Ten-Year Voriconazole Susceptibility Trends Among 1482 Clinical Aspergillus fumigatus Isolates: A Compositionally Controlled Multicenter Analysis of the VIVLI/SENTRY Platform (2010–2020)

Round 1

Reviewer 1 Report

The manuscript titled "Ten-Year Evolution of Voriconazole Susceptibility in Aspergillus fumigatus: A Multicenter Analysis of the VIVLI/SENTRY Antifungal Surveillance Platform (2010–2020)" written by Wagner Lopez et al. brings into light important data about Voriconazole resistance in Aspergillus fumigatus. As this is a pathogen with high significance in clinical infections, and species that are resistant to voriconazole are emerging, the article contributes greatly to raise awareness on the evolution of this resistance patterns as well as the geographical regions that are mostly affected.

The data is summarized and analyzed really well. The author's draw some important conclusions after the processing of a large number of entries. I also appreciated that the author's did not exclude entries due to partially missing data, if they still had enough information to include it in the study.

The manuscript is well-written and detailed. I only have two minor suggestions which I have detailed below.

• genera and species names should be written in italic. The lines are not numbered in the manuscript so I can't point specific examples, but in the whole manuscript the names of the microorganisms are not marked in italic. Please check and correct everywhere
• the tables are sometimes hard to follow because the entries are not delimited in the tables, but this could be because of the formatting required from the manuscript. If possible, please correct

Author Response

• Comment 1: genera and species names should be written in italic. The lines are not numbered in the manuscript so I can't point specific examples, but in the whole manuscript the names of the microorganisms are not marked in italic. Please check and correct everywhere
• Response 1: We thank the Reviewer for their comment. We have made a full correction to the document.

 

• Comment 2: the tables are sometimes hard to follow because the entries are not delimited in the tables, but this could be because of the formatting required from the manuscript. If possible, please correct
• Response 2: We thank the Reviewer for this helpful observation. We agree that the tables were difficult to follow and this was a formatting issue on our part rather than a constraint of the journal template. We have reformatted all tables following the MDPI style (horizontal rules only). We have added cell padding to clearly separate adjacent entries, inserted a top rule a rule below the header row and a bottom rule. A light horizontal rule has also been added between variable blocks in Table 2 to visually delimitate each grouping variable and its categories. Header rows are now set to repeat when a table spans a page break. We hope the revised tables are easier to read.

Author Response File: Author Response.pdf

Reviewer 2 Report

This is an observational study characterizing the evolution of voriconazole susceptibility in *A. fumigatus* isolates from the VIVLI/SENTRY ATLAS antifungal dataset (2010–2020); it also evaluates the stability of the demographic, geographic, and clinical specialty composition of the sampled isolates. The study is scientifically rigorous and the presentation is generally clear; however, there are some details that could be improved:

Title

Must include information on the sample composition

Materials and Methods
Clarify the meaning of (Family = “Aspergillus”)
Define all variables, including "clinical specialty"

Results
3.1. Study population
Specify which other *Aspergillus* species (*Aspergillus* spp.) are included; there are 120 isolates of other *Aspergillus* species, so it would be advisable to specify these species
Table 3: The authors include "clinical specialty" as a variable, but this was not mentioned in the variable definitions. The specific clinical specialties are not detailed in the Results section

Discussion
The study shows a trend toward reduced voriconazole susceptibility in *Aspergillus fumigatus*; however, the year 2011 stands out (Figure 1), showing a relatively marked decrease compared to the 2012–2013 period, for example. This aspect should be analyzed in the context of the sample composition
Were there no resistant isolates in 2011 and 2015? What is the possible explanation for this regarding the sample composition?

Write scientific names in italics

Author Response

Dear Editor & Reviewers

Thank you very much for your valuable feedback and constructive comments during the review process. We have carefully addressed each of your suggestions and incorporated the necessary revisions into the manuscript, which has significantly enhanced the overall quality and clarity of the work.

Comment 1 — The title must include information on the sample composition

Response. We agree, and the title now states both the source population and the analytic sample. It also states the study's actual conclusion rather than the crude observation:

Ten-Year Voriconazole Susceptibility Trends among 1,482 Clinical Aspergillus fumigatus Isolates: A Compositionally Controlled Multicenter Analysis of the VIVLI/SENTRY Platform (2010–2020)

Changes: Title page.

Comment 2 — Methods: clarify the meaning of (Family = "Aspergillus")

Response. "Family" is the name of a grouping field in the ATLAS Antifungals export, not a taxonomic family in the formal Linnaean sense; in this dataset it corresponds to the genus-level organism group. We have clarified this in Section 2.2 and in the footnote to Table 1:

"…2,111 were selected using the Family field of the ATLAS export. In this data dictionary, Family denotes the genus-level organism grouping rather than a taxonomic family in the formal Linnaean sense: Family = “Aspergillus” therefore selects all isolates identified to the genus Aspergillus, irrespective of species, and does not denote the taxonomic family Aspergillaceae."

Changes: Section 2.2, first paragraph; footnote to Table 1; in-line comment in script S1.

Comment 3 — Methods: define all variables, including "clinical specialty"

Response. The Reviewer is right: clinical specialty was tested in Table 3 but not defined. Section 2.3 has been rewritten so that every variable appearing in any table is defined there — sex, age group, study period, species, specimen source, geographic region, clinical specialty, and the voriconazole CLSI interpretation. The specialty definition now reads:

“Clinical specialty: taken from the Speciality field of the source dataset (the hospital service submitting the isolate) and grouped into Internal Medicine; Cardiothoracic/Pulmonary; Critical Care (intensive care unit); Hematology/Oncology & Transplant; Ambulatory/Outpatient; Surgery (general, neuro-, orthopaedic, trauma, burn, urological and obstetric/gynaecological); Pediatrics/Neonatology; Other; and Unknown/Missing. Grouping was required because the raw field contains 29 non-missing levels, of which 55% of cells had expected counts below 5, and because it contains duplicate levels differing only in spelling (“Ear, Nose, Throat”/“Ear, Nose, Throat (Otolaryngology)”; “Long Term Care”/“Long-Term Care”), which were merged.”

As set out above, the grouping is not merely cosmetic: it was required for the chi-square test to be valid. The rationale, including the merging of duplicate spellings in the raw field, is stated explicitly in Section 2.3.

Changes: Section 2.3 (rewritten); Table 3; script S1.

Comment 4 — Results 3.1: specify which species are included in "other Aspergillus spp." (n = 120)

Response. We have disaggregated this category. The 120 isolates comprised 23 taxa, most frequently A. nidulans (n = 48, including the species complex), A. ustus (n = 14, including the species complex), A. versicolor (n = 11), A. lentulus (n = 7), A. sydowii (n = 7) and A. tubingensis (n = 7). The complete listing is provided as Supplementary Table S2 and summarised in Section 3.1 and in a footnote to Table 2.

We have also drawn out a point the Reviewer's question exposes: this group contains cryptic species with intrinsically reduced azole susceptibility — A. lentulus, A. thermomutatus and A. ustus — none of which could be evaluated here, because the CLSI voriconazole interpretation is missing for 100% of non-fumigatus isolates. We now note this in the Discussion as an argument for extending epidemiological cutoff values to these species.

Changes: Section 3.1; footnote to Table 2; new Supplementary Table S2; Discussion (limitations).

Comment 5 — Table 3: "clinical specialty" is not defined, and the specialties are not detailed in the Results

Response. Both points are addressed. The definition is now in Section 2.3 (Comment 3 above). In addition, the distributions of clinical specialty and of specimen source — both tested in Table 3 but previously not displayed — have been added as new rows in Table 2, so that every variable tested in Table 3 is described descriptively beforehand. Section 3.1 now reports the specialty distribution, and Section 3.2 reports the direction of the shift:

"The 29 non-missing clinical-specialty categories were Internal Medicine; Cardiothoracic/Pulmonary; Intensive Care Unit; Ambulatory/Outpatient; Hematology/Oncology; Pediatrics/Neonate; Surgery; Infectious Disease; Ear, Nose, Throat; Neurology; Transplant; General/GI; Geriatrics; Neurosurgery; Family Practice; Ear, Nose, Throat (Otolaryngology); Emergency; Renal; Orthopedics; Obstetrics/Gynecology; Burn; Ophthalmology; Rehabilitation; Trauma; Dialysis; Dermatology; Long Term Care; Long-Term Care; and Urology. The two ear, nose and throat labels and the two long-term-care labels were retained as distinct source categories, as recorded in the dataset."

Changes: Section 2.3; Table 2 (new specialty and specimen-source blocks); Sections 3.1 and 3.2.

Comment 6 — Discussion: the 2011 point stands out relative to 2012–2013; analyse it in the context of the sample composition

Response. We are grateful for this comment, which turned out to be the thread that unravelled the paper's original conclusion. The 2011 estimate rests on 48 isolates, and the apparent drop corresponds to four non-susceptible isolates, all Intermediate, with no resistant isolate recorded that year. Its Wilson interval (80.45–96.71%) overlaps those of 2010, 2012 and 2013 entirely, so the point is not statistically distinguishable from its neighbours.

Examining the composition, as the Reviewer suggests, is revealing. That year's contributors were atypical: Brazil supplied 9 of 48 isolates (18.8%, against 1.3–5.0% in most other years) and Slovakia supplied 10, having contributed in 2010 and 2011 only and never again. We have added a Discussion paragraph making this point, and Figure 1 now distinguishes years with fewer than 100 isolates using open markers, with the legend noting that these estimates should be interpreted with caution.

Changes: Discussion, new paragraph; Figure 1 (open markers for n < 100) and legend; script S1

Comment 7 — Were there no resistant isolates in 2011 and 2015? Possible explanation in relation to sample composition

Response. That is correct: no isolate was categorised as Resistant in 2011 or 2015. The non-susceptible isolates in those years were all Intermediate (4/48 in 2011 and 9/262 in 2015). Two considerations explain this without invoking any biological change, and pursuing the second is what led to the reframing described at the start of this letter.

First, resistant isolates were rare throughout the first half of the series: only 5 were recorded in 2010–2015 as a whole (5/697, 0.72%). At that rate, observing zero in a given year is expected for 2011 (P ≈ 0.40 at n = 48) and unremarkable for 2015 (P ≈ 0.15 at n = 262). Of the 28 resistant isolates across the whole series, 19 (68%) occurred in 2018–2020.

Second, and more informatively, the countries that contribute resistant isolates entered the network late. Italy, the United Kingdom and Belgium — which together supply 16 of the 28 resistant isolates — first contributed in 2015, 2019 and 2018 respectively, and none was present in 2011. In 2015 Italy had only just joined and the other two had not; the nine non-susceptible isolates that year were all Intermediate, spread across seven countries, with no high-prevalence contributor represented. The zero counts are thus a direct consequence of who was in the network, which is precisely the mechanism the revised paper quantifies.

Changes: Discussion, new paragraph; new Section 3.4.

We hope the Reviewer agrees that the revised manuscript is substantially stronger for these comments. 

Author Response File: Author Response.pdf