Exploring the Impact of Diabetes Mellitus on Clinical Outcomes in Patients Following Severe Traumatic Brain Injury Using the TriNetX Database

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Many thanks for the opportunity to review the present study about the impact of DM on TBI patients. The strength of this work is the correlation between DM and outcome in a very large number of cases.
The manuscript is well written, the methodology and results sound. There are some issues that require attention before it can be considered for publication.
1) some sections of the manuscript are too long and provide redundant information, whereas other parts should be expanded (the many confounding factors that affect the results and might have introduced biases such as mechanism of TBI, politrauma vs isoltated head injury, etc.); please shorten it as possible and try to reduced confounding and biasing factors.
2) the authors should expand the reference list to capture articles regarding the challenges faced in major trauma centers dealing with high volume of patients and how a tight glucose control might improve outcomes
3) the authors should provide a possible theory justifying why there was reduced mortality in DM (regardless of the type) which seems counterintuitive.

Author Response

Comment 1: Some sections of the manuscript are too long and provide redundant information, whereas other parts should be expanded (the many confounding factors that affect the results and might have introduced biases such as mechanism of TBI, polytrauma vs isolated head injury, etc.); please shorten it as possible and try to reduced confounding and biasing factors.

Response 1: We thank the reviewer for their comments. We have further refined the manuscript to remove repetitive statements while still emphasizing the necessary points of our study. We have also more explicitly listed the type of traumatic brain injury as a potential confounding factor that could not be accounted for due to an inherent limitation of the TriNetX software on page 12 in lines 400-403.

Comment 2: The authors should expand the reference list to capture articles regarding the challenges faced in major trauma centers dealing with high volume of patients and how a tight glucose control might improve outcomes

Response 2: Regarding the second comment, we have revised the paragraph on page 11 in lines 351-360 to incorporate the recommendations.

Comment 3: The authors should provide a possible theory justifying why there was reduced mortality in DM (regardless of the type) which seems counterintuitive.

Response 3: Regarding the third comment, we have added several theories justifying the reduced mortality result on pages 8 and 9 in lines 197-227. We have also added a point in the limitations section on page 12 in lines 395-400 identifying our inability to stratify patients by the level of care received as another confounding factor that may have impacted the results.

Reviewer 2 Report

Comments and Suggestions for Authors

The authors of this study have conducted a large retrospective review of  patients with either form of DM (Type I and II) admitted to hospital with severe TBI and assessed their clinical outcomes and mortality. Their results indicate that patients with DM are more susceptible to a range of in-hospital complications such as cerebral infarction, seizure, sepsis and pneumonia, however the paradox is that patients with DM had a significantly lower mortality than non-DM patients. 

I would definitely recommend to highlight better the various limitations of this study, and I have listed here three action points that the authors should consider:   1) One aspect explaining this paradox between increased morbidity and reduced mortality might be the fact that the authors could not stratify patients per level of care received (e.g. Level I, II or III trauma center), hence we can't assess whether patients treated in Neurotrauma Services had better outcomes as suggested by Dasic et al. This is a critically important aspect that the author should highlight in their discussion. 
2) I certainly agree with the main take home message that a more aggressive management aimed at controlling glucose levels would possibly improve the clinical trajectory following TBI, however I would agree though that this should be the case regardless of the severity of the injury. In fact, authors should also highlight that not all severe TBI are the same and the type of injury (penetrating vs non penetrating, space occupying post-traumatic lesions vs brain swelling, etc.) could have introduced biases in the analysis.    3) Finally, the authors should also mention that each of the complications included as outcome measures can increase glycemic levels even in non-DM patients (Hirawasa et al., Xiu et al.), nonetheless such limitations are mostly compensated by the fact that the authors have dichotomized patients according to their DM status and DM subgroups, an approach that makes this work even more relevant.    Ref:  - Dasic, D.; Morgan, L.; Panezai, A.; Syrmos, N.; Ligarotti, G.K.I.; Zaed, I.; et al.  A scoping review on the challenges, improvement programs, and relevant output metrics for neurotrauma services in major trauma centers. Surg Neurol Int. 2022, 13, 171. doi: 10.25259/SNI_203_2022
- Hirasawa, H.; Oda, S.; Nakamura, M. Blood glucose control in patients with severe sepsis and septic shock. World J Gastroenterol. 2009, 15(33), 4132-6. doi: 10.3748/wjg.15.4132
- Xiu, F.; Stanojcic, M.; Diao, L.; Jeschke, M.G. Stress hyperglycemia, insulin treatment, and innate immune cells. Int J Endocrinol. 2014, 2014, 486403. doi: 10.1155/2014/486403

Author Response

Comment 1: One aspect explaining this paradox between increased morbidity and reduced mortality might be the fact that the authors could not stratify patients per level of care received (e.g. Level I, II or III trauma center), hence we can't assess whether patients treated in Neurotrauma Services had better outcomes as suggested by Dasic et al. This is a critically important aspect that the author should highlight in their discussion. 

Response 1: We thank the reviewer for their comments. We agree with their first comment and have more explicitly listed, on page 12 in lines 395-400, our inability to stratify patients by the level of care received as another confounding factor that may have contributed to the paradoxical results.

Comment 2: I certainly agree with the main take home message that a more aggressive management aimed at controlling glucose levels would possibly improve the clinical trajectory following TBI, however I would agree though that this should be the case regardless of the severity of the injury. In fact, authors should also highlight that not all severe TBI are the same and the type of injury (penetrating vs non penetrating, space occupying post-traumatic lesions vs brain swelling, etc.) could have introduced biases in the analysis.

Response 2: We also agree with the second comment and have more explicitly stated how the type of traumatic brain injury is another potential confounding factor that could not be accounted for due to an inherent limitation of the TriNetX software on page 12 in lines 400-403.

Comment 3: Finally, the authors should also mention that each of the complications included as outcome measures can increase glycemic levels even in non-DM patients (Hirawasa et al., Xiu et al.), nonetheless such limitations are mostly compensated by the fact that the authors have dichotomized patients according to their DM status and DM subgroups, an approach that makes this work even more relevant.

Response 3: Regarding the third comment, we have included the recommended point on page 12 in lines 403-406.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

In this revised version the authors answered sufficiently to the reviewers raised points, criticisms and suggestions. The study has been improved and it could now be reconsidered for publication.