Associations Between Neuropathy, Nephropathy and Hearing Loss in Individuals with Type 2 Diabetes

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

I read with great interest the manuscript entitled “Hearing Impairment Is Associated with Neuropathy but Not Nephropathy in Individuals with Type 2 Diabetes”.

This is a large, well-characterised cohort study (n=4,245) investigating associations between hearing loss (HL), diabetic neuropathy, nephropathy, and inflammatory markers in individuals with type 2 diabetes. The use of nationwide registry linkage, biomarker data, and polygenic risk score adjustment represents notable strengths. The statistical modelling is thorough and clearly structured.

The manuscript addresses a clinically relevant question and contributes meaningful data to an area with inconsistent prior findings. However, several methodological and interpretative issues require clarification before the conclusions can be fully supported.

 

Major Comments

1. HL is defined exclusively by ICD-10 diagnostic codes (pages 2–3, Outcome section). This approach captures symptomatic and clinically diagnosed HL only, likely representing more advanced cases. This has several implications: potential under-ascertainment of mild or subclinical HL, detection bias (patients with more comorbidities may have more healthcare contact), reduced comparability with prior studies using audiometry. The authors acknowledge this in the Discussion, but the limitation deserves stronger emphasis. Specifically: how many HL diagnoses preceded diabetes enrolment? Was timing between HL diagnosis and exposure assessment consistent? Could reverse causality (HL preceding neuropathy assessment) influence interpretation? Please: Clarify temporality more explicitly, provide median time from HL diagnosis to index date, discussi potential surveillance bias more critically.
2. Inflammatory markers were measured at enrolment (median 2.8 years before questionnaire), while HL was defined using diagnoses up to the index date (pages 3-4). This introduces temporal ambiguity: Some HL cases may have occurred before biomarker measurement; Neuropathy assessment (MNSI) occurred after some HL diagnoses; This is a major limitation in interpreting associations as biologically meaningful. I recommend: 1) Explicitly state proportion of HL diagnosed before enrolment; 2) Clarify chronological sequence in a diagram or table; 3) Temper causal language further.
3. Neuropathy was assessed by questionnaire only (self-reported MNSI, sensitivity 40%, specificity 92%). Issues: Low sensitivity → potential misclassification; No clinical confirmation; Questionnaire administered years after enrolment. Although acknowledged in limitations, the authors interpret neuropathy as a “key driver” of HL in T2DM (Discussion, page 9). This is too strong for cross-sectional data using self-reported neuropathy. Replace causal phrasing (“key driver”) with neutral association language. Discuss misclassification bias more carefully.
4. The sex interaction for neuropathy is statistically significant (P-interaction=0.020). However: Effect estimates overlap substantially; Absolute prevalence differences are not presented; No correction for multiple interaction testing. The Discussion attributes differences partly to occupational noise exposure (pages 9-10), but no noise exposure data are available. Please: present absolute predicted probabilities by sex; tone down speculative explanations, clarify whether interaction testing was prespecified.
5. Only the highest tertile of TNF-α remains associated with HL after full adjustment (Table 3, page 7). Concerns: association is modest (OR 1.40), no association in women in sex-stratified models, no consistent association across inflammatory markers.
6. Given these findings, the conclusion that TNF-α “underscores chronic low-grade inflammation as a potential contributor” is slightly overstated. Rephrase to reflect modest association, emphasize inconsistency across markers, clarify biological plausibility cautiously.

Minor Comments

1. Clarify whether UACR was single measurement or averaged.
2. LDL-C missingness is high (37.9%), provide sensitivity analysis without imputation.
3. Explain whether proportional odds assumption was considered (tertiles).
4. Figure 1 would benefit from clearer labeling of sample sizes in each subgroup.
5. Standardize reporting of p-values (<0.05 vs exact).

Author Response

Dear reviewer,

Thank you very much for your reading of our manuscript and for the thoughtful, constructive comments you provided. We truly appreciate the time and effort you have invested in helping us improve our work. The responses to all comments, along with the revised manuscript, are attached, and we hope that the changes address your concerns.

Best regards,

Joutiar Razay

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

This manuscript investigates the associations between symptomatic hearing loss (HL), neuropathy, and nephropathy in a large cohort of 4,245 individuals with Type 2 Diabetes Mellitus (T2DM). While the sample size is huge, the study suffers from significant methodological limitations regarding the definition of the primary outcome (hearing loss) and the temporal sequence of exposures. These limitations substantially weaken the validity of the findings. Unfortunately, the current version of the manuscript did not meet the requirement for publications.

Below are my major comments:

1. My first concern is that the authors conclude that HL is not associated with nephropathy. Although the authors acknowledged that the definition of hearing loss by ICD-10 21 codes was symptomatic, it does not necessarily explain the lack of relationship between nephropathy and HL. The ICD-10 inevitably misses subclinical or mild-to-moderate hearing loss, which is highly prevalent in T2DM. Prior studies suggested that nephropathy correlates with cochlear microangiopathy, which may manifest as early or mild threshold shifts (detectable by pure-tone audiometry) long before a patient seeks a diagnosis. By excluding subclinical (mild-to-moderate) HL, this study inevitably lost the sensitivity to detect the microvascular link. Thus, the null result is likely a Type II error driven by phenotype misclassification. The authors should be cautious to draw this conclusion and discuss this limitation. Also, I would suggest the authors remove “but Not 2 Nephropathy” from the title for rigorousness.
2. My second concern is the way to measure neuropathy in the study and its potential effect on the interpretation of the results. The authors used MNSI questionnaire as a measure of neuropathy, which is subjective and prone to be biased by their recall/perception, compared to objective clinical tests. Given the absence of pure tone audiometry, the strong correlation found between subjective HL and subjective neuropathy could be simply explained by "complaint-prone" behavior pattern or reporting bias rather than a strictly physiological link.
3. Based on the Line 88, the "index date" for neuropathy assessment was June 2016, while inflammatory markers were measured at enrollment (median 2.8 years prior). Notably, HL was identified by codes recorded up to the index date, as described in Line 98. Therefore, It is highly probable that for many subjects, the HL was diagnosed before the measurement of inflammatory markers or neuropathy. In this way, it is difficult to determine causal relationships among these symptoms.
4. The authors ran multiple logistic regression models for different exposures and subgroups. However, there is no correction for multiple comparisons, which could also potentially account for the significant findings, which might be a Type I error. Thus, the authors should apply an appropriate correction for multiple testing or rigorously justify the reasons for why decided not to perform such adjustments.
5. Line 240-242: The authors inferred that the dilute association between neuropathy and HL in men might be explained by greater noise exposure. However, the manuscript did not include any relevant measures for noise exposure. Thus, it is highly speculative and suspective to make such statement.
6. In the limitation section, the authors must explicitly state that their findings cannot be directly compared to PTA-based studies. The phenotype of "ICD-10 diagnosed HL" is fundamentally different from "audiometric threshold shift."

Author Response

Dear Reviewer,

Thank you very much for your reading of our manuscript and for the thoughtful and constructive comments you provided. We truly appreciate the time and effort you have invested in helping us improve our work. The responses to all comments, along with the revised manuscript, are attached, and we hope that the changes address your concerns.

Best regards,

Joutiar Razay

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Satisfied with authors answer

Author Response

1. Title modification:

Comment: The current title uses a definitive statement that may imply causality. You are required to revise the title to focus on the "Association" and place the exposure factors before the outcome, in accordance with observational study conventions. Suggested Title: "Association between neuropathy, nephropathy and Hearing Impairment in Individuals with Type 2 Diabetes" (or similar).

 

Response: We thank the editor for this helpful suggestion. In line with the recommendation, we have revised the title to emphasise associations rather than causality and to list the exposure factors before the outcome.

 

Action:

Title changed from “Hearing loss is associated with neuropathy in individuals with type 2 diabetes”

to

“Associations between neuropathy, nephropathy and hearing loss in individuals with type 2 diabetes”

2. (Supplementary Figure 1)

Comment: In the Methods section, you must provide a clearer chronological map.

 

Response: We thank the editor for this helpful suggestion. To provide a clearer chronological overview, we have revised the description in the Methods section. In addition, we have added clarifying text later in the appendix section (line 434).

 

Action:

Replaced (line 133-134)“The time from exposure measurements and diagnoses are shown in supplementary figure 1.”

With

(line 103-105): “The timing of the measurements of UACR, inflammatory biomarkers, and HL diagnoses relatively to the MNSI are visualized in Supplementary Figure 1.”

 

Added text (line 434): […](estimation of neuropathy).

 

 

3. Please explicitly state the proportion of HL cases diagnosed before or after the assessment of neuropathy and biomarkers.

 

Response: To already in the Method section state the proportion of HL cases diagnosed before or after the assessment of neuropathy and biomarkers we have moved forward the reference to Supplementary Figure 3. The presentation of the results are still provided in the Result section (lines 205-209).

Action:

We have in line 105 added following: ”... and specified in numbers in Supplementary Figure 3”

In addition, we have made the heading of Supplementary Figure 3 more clear:

Change from:

Supplementary Figure 3: Timing of hearing loss diagnoses in relation to DD2 enrolment and index date.

To:

Supplementary Figure 3: Timing of hearing loss diagnoses in relation to DD2 enrolment (measurement of inflammatory biomarkers) and index date (estimation of neuropathy).

 

 

 

 

4. Symptomatic HL was strongly

Comment: Avoid use "symptomatic HL was strongly"

Response:

We thank the editor for this comment regarding the wording of the abstract conclusion. To avoid over‑stating the strength of the association, we removed the word “strongly” completely.

 

Action:

Removed in the word “strongly” in line 33.

The conclusion in the abstract now reads “In subjects with T2DM, neuropathy was associated with symptomatic HL, and the association seemed to be stronger in females.”

 

 

 

 

 

 

5. Neutral, associative language

Comment: Throughout the Discussion, please replace causal phrasing (e.g., "leads to") with neutral, associative language (e.g., "is associated with," "may contribute to").

 

Reponse: We appreciate the editor’s reminder to avoid causal phrasing in the Discussion. As our study is cross-sectional and cannot establish causality, we carefully reviewed the Discussion and replaced any potentially causal wording with neutral, associative language.

 

 

6. exposure factors followed by the outcome

Comment: Restructure the manuscript to present the exposure factors first, followed by the outcome

Response: We thank the editor for this suggestion regarding the overall structure of the manuscript. We have now restructured the text so that the exposure factors are presented before the outcome throughout the manuscript, in line with this recommendation.

The following changes have been made:

 

Action:

• The conclusion in the abstract (lines 33–37) now presents the exposure factors first, followed by the outcome
• In the Methods section, the Outcome section (lines 107–115) has been moved to follow the description of the exposure factors.
• In the Results section (lines 211–214 and 233–234), the exposure factors are now presented before the outcome.
• In the Discussion (lines 259–264, 277, and 312–313), the exposure factors are now presented before the outcome.
• The Conclusion has also been revised accordingly (lines 375–380).

 

Author Response File: Author Response.docx