Treatment of Age-Related Hearing Loss in Persons with Alzheimer's Disease

13 February 2018

Companion papers detailing a multicenter double-blind randomized placebo-controlled trial investigating the effects of hearing aids (HA) on cognitive and functional outcomes for persons with Alzheimer's disease (AD) were recently published in the Journal of Alzheimer's Disease [1,2]. While there is ample evidence of comorbid cognitive and sensory decline as well as increased risk of dementia for persons with age-related hearing loss (ARHL), there has been very little research as to the effects of treatment of hearing loss for persons with Alzheimer's disease and related dementias (ADRD). It is commendable to undertake the challenge of implementing a double-blind randomized placebo-controlled trial. Such a trial would have the potential to provide definitive evidence regarding the efficacy of treating hearing loss as a key component of providing care for persons with ADRD.

Although the authors describe the treatment in the study as "hearing aids", the treatment for hearing loss, in fact, is restoring audibility, which is typically achieved through evidence-based fit hearing aids. Given that returning audibility is the treatment proposed in this study, one would expect a protocol to have been used in order to ensure that audibility was achieved for participants. At a minimum, one would expect data to be reported related to what level of audibility was returned to individuals which would have come in the form of measuring the output of the hearing aids in the individual's ear canal and reporting the speech intelligibility index achieved for various input levels (especially soft and moderate input levels). These data would be essential in interpreting the results of this study.

The authors indicate that "the gain rule used for amplification was based on a proprietary fitting algorithm (Adaptive Phonak Digital), derived from the National Acoustic Laboratories' nonlinear fitting method version 1." As Sanders et al. [3] revealed in their article entitled "Manufacturer's NAL-NL2 fitting fail real-ear verification", the method used in this study actually ensures a lack of increased audibility for soft and moderate inputs. In fact, the NALNL1 proprietary fitting version used by the authors is even poorer in terms of audibility than NAL-NL2 investigated by Sanders et al. [3]. This alone would predict the finding of no difference between the placebo (actually mild gain) and hearing aid (also mild gain due to underfitting created by this proprietary fitting algorithm) conditions used in this study.

The authors state that "it seemed important to conduct a rigorous interventional study to investigate potential benefits in terms of psychological symptoms, of auditory rehabilitation of elderly patients suffering both ARHL and AD". A rigorous interventional study would have required an evidence-based approach to fitting and verifying the audibility provided by the amplification used in the study, and this was not accomplished. The study reports essentially no difference between the group with hearing aids and the placebo control group (mild gain amplification) on cognitive status, functional abilities, or behavioral symptoms. Given the essential lack of audibility difference in the hearing aid and placebo condition due to the hearing aid fitting protocol, this null finding is to be expected.

The authors also used self-reported or care-giver reported use data in their analysis. Taubman et al. [4] among others have reported that individuals are not accurate in reporting use time by the day or week, both over and underestimating usage. The investigators used hearing aids that include a data logging feature that could have provided them with accurate use data given the number of visits that were included in the study. It is impossible to know if individuals who never or rarely wore their hearing aids were included in this analysis because the analysis relied on self-report.

If the authors are able to continue this work, we hope they will employ best practices when fitting and verifying hearing aids so this important work can be interpreted in light of a known treatment (audibility and hours of use) that has been applied as compared to a true placebo.

The authors acknowledged in the discussion that comprehensive aural rehabilitation was not included in the protocol and that this may have impacted the results. We agree that informational counseling, device use, and communication training may be essential to the success of this group of patients. However, no amount of counseling will make up for an inadequately fit hearing aid. This type of investigation must first ensure that an audible signal is being returned to the patient and then appropriate aural rehabilitation can be expected to have the greatest impact.

Finally, with regards to the primary outcome measure, the authors set a 'success' criterion of < 6 points of decline on the Alzheimer's disease Assessment Scale Cognitive Subscale (ADAS-Cog). This value was based on the expected decline of cognitive scores per the literature on ADAS-Cog in an AD population; however, both the active HA group and the placebo group performed better than this expected decline. This may be explained by both of these groups essentially having very mild gain hearing aids, which may have provided some minimal benefit but no difference between the group because of the fitting issues discussed above.

In summary, the findings from these papers should not deter families or providers from seeking to identify and treat ARHL in persons with ADRD. The potential benefit of HA treatment will come from an evidence-based HA fitting procedure that ensures that audibility is returned to the individual followed by an aural rehabilitation program that assists the patient and family/caregivers in understanding communication strategies and the reality of adjustment to routinely using amplification.

Sara K. Mamo1 and Catherine V. Palmer2, 3,4
1Department of Communication Disorders, University of Massachusetts Amherst
2Department of Communication Science and Disorders, University of Pittsburgh
3Department of Otolaryngology, University of Pittsburgh
4University of Pittsburgh Medical Center (UPMC)

[1] Adrait A, Perrot X, Nguyen MF, Gueugnon M, Petitot C, Collet L, Roux A, Bonnefoy M; ADPHA study group (2017) Do hearing aids influence behavioral and psychological symptoms of dementia and quality of life in hearing impaired Alzheimer's disease patients and their caregivers. J Alzheimers Dis 58, 109-121.
[2] Nguyen MF, Bonnefoy M, Adrait A, Gueugnon M, Petitot C, Collet L, Roux A, Perrot X; ADPHA study group (2017) Efficacy of hearing aids on the cognitive status of patients with Alzheimer's disease and hearing loss: A multicenter controlled randomized trial. J Alzheimers Dis 58, 123-137.
[3] Sanders J, Stoody T, Weber J, Mueller HG (2015) Manufacturers' NAL-NL2 fittings fail real-ear verification. Hear Rev 21, 24.
[4] Taubman L, Palmer C, Durrant J, Pratt S (1999) Accuracy of hearing aid use time as reported by experienced hearing aid wearers. Ear Hear 20, 299-305.


Restoring Audibility in Alzheimer’s Disease Patients with Hearing Loss: a Worthwhile but Difficult Goal

We thank S. Mamo and C. Palmer for their interest in our companion papers, assessing cognitive and psychobehavioral outcomes in hearing-impaired Alzheimer’s disease (AD) patients rehabilitated with bilateral hearing aids (HAs) [1,2]. In their letter, they stated that the inadequate procedure used to fit hearing aids—with a lack of “evidence-based approach”—could explain by itself alone the negative results reported in our clinical trial. As clinicians and researchers, it is very stimulating to see colleagues raising such an important issue. However, the position defended by S. Mamo and C. Palmer seems to us to be a very theoretical approach to the problem, far from the reality on the ground.

First of all, we would like to emphasize two key points about our study. On the one hand, our trial was conducted in AD patients, which resulted in greater practical constraints than in trials in non-demented elderly patients [3]. On the other hand, for the care management of these patients, we chose to favor a pragmatic “ecological” approach rather than an explanatory “laboratory” approach [4]. However, it should be noted that this choice was not made at the expense of the internal validity of our trial, the methodological soundness of which was acknowledged in a recent Cochrane systematic review [5]. In practice, this has resulted in necessary adaptations of standard audiological procedures. For hearing tests, we have refrained from doing a speech audiometry, in order to avoid introducing potential evaluation bias between patients with heterogeneous cognitive status, some of them being likely to achieve this test, others not [6]. For HA fitting procedure, we decided to limit assessment to simple tests that were feasible and/or acceptable by all AD patients at a mild to moderate stage [3]. During the third week, a free-field pure tone audiometry, without and with HAs, allowed the hearing care professional to adjust the gain of amplification with the patient, with particular emphasis on facilitating acceptance of HAs.

Regarding the gain rule used for amplification, as indicated in one of our papers, we were ‘based on’ a proprietary fitting algorithm, specifically designed to provide optimal fitting for Phonak HAs at the time, and ‘derived from’, but not similar to, the National Acoustic Laboratories’ nonlinear fitting method version 1 (NAL-NL1). Thus, this algorithm cannot be assimilated to a generic method, such as NAL-NL1 or NAL-NL2 [7]. Therefore, the findings of Sanders et al. [8], on which S. Mamo and C. Palmer were based to assert a “lack of increased audibility”, could not be directly applied to our study. Moreover, it is interesting to notice that in Sanders et al.’s study [8], only three over five manufacturer’s NAL-NL2 fitting failed real-ear verification and that these results were not confronted to real speech understanding measurements.

As for the “lack of audibility difference” between active and placebo groups, claimed by S. Mamo and C. Palmer given a supposed identical gain of amplification for both experimental groups, we disagree with this false reasoning. Indeed, amplification for placebo HAs was intended to compensate for the occlusion effect only (30 dB on average), whereas amplification for active HAs was intended to compensate for the individual hearing loss measured through pure-tone audiometry (50 dB on average). Thus, a basic difference in gain prescription existed per se and remained after HA fitting, regardless of the method used. Similarly, use this alleged lack of audibility/gain difference between groups to explain the negative results for our primary outcome measure (ADAS-Cog score) represents a speculative hypothesis, which is not based on any tangible argument.

Concerning the question of measurement of HAs use time, we have voluntarily decided not to use data logging, since this technology did not include actimeter. Indeed, the data logging feature gave only basic information (on-mode/off-mode status), without any possibility to check the actual use of HAs [9]. In other words, we could not have distinguished HAs turned on and actually worn by the patient from HAs turned on but forgotten on a table. This lack of specificity of data logging precluded its use for AD patients. We have rather used patient and caregiver reports, knowing that when the data were discordant between the patient and the caregiver (most often, the spouse), we retained the value indicated by the caregiver. As for a potential overestimation of HA use time, there was no a priori reason that this risk was different between the active and placebo groups. Besides, it is interesting to note that the objective and subjective measures of HA use time are usually highly correlated, suggesting a relative concordance between these two measures [10].

As for the advocacy of S. Mamo and C. Palmer to encourage us to employ evidence-based procedures for fitting and verifying HAs in AD patients, we remain cautious. On the one hand, to assume that procedures applicable in the hearing-impaired non-demented elderly patient are equally applicable in the hearing-impaired demented patient remains to be demonstrated [3,6]. On the other hand, it sounds somehow unrealistic to advocate for real-ear measurements of HAs gain in AD patients, whereas this type of measurements is currently performed on a routine basis only by less than half of hearing care professionals [11].

Finally, we agree with the concluding remarks of S. Mamo and C. Palmer, in the sense that the provision of hearing aids for hearing-impaired demented patients is not sufficient and that aural rehabilitation intervention will need to be optimized in order to have reasonable chance to be effective. In this perspective, we are convinced that care at an early stage of the disease, field experience of the hearing care professional and involvement of family or caregivers will be the three most important key factors for successful care.

Xavier Perrot1,2,3, on behalf of the ADPHA study group4
1Department of Audiology, Edouard Herriot Hospital, Hospices civils de Lyon, Lyon, France
2University Claude Bernard Lyon 1, Faculty of medicine Lyon Sud, Pierre Bénite, France
3Institute of Sciences and Techniques for Rehabilitation (ISTR), Lyon 1 University, University of Lyon, Lyon, France
4Marie-France Nguyen, Arnaud Adrait, Marine Gueugnon, Charles Petitot, Adeline Roux, Marc Bonnefoy, Lionel Collet and the ‘Alzheimer Disease, Presbycusis and Hearing Aids’ study group

[1] Adrait A, Perrot X, Nguyen MF, Gueugnon M, Petitot C, Collet L, Roux A, Bonnefoy M; ADPHA study group (2017) Do hearing aids influence behavioral and psychological symptoms of dementia and quality of life in hearing impaired Alzheimer’s disease patients and their caregivers. J Alzheimers Dis 58, 109-121.
[2] Nguyen MF, Bonnefoy M, Adrait A, Gueugnon M, Petitot C, Collet L, Roux A, Perrot X; ADPHA study group (2017) Efficacy of hearing aids on the cognitive status of patients with Alzheimer’s disease and hearing loss: A multicenter controlled randomized trial. J Alzheimers Dis 58, 123-137.
[3] Lemke U (2011) Hearing impairment in dementia – how to reconcile two intertwined challenges in diagnostic screening. Audiol Res 1, e15.
[4] Godwin M, Ruhland L, Casson I, MacDonald S, Delva D, Birtwhistle R, Lam M, Seguin R (2003) Pragmatic controlled clinical trials in primary care: the struggle between external and internal validity. BMC Med Res Methodol 3, 28.
[5] Ferguson MA, Kitterick PT, Chong LY, Edmondson-Jones M, Barker F, Hoare DJ (2017) Hearing aids for mild to moderate hearing loss in adults. Cochrane Database Syst Rev 9, CD012023.
[6] Burkhalter CL, Allen RS, Skaar DC, Crittenden J, Burgio LD (2009) Examining the effectiveness of traditional audiological assessments for nursing home residents with dementia-related behaviors. J Am Acad Audiol 20, 529–538.
[7] Keidser G, Brew C, Peck A (2003) Proprietary fitting algorithms compared with one another and with generic formulas. Hear J 56, 28–32.
[8] Sanders J, Stoody T, Weber J, Mueller HG (2015) Manufacturers’ NAL-NL2 fittings fail real-ear verification. Hear Rev 21, 24–30.
[9] Gaffney P (2008) Reported hearing aid use versus datalogging in a VA population. Hear Rev 15, 42.
[10] Solheim J, Hickson L (2017) Hearing aid use in the elderly as measured by datalogging and self-report. Int J Audiol 56, 472-479.
[11] Jorgensen LE (2016) Verification and validation of hearing aids: Opportunity not an obstacle. J Otol 11, 57–62.