Diabetes-Associated Changes in Cortical Auditory-Evoked Potentials in Relation to Normal Aging

Abstract

Objectives: (1) To characterize the influence of type 2 diabetes mellitus (DM) on cortical auditory-evoked potentials (CAEPs) separate from the effects of normal aging, and (2) to determine whether the disease-related effects are modified by insulin dependence.

Design: A cross-sectional study was conducted in a large cohort of Veterans to investigate the relationships among type 2 DM, age, and CAEPs in randomly selected participants with (N = 108) and without (N = 114) the disease and who had no more than a moderate hearing loss. Participants with DM were classified as insulin-dependent (IDDM, N = 47) or noninsulin-dependent (NIDDM, N = 61). Other DM measures included concurrent serum glucose, HbA1c, and duration of disease. CAEPs were evoked using a passive homogeneous paradigm (single repeating stimulus) by suprathreshold tones presented to the right ear, left ear, or both ears. Outcome measures were adjusted for the pure-tone threshold average for frequencies of 0.5, 1, and 2 kHz and analyzed for differences in age effects between participant groups using multiple regression.

Results: There is little variation across test ear conditions (left, right, binaural) on any CAEP peak in any of the groups. Among no-DM controls, P2 latency increases about 9 msec per decade of life. DM is associated with an additional delay in the P2 latency of 7 and 9 msec for the IDDM and NIDDM groups, respectively. Moreover, the slope of the function relating P2 latency with age is similar across participant groups and thus the DM effect appears constant across age. Effects on N1 latency are considerably weaker, with age effects of less than 4 msec per decade across all groups, and DM effects of only 2 (IDDM) or 3 msec (NIDDM). In the NIDDM group, the slope relating N1 latency to age is steeper relative to that observed for the no-DM group, providing some evidence of accelerated "aging" for this CAEP peak. DM does not substantially reduce N1-P2 amplitude and age relationships with N1-P2 amplitude are effectively absent. There is no association between pure-tone average at 0.5, 1, and 2 kHz and any aspect of CAEPs in this cohort.

Conclusions: In a large cohort of Veterans, we found that type 2 DM is associated with prolonged N1 and P2 latencies regardless of whether insulin is required to manage the disease and independent of peripheral hearing thresholds. The DM-related effects on CAEP latencies are threefold greater for P2 compared with N1, and there is little support that at the cortical level, IDDM participants had poorer responses compared with NIDDM participants, although their responses were more variable. Overall, these results indicate that DM is associated with slowed preattentive neural conduction. Moreover, the observed 7 to 9 msec P2 latency delay due to DM is substantial compared with normal age changes in P2, which are 9 msec per decade of life in this cohort. Results also suggest that whereas N1 latency changes with age are more pronounced among individuals with DM versus without DM, there was no evidence for more rapid aging of P2 among patients with DM. Thus, the damage responsible for the major DM-related differences may occur early in the DM disease process. These cross-sectional results should be verified using a longitudinal study design.

Figure 1

Theoretical models of possible effects of senescence and diabetes mellitus (DM).

Figure 2. Grand average CAEP waveforms can reveal major features of the data, but obscure important details of the sample

Grand average waveforms are shown for participants <50 years (dashed line) and <= 50 years (solid lines). Data are presented separately for binaural, left ear, and right ear presentation conditions (columns) and participant group (rows). The number of participants < 50 years is as follows: No DM=43, NIDDM=31, IDDM=21. The number of participants >=50 years is as follows: No DM=71, NIDDM=30, IDDM=26.

Figure 3. Raw data trend toward an increase in CAEP latencies with age and DM

Observed CAEP peak measures are plotted by age. Each CAEP metric is given in a separate row with presentation condition (binaural, left, and right ear) indicated by column. Participant group is denoted by symbol type and color as indicated in the figure. A loess smoothed curve is fit to the data for each participant group.

Figure 4. Raw CAEP data do not show a consistent trend with PTA

Observed CAEP peak measures are plotted here by PTA. Format is the same as for Fig. 4.

Figure 5. Summary of model results showing statistically adjusted, age trajectories for each DM group and the associated confidence interval

Model results overlay the observed data; the thick solid lines are the fitted mean response and open circles are the observed data. The shaded region is a 95% Bayesian confidence interval for the fitted mean, corresponding to the region within which we are 95% certain the true mean lies. The thin, solid lines show the 95% prediction intervals, corresponding to the region in which we expect a new CAEP measurement to lie. Note that confidence intervals are wider at the extremes of the age range where data are less plentiful, particularly for participants with DM.