Pre-registered controlled comparison of auditory function reveals no difference between hospitalised adults with and without COVID-19

Figures & data

Table 1. Summary of participant characteristics per participant group.

	Control group	COVID-19 group
N	40	57
Median age (and IQR), in years	57.5 (20.5)	58 (21)
Gender (female/male/other)	18/22/0	20/37/0
Mean BMI (and SD)	29.3 (6.8)	31.6 (6.3)
Majority ethnic group	White (95%)	White (81%)
Mean time since hospital admission (and SD), in months	9.1 (6.0)	10.7 (3.0)

Table 2. Summary of outcome measures and their basic characteristics.

Hypothesis ID	Measure	Basic characteristics
Cochlear function
1	Standard-frequency pure-tone audiometry (PTA) thresholds	Mean of thresholds at 0.25 to 8 kHz
2	Extended high-frequency (EHF) audiometry thresholds	Mean of 12.5 and 16 kHz thresholds
3	Standard-frequency distortion product otoacoustic emission (DPOAE) amplitudes	Mean of amplitudes at 0.5 to 8 kHz
4	EHF DPOAE amplitude	Amplitude at 10 kHz
Neural function
5	Acoustic reflex threshold (ART)	Threshold for broadband (BB) noise elicitor using 226 Hz probe tone
6	Auditory brainstem response (ABR) wave I amplitude	Peak-trough amplitude
7	ABR wave I–V inter-peak interval	Interval between wave I peak and wave V peak
Auditory perception
8	Digits-in-noise (DiN) signal-to-noise ratio for 71% correct (SNR71%)	Monaural threshold for identification of digits in speech-shaped noise
9	Speech, Spatial and Qualities of Hearing scale (SSQ12) change score	Sum of the 12 change scores
10	Tinnitus change score	Binary outcome: tinnitus onset/worsened vs tinnitus stable/absent

Table 3. Summary of the number of participants contributing data in each test and the number of participants who completed each questionnaire.

Measure	Participant group	Data from both ears	Data from one ear only	Total N
PTA (standard frequencies)	Control	40	0	40
	COVID-19	57	0	57
PTA (EHF)	Control	40	0	40
	COVID-19	57	0	57
DPOAE (standard frequencies)	Control	39	1	40
	COVID-19	52	4	56
DPOAE (EHF)	Control	39	1	40
	COVID-19	52	4	56
ART	Control	36	3	39
	COVID-19	53	3	56
ABR, wave I amplitude	Control	28	7	35
	COVID-19	36	18	54
ABR, wave I–V latency	Control	28	7	35
	COVID-19	35	19	54
Digits-in-noise	Control	40	0	40
	COVID-19	57	0	57
Questionnaires
DHI	Control			15
	COVID-19			23
FAS	Control			40
	COVID-19			57
HHIA	Control			18
	COVID-19			26
IIH	Control			40
	COVID-19			57
SSQ12	Control			40
	COVID-19			57
THI	Control			9
	COVID-19			8

Table 4. Model summaries for Hypotheses 1 to 9.

Hypothesis	Parameter	Sum of squares	df	Mean square	F	p	ηp^2
1. PTA (standard)	(Intercept)	257.05	1	257.05	2.61	0.110	0.03
	Gender	8.61	1	8.61	0.09	0.768	0.00
	Age (years)	3911.49	1	3911.49	39.66	<0.001	0.30
	Length of stay in hospital	3.06	1	3.06	0.03	0.860	0.00
	Participant group	151.22	1	151.22	1.53	0.219	0.02
	Residuals	9074.03	92	98.63
2. PTA (EHF)	(Intercept)	9478.43	1	9478.43	27.90	<0.001	0.23
	Gender	1217.99	1	1217.99	3.58	0.061	0.04
	Age (years)	53,194.40	1	53,194.40	156.57	<0.001	0.63
	Length of stay in hospital	931.72	1	931.72	2.74	0.101	0.03
	Participant group	96.39	1	96.39	0.28	0.596	0.00
	Residuals	31,257.49	92	339.76
3. DPOAE (Standard)	(Intercept)	321.25	1	321.25	19.00	<0.001	0.17
	Gender	2.05	1	2.05	0.12	0.729	0.00
	Age (years)	906.90	1	906.90	53.63	<0.001	0.37
	Length of stay in hospital	18.16	1	18.16	1.07	0.303	0.01
	Participant group	7.31	1	7.31	0.43	0.512	0.00
	Residuals	1538.95	91	16.91
4. DPOAE (EHF)	(Intercept)	5.63	1	5.63	0.23	0.635	0.00
	Gender	28.32	1	28.32	1.14	0.288	0.01
	Age (years)	954.35	1	954.35	38.54	<0.001	0.30
	Length of stay in hospital	1.40	1	1.40	0.06	0.812	0.00
	Participant group	6.59	1	6.59	0.27	0.607	0.00
	Residuals	2253.64	91	24.77
5. ART	(Intercept)	27,404.71	1	27,404.71	236.49	<0.001
	Gender	0.02	1	0.02	0	0.989	0
	Age (years)	947.79	1	947.79	8.18	0.005	0.08
	Length of stay in hospital	81.94	1	81.94	0.71	0.403	0.01
	Participant group	24.80	1	24.80	0.21	0.645	0
	Residuals	10,429.37	90	115.88
6. ABR, wave I amplitude	(Intercept)	887,135.37	1	887,135.37	137.88	<0.001	0.62
	Gender	63,888.93	1	63,888.93	9.93	0.002	0.11
	Age (years)	289,719.07	1	289,719.07	45.03	<0.001	0.35
	Length of stay in hospital	7670.80	1	7670.80	1.19	0.278	0.01
	Participant group	1291.07	1	1291.07	0.20	0.655	0.00
	Residuals	540,452.08	84	6433.95
7. ABR, wave I–V interval	(Intercept)	58.15	1	58.15	963.38	<0.001	0.92
	Gender	0.17	1	0.17	2.81	0.097	0.03
	Age (years)	0.52	1	0.52	8.64	0.004	0.09
	Length of stay in hospital	0.07	1	0.07	1.10	0.297	0.01
	Participant group	0.07	1	0.07	1.12	0.293	0.01
	Residuals	5.07	84	0.06
8. DiN	(Intercept)	1408.43	1	1408.43	567.35	<0.001	0.86
	Gender	5.26	1	5.26	2.12	0.149	0.02
	Age (years)	37.54	1	37.54	15.12	<0.001	0.14
	Length of stay in hospital	0.09	1	0.09	0.04	0.848	0
	Participant group	1.88	1	1.88	0.76	0.387	0.01
	Residuals	228.39	92	2.48
9. SSQ12	(Intercept)	28.89	1	28.89	2.34	0.13	0.03
	Gender	48.44	1	48.44	3.92	0.051	0.04
	Age (years)	5.57	1	5.57	0.45	0.504	0
	Length of stay in hospital	7.34	1	7.34	0.59	0.443	0.01
	Participant group	59.25	1	59.25	4.79	0.031	0.05
	Residuals	1125.56	91	12.37

Table 5. Model summary for Hypothesis 10, change in tinnitus.

Hypothesis	Parameter	Fit	B	z	p	β
10. Change in tinnitus	(Intercept)		−39.12	0	0.995	−39.37
	Gender		18.6	0	0.997	18.6
	Age (years)		−0.01	0	0.893	−0.07
	Length of stay in hospital		0.01	0	0.705	0.12
	Participant group		18.62	0	0.996	18.62
	AIC	35.18
	BIC	48.05
	Tjur’s R2	0.07
	Sigma	0.52
	Log loss	0.13

Figure 1. Panel A: Air conduction pure-tone thresholds. Grey lines and points represent individual participants. Bold, coloured lines show the means for each group at each frequency. Shaded ribbons around the bold lines show 1 SD from the mean. Panel B: Mean air conduction thresholds. Mean of standard (0.25–8 kHz) and extended high (12.5 kHz) frequencies. Jittered, coloured points show the raw data. Boxplot whiskers show 1.5 times the interquartile range. The hollow point inside boxplots shows the mean. Distribution curves show the probability density.

Figure 2. DPOAE levels: Means of standard (0.5 to 8 kHz) and extended high (10 kHz) frequencies. Jittered, coloured points show the raw data. Boxplot whiskers show 1.5 times the interquartile range. The hollow point inside boxplots shows the mean. Distribution curves show the probability density.

Figure 3. Mean ART. Jittered, coloured points show the raw data. Boxplot whiskers show 1.5 times the interquartile range. The hollow point inside boxplots shows the mean. Distribution curves show the probability density.

Figure 4. Panel A: Grand average waveforms for each group. Panel B: Wave I amplitude. Panel C: Wave I–V inter-peak interval. For panels B & C jittered, coloured points show the raw data. Boxplot whiskers show 1.5 times the interquartile range. The hollow point inside boxplots shows the mean. Distribution curves show the probability density.

Figure 5. SNR threshold for digits-in-noise test. Jittered, coloured points show the raw data. Boxplot whiskers show 1.5 times the interquartile range. The hollow point inside boxplots shows the mean. Distribution curves show the probability density.

Figure 6. Total change scores on the SSQ12 questionnaire. Positive scores indicate a worsening of experience since hospitalisation. The range on the y-axis represents the minimum and maximum total scores possible. Jittered, coloured points show the raw data. Boxplot whiskers show 1.5 times the interquartile range. The hollow point inside boxplots shows the mean. Distribution curves show the probability density.

Data availability statement

Supplementary analyses, materials, code for analysis and de-identified data are available in the online repository for the project, which can be found at https://osf.io/rc5fu/, DOI: 10.17605/OSF.IO/RC5FU.