The authors investigated the therapeutic effect and safety of tyndallized Lactobacillus rhamnosus (RHT 3201) in Korean children with atopic dermatitis(AD) by a randomized trial. They concluded that oral administration of RHT 3201 showed the therapeutic effect on children with moderate AD, and the effects correlated with decrement of eosinophil cationic protein(ECP) and interleukin (IL)-31. These results seem to be interesting, however, there are some serious problems in this manuscript. Please see below.

Major comments:

1 The authors should show the result of efficacy analysis in intention to treat population, in addition to the per-protocol set analysis.

2 When the authors performed sub-group analysis, they should show demographic characteristics of the sub-groups, and explain the reason why they select the group, “subjects with AD for more than 50 months”.

3 Topical steroids and emollients are crucial for the treatment of AD and may affect the improvement of symptoms including SCORAD score. The authors said minimal use of topical steroids and emollients were permitted in this study (line 215-217), however, concrete protocol should be described in detail.

The total days and rank of topical steroid and the total days and amount of use of emollient were not shown in RESULTS. (line 149-153)
There was no explanation about the statistical analysis with adjustment for the amount of emollient and topical steroid used in RESULTS. (line 245-246)

4 In this study, the authors measured ECP levels, eosinophil counts, and IL-31 levels as immunologic markers for AD. They found the reduction of these markers in RHT 3201 group with the improvement of AD symptoms. However, these markers were increased in the placebo group, although the SCORAD score was significantly decreased in both groups. The authors should mention this point in discussion.

5 (line 178-181) For safety analysis, the details of adverse events in each group should be shown in a table. Moreover, the authors should describe the reason why they determine those events as definitely not related to the study treatments.

Minor comments:

1 (line145) The results of subgroup analysis of total SCORAD were mentioned in line 137, which showed no profound reduction (-7.32) in the placebo group compared to the change (-8.37) in the analysis of total subjects.

2 (line 145) The objective SCODAD score of placebo group was no more reduced in subgroup analysis (-1.88) than in total subject analysis (-5.25).

3 (line 193) The authors misspell the term “TNF-α” as “TNF-a”.

4 (line 224 & 244) The authors misspell the term “IFN-γ” as “INF-γ”.

5 (line 236-242) Are there any previous reports that examined the therapeutic effects of heat-killed probiotics on AD in humans? If it exists, they should mention about the study.

6 (line 248) Does this “RHT” mean “RHT 3201”?

7 Table 2 is not cited in the text.

I really appreciate your revising the manuscript according to our comments.
It has been improved, however, I still have few minor comments to be considered.

1 (line 151-153) The extent of disease and intensity of disease were more profoundly reduced in subjects with AD for the RHT-3201-treated group and the extent of AD was significantly different in both groups (Fig 3e, 3f)

This sentence is confusing, and you need show data which is the basis of the description.
When you say “more profoundly reduced” in the first part of the sentence, you compare reductions in total subjects with those in subjects with AD for over 50 months duration, however, you did not show analysis results which compare the data of these subjects, such as -15.74 vs -19.91 and -1.97 vs -2.00.

In the latter part of the sentence, you compared changes of SCORAD score between RHT-3201 group and placebo group.
→“The extent of disease was significantly reduced in the RHT-3201-treated group (Fig 3e), but the reduction in disease intensity was not statistically different in both groups (Fig 3f)”.

2 (line 205-206) We also found that the changes in SCORAD total score and objective score were more profound by subgroup analysis in subjects with AD duration of > 50 months.

Same as above.

3 (line 218-219) the primary outcome was an improvement SCORAD score between the treated and control groups

→ the primary outcome was the difference of an improvement in SCORAD score between the treated and control groups

4 (line 271) INF-γ