The Effect of Short-Term Ventilation Tubes versus
Watchful Waiting on Hearing in Young Children
with Persistent Otitis Media with Effusion: A
Randomized Trial
Maroeska M. Rovers, Huub Straatman, Koen Ingels, Gert-Jan van der Wilt,
Paul van den Broek, and Gerhard A. Zielhuis
Objective: To study the effect of short-term ventilation tubes in children aged 1 to 2 yr with screeningdetected, bilateral otitis media with effusion (OME)
persisting for 4 to 6 mo, as compared with watchful
waiting.
signs and symptoms of an acute infection (Bluestone, 1999; Senturia, Bluestone, Klein, Lim, &
Paradise, 1980). The adverse effects of OME on
cognitive and linguistic development are generally
considered to be the result of conductive hearing loss
associated with OME (Vernon-Feagans, 1999). Persistent hearing loss of 25 to 30 dB HL for more than
3 mo has been suggested as an indication for treatment (Freemantle et al., 1992). The most common
treatment for OME is still the insertion of ventilation tubes, introduced by Armstrong in 1954 (Armstrong, 1954).
Studies on the effectiveness of ventilation tubes
for persistent OME showed that short-term improvement in hearing can be attained (Black, Sanderson, Freeland, & Vessey, 1990; Mandel, Rockette,
Bluestone, Paradise, & Nozza, 1989; Maw & Herod,
1986). The average hearing levels 6 mo after the
insertion of ventilation tubes were in the order of 15
to 20 dB HL, with a mean improvement of 5 to 15 dB
compared with baseline (pretreatment) measurement. After 12 mo or longer, however, hearing levels
in the treated groups equalled those in nontreated
control groups, probably due to spontaneous recovery in the controls and recurrent OME in the treated
children (Black et al., 1990; Dempster, Browning, &
Gatehouse, 1993; Maw & Herod, 1986). Only the
study by Gates (Gates, Avery, Prihoda, & Cooper,
1987) provides information on the mean time spent
with effusion: 35% and 49% of the time during a
follow-up of 12 mo, in the treated and the control
groups, respectively.
It should be noted that there was wide variation
between the previous studies regarding design and
population characteristics, such as the mean duration of OME before randomization, inclusion and
exclusion criteria, treatment protocol, types of referent treatment, and outcome measures (other than
the mean hearing level). Moreover, all these studies
were performed on children between 4 to 8 yr of age
(Black et al., 1990; Dempster et al., 1993; Gates et
al., 1987; Maw & Herod, 1986). As OME tends to
occur more frequently in younger children, many of
Design: Multi-center randomized controlled trial (N
ⴝ 187) with two treatment arms: short-term ventilation tubes versus watchful waiting. Young children underwent auditory screening; those with persistent (4 to 6 mo) bilateral OME were recruited.
Results: The mean duration of effusion over 1-yr
follow-up was 142 days (36%) in the ventilation tube
(VT) group versus 277 days (70%) in the watchful
waiting (WW) group. After 6 mo of follow-up, the
pure-tone average in the VT group was 5.6 dB A
better than that in the WW group. After 12 mo, most
of the advantage in the VT group had disappeared.
After the insertion of ventilation tubes, the children
with poorer hearing levels at randomization improved more than the children with better hearing
levels. The largest difference in hearing levels was
found between the children in the VT group whose
ventilation tubes remained in situ and the children
in the WW group. In the VT children with recurrence of OME, the hearing levels again increased,
but remained slightly lower than those in the infants with persistent OME in the WW group.
Conclusions: Ventilation tubes have a beneficial
effect on hearing in the short run (6 mo); this effect,
however, largely disappears in the long run (12 mo).
This is probably due to partial recurrent OME in
the VT group and to partial spontaneous recovery
in the WW group.
(Ear & Hearing 2001;22;191–199)
Otitis media with effusion (OME) is one of the
most common diseases in childhood, and refers to an
accumulation of fluid in the middle ear cavity behind an intact tympanic membrane without the
Departments of Otorhinolaryngology (M.M.R., K.I., P.v.d.B.),
Epidemiology (H.S., G.A.Z.), and Medical Technology Assessment
(G.-J.v.d.W.), University Medical Centre, Nijmegen, The
Netherlands.
0196/0202/01/2203-0191/0 • Ear & Hearing • Copyright © 2001 by Lippincott Williams & Wilkins • Printed in the U.S.A.
191
192
EAR & HEARING / JUNE 2001
these children will receive ventilation tubes (Engel,
Anteunis, & Hendriks, 1999; Stephenson & Haggard, 1992). An important rationale for treatment is
to reduce the time with bilateral hearing loss at an
age that is critical for speech-language development.
This presents the need to study the effect of ventilation tubes on hearing levels in infants of 1 to 2 yr
of age, and to find out whether the conclusions
drawn from studies on older children are valid in
younger children.
We performed a multi-center randomized controlled trial in which infants with persistent OME
were assigned to either early insertion of short-term
ventilation tubes or a watchful waiting period. Data
were obtained on aspects such as hearing sensitivity, language development and quality of life. The
latter two aspects are described in separate papers
(Rovers, Krabbe, Straatman, Ingels, van der Wilt, &
Zielhuis, 2001; Rovers, Straatman, Ingels, van der
Wilt, van den Broek, & Zielhuis, 2000).
In addition to the intention to treat analyses
normally performed in randomized controlled trials,
we examined the hearing levels over four different
periods in these randomized children, namely 1)
before assignment to the ventilation tube (VT) or
watchful waiting (WW) group; 2) during the period
that the tubes were in situ and functioning; 3)
during the period that the tubes were extruded and
the child remained OME free; and 4) during the
period that a child in the VT group was diagnosed
with recurrent OME. The research questions related
to these periods were: 1) How large was the hearing
deficit due to OME? 2) Did hearing levels return to
normal after treatment with ventilation tubes? and
3) Did hearing levels remain normal or at least
better than those in the WW group after the tubes
had been extruded?
PATIENTS
AND
METHODS
Patients
The trial population was embedded in a cohort,
which included 30,099 children born in the Eastern
part of the Netherlands between January 1, 1996,
and April 1, 1997. All of these children were invited
for a routine behavioral hearing screening at the age
of 9 mo; 2457 (9%) of them did not respond to this
invitation, and another 1474 (5%) children were lost
to follow-up during the three subsequent screening
tests. The test was performed in a quiet environment in which the parent sits on a chair with the
child on his or her lap. The auditory stimuli were
presented behind the parent and child, on the right
and left, exactly at one meter distance and at an
angle of 45°. The frequencies of the sounds presented were between 250 and 8000 Hz, and the
intensity of the sound was 35 dB SPL. Each sound
was presented for 5 sec. If the child did not react to
one of the sounds, the sound was presented again in
another standardized way. The screening test was
passed if a child did react correctly to all four sounds
presented to the right and the left ear. If a child
failed the first screening test, he or she was rescreened 1 mo later. If the child also failed this
second test, he or she was recalled for a more
comprehensive test 1 mo later. During this third test
sounds were presented louder than previously. The
aim of this was to measure the extent of the hearing
loss, although no exact hearing thresholds were
obtained. The children were referred when there
was no response to sounds at 35 dB (Rovers, Zielhuis, Straatman, Ingels, van der Wilt, & Kauffman-de Boer, 1999). For the purpose of the study,
those who failed three times were referred to one of
the 13 participating ear, nose, and throat (ENT)
outpatient clinics in the region for diagnosis and
follow-up (N ⫽ 1081). The parents of infants diagnosed with persistent (4 to 6 mo) bilateral OME
(confirmed by tympanometry and otoscopy) in subsequent visits to the ENT surgeon (N ⫽ 386) were
asked whether they would give consent for their
child to participate in the randomized controlled
trial. Children with Down syndrome, sensorineural
hearing loss, cystic fibrosis, asthma and cleft palate
were excluded. The children for whom informed
consent was obtained (N ⫽ 187) were randomly
allocated to one of two groups: either treatment with
ventilation tubes (N ⫽ 93) (Bevel Bobbins® Entermed BV, Woerden, The Netherlands), or watchful
waiting (N ⫽ 94). The two groups were followed for 1 yr
with three monthly tympanometry and otoscopy measurements, and audiometry every 6 mo.
Information on prognostic factors, such as common colds since birth, attending daycare, etc., was
obtained using a questionnaire that was completed
by the parents during the first visit to the ENT clinic
(Rovers, Zielhuis, Straatman, Ingels, van der Wilt,
& van den Broek, 1999). Information on other clinical symptoms, such as adenoidectomy before randomization was obtained by means of a questionnaire that was completed by the ENT surgeon.
We obtained approval for this study from the
Ethical Committees at the 13 participating hospitals.
Methods
To increase comparability at baseline, a balanced
allocation procedure was employed with five balancing factors: sex, age, season at randomization, educational level of the mother, and hospital.
During the trial, tympanometry and audiometry
were performed by experienced audiologists (who
were not blinded to the assignment of a child);
EAR & HEARING, VOL. 22 NO. 3
otoscopy was performed by the ENT surgeon. Tympanometry typically was performed before audiometry, but if a child appeared to be anxious, audiometry was performed first to ease the child. The
tympanograms were classified according to Jerger
(1970), and OME was defined according to the
MOMES protocol (Engel, Anteunis, Volovics, Hendriks, & Marres, 1999); also see Figure 1. To study
the percentage children with effusion, we only differed between the presence/absence of bilateral effusion and not between tubes in place and functioning versus occluded tubes. Hearing assessment was
performed in standard rooms in the 13 participating
hospitals. If no special audiometric test booth was
available, a quiet and convenient room was selected.
A portable visual reinforcement audiometry set was
used, calibrated according to the substitution
method in dB A (Rovers, Snik, Ingels, van der Wilt,
& Zielhuis, Reference Note 1) as advocated by Beynon and Munro (1993). The noise floor in the test
rooms was assessed by measuring sound field hearing thresholds in adults with normal hearing
(thresholds of between ⫺5 and ⫹10 dB HL). Averaged over the four frequencies (500, 1000, 2000, and
4000 Hz), the noise level varied per test room from
8.8 dB A to 19.1 dB A, with an average of 13.0 dB A
(Rovers et al., Reference Note 1). One should bear in
mind that a visual reinforcement audiometry set in
fact measures a minimal response level instead of a
real mean hearing level.
To study the (natural) course of OME during the
trial we examined the mean time spent with OME as
well as the mean duration that the children had
hearing thresholds of poorer or equal to 35 dB A. If
a child had OME (hearing thresholds poorer or equal
to 35 dB A) on two successive occasions, the days
between these measurements were counted as days
with effusion (poor hearing thresholds). If on the
193
other hand a child had OME on the first occasion but
not on the second, only half of the days were counted
as days with effusion.
To study the efficacy of the ventilation tubes, we
distinguished three different situations: 1) ventilation tubes in situ and functioning; 2) extruded ventilation tubes and no bilateral OME; and 3) extruded
ventilation tubes but bilateral effusion. In these
analyses we distinguished between tubes in place
and functioning, occluded tubes and extruded tubes,
whereas the other analyses were based on the intention to treat principle. Functioning and nonfunctioning tubes were distinguished on the basis of tympanometry: if a child had ventilation tubes inserted the
ear canal, volume had to be larger than 1.5 cm3. The
difference between occluded and extruded tubes had
to be based on otoscopy, because this was the only
way to determine whether the tube was occluded or
extruded. Children diagnosed with otorrhea (diagnosed
via otoscopy) were excluded from these analyses.
To compare the extrusion rate of the ventilation
tubes with those reported in the literature, we also
studied the percentage of tubes that were in situ at
3, 6, 9, and 12 mo of follow-up.
Statistical Analysis
All the analyses were performed according to the
intention to treat principle. Differences in mean
hearing thresholds in the better ear (at 500, 1000,
2000, and 4000 Hz) at 0, 6, and 12 mo of follow-up
between the VT group and WW group were tested
with the Student t-test for independent groups. The
choice of the better ear was motivated by the etiological model of this trial: to study the effect of
treatment with ventilation tubes on developmental
outcomes mediated by the conductive hearing deficit
in the better ear. However, as the consequences of
Figure 1. Algorithm to diagnose otitis media with effusion (MOMES protocol). Type A: compliance > 0.2 mL, pressure > ⴚ100
daPa; Type B: compliance < 0.2 mL, pressure < 200 daPa; Type C1: compliance > 0.2 mL, ⴚ100 daPa < pressure < ⴚ200 daPa;
Type C2: compliance > 0.2 mL, pressure < ⴚ200 daPa.
194
EAR & HEARING / JUNE 2001
early onset hearing loss may have to do with loss of
binaural hearing ability, we also studied differences
between left and right ears or between better and
worse ears.
To adjust for potential confounders and to explore
possible effect modifiers with treatment, we performed regression analyses. Our a priori model was:
M(hearing level at T12 ⫺ hearing level at T0) ⫽ ␤0
⫹ ␤i ⫻ treatment ⫹ ␤j ⫻ confounders
⫹ ␤k ⫻ treatment ⫻ effect modifiers ⫹ ␧i
Potential confounders and/or effect modifiers
were age at randomization (⬍20 mo/ⱖ20 mo), adenoidectomy before randomization (yes/no), hospital
(1 to 13), season at randomization (1 to 4), number of
upper respiratory tract infections since birth (⬎4/
ⱕ4), and hearing loss at baseline (dB A).
We also designed a logistic regression model with
the same parameters to analyze the effect of ventilation tubes on the probability that children would
improve by more than 15 dB. To study the modifying
effect of treatment for specific clinical complaints
such as upper respiratory tract problems, clusters of
children were formed by means of a principal component analysis with two principal components using the following information from the medical history: 1) adenoidectomy; 2) number of common colds;
3) persistence of OME before randomization; and 4)
periods of earache and/or fever indicating acute
otitis media. Three clusters were formed: 1) children
without complaints (N ⫽ 67); 2) children with some
complaints (N ⫽ 57); and 3) children with frequent
complaints (N ⫽ 55), see also Table 1.
RESULTS
The Children in the Trial
A total of the 187 children were randomized: 93
children entered the ventilation tube (VT) group and
94 children entered the WW group. During the trial
period, 11 children dropped out (three children from
the VT group and eight from the WW group). In
addition, 10 children in the WW group received
treatment with ventilation tubes. Eight children in
the VT group underwent surgery for ventilation
tube insertion twice, because the ventilation tubes
were extruded before the 6 mo follow-up. The number of alternative/additional treatments were either
equally distributed (adenoidectomy) over both
groups, or the children in the VT group received
slightly more (antibiotics, nose drops) as compared
with the WW group.
The mean age of the children at randomization
was 19.5 mo (SE ⫽ 1.7) in the VT group and 19.4 mo
(SE ⫽ 1.9) in the WW group.
Patient characteristics and mean hearing levels
during follow-up are shown in Table 2.
Tympanometry and Otoscopy
At 3-mo follow-up, 13 (15%) of the children in the
VT group were diagnosed as having bilateral OME.
At 6, 9, and 12 mo follow-up, 29, 27, and 27% of the
children in the VT group were diagnosed as having
bilateral OME, respectively. In the WW group the
percentages were 77, 66, 57, and 53 at 3, 6, 9, and 12
mo follow-up, respectively; see also Figure 2.
In the WW group, 25 (27%) children were diagnosed as having bilateral OME at all visits, and 10
(11%) were diagnosed as having bilateral OME only
TABLE 1. Description of the three clusters formed by means of principal component analysis (explained variation ⴝ 57%).
Cluster 1
(No Complaints)
Treatment group
Ventilation tube group
Watchful waiting group
Sex
Male
Female
Adenoidectomy
Yes
No
Fever
Yes
No
Earache
Yes
No
Mean number of visits with common colds
Hearing level at baseline
Cluster 2
(Some Complaints)
Cluster 3
(Frequent Complaints)
31 (46.3%)
36 (53.7%)
31 (54.4%)
26 (45.6%)
28 (50.9%)
27 (49.1%)
36 (53.7%)
31 (46.3%)
33 (57.9%)
24 (42.1%)
35 (63.6%)
20 (36.4%)
0 (0%)
67 (100%)
5 (8.8%)
52 (91.2%)
11 (20.0%)
44 (80.0%)
0 (0%)
67 (100%)
8 (14.0%)
49 (86.0%)
49 (89.1%)
6 (10.9%)
0 (0%)
67 (100%)
0.6
44.4
31 (54.4%)
26 (45.6%)
0.7
44.9
55 (100%)
0 (0%)
0.7
45.4
EAR & HEARING, VOL. 22 NO. 3
195
TABLE 2. Patient characteristics and mean hearing levels in the better (worse) ear during follow-up.
Treatment group
Ventilation tube group
Watchful waiting group
Sex
Male
Female
Educational level mother
High
Middle
Low
Season at randomization
Spring
Summer
Autumn
Winter
Age at randomization
⬍20 mo
ⱖ20 mo
District
1
2
3
4
Number
Ventilation
Tubes
Watchful
Waiting
Mean (dB) at
T⫽0
Mean (dB) at
T⫽6
Mean (dB) at
T ⫽ 12
⌬6 (dB)
⌬12 (dB)
93
94
93
—
—
94
46.4 (50.1)
43.4 (47.0)
35.8 (39.8)
38.7 (43.6)
33.2 (37.2)
34.7 (38.4)
10.2 (10.1)
4.6 (3.2)
13.1 (12.9)
8.5 (8.3)
110
77
55
38
55
39
43.8 (47.7)
46.2 (49.8)
37.1 (41.4)
37.5 (42.0)
34.2 (37.9)
33.6 (37.6)
6.7 (6.2)
8.4 (7.3)
9.5 (9.6)
12.8 (12.2)
46
92
49
23
45
25
23
47
24
43.5 (47.6)
46.8 (50.2)
42.7 (46.5)
38.8 (43.9)
37.1 (41.1)
36.2 (40.5)
35.3 (39.9)
33.6 (37.3)
33.0 (36.5)
5.7 (4.5)
9.1 (8.4)
5.8 (5.2)
8.7 (8.1)
13.0 (12.7)
8.9 (9.1)
43
47
59
38
20
23
31
19
23
24
28
19
45.9 (49.5)
41.4 (44.6)
45.2 (48.7)
47.4 (51.8)
38.2 (42.5)
37.6 (42.3)
37.6 (42.1)
35.3 (39.3)
31.7 (35.2)
35.2 (38.7)
34.4 (38.4)
34.2 (38.5)
8.3 (7.7)
2.5 (0.9)
7.2 (6.1)
12.3 (12.7)
14.7 (14.8)
5.2 (4.9)
10.9 (10.3)
13.0 (13.0)
145
42
72
21
73
21
44.2 (47.8)
47.2 (51.0)
37.6 (41.7)
36.2 (41.3)
34.5 (38.7)
31.7 (34.6)
6.4 (5.8)
11.1 (9.7)
9.6 (9.6)
15.4 (16.5)
11
125
25
26
7
62
13
11
4
63
12
15
45.5 (49.7)
43.3 (46.8)
47.4 (51.9)
50.2 (53.4)
38.6 (41.9)
37.2 (41.3)
34.2 (37.9)
40.1 (46.0)
32.8 (37.9)
33.9 (37.6)
34.4 (37.7)
34.1 (38.9)
6.8 (2.7)
6.1 (5.2)
12.9 (13.0)
9.2 (6.5)
14.1 (13.5)
9.6 (9.4)
12.0 (13.7)
15.4 (13.7)
The underlined difference scores were statistically significant (p ⬍ 0.05).
once. In the VT group, three children (3%) were
diagnosed as having bilateral OME at all visits, and
44 (47%) of the children were diagnosed as having
bilateral OME only at randomization.
The estimated mean number of days with bilateral OME in the VT group was 142 days (36%)
versus 277 days (70%) in the WW group.
Audiometry
Figure 3 shows the changes in mean hearing loss
in the better ear and worse ear. The mean hearing
levels improved in the two groups, but the VT group
improved more, especially during the first 6 mo. The
improvement at 6 mo follow-up in the VT group in
the better ear was 10.2 dB (95% confidence interval
[CI] 7.3 to 13.1 dB) versus 4.6 dB (95% CI 2.0 to 7.3
dB) in the WW group (p ⫽ 0.006). At 12 mo follow-
Figure 2. Percentage of children with bilateral otitis media
with effusion (OME) during follow-up. VT ⴝ ventilation tube;
WW ⴝ watchful waiting.
up, the improvements were 13.1 dB (95% CI 10.42 to
15.8 dB) and 8.5 dB (95% CI 5.5 to 11.5 dB),
respectively (p ⫽ 0.03). The improvements in the
worse ear are similar to those in the better ear at 6
and 12 mo follow-up; only the absolute means are
slightly higher in the worse ear. The mean hearing
deficit in the VT group decreased from about 45 dB
A to about 35 dB A; the estimated mean duration
with a hearing deficit poorer than or equal to 35 dB
A were 50% and 60% of the total follow-up time in
the VT group and the WW group, respectively. It
should be noted, however, that the noise floor was
about 20 dB HL, which means that the mean air
bone gap was about 15 dB. The results changed very
little when the measurements at 4000 Hz were
omitted.
Figure 3. The pure-tone average (PTA) in the better and worse
ear at 0, 6, and 12 mo follow-up. VT ⴝ ventilation tube; WW
ⴝ watchful waiting.
196
EAR & HEARING / JUNE 2001
Unexpectedly, there was a difference of 3 dB in
mean hearing levels between the VT group and
the WW group at baseline, despite the balanced
randomization procedure. This may have been
partly due to the fact that, immediately after
finding out the result of the randomization allocation, the parents of 19 children, who had given
informed consent, withdrew their child from the
study. These children were not included in the
analyses; 15 had been randomized to the VT group
and four to the WW group. Audiometry data on
these children showed that the four children randomized to the WW group had poorer hearing
levels than the remaining children in the WW
group. To adjust for this baseline difference, hearing level at randomization was added as a covariate in the regression analyses.
Efficacy Analyses
Table 3 shows the results of the efficacy analyses. The largest difference in mean hearing
thresholds was found between the children with
functioning tubes and the children in the WW
group. In the latter the mean hearing thresholds
were 38.7 dB A and 34.7 dB A at 6 and 12 mo of
follow-up, respectively; the difference between the
groups was about 5 dB (p ⫽ 0.0001 and 0.0003 at
6 and 12 mo, respectively). Hearing loss increased
between 6 and 12 mo follow-up in the children
with recurrent OME in the VT group, but still
remained slightly lower (2 to 3 dB, p ⫽ 0.3 and
0.07, respectively) than in the WW group. The
differences became larger when the latter hearing
levels were compared with those in children in the
WW group who did not recover spontaneously:
mean hearing loss in these children was 41.1 dB A
(SE ⫽ 1.2) and 38.7 dB A (SE ⫽ 1.1) after 6 and 12
mo follow-up, respectively. The hearing losses in
the WW children who recovered spontaneously
were almost identical to those in the children with
functional tubes: 33.6 dB A (SE ⫽ 1.3) and 30.1 dB
A (SE ⫽ 0.8) after 6 and 12 mo follow-up,
respectively.
At 3, 6, 9, and 12 mo of follow-up, 92%, 76%, 56%,
and 30% of the tubes were still in situ, respectively.
Regression Analyses
In the univariate analysis with hearing improvement at 6 mo as the dependent variable, season and
hearing level at randomization appeared to be confounders. In the multivariate analysis, only hearing
level at randomization was a confounder (disturbing
the relation between treatment and hearing level)
and an effect modifier (children with a higher hearing level at randomization improved more than
children with a lower hearing level at randomization). The final linear models for the better and
worse ear are shown in Table 4A. From this model it
was possible to calculate the improvement in hearing levels in the two groups over the various categories of hearing thresholds at randomization. For
example, when the hearing loss in the better ear at
randomization was 50 dB, a child in the VT group
improved by 15.9 dB while a similar child in the WW
group improved by 11.4 dB.
In the univariate analysis with improvement at
12 mo as dependent variable, hearing level at randomization, season and age at randomization appeared to be important covariates. In the multivariate linear model, hearing level and age at
randomization acted as confounders (see Table 4B).
The difference in hearing improvement at 12 mo
follow-up between the VT group and WW group, adjusted for the hearing level and age at randomization,
was 1.6 dB (95% CI ⫺0.6 to 3.6) for the better ear and
1.5 dB (95% CI ⫺1.0 to 4.0) for the worse ear.
Ten children in the WW group received treatment
with ventilation tubes during follow-up. A sensitivity analysis showed that the treatment effects in the
intention-to-treat analysis did not change when we
analyzed these 10 children as treated or when they
were omitted from the analysis. A further 11 children dropped out in the trial period (eight from the
WW group and three from the VT group). There was
a difference in the mean hearing level at randomization in the three children withdrawn from the VT
TABLE 3. The efficacy of the ventilation tubes in the VT group.*
Months of
Follow-Up
6
12
VT In Situ
Bilateral OME
Number of
Children**
Mean Hearing
Deficit (dB A)
SE (dB A)
Yes
No
No
Yes
No
No
No
No
Yes
No
No
Yes
49
—
8
26
7
5
32.5
—
36.1
29.9
32.5
32.9
0.8
—
2.8
0.8
2.9
2.1
* For reference at 6 and 12 mo follow-up, hearing deficits in the WW group were 38.7 dB A (SE ⫽ 1.0) and 34.7 dB A (SE ⫽ 0.8), respectively.
** The total number was lower than the 94 due to missing values for tympanograms and/or otoscopy (N ⫽ 23 and 37 at 6 and 12 mo follow-up, respectively) as well as due to the exclusion
of children with ottorhea (N ⫽ 12 and 10 children at 6 and 12 mo follow-up, respectively).
VT ⫽ ventilation tube; WW ⫽ watchful waiting; OME ⫽ otitis media with effusion.
EAR & HEARING, VOL. 22 NO. 3
197
TABLE 4A. The final model with ⌬hearing level in the better
(worse) ear as dependent variable at 6 mo follow-up.
Co-variable
Estimate
SE
Intercept (␣)
Group
Watchful waiting
Ventilation tubes
(ref)
Hearing at baseline
(dB)
Interaction
(Group ⫻ hearing
baseline)
39.1 (41.9)
4.7 (5.7)
⫺10.5 (⫺10.7)
—
6.0 (7.6)
—
0.084 (0.15)
—
⫺1.1 (⫺1.0)
0.1 (0.1)
0.0001 (0.0001)
0.3 (0.3)
0.1 (0.2)
0.023 (0.04)
p-value
Maximal Contrast
(N ⫽ 166, with 82 children in the watchful waiting group and 84 in the ventilation tube
group).
TABLE 4B. The final model with ⌬hearing level in the better
(worse) ear as dependent variable at 12 mo follow-up.
Co-variable
Estimate
SE
Intercept (␣)
Group
Watchful waiting
Ventilation tubes (ref)
Hearing at baseline (dB)
Age at baseline
⬍20 months
ⱖ20 months (ref)
32.7 (32.6)
2.9 (3.5)
⫺1.6 (⫺1.5)
—
⫺1.0 (⫺1.0)
1.1 (1.3) 0.16 (0.25)
—
—
0.1 (0.1) 0.0001 (0.0001)
2.6 (4.2)
—
analysis without those having had adenoidectomies
revealed no differences between the clusters.
p-value
1.4 (1.5) 0.06 (0.01)
—
—
(N ⫽ 166, with 80 children in the watchful waiting group and 86 in the ventilation tube
group).
Spontaneous resolution and recurrence of OME in
the WW group obscured the contrast in underlying
disease between the VT group and the WW group. To
test our etiological model (effusion causes hearing
deficits) we compared the children with effusion during the whole follow-up period (N ⫽ 28) with the
children who were effusion-free during the whole follow-up period (N ⫽ 54); note that in this analysis the
original randomized intervention was disregarded.
After adjustment for hearing level at randomization, the mean hearing level in the children
who were diagnosed as only having bilateral effusion at randomization showed a greater improvement of 6.8 dB compared with the children with
effusion during the whole follow-up (p ⫽ 0.007).
The absolute hearing levels in the children who
were effusion-free during the follow-up were 34.1
dB A (SE ⫽ 1.0) and 30.8 dB A (SE ⫽ 0.7) at 6 and
12 mo, respectively. In the children with bilateral
effusion during the entire follow-up the hearing
levels were 40.8 dB A (SE ⫽ 1.7) and 39.3 dB A (SE
⫽ 1.5), respectively.
DISCUSSION
group (mean hearing level of 59.2 dB A). However,
these three children will have little impact on the
overall results.
Logistic Regression
The final logistic model is described by:
Logit P(⬎15 dB A) ⫽ 0.39 ⫺ 1.11 ⫻ treatment.
Therefore, the probability of a greater than 15 dB
improvement is 0.40 (95% CI 0.31 to 0.51) and 0.18
(95% CI 0.11 to 0.28) in the VT group and WW
group, respectively. In other words, the children in
the VT group were 2.2 times more likely to improve
by at least 15 dB than the children in the WW group.
Principal Component Analyses
The subgroup of children with some complaints
(upper respiratory tract infections, earache, and
fever, which might indicate acute otitis media) appeared to benefit more from treatment with ventilation tubes (improvement 8.5 dB) than the children
with no complaints or frequent complaints. It should
be noted that 11 children in the latter cluster had
undergone adenoidectomy before randomization. As
these young children probably underwent adenoidectomy because of illness, the results suggest a
positive effect of adenoidectomy on hearing improvement during the trial period. A principal component
Some aspects of the findings in this study
confirm those reported in other randomized controlled studies on surgery for OME in older children (Black et al., 1990; Dempster et al., 1993;
Gates et al., 1987; Maw & Herod, 1986 ). The effect
of ventilation tubes on group average hearing
levels was evident at 6 mo follow-up, but most of
the advantage had disappeared by the 1 yr followup. In agreement with the results of the study by
Gates et al. (1987) effusion was present for about
one third of the follow-up period in the VT group.
The percentage of tubes that remained both functioning and in place (92%, 76%, 56%, and 30% at 3,
6, 9, and 12 mo follow-up, respectively) are in
agreement with the findings of Gibb and Mackenzie (1985) who studied Reuter Bobbins®. In this
study we chose to use short-term tubes (Bevil
Bobbins®) because in Europe these are routinely
used in infants receiving tubes for the first time.
The results of this study are therefore only generalizable to short-term tubes. It is possible that the
advantage of ventilation tubes may extend beyond
6 mo if a longer term tube had been used.
It should be emphasised that the children in our
study were much younger than those in previous
trials. The mean age at randomization in our
study was 19 mo compared with 4 yr and older in
198
EAR & HEARING / JUNE 2001
all the other studies. Therefore, comparisons with
the literature might not always be meaningful.
The children in our study were referred to an
ENT Department after they had failed a population
based auditory screening test, whereas in the other
studies the children were recruited by means of
referral by GPs, speech and hearing centers, or other
hospital departments because of hearing problems
and/or behavior problems. This might imply that the
children in the present study had milder hearing
loss than those in previous studies.
Because of the young age of the children and the
test situation (quiet rooms instead of soundproof
booths), the noise floor was about 20 dB. Hearing
thresholds better than this noise floor were not
detectable, which might have resulted in an underestimation of the hearing benefit. On the other
hand, the noise floor is acceptable for multi-center
clinical trials on hearing sensitivity in OME-prone
children: most regular hospitals do not have audiometry sets or test rooms that are suitable for
testing young children. Therefore, a portable VRA
was the only alternative to referring all the children to the nearest Audiological Centre, which
would have resulted in an increased burden for
the parents and hence would have decreased the
accrual rate.
In conclusion, this study has shown that ventilation tubes improve the hearing levels in infants
with persistent bilateral OME. The magnitude
and duration of this effect, however, is limited.
Whether this provides sufficient justification for
treatment in this age group also depends on other
effects, such as possible improvement in language
development or in quality of life. These effects
need to be studied before conclusions can be drawn
about the (cost) effectiveness of treatment with
ventilation tubes in infants who are diagnosed as
having persistent OME after referral from a
screening program.
ACKNOWLEDGMENTS
We thank the parents and children who took part in this study,
the audiologists who performed tympanometry and audiometry,
the ENT surgeons* who performed otoscopy and provided medical
management, Joost Engel and Ad Snik for their valuable suggestions concerning this paper, and Sonja van Oosterhout for both
trial management and data entry.
This project was funded by the Dutch Investigative Medicine
Fund of the National Health Insurance Board.
*ENT surgeons and centers: J. van Leeuwen, Rijnstate Hospital,
Arnhem; G. Pluimers, H. Vencker, J. Mol, and A. Frima-van
Aarem, Eemland Hospital, Amersfoort; J. Engel and S.J. de
Vries, Canisius Wilhelmina Hospital, Nijmegen; K. Ingels, University Hospital St. Radboud, Nijmegen; M. Nijs, P. Gerritsma,
and P. van den Ven, Slingeland Hospital, Doetinchem; S.J.
Rietema, Hospital Gelderse Valei, Ede; E. Teunissen and G. van
de Meerakker, Carolus Liduina Hospital, ‘s Hertogenbosch; E.
Kwadijk and A. de Visscher, GZG, ‘s Hertogenbosch; H. Cats,
Willem-Alexander Hospital, ‘s Hertogenbosch; T. Bruggink and S.
Hofstad, Hospital Rivierenland, Tiel; G. Peeters and D. Elprana,
St. Maartensgasthuis, Venlo; J. Alvarado-van Os, Beatrix Hospital, Winterswijk; F. Mud and Ch. Sepmeijer, Hospital Het
Nieuwe Spittaal, Zutphen.
Address for correspondence: Maroeska M. Rovers, Department of
Otorhinolaryngology, University Medical Centre Nijmgen, P.O.
Box 9101, 6500 HB Nijmegen, The Netherlands.
Received February 18, 2000; accepted December 19, 2000
REFERENCES
Armstrong, B. W. (1954). A new treatment of chronic secretory
otitis media. Archives of Otolaryngology, 59, 653– 654.
Beynon, G., & Munro, K. (1993). A discussion of current sound field
calibration procedures. British Journal of Audiology, 27, 427–435.
Black, N. A, Sanderson, C. F. B, Freeland, A. P., & Vessey, M. P.
(1990). A randomised controlled trial of surgery for glue ear.
British Medical Journal, 300, 1551–1556.
Bluestone, C. D. (1999). Definitions, terminology, and classification. In R. M. Rosenfeld & C. D. Bluestone (Eds.), Evidence-Based
Otitis Media (pp. 85–103). Hamilton, Ontario: B.C. Decker, Inc.
Dempster, J. H, Browning, G. G., & Gatehouse, S. G. (1993). A
randomized study of the surgical management of children with
persistent otitis media with effusion associated with a hearing
impairment. Journal of Laryngology and Otology, 107, 284 –289.
Engel, J. A. M, Anteunis, L. J. C, & Hendriks, J. J. T. (1999).
Treatment with grommets in the Netherlands: Incidence in
children from birth to 12 years. In M. Tos, J. Thomsen, & V.
Balle (Eds.), Otitis Media Today (pp. 451– 455). The Hague,
The Netherlands: Kugler Publications.
Engel, J., Anteunis, L., Volovics, A., Hendriks, J., & Marres, E.
(1999). Prevalence rates of otitis media with effusion from 0 to 2
years of age: Healthy born versus high-risk-born infants. International Journal of Pediatric Otorhinolaryngology, 47, 243–251.
Freemantle, N., Long, A., Mason, J., Sheldon, T., Song, F.,
Watson, P., & Wilson, C. (1992). The treatment of persistent
glue ear in children. Effective Health Care, 4, 1–16.
Gates, G. A, Avery, C. A., Prihoda, T. J., & Cooper, J. C. (1987).
Effectiveness of adenoidectomy and tympanostomy tubes in
the treatment of chronic otitis media with effusion. New
England Journal of Medicine, 317, 1444 –1451.
Gibb, A. G., & Mackenzie, I. J. (1985). The extrusion rate of
grommets. Otolaryngology-Head and Neck Surgery, 93, 695– 699.
Jerger, J. (1970). Clinical experience with impedance audiometry.
Archives of Otolaryngology, 92, 311–324.
Mandel, E. M., Rockette, H. E., Bluestone, C. D., Paradise, J. L.,
& Nozza, R. J. (1989). Myringotomy with and without tympanostomy tubes for chronic otitis media with effusion. Archives
of Otolaryngology—Head and Neck Surgery, 115, 1217–1224.
Maw, R., & Herod, F. (1986). Otoscopic, impedance, and audiometric findings in glue ear treated by adenoidectomy and
tonsillectomy. Lancet, 6, 1399 –1402.
Rovers, M. M, Krabbe, P. F. M, Straatman, H., Ingels, K., van der
Wilt, G. J., & Zielhuis, G. A. (2001). The effect of ventilation tubes
on quality of life in infants with persistent otitis media with
effusion: A randomised trial. Archive of Disease in Childhood, 84,
45– 49.
Rovers, M. M, Straatman, H., Ingels, K., van der Wilt, G. J., van
den Broek, P., & Zielhuis, G. A. (2000). The effect of ventilation
EAR & HEARING, VOL. 22 NO. 3
tubes on language development in infants with otitis media
with effusion: A randomized trial. Pediatrics, 106, E42:1– 8.
Rovers, M. M, Zielhuis, G. A., Straatman, H., Ingels, K., van der
Wilt, G. J., & van den Broek, P. (1999). Prognostic factors for
persistent otitis media with effusion in infants. Archives of
Otolaryngology—Head and Neck Surgery, 125, 1203–1207.
Rovers, M. M, Zielhuis, G. A., Straatman, H., Ingels, K., van der
Wilt, G. J., & Kauffman-de Boer, M. (1999). Comparison
between the CAPAS test and the Ewing test for hearing
screening in infants. Journal of Medical Screening, 6, 188 –192.
Senturia, B. H, Bluestone, C. D., Klein, J. O., Lim, D. J., & Paradise,
J. L. (1980). Report of the ad hoc committee on definition and
classification on otitis media and otitis media with effusion.
Annals of Otology, Rhinology and Laryngology, 89, 3– 4.
199
Stephenson, H., & Haggard, M. (1992). Rationale and design of
surgical trials for otitis media with effusion. Clinical Otolaryngology, 17, 67–78.
Vernon-Feagans, L. (1999). Impact of otitis media on speech,
language, cognition, and behavior. In R. M. Rosenfeld & C. D.
Bluestone (Eds.), Evidence-Based Otitis Media (pp. 353–373).
Hamilton, Ontario: B.C. Decker, Inc.
REFERENCE NOTE
1 Rovers, M. M, Snik, A. F. M, Ingels, K., van der Wilt, G. J., &
Zielhuis, G. A. (Submitted). Visual reinforcement audiometry
in epidemiological studies on otitis media with effusion in
infants.