Introduction
Peripheral venous catheters (PVCs) are the most commonly used invasive devices in hospital settings, with up to 70% of hospitalized patients receiving a PVC Reference Zingg and Pittet1 and approximately 2 billion PVCs sold worldwide annually. Reference Rickard, Marsh and Webster2 While generally considered low-risk devices, PVCs can occasionally lead to serious adverse events, including mechanical, vascular, and infectious complications.
Research on strategies to reduce PVC-associated local complications has identified several risks: Key factors influencing complication rates include hand hygiene compliance, vein selection, puncture technique, skin disinfection, and regular inspection of the insertion site. Reference Guenezan, Marjanovic and Drugeon3,Reference Buetti, Abbas and Pittet4 Among available skin disinfectants, chlorhexidine stands out as a broad-spectrum antibacterial agent effective against both Gram-positive and Gram-negative bacteria, with additional activity against yeasts and dermatophytes. Reference McDonnell5
Previous studies have demonstrated that 2% chlorhexidine gluconate in 70% alcohol (CHG) outperforms other alcohol-based disinfectants in reducing infection-related complications. Reference Adams, Quayum, Worthington, Lambert and Elliott6 Our study aimed to evaluate whether implementing CHG as a routine disinfectant in a real-world setting would impact local PVC site complication rates.
Methods and findings
Study setting and intervention
Prior to 2023, the tertiary care hospital St. Anna (Lucerne, Switzerland) used isopropanol alcohol (IPA) as the standard skin disinfectant for PVC insertion. On February 1, 2023, following recommendations from the hospital hygiene department, all existing Kodan® forte bottles (containing isopropyl alcohol, propanol, and 2-phenylphenol) were replaced with Softasept® chlorhexidine 2% solution. This transition was preceded by staff communication, updates to standard operating procedures, and appropriate training.
Monitoring and documentation
The nursing staff maintained consistent, prospective monitoring practices throughout both periods, documenting local findings as long as the PVC was in place (suppuration, redness, vein hardening) three times daily in the electronic health record system. Individual PVC episodes were recognized and defined through specific attributes such as insertion and removal data; side (left vs right); as well as anatomic position (eg, wrist, forearm, cubital fossa). Postremoval PVC surveillance was by definition not possible. Ward of insertion was recorded as well in the data set—all PVC inserted outside the institution were excluded from the study.
The hospital hygiene staff conducted periodic checks to ensure proper implementation and documentation. Check on the implementation of disinfectant distribution and use was not performed pre intervention, as CHG was not available for PVC site disinfection in the designated wards.
Statistical analysis
We compared patient characteristics and outcome rates between the two periods using χ 2 tests for categorical data and Wilcoxon tests for continuous data. The primary outcome was a composite of local complications (redness, suppuration, and hardening). Sensitivity analysis included the composite redness and suppuration only, which can be viewed as a more reliable marker of possible PVC infection. We calculated the crude PVC outcome rate per 1 000 lay days, as well as the incidence rate ratio using Poisson regression, analyzing monthly complication rates before and after the intervention. Statistical significance was set at P < .05, using R version 4.1.2.
This quality assurance study was exempt from ethical review under Swiss law. According to the Swiss Human Research Act, quality assurance and quality control studies are classified as basic research activities protected under Article 20 of the Federal Constitution of the Swiss Confederation. Such studies do not require prior regulatory approval.
Results
Study population
After exclusion of 667 PVC inserted outside the institution, our analysis included 35,032 PVC episodes: 23 895 in the IPA group (January 2021–January 2023) and 11,137 in the CHG group (February 2023–January 2024). Both groups (Table 1) showed statistically not significantly different patient characteristics: Median PVC dwell time: 2 days in both groups. Sex distribution was 54.1% women (IPA) versus 54.8% (CHG) and the median age: 66 [IQR 51–77] years (IPA) versus 66 [IQR 51–77] years (CHG).
Baseline characteristics and outcomes (composite and single events) per study period
Primary outcome
The composite outcome rate of local complications for the IPA group was 317/23 895 (1.3%) PVC episodes and in the CHG group 118/11 137 (1.1%) PVC episodes (P = .040). Per PVC days, the crude rate for the composite outcome was 5.28/1 000 lay days and 4.37/1 000 for CHG (P = .085).
The detailed comparison of individual outcomes are listed in Table 1.
The incidence rate ratio after CHG implementation was 0.80 (95% CI, 0.64–0.98; P = .037), indicating a statistically significant reduction in complications (Figure 1).
Incident rates per month with 95% CI and trendline per period (before/after intervention).
In the sensitivity analysis (Supplementary Table and Supplementary Figure) focussing on the outcome suppuration and redness only, the IRR was 0.86 (95% CI 0.67–1.09, P = .21)
Discussion
Our findings demonstrate that switching from IPA to CHG was associated with a marginally statistically significant reduction in local PVC complications, aligning with previous clinical trial data on CHG efficacy Reference Guenezan, Marjanovic and Drugeon3,Reference Small, Adams, Casey, Crosby, Lambert and Elliott8 . Taking into consideration the results of the sensitivity analysis, the statistically nonsignificant difference in the rate of outcome/1 000 PVC days and several limitations associated with this before/after study, our careful interpretation is, that the introduction of CHG disinfection was not associated with a major change of incidence in local skin complications.
While nursing adherence to insertion site monitoring was not directly measured, the consistent monthly PVC rates suggest stable compliance throughout the study period. To the best of our knowledge, there was no substantial change in other prevention methods throughout the study period, such as staff education on proper insertion, hand hygiene compliance, and sterile dressing options. Accordingly, we do not expect significant confounding due to other prevention methods introduced throughout the study period.
Study limitations include: i) Confounding by the study design (active change to a potentially better disinfectant with more allergies may lead to under/overreporting of outcomes); ii) inability to link microbiological results and mortality data due to separate database systems: these markers could have represented stronger biological correlation of poor or limited skin antisepsis; iii) lack of knowledge of concomitant antimicrobial therapy that may occult PVC outcomes (even though we do not believe, this is an effect modifier); iv) the role of the COVID pandemic and other time-dependent changes in infection-prevention practice that may have confounded results; and v) potential limitations in generalizability to facilities with different service ranges or microbial spectra.
Conclusion
The implementation of CHG as the standard skin disinfectant was not associated with a major change of incidence in local skin complications in our before/after healthcare intervention. We rather consider the result or our study as hypothesis-generating than practice changing.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/ash.2025.10170.
Acknowledgements
We would like to thank to all nursing Stuff and to the department of the prevention of infectious diseases in the St. Anna Hospital, Luzern, Switzerland.
All authors contributed to writing and critical review of the manuscript, approved the final version, and made the decision to submit the manuscript for publication.
Financial support
Institutional funding. RS has received an unrestricted research grant from BBraun, Sempach, Switzerland (no role in design or writing of the current study).
Competing interests
All authors report no conflicts of interest relevant to this article.
Open access
