Comments on abstract, title, references

The aim of the study did not add a measurable or well-defined objective. The authors just reported, "This study aimed to determine the role of ONSD measurement by US for diagnosis of high ICP in TBI patients."

Instead, they may have stated clearly how they assess it by diagnostic test accuracy, likelihood ratio. Or, they aimed at searching for the appropriate cut-off value in their population.

Comments on introduction/background

The authors provided a fair background. But, I could not find the gap clearly.

Paragraph 4 and 5 in the introduction section showed the advantages of US measurement of ONSD including its effective diagnostic role.

With an obscure gap in knowledge, the study could not address a clear research question.

Comments on methodology

The selection of patients seems to be appropriate (opportunity sampling). Inclusion and exclusion criteria appear to be adequate for the traumatic brain injury (TBI). The study participants are unique and including them specifically is a good point.

However, the unclear research question usually results in improper methodology.

The reliability of the study is an issue. As they missed the double-check and the bilateral eye measurements. The clinician seems to be an expert, but interobserver and intraobserver were not taken into consideration. He should take 2-3 measures on each eye and take the average or/and other expert measures.

The evaluation of the CT for ICP should not be considered the gold standard if the aim of the study is to measure sensitivity and specificity. One hour period between the ONSD measurement and the CT imaging may affect the results due to the usual acute changes of ICP in such patients.

I could not identify the demographic and clinical data that were supposed to be collected from patients. Furthermore, I could not find the IRB or local ethical committee No. or the trial registry No.

The authors did not provide enough details for the replication of their study.

The sample size was not calculated and only a set of participants over a certain period of time.

Comments on data and results

The sample size is small for diagnostic test accuracy. The gold standard test (invasive ICP monitor) was used only in 22 cases.

They did not use the likelihood ratio which may be the best to show the clinical significance for each individual in the clinical life.

The unclear research question and methodology section resulted in a somewhat flawed results section.

Table titles had many abbreviations such as (ICP, ONSDUS, ONSDCT). I suggest avoiding appreciations in the titles of the tables.

In table 1 and table 2, I suggest separating the demographic and the measurement data.

I recommend avoiding the statistical analysis for demographic data according to SAMPL guidelines, with the presentation of mean and standard deviation as mean (SD).

It is unnecessary to present data with both mean and median, especially as the statistical analysis was done with a t-test.

The cut-off value of the increased ICP with Us was 0.55 cm according to the authors; however, this was the mean value in the negative group by the CT criteria. Therefore, this cut value is questionable. The statistical approach may need revision as regards this point. (I do not have the enough experience in such statistical methods to check)

When the invasive ICP measure used, which is the gold standard, the cut-off value increased to 0.61 cm for ONSDUS. But extracting such value from only 22 cases is highly questionable.

The repetition between the texts and the tables is evident; most, if not all, data is illustrated twice in the table and in the text.

I suggest adding the clinical outcome or the severity of the cases along with the measurements to ensure the clinical applicability.

Comments on discussion and conclusions

The aim-results correlation is not clearly stated in the discussion.

They presented a general background that should be presented in the introduction, not in the discussion section. Paragraphs 1 and 2 are background information to be presented in the introduction section NOT in the discussion section.

The authors reported the relevant studies adequately. However, they replicated the same major limitation - the small sample size -, which is a key component when examining test accuracy.

The explanation of similar results with the small studies is adequate.

The limitations are major but not fatal. They seem not to add risks to the patients by the new test, but the conclusion is still not supported by a reliable methodology or adequate sample size.

The authors may have added a point as regards future research (to categorize the patients as moderate and severe TBI) and this is a good point.

I think that conclusion might be supported by the references, but I can't find that the study added to the current knowledge.