Keywords health-related quality of life - cognition - cochlear implantation - anxiety - depression
Introduction
Hearing impairment (HI) is the 4th most common chronic disease worldwide;[1 ] it affects 1.2 billion people, and this rate is steadily increasing. The prevalence
of severe to profound HI (pure tone average [PTA] at 0.5–4 kHz ≥70 dB in the best
ear or a speech recognition [SR] score ≤30%) is of 0.2% in the Swedish adult population,
which is of ∼ 22,000 individuals.[2 ] One of the main impacts of HI is the decreased ability to communicate. Hearing impairment
has a negative influence on health-related quality of life (HRQoL),[3 ]
[4 ] and the severity of the hearing loss correlates with the reduction in HRQoL.[3 ] It has become more and more common to measure the impact of health interventions
in terms of HRQoL, including an evaluation of the outcome of the cochlear implantation
(CI), and there are several definitions of HRQoL, as well as several different methods
to measure it.
Hearing impairment is significantly associated with depression, especially among women.[5 ] In a recent study,[6 ] anxiety and depression were found to be more common and more severe in individuals
with severe to profound HI (PTA > 70 dB in the better ear at 0.5 kHz, 1 kHz, 2 kHz
and 4 kHz) than in the general population.[6 ] Hearing-impaired older adults have a higher incidence of hospitalization than those
with normal hearing,[7 ] and HI has also been shown to be a risk factor for mortality.[7 ]
[8 ]
Cochlear implantation is a safe and effective intervention for severe to profound
HI in adults.[9 ] The general indication for CI is an inability to communicate using spoken language
despite having optimally-programmed hearing aids (HAs). There are no definitive criteria
for CI, but the current benchmark is PTA (0.5 kHz, 1 kHz, 2 kHz, 4 kHz) ≥70 dB HL
and SR ≤50% correct repeated words in the better ear with HAs; the criteria can be
more lenient in case of rapid progression.[10 ]
Several studies show a significant improvement in SR after CI,[11 ]
[12 ]
[13 ] as well as a significant improvement in HRQoL, measured in different ways: SF-36
(The Short Form 36 questions),[11 ]
[12 ]
[13 ] NCIQ (Nijmegen Cochlear Implant Questionnaire),[11 ]
[14 ]
[15 ] GBI (Glasgow Benefit Inventory)[15 ]
[16 ] and Health Utilities Index 2 and 3 HUI2-3 (Health Utilities Index 2-3)[12 ]
[13 ]. Based on the register data, one study[6 ] found a lower risk of negative impact on HRQoL among CI users compared with individuals
with severe to profound HI without CI.
Hirschfelder et al[17 ] have shown a correlation between SR and HRQoL (using the NCIQ and SF-36) in adults
with severe HI who were submitted to CI, but other studies have failed to replicate
this result (using the NCIQ and SSQ [Speech, Spatial and Qualities][12 ] and the NCIQ, GBI and HUI3 questionnaires[15 ]). A meta-analysis[18 ] of 14 articles, involving a total of 679 CI patients, resulted in a low correlation
between patient-reported outcome measures, including HRQoL and SR, which, according
to the authors, supports the need for a more regular use of HRQoL instruments to assess
CI outcomes.[18 ]
There are several publications reporting the long-term outcome of SR and HRQoL among
adult CI users. No significant growth or decline in speech perception is observed
after 6–24 months post-CI.[19 ]
[20 ] When investigating the elderly group of CI users (≥70 years), the same stable results
regarding SR and HRQoL are shown.[21 ]
[22 ]
Previous results suggest different predictive factors for post-CI performance, such
as age,[23 ]
[24 ] gender,[25 ] duration of deafness,[26 ] residual hearing,[27 ] cognitive parameters,[23 ] etiology[24 ] and psychosocial factors.[28 ]
Objective
The first aim of the present study was to examine the outcomes one and three years
after CI in adults with bilateral severe to profound postlingual HI with respect to
hearing, HRQoL, level of depression and anxiety, and to determine the correlation
between the outcomes.
The second aim was to investigate whether age, gender, etiology, side of implantation,
residual hearing, level of anxiety and depression and cognitive performance could
predict the outcome.
Subjects and Methods
Subjects
Participants were recruited from the cochlear implant programme at the Department
of Otolaryngology, Linköping, a tertiary teaching university hospital in Sweden, from
1 April 2011 to 31 May 2013. Eligibility for the study was defined as: 1) age ≥18
years at the time of the CI; 2) a bilateral severe to profound postlingual HI defined
according to PTA and/or SR; 3) unilateral CI; and 4) at least three years of experience
with CI.
Out of a total of 60 implanted individuals, 1 died before the 1-year follow-up because
of causes not related to the CI, 2 were not able to participate because of language,
1 was excluded because of a severe visual impairment (the subject could not fill out
the questionnaires), 1 was excluded because of cognitive dysfunction to a degree that
made it impossible to answer the questionnaires, and 3 CI users did not want to answer
any questionnaires. One CI user was reoperated, and did not answer the questionnaire.
Thus, 51 CI users were asked for informed consent to participate in the study, and
40 (78%) accepted the invitation. A total of 16 (40%) participants were men. The mean
age at implantation was 71 years (standard deviation [SD]: 11.3 years), and the median
of age was 72 years (range: 19–88 years). Women were on average five years younger
than men at the CI surgery. The majority (n = 33) of the subjects were retired; 1 participant classified him/herself as active,
and 6 participants left the question about employment unanswered. The right side was
implanted in 23 participants and the left side, in 17.
The etiologies reported are based on the clinical evaluation before implantation,
and one participant could have more than one etiology for CI candidacy. Etiology (number
denoted): hereditary (12), noise-induced (8), unknown (6), auditory neuropathy (4),
Ménière disease (2), otitis media (2), parotitis (2), meningitis (1), rubella (1),
sudden deafness (2), vestibular schwannoma (1) (the CI was implanted in the opposite
ear), branchio-oto-renal syndrome (BOR) (1), probably autoimmune base (1), Usher 2
(1), and peripheral polyneuropathy (1). In addition to these etiologies, presbyacusis
may have had an impact.
Methods
The study design was prospective and longitudinal. The participants were followed
for 3 years. Approval from the regional ethical review board at Linköping University
was obtained (Number/ID:2011/87-31). Written informed consent was received from all
40 participants.
Review of Medical Records
Information about age at surgery, gender, etiology, side of implantation, audiological
measurements (including air conduction thresholds and SR) and cognitive performance
was collected from medical records.
In-quiet x-ish speech audiometry (SR) is based on lists with 50 phonemically balanced
monosyllabic words.[29 ] It is a standardized and validated method with a variability of ∼ 10%, and the result
is scored in percentage of correct words repeated. The PTA was defined as the average
of the air conduction pure-tone thresholds at 0.5 kHz, 1 kHz, 2 kHz, and 4 kHz; when
the threshold exceeded the maximum output level of the audiometer, 130 dB were used.[30 ]
To measure the reality-based hearing function, that is, the SR that is nearest to
each individual's hearing function in everyday life, we chose the best possible speech
recognition (BPSR) that each participant could perform, which was based on the best
SR score obtained at 65 dB hearing level in a sound field pre-CI and one and three
years post-CI, unilaterally or bilaterally, with or without optimally-adjusted HA(s).
To clarify, from the different measurements of SR, we chose each participant's highest
SR, supposing that it corresponds best to their daily hearing situation. The BPSR
pre-CI and one and three years post-CI, as well as the SR on the side to be submitted
to CI and the side submitted to CI were analyzed.
HUI3 and HADS Questionnaires
In the present work, HRQoL is defined as the value assigned to one's duration of life
as modified by the impairments, functional states, perceptions, and social opportunities
that are influenced by disease, injury, treatment or policy.[31 ] To asses HRQoL, the HUI3 questionnaire was used, and the Hospital Anxiety and Depression
Scale (HADS) was used to assess the levels of depression (HAD-D) and anxiety (HAD-A).
The questionnaires were sent by mail pre-CI and one and three years post-CI to all
participants.
The HUI3 is a generic test, and it has been used in many different HRQoL studies,
in which the objective is to monitor health changes due to treatment. It has strong
theoretical foundations, is valid and reliable.[32 ] It is a self-assessment questionnaire with 15 items corresponding to 8 dimensions
(vision, hearing, speech, ambulation, dexterity, emotion, cognition and pain). The
HUI3 attribute levels are determined based on responses to one or more questions, and they represent functional
classes. Single-attribute utility functions describe the burden of morbidity for a person, attribute by attribute. The HUI3 provides
both single-attribute and overall utility functions in a scale in which dead = 0.00
and perfect health = 1.00.[32 ]
The HADS is a brief, self-assessment mood scale with 14 items, with scores that range
from 0 to 21 for either anxiety or depression, and a cut-off point ≥8. The subscale
scores (HAD-A and HAD-D) can be used to measure the severity, and their repeated application
will provide information on changes.[33 ]
[34 ] The HADS is reliable and its validity is good, and results based on the instrument
show a high correlation with other instruments that measure psychological illness.[35 ]
Cognitive Tests
The cognitive capacity was assessed by means of five different tests, each one assessing
some specific cognitive ability, and they were applied pre-CI by means of a computer.
The reading span test was used as a measure of working memory capacity. The task was to decide whether
the sentences were semantically absurd or not, and after reading them, recall either
the first or the final words of the sentences in correct serial order. The physical matching task was used to assess the participant's general reaction time and to determine
whether two simultaneously presented letters had the same physical shape. The lexical decision-making test was used to measure the lexical access speed. The task was to decide whether
a combination of three letters constituted a real word or not. The rhyme-judgment test was used to measure the phonologic ability. The task was to decide whether two
words presented simultaneously rhymed or not. A test of antonyms was used to assess one aspect of the participant's verbal ability. The test used
is part of the Wechsler Adult Intelligence Scale (WAIS) battery, and is one of the
most frequently used tests of intelligence. The results were calculated in terms of
percentage of correct responses and the time required to perform the task.
In addition, in the HUI3 there is a dimension of cognition, (HUI3 CU), which measures
one aspect of long-term memory.
Data Analysis
The results of the audiological measurements, HUI3 and HADS pre-CI and one and three
years post-CI were analyzed. Consequently, every CI recipient was his or her own control.
The statistical calculations were performed with the International business Machines
Statistical Package for the Social Sciences (IBM SPSS, IBM Corp., Armonk, NY, US)
software, version 23. When investigating the occurrence of significance, a nonparametric
test, the Wilcoxon test (matched-pair signed rank), was used due to the heterogeneous
group of variables with both ordinal and interval data. The significance level was
set at p < 0.05. The SR and HUI3 scores were treated as interval data and the HUI3 levels
and the sum of the HADS scores, as ordinal data. The Spearman test was used when analyzing
correlation; in Spearman rho (ρ) correlation coefficient, ρ > 0.6 was chosen as a
high correlation.
Results
Audiometric Tests
The pre-CI assessment of residual hearing showed a large variation due to the variation
in hearing ability. Pre-CI, the mean PTA was of 88 dB (SD: 16.2; range: 55–130 dB)
on the best side, and of 100 dB (SD: 17.5; range: 60–130 dB) on the side to receive
the implant. For men, the pre-CI PTA was of 96 dB (SD: 11.4) on the side to receive
the implant; for women, it was of 102 dB (SD: 20.5). Only 10 out of 40 participants
could produce a result on the speech audiometric test on the side to receive the implant
without HA, but 18 participants could produce a result with an optimally adjusted
HA. Pre-CI, the mean SR was of 7.9% (SD: 10.6%; range: 0–34.0%) with HA on the side
to receive the implant, and the mean BPSR was of 23.8% (SD: 18.0%; range: 0–58.0%).
One-year post-CI, on the CI side, the mean SR was of 54.7% (SD: 24.3%; range: 0–98.0%),
showing a significant improvement of 46.8% (SD: 24.3; p < 0.000) but without a significant difference 3 years post-CI: 54.3% (SD: 21.4%;
range: 0–98.0%). The mean BPSR one year post-CI was of 62.8% (SD: 21.5; range: 0–98.0%;
p < 0.001), with no significant difference 3 years post-CI: 59.6% (SD: 21.7%; range:
0–96.0%; n = 38). The BPSR one and three years post-CI were correlated (ρ= 0.87).
A reduction greater than 10%, that is, 5 words, in the BPSR, was observed in 1 patient
1 year post-CI compared with the pre-CI period, and reductions greater than 10% were
observed in 8 patients 3 years post-CI compared with 1 year post-CI, but no explanations
for this were identified. No significant correlation was found between the residual
hearing (BPSR) pre-CI and the hearing outcome (BPSR) one (ρ= 0.34) and three (ρ= 0.30)
years post-CI.
The BPSR for the gender subgroups pre-CI and one and three years post-CI can be seen
in [Figs. 1 ] and [2 ]. At the follow up after three years, there was missing data on two individuals in
the male group, and no missing data in the female group, but BPSR = 0 for two individuals.
No significant gender difference in BPSR was found pre-CI, but women improved significantly
(p = 0.009) more than men 1 year post-CI, on average 16.8% better, as shown in [Table 1 ]. Similar results were observed when evaluating the SR only on the side to receive
the implant or on the implanted side: women improved significantly (p = 0.036) more than men one year post-CI. By the 3 years post-CI follow up, the men's
mean BPSR was of 54.6% (SD: 27.8%; range: 0–96.0%; n = 14) and the women's was of 63.1% (SD: 15.7%; range: 24.0–88.0%; n = 24), which is not a significant (p = 1.53) gender difference, since the men's score remained unchanged while that of
the women decreased by 5.1%.
Table 1
Outcomes of BPSR, SR on the CI side, HUI3 HU and HUI3 OU separated by gender pre-
and one and three years post-CI mean (SD)
Outcome
Men
Women
Pre-CI
One year
post-CI
Three years
post-CI
Pre-CI
One year
post-CI
Three years
post-CI
BPSR (%)
25.5 (18.1)
(n = 16)
54.6 (25.9)
(n = 16)
54.6 (27.8)
(n = 16)
22.6 (18.2) (n = 24)
68.3 (16.4)
(n = 24)
63.1 (15.7)
(n = 22)
SR on the CI side (%)
8.1 (11.3)
(n = 16)
44.4 (29.6)
(n = 16)
48.7 (27.2)
(n = 14)
7.7 (10.4)
(n = 24)
61.3 (17.8)
(n = 22)
57.9 (16.6)
(n = 22)
HUI3 HU
0.29 (0.25)
(n = 16)
0.66 (0.18)
(n = 16)
0.85 (0.10)
(n = 15)
0.29 (0.31)
(n = 23)
0.62 (0.23) (n = 22)
0.87 (0.80)
(n = 18)
HUI3 OU
0.48 (0.20)
(n = 16)
0.69 (0.16)
(n = 16)
0.63 (0.24)
(n = 15)
0.43 (0.24)
(n = 22)
0.63 (0.20)
(n = 22)
0.59 (0.18)
(n = 17)
Abbreviations: BPSR, best possible speech recognition; CI, cochlear implantation;
HUI3 HU, Health Utilities Index 3, hearing utility; HUI3 OU, Health Utilities Index
3, overall utility; SD, standard deviation; SR, speech recognition.
Fig. 1 Best possible speech recognition (BPSR) pre- and one and three years post-cochlear
implantation (CI) in the male subgroup.
Fig. 2 Best Possible Speech Recognition (BPSR) pre- and one and three years post-CI in the
female subgroup.
The results of the BPSR and SR one and three years after the procedure on the side
that received the CI, as well as the differences in the results between the two periods
were analyzed in subgroups by age (separated by the median age of 72), and no significant
differences were found. Moreover, when excluding the outliers (one participant was
much younger than the rest), the result remained the same (data not shown).
The results of hearing (BPSR pre-CI and one year post-CI) were also analyzed after
transforming the data in “rationalized” arcsine units (RAUs),[36 ] but no significant correlations emerged. The RAUS statistical method is used in
analyses related to scales, such as SR scores, in which a change is not comparable
in both ends of the scale: the experienced change from SR 60% to 70% is not the same
as from 15% to 25%, for example.
The use of HAs on the contralateral side could be noted 1 year post-CI among 30 of
the 40 participants, 5 subjects did not use HA, 1 meant to try an HA, and data were
missing in 3 cases. One participant was submitted to a second CI after being included
in the study. Three years post-CI, 25 participants used HAs on the contralateral side,
1 used a vibrant soundbridge, 3 did not use HAs, and data were missing in 11 cases
(2 participants moved).
HRQoL
Pre- and one-year post-CI, 38 participants completed the HUI3 questionnaires sufficiently
to enable the calculation of the overall utility multi-attribute score (HUI3 OU). Three years post-CI, 32 participants completed the
questionnaires.
The HUI3 Hearing level (HUI3 HL) 6, “Unable to hear at all,” was the most common level pre-CI. One and three
years post-CI, HUI3 HL 3 was the most common (“Able to hear what is said in a conversation
with one other person in a quiet room with a hearing aid” and “Able to hear what is
said in a group conversation with at least three other people, with a hearing aid”).
According to the classification system to categorize HUI3 attribute levels by disability (none, mild, moderate and severe), 67.5% of the participants were classified as severe,
and 30%, as moderate pre-CI, while only 17.5% were classified as severe, and 72.5%,
as moderate 1 year post-CI; 3 years post-CI, only 10% were classified as severe, and
70%, as moderate.
The single-attribute utility score of hearing (HUI3 HU) significantly improved one year post-CI, as well as the utility
scores of emotion and pain, and the multi-attribute overall utility score , (HUI3 OU; see [Table 2 ]). The HUI3 HU significantly improved again between one and three years post-CI.
Table 2
Mean single- and multi-attribute-scores of the HUI3 pre- and one and three years post-CI
HUI3
Pre-CI
One year
post-CI
Three years post-CI
Vision, HUI3 VU
0.88
(n = 39)
0.89
(n = 40)
0.97**
(n = 32)
Hearing, HUI3 HU
0.29
(n = 39)
0.63*
(n = 38)
0.86**
(n = 33)
Speech, HUI3 SU
0.97
(n = 39)
0.98
(n = 38)
0.99
(n = 33)
Ambulation, HUI3 AU
0.96
(n = 40)
0.96
(n = 40)
0.96
(n = 33)
Dexterity, HUI3 DU
0.97
(n = 40)
0.97
(n = 40)
0.98
(n = 33)
Emotion, HUI3 EU
0.88
(n = 40)
0.95*
(n = 40)
0.96
(n = 33)
Cognition, HUI3 CU
0.98
(n = 40)
0.99
(n = 40)
0.99
(n = 33)
Pain, HUI3 PU
0.90
(n = 39)
0.94*
(n = 40)
0.96
(n = 33)
Overall, HUI3 OU
0.45
(n = 38)
0.66*
(n = 38)
0.61
(n = 32)
Abbreviations: CI, cochlear implantation; HUI3, Health Utilities Index 3.
Notes: *Significant difference between pre-CI and one year post-CI scores, p < 0.05. **Significant difference between one and three years post-CI scores, p < 0.05.
A high correlation was observed between the HUI3 HU and HUI3 OU pre-CI (ρ= 0.76);
however, this was not repeated one and three years post-CI. A correlation could also
be observed between the HUI3 OU pre-CI and 3 years post-CI (ρ = 0.58), and 1 and 3
years post-CI (ρ =0.62), that is, the degree of HRQoL pre-CI correlates with the degree
of HRQoL post-CI.
No significant gender differences were observed in the HUI3 HU pre-, one and three
years post-CI, but it is worth mentioning that men scored higher one year post-CI,
and women scored higher three years post-CI.
Levels of Anxiety and Depression
All 40 participants filled out the HADS pre- and one year post-CI; however by the
3 years of follow-up only 31 participants had filled out the HAD-A, and 30 had filled
out the HAD-D. The overall results are shown in [Fig. 3 ]. The most frequent score on the HAD-A and HAD-D pre-, one and three years post-CI
was zero. The scores on the HAD-A did not significantly change pre-, one (p = 0.260) and three years (p = 0.914) post-CI. The HAD-D significantly improved one year post-CI (p = 0.036), and significantly deteriorated three years post-CI (p = 0.031), so there was no significant difference between scores on the HAD-D pre-
and three years post-CI (p = 0.762). Every participant (n = 5) with a subscale score ≥8 pre-CI had the score normalized 1 year post-CI, but
2 participants had a subscale score ≥8 on the HAD-A 1 year post-CI, and another 2
participants had a subscale score ≥8 on the HAD-A 3 years post-CI. In general women
scored higher (but not significantly) on the HAD-A and HAD-D than men pre- and 1 year
post-CI. By the three years of follow-up, women scored higher on the HAD-A subscale
and men, on the HAD-D subscale. A high correlation could be observed between the HAD-A
and HAD-D scores pre-CI (ρ= 0.67) and one year (ρ= 0.63) and three years post-CI (ρ=
0.75).
Fig. 3 Boxplot showing the Hospital Anxiety and Depression Scale (HADS) anxiety (HAD-A)
and depression (HAD-D) subscales pre- and one and three years post-CI.
Cognition
One participant was investigated very rapidly, due to an acute infektion, and therefore
did not perform the cognitive test battery, nor did 2 participants with severe vision
impairment, resulting in a sample composed of 37 participants. The results were presented
in percentage of correct responses and the time it took to generate a response.
There was no correlation between the cognitive dimension (HUI3 CU) pre-CI and one
(ρ= 0.19) and three-years (ρ= 0.37) post-CI, but the scores for this dimension one
and three years post-CI were correlated (ρ= 0.82).
None of the results of the cognitive subtests in the WAIS battery correlated with
the HUI3 CU pre-, one, and three years post-CI. The results of the cognitive performance
were further analyzed in subgroups of gender and age (median age: 72 years), and they
had the same performance ([Table 3 ]). None of the results of the cognitive subtests in the WAIS battery correlated with
the HUI3 CU pre-, one, and three years post-CI ([Table 4 ]).
Table 3
Mean (SD) performance for the total group and the subgroups separated by gender and
age pre-CI
Cognitive Performance
Total
Men
Women
Age ≤ 72
Age > 72
Reading span
(% of right answers)
47.3 (10.7)
(n = 36)
45.8 (12.0)
(n = 14)
48.3 (10.0)
(n = 22)
47.0 (11.6)
(n = 18)
47.7 (10.1)
(n = 18)
Physical matching
(% of right answers)
85.9 (16.4)
(n = 37)
85.9 (17.7)
(n = 15)
85.9 (15.9)
(n = 22)
88.7 (12.4)
(n = 18)
83.3 (19.4)
(n = 19)
Lexical decision-making
(% of right answers)
95.7 (5.1)
(n = 37)
95.6 (5.6)
(n = 15)
95.7 (4.9)
(n = 22)
95.9 (5.1)
(n = 18)
95.5 (5.3)
(n = 19)
Rhyme judgment
(% of right answers)
77.0 (15.9)
(n = 37)
70.9 (15.9)
(n = 15)
81.1 (14.8)
(n = 22)
75.9 (15.8)
(n = 18)
78.0 (16.3)
(n = 19)
Antonyms
(numbers right)
13.4 (4.9)
(n = 36)
13.5 (5.3)
(n = 14)
13.4 (4.8)
(n = 22)
13.8 (5.1)
(n = 18)
13.0 (4.9)
(n = 18)
Abbreviations: CI, cochlear implantation; HUI3, Health Utilities Index 3; SD, standard
deviation.
Table 4
Correlation between cognitive performance and pre-, one- and three years post-CI outcome
Cognitive subtest
BPSR
HUI3 HU
HUI3 CU
HUI3 OU
Pre-CI
One year post-CI
Three years post-CI
Pre-CI
One year post-CI
Three years post-CI
Pre-CI
One year post-CI
Three years post-CI
Pre-CI
One year post-CI
Three years post-CI
Reading span (% of right answers)
0.15
0.21
0.26
0.33
−0.12
−0.11
0.24
−0.02
0.08
0.28
0.08
0.28
Physical matching (% of right answers)
0.03
0.30
0.31
0.08
0.08
0.01
0.02
−0.03
−0.09
−0.06
0.10
−0.08
Lexical decision-making (% of right answers)
0.22
0.12
0.07
0.28
−0.11
−0.14
−0.11
−0.09
0.00
0.04
−0.27
−0.03
Rhyme judgment (% of right answers)
0.24
0.21
0.12
0.27
0.20
0.14
0.07
0.18
0.27
0.08
0.18
0.16
Antonyms (% of right answers)
0.25
0.20
0.15
0.23
0.23
0.07
0.07
0.33
0.39
0.13
0.33
0.25
Abbreviatons: BPSR, best possible speech recognition; CI, cochlear implantation; CU,
Cognition Utility; HU, Hearing Utility; HUI3, Health Utilities Index 3; OU, Overall
Utility.
Note: Spearman's rho (ρ) correlation coefficient. No high correlations were observed.
Correlations Between the Outcomes
There was no correlation between the SR, HUI3 and HADS, including measurements taken
pre- and one and three-years post-CI (the SR evaluated was as BPSR and as SR only
on the CI-side and considering the maximum score as well as the gain, that is, the
difference between the BPSR one year post- and pre-CI). In other words, the pre-CI
scores of the SR, HUI3 HAD-A and HAD-D could not predict the post-CI outcomes of the
same variables after examining the total group as well as the gender and age subgroups.
A subgroup analysis of the high performers by the one-year follow-up, that is, the
best quartile (BPSR ≥82%, n = 9 [3 men and 6 women]) revealed a high and reverse correlation between the BPSR
pre-CI and the HUI3 OU one year post-CI (ρ= -0.68). An analysis of the participants
with the largest reduction in BPSR three years post-CI showed a higher mean HUI3 HU
than the total (data not shown).
No correlation was found between the HADS and HUI3, pre- and one year post-CI, but
three years post-CI, there was a negative correlation between the HUI3 OU and the
HAD-D (ρ= -0.81) and the HAD-A (ρ= -0.61). Another negative correlation was found
between the HAD-D pre-CI and the HUI3 OU three years post-CI (ρ= -0.61). No correlation
was found between the cognitive performance pre-CI and gender, age at surgery, HADS,
HUI3 (HU and OU) pre-, one and three years post-CI.
Analyses of Background Factors
No correlation was found between age at surgery, etiology, side of implantation, residual
hearing, cognitive performance and the CI outcomes of hearing (SR), HRQoL (HUI3 OU)
and the levels of anxiety and depression on the HADS. The low correlation between
the cognitive performance and pre- and one and three years post-CI outcome is shown
in [Table 3 ]. As mentioned before, women improved significantly more than men in SR one year
post-CI.
Discussion
The results of this one- and three-year follow-up study show that CI improves hearing
(SR) and HRQoL (HUI3), a finding that is in line with that of several previous studies.[11 ]
[14 ]
[15 ] Furthermore, the results show no correlation between the level of hearing and the
level of HRQoL pre- and post-CI. Neither did it matter whether the absolute number
of percentage points of improvement in SR or the gain was used. Thus, the improvement
in hearing does not necessarily follow the degree of improvement in HRQoL. This is
an interesting finding, as one would expect HRQoL to be better with better hearing.
One possible explanation is that the outcome of the CI is more complex and multi-dimensional
than the sheer improvement in SR. According to the current results, both hearing and
HRQoL are important measures of the outcome of the CI. As improvements in implant
technology and methods of surgery lead to the expansion of the indication criteria
for CI (patients with more residual hearing are implanted), it is important to investigate
not only the hearing outcomes, but also other factors, such as the HRQoL.
In previous studies, many different audiological methods and HRQoL instruments have
been used, which could explain the contradicting findings regarding the question of
whether post-CI results for hearing and HRQoL correlate or not. The method of testing
SR could also play an important role. In the present study, a speech audiometric in-quiet
test was used to measure the SR. The use of an in-noise speech test could probably
describe the functionality better, but we don't have enough data for that analysis,
because only a few participants could produce a result in the in-noise speech test.
Another issue related to methodology in the published studies is the fact that remarkably
often the definition of QoL/HRQoL is missing, and it is thus unclear what is being
measured. In the end, a conventional definition of QoL/HRQoL applicable to hearing
studies would be desirable.
It is generally difficult to compare the results obtained in different countries or
language areas, as speech audiometry is dependent on the language used. Considering
clinical studies, particularly those focused on individuals with severe to profound
HI, there are participants who are not able to perform open-set speech audiometric
tests due to the severity of their HI. This was the case in the present study. Because
we aimed to measure which scenario was the closest to the subjects' real-life hearing
capacity, we chose what we call BPSR. Instead of solely focusing on the improvement
in hearing acuity on the operated CI side, the hearing outcome was based on the participants'
best speech recognition scores, which we regard as an ecological situation.
The HUI3 has been recommended in several recent studies when measuring HRQoL in populations
with hearing complaints,[37 ]
[38 ] including CI users.[39 ] Differences ≥0.03 in the mean HUI3 OU are clinically important.[40 ] In the present study, the HUI3 OU difference 1 year post-CI was of 0.21, which indicates
a large-scale positive change. The same difference in previous studies was of 0.117,[41 ] 0.15[42 ] and 0.17.[39 ] Interestingly, not only were the hearing single-attribute (HUI3 HU) and multi-attribute
scores (HUI3 OU) significantly improved, but the single-attribute scores of pain (HUI3
PU) and emotion (HUI3 EU) were also improved, as previously described.[11 ] Moreover, CI users score higher in HRQoL with respect to pain and emotion compared
with people with normal hearing.[43 ]
Concerning the different hearing measurements, SR and HUI3 HU, both improved significantly
one year post-CI, but without correlation. The SR remained unchanged three years post-CI,
but the HUI3 HU significantly improved again. This may foment further discussion regarding
how hearing/hearing residuals should be assessed in the CI-candidacy procedure, and
how the hearing outcome should be evaluated. Similarly, one can discuss the different
ways of assessing cognition, as the cognitive tests and the HUI3 CU did not correlate.
No previously published study has focused on monitoring the CI procedure with the
HADS instrument, as far as we know, and only a few studies on HI have used the HADS.[6 ]
[44 ] In conclusion, the level of anxiety and depression is low in this group both before
and after CI. Compared with Carlsson et al,[6 ] the level of depression in this population is not as high, and does not correlate
with the degree of HI, which might depend on methodological differences. Whether the
HADS should be considered as ordinal or interval data is a topic of discussion. In
the present study, the sum of the HADS scores was treated as ordinal data because
“not often” is not doubly as depressive as “sometimes.”
Unlike previous studies, the present study could not show that age, cognitive performance
or residual hearing can predict the CI outcome. We did not find a correlation between
age and BPSR, age and cognitive performance, and age and HUI3 post-CI, which is in
accordance with several studies,[14 ]
[27 ]
[45 ]
[46 ] and, consequently, supports the indication of CI at all ages. The lack of significance
between cognitive performance and SR/SR post-CI may possibly be due to fact that the
tests that estimate speech perception are not cognitively demanding enough to be decisive.
Even though residual hearing cannot predict the CI outcome, it is worth noting that
individuals with low SR pre-CI can achieve the highest gains in SR. The other background
factors, such as etiology, operated side, and level of anxiety or depression, did
not have any significance on the CI outcomes. The only predictor with high correlation
to a high SR one year post-CI in the present study was the female gender. The role
of gender needs to be explored more in depth in future studies.
There are only few published results regarding gender differences in the CI outcomes.
Lenarz et al,[19 ]
[25 ] found that men performed better than women in more complex listening situations
such as in-noise speech tests, but found no differences in the monosyllabic test.[25 ] In the present study, more women than men were implanted, at a lower age, and had
less residual hearing with respect to a higher PTA (and a larger SD). As we could
not observe a correlation between age or residual hearing with the hearing outcomes
post-CI, no association between the women's lower age and higher PTA pre-CI and the
women's significantly better SR one year post-CI was found. Despite this, men scored
higher and improved more in the HUI3 HU than women; in other words, men experienced
more improvement in their hearing than women. It is a novel finding that women perform
significantly better in SR. Bodmer et al[27 ] did not find a correlation between gender and SR post-CI, but, unlike the present
study, they retrospectively analyzed 444 adults (with an average age of 47 years and
45% of men) with pre- and postlingual HI.[27 ] Women with profound HI tended to display lower HRQoL than men,[4 ] a result that is in accordance with those of the present study. Men scored higher
in the HUI3 OU pre- and one year post-CI. There seem to be some gender differences
in CI outcomes that might be important to consider when dealing with CI users/CI candidates,
and more research is needed to look further into the gender differences.
The strengths of the present study are the inclusion of the definition of HRQoL, the
pre-CI assessment of cognitive ability, the monitoring with HADS, the one- and three-year
follow-up, and the gender perspective. The shortcomings of the present study include
the number of study subjects and the relative large variation in etiologies.
Conclusion
The results of the present study showed a significant benefit of CI in adults with
postlingual HI regarding SR and HRQoL one year post-CI, but without mutual correlation.
Three years post-CI, the SR and HUI3 OU were unchanged, whereas the HUI3 HU improved
significantly. The present study also confirmed that the HUI3 can be used to assess
HRQoL in a population of people with CI, and that the HADS can be used to follow up
on levels of depression and anxiety. In the present study, age, etiology, side of
implantation, residual hearing, level of anxiety and depression and cognitive performance
did not predict the CI outcome concerning SR and HRQoL, but the female gender did.
