Healthcare-Associated Infections, Antibiotic Use, and Invasive Devices: A Repeated Point Prevalence Survey

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

1. The total number of HAI cases across all three survey periods is only 17 (4, 10, and 3 cases respectively). This is too small to support meaningful statistical analysis, and the wide confidence intervals in Table 5 (e.g., OR 9.20, 95% CI 3.37-25.11) reflect this limitation. The authors should clearly acknowledge this and avoid drawing strong conclusions from these numbers.
2. The ward composition changed across survey periods, with Cardiology and Pediatric Surgery units added in P2 and P3 but absent in P1. This affects the comparability of HAI prevalence across periods. The authors should discuss how this structural difference may have influenced the observed trends, particularly the peak in P2.
3. The reduction in “unknown” McCabe scores is interpreted as an improvement in clinical documentation quality. However, this change could also reflect differences in how data collectors filled in the forms across survey periods rather than a real improvement in clinical practice. A more cautious interpretation is needed.
4. The overall antimicrobial use prevalence (~48-50%) is substantially higher than the European average of 35.5% reported by ECDC. The authors acknowledge this but do not provide a sufficient explanation. A more detailed discussion of potential reasons, such as case-mix, prescribing culture, or prophylaxis practices, is needed.
5. The increase in glycopeptide use from 1.5% in P1 to 7.8% in P3 is a notable finding that deserves more discussion. The authors briefly mention coverage for resistant Gram-positive organisms, but the clinical and stewardship implications of this trend should be addressed more carefully.
6. The authors describe temporal trends in HAI prevalence and antimicrobial use across three time points, but the study is fundamentally cross-sectional at each point. The discussion and conclusion sections should more consistently reflect this limitation and avoid language that implies causal or longitudinal inference.
7. This is a single-center study with a relatively small patient population. The conclusions should be toned down accordingly, and statements about generalizability or reproducibility of the model should be qualified.
8. Figure 2 uses red and blue colors to distinguish McCabe score categories. These colors may be difficult to distinguish in grayscale printing. The authors should consider using patterns or higher-contrast color combinations in line with the journal's requirements.

Author Response

Comments and Suggestions for Authors

Comment 1. The total number of HAI cases across all three survey periods is only 17 (4, 10, and 3 cases respectively). This is too small to support meaningful statistical analysis, and the wide confidence intervals in Table 5 (e.g., OR 9.20, 95% CI 3.37-25.11) reflect this limitation. The authors should clearly acknowledge this and avoid drawing strong conclusions from these numbers.

Response to Comments. We thank the Reviewer for the constructive comment, which identified a crucial limitation of our study. We fully acknowledge that the total of 17 HAIs (4 in P1, 10 in P2, 3 in P3) limits the robustness of the statistical analysis, as demonstrated by the wide confidence intervals in Table 5:

In the Results section, lines 341-351, the text was changed as follows “To identify factors potentially associated with HAI acquisition, a univariate analysis was performed (Table 6). The presence of a CVC was more frequent among infect-ed than non-infected patients (58.8% vs. 13.4%; OR, 9.20; 95%CI, 3.37–25.11; p < 0.001), although the small number of registered HAIs (n = 17 across all surveys), thus our results should be interpreted as exploratory. UC use was also more common among infected patients, although not statistically significant (52.9% vs. 36.7%; OR, 1.94; 95%CI, 0.73–5.13; p = 0.20), possibly reflecting the widespread utilization of this device across the study population. Length of stay was markedly longer among patients with HAIs compared to those non-infected (mean+SD, 34.6±26.2 days vs. 9.9±13.1 days; p < 0.001), likely associated with our cross-sectional study design rather than a proven causal effect.”

In the Discussion section, on lines 458-461, the following text was added “The analysis of associated factors was conducted exclusively through univariate analysis due to the limited number of observed events (17 of total HAIs), as also confirmed by the wide confidence intervals observed.”

In the Conclusions section, on lines 471-478, the following text was added “Our serial PPS provided a comprehensive overview of the local epidemiology of HAIs and antimicrobial consumption in a tertiary-care University Hospital in Southern Italy, highlighting CVCs as a potential modifiable target for prevention bundles, along with clinical complexity and length of hospitalization, as exploratory associations with infection risk, derived from univariate analyses based on a limited number of HAI events. The integration of periodic epidemiological surveillance with AMS may be a useful approach for local clinical governance and patient safety. However, our results are exploratory and should be confirmed in larger prospective multicentric studies.”

 

Comment 2. The ward composition changed across survey periods, with Cardiology and Pediatric Surgery units added in P2 and P3 but absent in P1. This affects the comparability of HAI prevalence across periods. The authors should discuss how this structural difference may have influenced the observed trends, particularly the peak in P2.

Response to Comments 2. We thank the Reviewer for this important observation. We agree that ward composition differed across survey periods, as Cardiology and Pediatric Surgery units were included in P2 and P3 but not in P1, and this may have affected the crude HAI prevalence estimates. Because HAI prevalence varies substantially by ward specialty and case-mix, differences in ward distribution across surveys can influence apparent temporal trends and may partly explain the higher prevalence observed in P2 rather than reflecting a true increase in infection burden. We have now added this limitation to the Discussion and clarified that the observed peak in P2 should be interpreted in light of the changed ward structure and the absence of adjustment for ward-level case-mix.

In the Discussion section, on lines 456-469, the following text was added “The main limitation of this study is the cross-sectional nature of PPS, which does not allow causal inference nor calculation of incidence rates. Furthermore, the generalizability of our findings may be limited due to the monocentric nature of this study. The analysis of associated factors was conducted exclusively through univariate analysis due to the limited number of observed events (17 of total HAIs), as also confirmed by the wide confidence intervals observed. Moreover, ward composition varied over PPS, by introducing Cardiology and Pediatric Surgery in P2 and P3, thus influencing case-mix and patients’ complexity. However, the observed prevalence peak in P2 persisted in sensitivity analysis restricted to consistently monitored wards, suggesting that factors beyond ward composition may have contributed to the observed temporal variation, despite the absence of a statistically significant linear trend. Therefore, our results are exploratory, although the serial survey repetition might provide a more in-formative cross-sectional perspective than a single PPS, and might support the relevance of the observed trends for clinical planning and governance.”

 

Comment 3. The reduction in “unknown” McCabe scores is interpreted as an improvement in clinical documentation quality. However, this change could also reflect differences in how data collectors filled in the forms across survey periods rather than a real improvement in clinical practice. A more cautious interpretation is needed.

Response to Comment 3. We thank the Reviewer for this valuable comment. We agree that the reduction in “unknown” McCabe scores should be cautiously interpreted, as it may reflect not only improved data completeness but also differences in the way data collectors completed the forms across survey periods. We have revised the manuscript accordingly and now present this finding as a likely improvement in data completeness and coding consistency, rather than as a direct indicator of clinical practice change.

In the Discussion section, on lines 402-411, the following text was added “During our three-period PPS, a significant progressive reduction in “unknown” cases and in distribution of the McCabe score was described, suggesting an improvement in data completeness and a greater consistency in data collection procedures over time, although differences in form completion might contribute to this trend. Patients with poor prognostic categories were consistently concentrated in high-complexity wards (Oncology, General or Internal Medicine, Emergency Cardiac Surgery, and UTIPO), in line with the association between clinical severity and HAI risk already described in literature [5,25,29,30]. However, in our study this relationship is based on univariate analyses with a limited number of events and should therefore be interpreted as exploratory and hypothesis-generating rather than causal.”

 

Comment 4. The overall antimicrobial use prevalence (~48-50%) is substantially higher than the European average of 35.5% reported by ECDC. The authors acknowledge this but do not provide a sufficient explanation. A more detailed discussion of potential reasons, such as case-mix, prescribing culture, or prophylaxis practices, is needed.

Response to Comment 4. We thank the Reviewer for this important comment. We agree that the higher prevalence of antimicrobial use observed in our hospital, compared with the European average reported by the ECDC, deserves more detailed discussion. This finding may be explained at least in part by differences in case-mix, since our institution is a tertiary-level university hospital that manages complex patients, high-acuity wards, and a higher proportion of subjects exposed to invasive devices and severe comorbidities, all factors associated with a greater likelihood of antibiotic prescribing. In addition, local prescribing habits and the relatively frequent use of prophylaxis may have contributed to the observed prevalence. We have therefore expanded the Discussion to better contextualize these findings, emphasizing that antimicrobial use reflects both the complexity of the patient population and the local prescribing culture, and that antimicrobial stewardship strategies remain essential to further optimize prescribing.

In the Discussion section, on lines 420-428, the following text was added “The overall antimicrobial use prevalence (48–50%) observed in our hospital was higher than the European average reported by ECDC (35.5%) [7,35], likely because our tertiary-level institution manages complex patients, has several high-acuity wards, and a substantial proportion of subjects are exposed to invasive devices and severe comorbidities, which are all factors associated with a greater likelihood of antimicrobial prescribing. In addition, local prescribing habits and prophylactic practices, particularly in surgical and other high-risk settings, may have contributed to the observed prevalence. Therefore, these results should be interpreted according to hospital’s case-mix and institutional characteristics, while reinforcing the need for continued AMS efforts.”

 

Comment 5. The increase in glycopeptide use from 1.5% in P1 to 7.8% in P3 is a notable finding that deserves more discussion. The authors briefly mention coverage for resistant Gram-positive organisms, but the clinical and stewardship implications of this trend should be addressed more carefully.

Response to Comment 5. We thank the Reviewer for this important comment. We agree that the increase in glycopeptide use from P1 to P3 warrants closer attention. This pattern may reflect greater clinical complexity, the higher burden of device-related infections, and the need for coverage against resistant Gram-positive pathogens in selected cases. We have revised the Discussion to highlight both the clinical rationale for this trend and its stewardship implications, emphasizing the need to preserve glycopeptides for appropriate indications.

On lines 433-440, the following text was added “The progressive increase in glycopeptide use from P1 to P3 is a relevant finding, as these agents are generally reserved for suspected or confirmed resistant Gram-positive infections and device-related infections. In our setting, this trend may reflect the higher clinical complexity of hospitalized patients, the burden of invasive device-associated infections, and cautious empirical prescribing in selected high-risk scenarios. Nevertheless, glycopeptides are reserve antibiotics, and their increasing use should be interpreted carefully considering stewardship principles, to avoid unnecessary exposure and preserve their efficacy.”

 

Comment 6. The authors describe temporal trends in HAI prevalence and antimicrobial use across three time points, but the study is fundamentally cross-sectional at each point. The discussion and conclusion sections should more consistently reflect this limitation and avoid language that implies causal or longitudinal inference.

Response to Comment 6. We thank the Reviewer for this valuable observation. We fully agree that the design of the study is cross-sectional at each time point, and therefore does not support causal inference or longitudinal patient-level analysis. Accordingly, we have revised the Discussion and Conclusions to ensure that the wording consistently reflects the repeated cross-sectional nature of the surveys, and we have removed or softened statements that could be interpreted as implying causality. The manuscript now emphasizes temporal variation across survey periods as an observational finding, interpreted cautiously in the context of the study design.

On lines 357-361, the following text was added “This study, conducted through three consecutive PPS, provides a repeated cross-sectional overview of the epidemiology of HAIs, antimicrobial consumption, and key risk factors in a University Hospital in Southern Italy. This design allowed to observe temporal variations in HAI prevalence, clinical case-mix, prescribing practices, and data quality, without supporting causal inference.”

On lines 466-469, the following text was added “Therefore, our results are exploratory, although the serial survey repetition might pro-vide a more informative cross-sectional perspective than a single PPS, and might support the relevance of the observed trends for clinical planning and governance.”

 

Comment 7. This is a single-center study with a relatively small patient population. The conclusions should be toned down accordingly, and statements about generalizability or reproducibility of the model should be qualified.

Response to Comment 7. We thank the Reviewer for this valuable observation. We fully agree that the single-center design and the limited sample size require a more cautious interpretation of the findings. Accordingly, we have revised the manuscript to avoid broad claims regarding generalizability, reproducibility, or implementation in other settings. The results are now presented as exploratory and context-specific, and any implications for broader application are explicitly qualified as provisional.

On lines 445-448, the following text was included “Accordingly, integrating periodic epidemiological surveillance with AMS may provide a useful local framework for clinical governance, although its applicability to other settings should be confirmed in multicentric studies.”

On lines 476-478, the following sentence was added “The integration of periodic epidemiological surveillance with AMS may be a useful approach for local clinical governance and patient safety. However, our results are exploratory and should be confirmed in larger prospective multicentric studies.”

 

Comment 8. Figure 2 uses red and blue colors to distinguish McCabe score categories. These colors may be difficult to distinguish in grayscale printing. The authors should consider using patterns or higher-contrast color combinations in line with the journal's requirements.

Response to Comment 8. We thank the Reviewer for this helpful comment. We agree that the original color scheme (red and blue) may not be optimal for grayscale printing. Accordingly, we have revised Figure 2 by adopting a colorblind safe combination, which provides improved contrast and ensures better distinguishability when printed in black and white. The updated figure has been modified in line with the journal’s requirements.

Reviewer 2 Report

Comments and Suggestions for Authors

The study highlights the ongoing issue of MDR pathogens like KPC-producing Klebsiella pneumoniae and the significant role of central venous catheters as a modifiable risk factor. Strengths include alignment with ECDC protocols, emphasis on antimicrobial stewardship, and local relevance. However, methodological and reporting issues, particularly related to small event numbers and data presentation, need to be addressed before publication.

1. Only 17 HAIs were identified among 456 patients (prevalence ~3.7%), limiting the robustness of risk factor analyses. CVC use emerged as a strong predictor (OR 9.20, 95% CI 3.37–25.11, p<0.001), but the wide confidence interval and lack of multivariate adjustment are significant limitations. It’s important to discuss how low event numbers affect precision and consider using logistic regression with penalized methods or clearly frame results as exploratory. The claim of “significant improvement in clinical risk stratification” is well-supported for documentation quality but should be cautious regarding causal links to infection control outcomes.
2. The definition of HAIs per ECDC guidelines lacks detail on infection criteria and distinctions between hospital-acquired and present-on-admission cases. More justification is needed for certain cases in Table 3 to meet ECDC PPS criteria, and clarification on how device-associated infections were attributed is essential for comparability with ECDC benchmarks.
3. Changes in ward composition between surveys (addition of Cardiology and Pediatric Surgery) may confound prevalence trends (3.1% → 6.1% → 1.9%). A sensitivity analysis focused on consistently monitored wards would strengthen conclusions about temporal trends versus case-mix changes.
4. While the manuscript notes MDR pathogens, it offers limited information on susceptibility testing methods and resistance mechanisms. Discussing local antibiograms and the proportion of HAIs caused by WHO priority pathogens would enhance the paper’s relevancy.
5. The results should more clearly quantify findings (e.g., exact HAI prevalences and p-values) and the conclusion should be revised to better align with the data presented to avoid overstating the association between clinical severity and infection risk.
6. The description of statistical tests needs to be completed with specifics about post-hoc tests or adjustments for multiple comparisons.
7. Integrate local findings with recent ECDC PPS data and Italian national reports. Discuss possible reasons for lower observed prevalences compared to national averages.
8. In the discussion section, the trend in carbapenem reduction is encouraging but modest; it should be contextualized against national AMS efforts and seasonal influences.

Author Response

Comments and Suggestions for Authors

The study highlights the ongoing issue of MDR pathogens like KPC-producing Klebsiella pneumoniae and the significant role of central venous catheters as a modifiable risk factor. Strengths include alignment with ECDC protocols, emphasis on antimicrobial stewardship, and local relevance. However, methodological and reporting issues, particularly related to small event numbers and data presentation, need to be addressed before publication.

Comment 1. Only 17 HAIs were identified among 456 patients (prevalence ~3.7%), limiting the robustness of risk factor analyses. CVC use emerged as a strong predictor (OR 9.20, 95% CI 3.37–25.11, p<0.001), but the wide confidence interval and lack of multivariate adjustment are significant limitations. It’s important to discuss how low event numbers affect precision and consider using logistic regression with penalized methods or clearly frame results as exploratory. The claim of “significant improvement in clinical risk stratification” is well-supported for documentation quality but should be cautious regarding causal links to infection control outcomes.

Response to Comment 1. We sincerely thank the Reviewer for the positive feedback regarding study's alignment to ECDC protocols, emphasis on antimicrobial stewardship, and local relevance. We fully agree with the methodological concerns raised and appreciate the constructive suggestions. We acknowledge that the low event numbers (17 HAIs among 456 patients, prevalence ~3.7%) limit the precision of risk factor analyses, as reflected by the wide confidence intervals for CVC use (OR 9.20, 95% CI 3.37–25.11). Due to insufficient events per variable (EPV <10), multivariate modeling was not feasible. We have therefore explicitly framed these results as exploratory and highlighted the statistical limitations throughout the manuscript. Moreover, we agree that the observed improvement in McCabe score documentation (unknown cases from 24.6% to 6.8%, p<0.001) reflects better data quality rather than direct infection control outcomes. We have revised this interpretation to emphasize data completeness without implying causal links to HAI reduction.

On lines 343-346, the following text was added “The presence of a CVC was more frequent among infected than non-infected patients (58.8% vs. 13.4%; OR, 9.20; 95%CI, 3.37–25.11; p < 0.001), although the small number of registered HAIs (n = 17 across all surveys), thus our results should be interpreted as exploratory.”

On lines 458-461, the following sentences were added “The analysis of associated factors was conducted exclusively through univariate analysis due to the limited number of observed events (17 of total HAIs), as also confirmed by the wide confidence intervals observed.”

On lines 402-405, the following text was included “During our three-period PPS, a significant progressive reduction in “unknown” cases and in distribution of the McCabe score was described, suggesting an improvement in data completeness and a greater consistency in data collection procedures over time, although differences in form completion might contribute to this trend.”

 

Comment 2. The definition of HAIs per ECDC guidelines lacks detail on infection criteria and distinctions between hospital-acquired and present-on-admission cases. More justification is needed for certain cases in Table 3 to meet ECDC PPS criteria, and clarification on how device-associated infections were attributed is essential for comparability with ECDC benchmarks.

Response to Comment 2. We thank the Reviewer for this important methodological observation. We agree that greater detail on ECDC HAI definitions and case attribution is essential for transparency and comparability. All HAIs were classified according to ECDC Point Prevalence Survey protocol v6.1 (2022-2023), specifically including only hospital-onset infections (onset ≥ day 3 of current admission, day 1 = admission day). Present-on-admission cases were systematically excluded. All the HAI episodes were reviewed case by case and fulfilled the ECDC PPS v6.1 site-specific and device-associated infection criteria.

On lines 111-121, the following text was added “Data collection was conducted by trained healthcare personnel through a comprehensive review of medical records, using standardized data collection forms aligned with the ECDC protocol version 6.1 (2022) [25]. HAIs were defined according to ECDC PPS criteria (protocol v6.1) [25], and were classified as infections with onset after >3 days from hospital admission (day of admission = day 1). Infections present on admission were systematically excluded. Case definitions required suggestive clinical signs and symptoms, microbiological confirmation, and/or strong clinical/radiological evidence consistent with ECDC site-specific criteria. Device-associated infections were defined when: a) a device was present within 7 days prior to symptom onset, b) no alternative infection sources were identified; and c) same microorganism from device tip or clinical improvement following device removal were observed.”

On lines 250-251, the following sentence was added “All infections met the ECDC PPS v6.1 case definitions and occurred at least after ≥3 days from the admission.”

 

Table 4. Characteristics of identified HAIs across Survey periods P1 (November 2024), P2 (June 2025), and P3 (November 2025).

Ward	Infection Type (ECDC Code)	Isolated Pathogen	Mean LOS, days	Device
P1, N = 4 cases
Emergency Cardiac Surgery	CRI3	Acinetobacter baumannii	49	CVC
	SSI-D	Klebsiella pneumoniae	84	-
	SSI-D	Staphylococcus aureus	32	-
General Surgery	CRI3	Escherichia coli	28	CVC
P2, N = 10 cases
Emergency Cardiac Surgery	SSI-S	Staphylococcus epidermidis	85	CVC
	SSI-S	Staphylococcus haemolyticus	13	-
Cardiology	UTI-C	Polymicrobial: E. faecalis, S. haemolyticus, C. freundii, Candida spp.	70	CVC or UC
	PN5	Negative (clinical diagnosis)	34	UC, CVC
General Medicine	CA-UTI	Enterococcus faecalis	18	UC
	BSI	Staphylococcus haemolyticus	27	-
UTIPO	CRI3	Serratia marcescens	8	CVC
	VAP + SSI	Pseudomonas aeruginosa (VAP) / Enterococcus faecalis + K. pneumoniae (SSI)	30	ET tube, UC
	VAP) + UTI + CDI	Corynebacterium, E. coli (VAP) / Klebsiella pneumoniae (UTI) / Clostridioides difficile	14	ET tube, UC, CVC
	CRI3	Klebsiella pneumoniae	10	CVC, UC
P3, N = 3 cases
Cardiology	UTI-B	Negative (clinical diagnosis)	8	Prior UC
Neurosurgery	CRI3	Klebsiella pneumoniae	60	CVC, UC
UTIPO	BSI	Candida parapsilosis	18	CVC, UC

Abbreviations. LOS, length of stay; CRI3, catheter-related bloodstream infection type 3 (microbiologically confirmed); SSI-D, deep surgical site infection; SSI-S, superficial surgical site infection; UTI-C, symptomatic urinary tract infection with microbiological confirmation; CA-UTI, catheter-associated urinary tract infection; BSI, bloodstream Infection; VAP, ventilator-associated pneumonia; UTI, urinary tract infection; CDI, Clostridioides difficile infection; UTI-B, asymptomatic bacteriuria; CVC, central venous catheter; UC, urinary catheter; UTIPO, Post-Operative Intensive Care Unit; ET, endotracheal tube.

Comment 3. Changes in ward composition between surveys (addition of Cardiology and Pediatric Surgery) may confound prevalence trends (3.1% → 6.1% → 1.9%). A sensitivity analysis focused on consistently monitored wards would strengthen conclusions about temporal trends versus case-mix changes.

Response to Comment 3. We thank the Reviewer for this excellent methodological suggestion. We agree that changes in ward composition (addition of Cardiology and Pediatric Surgery in P2/P3) may confuse crude prevalence trends, and a sensitivity analysis on consistently monitored wards is appropriate. This analysis demonstrates that the observed temporal pattern persists even after controlling for ward composition changes, strengthening the validity of prevalence trends beyond case-mix effects.

On lines 95-98, the following sentence was added “For surveys P2 and P3, Cardiology Unit and Pediatric Surgery were also included. However, sensitivity analysis was performed by including only wards consistently monitored across all survey periods.”

On lines 204-209, the following text was added “To address potential confounding by ward composition changes in P2 and P3, HAI prevalence was recalculated for wards monitored across all periods (376/456 patients, 82.5%), showing similar trends: P1, 3.1% (4/130); P2, 6.6% (8/121); P3, 1.6% (2/125). These findings indicate that the non-linear prevalence pattern with a peak in P2 remained even when controlled by ward composition (Cochran-Armitage trend test across periods, p = 0.55) (Table 2).”

 

Table 2. HAI prevalence in consistently monitored wards

Survey Period	Patients, n (%)	HAIs, n	Prevalence, % (95% CI)
P1	130 (34.6)	4	3.1 (0.9-7.7)
P2	121 (32.2)	8	6.6 (2.9-12.6)
P3	125 (32.2)	2	1.6 (0.2-5.7)
Total	376 (82.5)	14	3.7 (2.1-6.2)

Wards included: UTIPO, Neurosurgery, Obstetrics, Gynecology, Orthopedics/Traumatology, Emergency Cardiac Surgery, Nephrology, General Surgery, Oncology, General Medicine.

 

On lines 463-466, the following text was added “However, the observed prevalence peak in P2 persisted in sensitivity analysis restricted to consistently monitored wards, suggesting that factors beyond ward composition may have contributed to the observed temporal variation, despite the absence of a statistically significant linear trend.”

 

Comment 4. While the manuscript notes MDR pathogens, it offers limited information on susceptibility testing methods and resistance mechanisms. Discussing local antibiograms and the proportion of HAIs caused by WHO priority pathogens would enhance the paper’s relevancy.

Response to Comment 4. We thank the Reviewer for this valuable comment. We agree that providing more detail on microbiological methods and resistance patterns is important to enhance comparability with ECDC benchmarks and the WHO priority pathogen framework. All microbiological investigations were performed in our hospital’s accredited clinical microbiology laboratory using standardized culture and identification procedures, with antimicrobial susceptibility testing interpreted according to current EUCAST breakpoints. Multidrug-resistant (MDR) phenotypes and carbapenemase-producing strains were classified following the laboratory’s internal protocols, harmonized with national surveillance definitions. We have now clarified these methods in the Materials and Methods section and expanded the Results and Discussion to better characterize the resistance profile of HAI-related pathogens, to relate them to local cumulative antibiogram data, and to position the detected organisms within the WHO priority pathogen list.

On lines 131-139, the following subsection was included.

“2.4 Microbiological assays

All microbiological tests were performed at our Clinical Microbiology Laboratory, according to standardized procedures. Clinical specimens (e.g., blood, urine, respiratory secretions, or surgical samples) were processed using conventional culture methods and automated identification systems. Antimicrobial susceptibility testing was carried out by automated broth microdilution and interpreted according to current European Committee on Antimicrobial Susceptibility Testing (EUCAST) breakpoints [26]. MDR phenotypes and carbapenemase-producing Enterobacterales (e.g., KPC strains) were classified in line with the laboratory’s internal protocols, harmonized with national surveillance standards [27].”

On lines 282-289, the following paragraph was added “Overall, Gram-negative bacilli predominated among HAI-related pathogens, with frequent detection of KPC and other multidrug-resistant Enterobacterales, Acinetobacter baumannii, Pseudomonas aeruginosa, and Serratia marcescens, falling into the critical priority group for antibiotic resistance. Gram-positive isolates mainly included coagulase-negative staphylococci and Enterococcus faecalis, classified as “high” or “medium” priority pathogens, together with cases of Clostridioides difficile infection and invasive candidiasis, reflecting the high-risk microbiological profile of our setting despite the relatively low number of HAI events.”

On lines 396-401, the following paragraph was included “In our hospital, local cumulative antibiograms documented high resistance rates to third-generation cephalosporins and fluoroquinolones among Enterobacterales, and a substantial burden of KPC, in line with national surveillance data [32], and supported by the microbiological findings of pathogens belonging to the critical (carbapenem-resistant Enterobacterales, A. baumannii, and P. aeruginosa), or high/medium priority groups (e.g., Enterococcus faecalis, staphylococci and C. difficile).”

 

Comment 5. The results should more clearly quantify findings (e.g., exact HAI prevalences and p-values) and the conclusion should be revised to better align with the data presented to avoid overstating the association between clinical severity and infection risk.

Response to Comment 5.  We thank the Reviewer for this important remark. We agree that the main quantitative findings, including exact HAI prevalences and p-values, should be presented more explicitly in the Results, and that the Discussion and Conclusions must avoid overstating the association between clinical severity and infection risk. We have therefore: (1) added the exact HAI prevalences and corresponding p-values in the Results section, alongside the information already reported in the tables; (2) explicitly referred to the magnitude and statistical significance of differences in McCabe score distribution and length of stay; and (3) revised the wording in the Discussion and Conclusions to describe these relationships as exploratory, univariate associations that are consistent with previous evidence, without implying causality. The revised manuscript now more closely aligns conclusions with the underlying data and clearly acknowledges the limitations related to sample size, cross-sectional design, and univariate analysis.

On lines 248-251, the text was changed as “The point prevalence of HAIs exhibited seasonal and temporal variability throughout the study, following a non-linear trend across the three survey periods: 3.1% (4/130) in P1, 6.1% (10/165) in P2, and 1.9% (3/161) in P3 (Table 4). All infections met the ECDC PPS v6.1 case definitions and occurred at least after ≥3 days from the admission.”

On lines 215-218, the following results were included “The distribution of clinical severity (McCabe score) significantly differed across the three survey periods (χ² = 22.2, df = 6, p = 0.001; Table 3). This variance was mainly driven by a reduction in the ‘unknown’ category from 24.6% in P1 to 6.8% in P3 (p = 0.01) and by an increase in non-fatal disease from 54.6% to 69.0% (p = 0.03).”

On lines 342-351, the following text was added “To identify factors potentially associated with HAI acquisition, a univariate analysis was performed (Table 6). The presence of a CVC was more frequent among infected than non-infected patients (58.8% vs. 13.4%; OR, 9.20; 95%CI, 3.37–25.11; p < 0.001), although the small number of registered HAIs (n = 17 across all surveys), thus our results should be interpreted as exploratory. UC use was also more common among infected patients, although not statistically significant (52.9% vs. 36.7%; OR, 1.94; 95%CI, 0.73–5.13; p = 0.20), possibly reflecting the widespread utilization of this device across the study population. Length of stay was markedly longer among patients with HAIs compared to those non-infected (mean+SD, 34.6±26.2 days vs. 9.9±13.1 days; p < 0.001), likely associated with our cross-sectional study design rather than a proven causal effect.”

On lines 405-409, the following text was added “Patients with poor prognostic categories were consistently concentrated in high-complexity wards (Oncology, General or Internal Medicine, Emergency Cardiac Surgery, and UTIPO), in line with the association between clinical severity and HAI risk al-ready described in literature [5,28,33,34].”

On lines 181-185, the following results were added “Indeed, overall, infected patients had a significantly longer LOS compared to non-infected subjects (mean±SD: 34.6±26.3 days vs. 9.9±12.8 days; p < 0.001), indicating that prolonged hospitalization is associated with the occurrence of HAIs in this cohort, although the cross-sectional design does not allow causal inference.”

On lines 471-478, conclusions were rephrased as follows “Our serial PPS provided a comprehensive overview of the local epidemiology of HAIs and antimicrobial consumption in a tertiary-care University Hospital in Southern Italy, highlighting CVCs as a potential modifiable target for prevention bundles, along with clinical complexity and length of hospitalization, as exploratory associations with infection risk, derived from univariate analyses based on a limited number of HAI events. The integration of periodic epidemiological surveillance with AMS may be a useful approach for local clinical governance and patient safety. However, our results are exploratory and should be confirmed in larger prospective multicentric studies.”

 

Comment 6. The description of statistical tests needs to be completed with specifics about post-hoc tests or adjustments for multiple comparisons.

Response to Comment 6. We thank the Reviewer for this important clarification request. In our study, the primary inferential comparisons across the three survey periods were performed using global tests (Chi-square or Fisher’s exact test for categorical variables, and ANOVA or non-parametric tests for continuous variables). When an overall test was statistically significant, we explored pairwise differences between periods using the same tests. No formal post-hoc procedures (e.g., Tukey, Bonferroni) or adjustments for multiple comparisons were applied, given the exploratory nature of the study and the limited number of HAI events. Accordingly, pairwise p-values are reported as descriptive and are not used to define independent thresholds of statistical significance. We have now clarified this approach in the Statistical Analysis section. No formal post-hoc procedures or adjustments for multiple comparisons were applied; pairwise p-values are reported as descriptive, in line with the exploratory nature of the study.

The paragraph has been changed as follows.

“Statistical Analysis

Data were collected in a spreadsheet and analyzed using IBM SPSS Statistics soft-ware, version 23.0. Categorical variables were expressed as frequencies and percentages, whereas continuous variables were reported as mean ± standard deviation (SD). Comparisons of categorical variables across the three study periods (P1, P2, P3) were performed using by Chi-square (χ²) or Fisher’s exact test, as appropriate. Continuous variables were compared across the three periods using one-way analysis of variance (ANOVA). Normality and homogeneity of variance assumptions were assessed before applying parametric tests. For comparisons between two groups (e.g., patients with HAIs vs. non-infected patients), the Mann–Whitney U test was employed. Univariate analysis was performed to identify potential risk factors associated with HAI occurrence, and odds ratios (ORs) with 95% confidence intervals (95% CI) were calculated from 2×2 contingency tables to quantify the association between exposure to invasive devices and infection. Global tests across the three survey periods were considered the primary inferential comparisons. When an overall test was significant, pairwise comparisons between periods were explored using the same statistical test. No formal post-hoc procedures or adjustments for multiple comparisons were applied; pairwise p-values are reported as descriptive, in line with the exploratory nature of our study. Stratified analyses were performed by ward and survey period to assess differences in clinical characteristics, infection prevalence, and antimicrobial consumption. All statistical tests were two-tailed, and a p-value < 0.05 was considered statistically significant.”

 

Comment 7. Integrate local findings with recent ECDC PPS data and Italian national reports. Discuss possible reasons for lower observed prevalences compared to national averages.

Response to Comment 7. We thank the Reviewer for this valuable suggestion. We agree that integrating our local findings with recent ECDC PPS data and Italian national reports is essential to contextualize the observed HAI prevalences and to discuss why they appear lower than national averages. We have therefore expanded the Discussion to (i) more explicitly compare our point estimates with the ECDC 2022–2023 PPS and Italian surveillance figures, and (ii) discuss possible explanations, including hospital bed-mix, ward selection, infection prevention and control (IPC) practices, and methodological aspects of the PPS design. The revised text now emphasizes that our lower prevalences may reflect a combination of local IPC measures and structural characteristics of the hospital, but that under-ascertainment and sampling limitations cannot be excluded.

On lines 357-377, the following text was added “This study, conducted through three consecutive PPS, provides a repeated cross-sectional overview of the epidemiology of HAIs, antimicrobial consumption, and key risk factors in a University Hospital in Southern Italy. This design allowed to observe temporal variations in HAI prevalence, clinical case-mix, prescribing practices, and data quality, without supporting causal inference. According to the most recent data from the ECDC PPS from 2022–2023, the average prevalence of HAIs in European acute care hospitals is ~7.1%, with marked heterogeneity across Countries and settings [5,28,29]. In the Italian context, national surveillance studies report similar HAI prevalence in acute care hospitals compared to the European average (~8.8% of patients with at least one HAI) [30]. In our study, observed prevalences across all study periods were lower than the European and Italian averages, with P2 approaching the European benchmark. This discrepancy may be partly explained by structural and organizational factors, such as the bed-mix of the hospital (one post-operative ICU and a limited number of high-risk units compared with large national samples), the inclusion of a selected set of wards rather than all hospital wards, and the implementation of local infection prevention and control measures and antimicrobial stewardship activities during the study period. On the other hand, the point-prevalence design, the relatively small sample size, and the possibility of residual under-ascertainment of HAIs cannot be ruled out, and may also have contributed to lower estimates. Therefore, these results should be interpreted cautiously as reflecting the local epidemiological situation in this single center, rather than indicating a systematically lower national risk.”

 

Comment 8. In the discussion section, the trend in carbapenem reduction is encouraging but modest; it should be contextualized against national AMS efforts and seasonal influences.

Response to Comment 8. We thank the Reviewer for this insightful comment. We agree that the observed reduction in carbapenem use (from 12.0% to values below 10%) is encouraging but modest, and should be interpreted within the broader context of national antimicrobial stewardship initiatives and seasonal prescribing patterns. We have therefore expanded the Discussion to explicitly link our findings to Italian and European AMS efforts and to the temporal placement of the three PPS (two winter surveys and one summer survey), emphasizing the descriptive and exploratory nature of the observed trend.

On lines 318-321, the following text was added “This modest but consistent reduction was in line with national AMS efforts and Italian stewardship recommendations to limit carbapenem exposure, particularly in high-risk settings, although the small sample size and the point-prevalence design do not allow definitive conclusions.”

 

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

the authors have adequately addressed all comments

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript shows significant improvement over typical first submissions. The authors are cautious in their interpretations, acknowledging the cross-sectional design, limited HAI events, and the exploratory nature of their findings.