. Jacobsen et al. showed that using an individualized sleep system, in comparison with the sleep system which an individual usually uses, causes the shoulder and back pain of the individual to decrease and his sleep quality to increase [13]. "
    [Show abstract] [Hide abstract] ABSTRACT: A proper sleep system can affect the spine support in neutral position. Most of the previous studies in scientific literature have focused on the effects of customary mattresses on the spinal alignment. To keep the spine in optimal alignment, one can use sleep surfaces with different zonal elasticity, the so called custom-made arrangements. The required stiffness of a sleep surface for each individual can be obtained by changing this arrangement applying the experimental method and modeling. In experimental part, the coordinate positions of the markers mounted on the spinous processes of the vertebrae of 25 male volunteers were registered in frontal plane through the optical tracking method and so the spinal alignment was obtained in lateral sleep position on soft and firm surfaces and on the best custom-made arrangement. Thereupon the π-P₈ angles were extracted from these alignments and then were compared with each other. In modeling part the anthropometric data of four different types of volunteers were used. And then the models built in BRG.LifeMOD (ver. 2007, Biomechanics Research Group, Inc., USA) based on these data and in accordance with the experimental tests, were analyzed. The one way ANOVA statistical model and the post hoc tests showed a significant difference in the π-P₈ angles between soft & custom-made and soft & firm mattresses at the p = 0.001 level and between firm & soft mattresses at the p = 0.05 level. In modeling part, the required stiffness of the sleep surface for four weight-dimensional groups was acquired quantitatively. The mattress with a custom-made arrangement is a more appropriate choice for heavier men with pronounced body contour. After data fitting, it was observed that the variations of spinal alignment obtained from both methods have the same trend. Observing the amount of required stiffness obtained for the sleep surface, can have a significant effect on keeping the spine healthy.
    Full-text · Article · Nov 2011
Show more