Clinical-epidemiological predictors of phlebitis associated with peripheral intravenous catheters in Spanish hospitals: results of a national cohort study

Abstract

Background

Peripheral intravenous catheters (PIVC) are essential medical devices, yet they frequently lead to complications such as phlebitis, infiltration, and occlusion. Identifying risk factors is key to reducing these complications.

Objectives

To estimate the incidence of PIVC‑associated phlebitis in Spanish hospitals and identify clinical and epidemiological risk factors.

Methods

We performed a prospective cohort study in 80 Spanish hospitals from 1 to 28 February 2023. Adult inpatients (≥18 years) receiving PIVCs in non‑ICU, non‑emergency, non‑pediatric wards were eligible. A total of 13,812 PIVCs in 9387 patients were followed daily by trained nurses until catheter removal, phlebitis onset (Maddox grade ≥ 2), or 15 days. We calculated cumulative incidence and incidence density per 100 catheter‑days. Multivariable Cox proportional hazards models estimated hazard ratios (HRs) for predictors—sex, age group, number of infused medications, hospital size, and dwell time.

Results

Phlebitis occurred in 1302 PIVCs (cumulative incidence 9.43 %; incidence density 0.14 per 100 person‑hours). Independent risk factors were female sex (HR 1.32, 95 % CI 1.21–1.45), age 65–79 years (HR 1.25, 95 % CI 1.12–1.40), administration of ≥ 2 medications (HR 1.50, 95 % CI 1.35–1.67), and hospital size ≥ 1000 beds (HR 1.30, 95 % CI 1.12–1.52). Phlebitis risk peaked 48–96 h post‑insertion.

Conclusions

Phlebitis incidence remains above recommended levels. Multivariable analysis identified female sex, a higher number of infused medications, older patient age, larger hospital size, and longer PIVC dwell time as the main independent predictors of phlebitis. Registration: Not registered.

What is already known.

• •Phlebitis is a common complication of peripheral intravenous catheters (PIVCs).
• •Reported phlebitis rates often exceed the 5 % threshold recommended by guidelines.
• •Risk factors like age, gender, and catheter site remain inconsistently reported.

Alt-text: Unlabelled box

What this paper adds.

• •Phlebitis incidence in Spanish hospitals is 9.43 %, exceeding recommended levels.
• •Female sex, age 65–79, and multiple drug use significantly increase phlebitis risk.
• •Risk of phlebitis peaks between 48 and 96 h post-insertion but decreases afterward.

Alt-text: Unlabelled box

1. Background

One of the most commonly used vascular devices during hospitalization is the peripheral intravenous catheter (PIVC). However, these devices are not without complications. Complications such as phlebitis, extravasation or infiltration, and occlusion have a significant impact on the patient's clinical outcomes and hospital length of stay (Marsh et al., 2024).

Although it is recommended that the incidence of phlebitis should remain below 5 % (Gorski et al., 2021), several studies have reported values well above this threshold in PIVCs used in hospital settings (García-Expósito et al., 2021; Lv y Zhang, 2020).

To achieve an incidence of PIVC-associated phlebitis below 5 %, it is essential to identify and understand the direct causes of these complications. Relevant factors include patient-dependent variables, environmental conditions, and circumstances such as female sex or advanced age (Lv y Zhang, 2020; Marsh et al., 2018), as well as clinical factors, such as the patient’s pathology (Cernuda Martínez et al., 2024). Other device-related factors include catheter size (note that a lower gauge number corresponds to a larger diameter, so twenty-two gauge (22 G) or twenty-four gauge (24 G) are relatively small-bore), with 22 G or 24 G devices being associated with a lower incidence of phlebitis—likely because their smaller diameter reduces mechanical trauma and chemical irritation of the vein wall (Marsh et al., 2018). Additionally, a risk factor associated with failure is the insertion site, such as the forearm in the antecubital fossa (Ferraz-Torres et al., 2021; Lv y Zhang, 2020), the wrist, or the dorsum of the hand (Comparcini et al., 2017; Simões et al., 2022), as well as the catheter's dwell time (Cernuda Martínez et al., 2024; Lv y Zhang, 2020).

Regarding the dwell time of the device, significant variability has been observed in recent years, along with changes in intervention protocols. A few years ago, procedures recommended the systematic replacement of the device (Chiu et al., 2015; Freixas et al., 2013; Mestre et al., 2013; Saliba et al., 2018) to prevent potential complications associated with maintaining the catheter beyond 72 h. Currently, evidence indicates a higher risk of phlebitis during the first 72 h of the device's lifespan (Cernuda Martínez et al., 2024).

This variability highlights the need to establish a standardized approach to reduce the heterogeneity in PIVC management and minimize associated complications.

Phlebitis associated with PIVC increases nursing workload (with over 27,000 days per year spent on unnecessary insertions) and incurs an estimated annual healthcare cost of A$437.5 million (Morgan et al., 2022), delays time-sensitive interventions and contributes to therapy interruptions, worsening patient outcomes and driving up costs (Marsh et al., 2021), and repeated insertions can lead to peripheral venous depletion, necessitating central venous access devices with even greater risks and expenses (Santos-Costa et al., 2022).

Given the lack of long-term Spanish PIVC follow‑up cohorts that evaluate both the incidence of PIVC‑associated phlebitis and the clinical and epidemiological risk factors driving its development, this study is essential to generate data that could elucidate whether patients in Spain exhibit distinct risk profiles shaped by genetic, socioclinical, and healthcare‑system differences.

The aim of this study was to determine the incidence rate of PIVC-related phlebitis in Spain during 2023, identify the main clinical and epidemiological risk factors contributing to its development, and establish the timeframe during which the event is most likely to occur.

2. Methods

2.1. Study design, population, and setting

A survival observational study was conducted in 80 Spanish hospitals, involving 9387 hospitalized individuals in February 2023 who had a total of 13,812 PIVCs inserted during their hospital stay. The choice of a survival study design was appropriate to analyze the time to event (phlebitis) in a prospective manner, allowing for systematic monitoring of PIVCs over time and identification of risk factors. This design was particularly suitable for capturing the dynamic nature of PIVC-related complications within a real-world clinical setting. All participating hospitals were affiliated with the national Flebitis Zero project, which systematically implements protocols for the prevention, documentation, and monitoring of PIVC-associated phlebitis episodes. Participants were enrolled between 1 February 2023 and 28 February 2023. Each catheter was then monitored daily for up to 15 days, ending on 15 March 2023 or until removal or phlebitis onset, whichever occurred first.

The multicenter prospective cohort design adopted in this study (conducted across 80 Spanish hospitals with daily monitoring of 13,812 PIVC in 9387 adult patients) is justified by the need to accurately estimate the incidence of peripheral intravenous catheter (PIVC)–associated phlebitis and to identify clinical‐epidemiological risk factors under real‐world clinical practice conditions. By following the temporal sequence of phlebitis onset after exposure to various factors, this design minimizes bias. Moreover, the inclusion of a large, nationally representative sample with clear inclusion and exclusion criteria, standardized outcome assessment using the Maddox scale, and multivariable analysis via Cox models yields robust and generalizable results, providing evidence to inform the optimization of PIVC prevention and management protocols in a context where phlebitis represents a significant complication with implications for morbidity and healthcare costs.

The inclusion criteria were: (i) being 18 years of age or older; (ii) having received intravenous pharmacological therapy through the PIVC at some point during the hospital stay; and (iii) providing consent to participate in the study. The exclusion criteria were: (i) individuals with PIVCs inserted in emergency, pediatrics, or intensive care units; (ii) individuals with PIVCs inserted in hospital units other than the one they were in during the study; and (iii) individuals with PIVCs inserted prior to the start date of the study.

This study was designed in accordance with the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines, ensuring transparency, methodological rigor, and comprehensive reporting of observational research (Elm et al., 2007). Adherence to these guidelines enhances the validity and reproducibility of the findings.

2.2. Study variables and data collection

The primary outcome of this study was the presence or absence of phlebitis in the vein where the PIVC was inserted. To assess this, the Jackson infusion phlebitis visual scale, based on the Maddox scale (Jackson, 1998), was used. This scale consists of six grades of phlebitis (Table 1). Phlebitis severity was evaluated daily by the study coordinator in each unit in collaboration with the unit nurse. The phlebitis grade was determined by consensus between both assessors, based on the criteria of the Maddox scale. A grade of 2 or higher was classified as phlebitis, requiring immediate PIVC removal (Ferraz-Torres et al., 2021). The time elapsed, in days, from PIVC insertion to phlebitis onset was recorded. The number of PIVCs inserted per patient during the follow‑up period and the duration of each catheter, measured in days, were recorded.

Table 1.

Criteria for judging phlebitis based on Maddox phlebitis grading scale.

Grade	Description
0	No pain, erythema, swelling, or palpable cord.
1	Pain without erythema, swelling, or palpable cord at the insertion site.
2	Pain with erythema and/or swelling without palpable cord at the insertion site.
3	Pain, erythema, swelling, induration, or a palpable venous cord <6 cm above the insertion site.
4	Pain, erythema, swelling, induration, or a palpable venous cord 6 cm or more above the insertion site.
5	Frank venous thrombosis with all previous signs and impaired or halted infusion.

Additionally, other variables were collected. Clinical and epidemiological variables included sex, categorized age (<50 years; 50–64 years; 65–79 years; ≥80 years), diabetes mellitus, hypertension, neoplasia, obesity, and hospital size (<200 beds; 200–500 beds; 500–1000 beds; >1000 beds). Device-specific variables included catheter gauge (16 G; 18 G; 20 G; 22 G; 24 G), insertion site (forearm; arm; cubital fossa; hand; wrist; lower limb; other), PIVC accessories (extension, three-way stopcock, bioconnector and/or cap), catheter securement method, number of medications administered, and dressing type (gauze, non-bordered polyurethane, partially bordered polyurethane, or fully bordered polyurethane).

Data were collected by the designated study coordinator in each participating unit throughout the 15‑day follow‑up period. Each coordinator (familiar with PIVC management and trained in standardized data recording) extracted information from the patient’s medical chart and performed daily, on‑site assessments of catheter maintenance. All entries from the paper data collection sheets were then transcribed into a secure online form, which automatically generated the study’s central electronic database.

2.3. Data analysis

Quantitative variables were described using central tendency (mean) and dispersion (standard deviation) measures, while categorical variables were expressed as absolute and relative frequencies ( %). Associations between phlebitis incidence and socio-demographic or clinical characteristics were evaluated using the chi-squared test. Incidence density was calculated as the ratio of phlebitis cases to the total number of PIVC days, multiplied by 1000.

Kaplan-Meier methods were employed to estimate the probability of remaining phlebitis-free over 15 days (360 h), with the occurrence of phlebitis considered the event of interest. Survival rates were compared using the Mantel-Cox log-rank test.

Univariate Cox proportional hazards regression models were used to assess the relationship between clinical and socio-demographic variables and phlebitis incidence. Following the STROBE guidelines, multivariable Cox regression models were used to control for confounding and thus identify independent risk factors for phlebitis. We acknowledge that inclusion of hospitals participating in the ‘Flebitis Zero’ initiative may introduce a selection bias (these centres are likely more sensitized to phlebitis prevention) yet this selection also enabled evaluation of practices in a setting committed to patient‐safety improvement.

A predictive model was developed using Cox regression, selecting variables with the lowest Akaike Information Criterion (AIC), lowest Schwarz Criterion (BIC), and highest Harrell's C-index. Parsimony was prioritized when models yielded similar results.

Model assumptions, including proportionality (using Schoenfeld residuals) and linearity, were tested and considered valid if the p-value exceeded 0.05. Statistical significance for all analyses was set at p < 0.05, except for assumption checks.

Analyses were performed using Stata v.15 (Stata Corp., College Station).

2.4. Ethical considerations

This research was performed in accordance with the principles outlined in the Declaration of Helsinki and its later amendments, as well as all relevant data protection regulations. The nurses responsible for PIVC insertion informed the patient or their legal representative about the study and obtained oral consent for participation. The study was approved by the Ethics Committee of the Principality of Asturias (Spain) under authorization code 121/14. All data were collected after removing any patient-identifiable information to ensure confidentiality. The personal data of all participants will be processed, communicated and transferred in full compliance with the provisions of the Spanish Organic Law 3/2018, of 5 December, on Personal Data Protection and the Guarantee of Digital Rights.

3. Results

The study included a total of 13,812 PIVC in 9387 participants, with a mean age of 70.76 years (SD=17.32). Of these, 4783 (51 %) were women. The mean number of PIVCs inserted per patient was 1.47, and the average duration of each PIVC was 2.90 days (SD = 2.62).

During the follow-up period, a total of 12,547 PIVCs (90.8 % of the 13,812 catheters) were censored before phlebitis onset for the following reasons: hospital discharge (4739; 34.2 %), end of treatment (2528; 18.3 %), extravasation (1744; 12.6 %), accidental removal (1554; 11.2 %), occlusion (837; 6.0 %), switch to another device (463; 3.3 %), pain (561; 4.1 %), and institutional protocol (121; 0.9 %).

Of the 13,812 PIVCs inserted, 11,388 (82.453 %) presented a Maddox grade of 0; 1122 PIVCs (8.12 %) had a Maddox grade of 1; 1042 PIVCs (7.54 %) had a Maddox grade of 2; 213 PIVCs (1.54 %) presented a Maddox grade of 3; 39 PIVCs (0.28 %) were assigned a Maddox grade of 4, and 8 PIVCs (0.06 %) were classified as having a Maddox grade of 5.

The incidence density of phlebitis was 0.14 cases per 100 person-hours. Of the 13,812 PIVCs inserted during the study period, 1302 cases of phlebitis were observed (Maddox grade 2 or higher), corresponding to a cumulative incidence of 9.43 %. Table 2 provides the cumulative incidence of phlebitis according to the clinical and epidemiological characteristics of PIVC carriers and hospital bed capacity, along with p-values obtained from chi-squared tests for the different categories of each variable.

Table 2.

Cumulative incidence of phlebitis according to clinical-epidemiological characteristics and hospital size.

Variable	n (%)	% Phlebitis	p-value
Sex
Male	6666 (48.13 %)	9.39 %	0.903
Female	7184 (51.87 %)	9.45 %
Diabetes
Yes	3490 (25.20 %)	9.34 %	0.849
No	10,360 (74.80 %)	9.45 %
Hypertension
Yes	6523 (47.10 %)	9.50 %	0.754
No	7327 (52.90 %)	9.35 %
Obesity
Yes	1783 (12.87 %)	9.70 %	0.664
No	12,067 (87.13 %)	9.38 %
Neoplastic disease
Yes	1843 (13.31 %)	10.04 %	0.331
No	12,007 (86.69 %)	9.33 %
Age
Under 50	1529 (11.04 %)	6.74 %	<0.000*
50 to 64	2360 (17.04 %)	9.19 %
65 to 79	4415 (31.88 %)	11.12 %
80 or older	5546 (40.04 %)	8.91 %
Catheter gauge
16G	33 (0.24 %)	6.06 %	<0.000*
18G	1039 (7.50 %)	3.85 %
20G	5154 (37.21 %)	9.64 %
22G	7221 (52.14 %)	10.14 %
24G	403 (2.91 %)	8.44 %
Number of drugs
0 drugs	7720 (55.74 %)	7.42 %	<0.000*
1 drug	4799 (34.65 %)	11.15 %
2 drugs	1051 (7.59 %)	15.13 %
3 drugs	223 (1.61 %)	13.45 %
4 or more drugs	57 (0.41 %)	14.04 %
Insertion site
Forearm	6587 (47.56 %)	11.34 %	<0.000*
Arm	1409 (10.17 %)	8.09 %
Cubital fossa	1709 (12.34 %)	6.61 %
Lower limbs	48 (0.35 %)	4.17 %
Hand	3055 (22.06 %)	7.00 %
Wrist	1026 (7.41 %)	11.11 %
Other	16 (0.12 %)	6.67 %
Hospital beds
200 beds or fewer	3278 (23.67 %)	8.41 %	<0.000*
201 to 500 beds	6217 (44.89 %)	7.51 %
501 to 1000 beds	3101 (22.39 %)	13.38 %
Over 1000 beds	1254 (9.05 %)	10.34 %
Accessory
Extension + three-way stopcock	1793 (12.95 %)	8.37 %	0.008*
Three-way stopcock + bioconnector	229 (1.65 %)	10.92 %
Extension	1721 (12.43 %)	8.95 %
Extension + bioconnector + three-way stopcock	1230 (8.88 %)	11.63 %
Extension + bioconnector	4004 (28.91 %)	10.24 %
Bioconnector	3597 (25.97 %)	9.06 %
Three-way stopcock	1021 (7.37 %)	7.64 %
Cap	255 (1.84 %)	7.45 %
Dressing type
Gauze	46 (0.33 %)	13.04 %	<0.000*
Non-bordered polyurethane	689 (4.97 %)	14.80 %
Partially bordered polyurethane	8238 (59.48 %)	8.63 %
Fully bordered polyurethane	4877 (35.21 %)	9.97 %
Strips
Away from insertion site	11,811 (85.28 %)	9.41 %	0.825
Over insertion site	1062 (7.67 %)	9.89 %
No strips	977 (7.05 %)	9.11 %
Catheter hours
Under 48 h	6092 (43.99 %)	8.81 %	0.041*
48 to 71 h	2690 (19.42 %)	10.22 %
72 to 96 h	1876 (13.55 %)	10.66 %
Over 96 h	3192 (23.04 %)	9.18 %

*: statistically significant value (p < 0.05).

Fig. 1, Fig. 2, Fig. 3, Fig. 4 illustrate the probability of remaining phlebitis-free over 15 days (360 h), both overall and stratified by clinical and sociodemographic variables. Fig. 1 demonstrates that catheter survival does not significantly vary with insertion time, indicating that time alone is not a predictor of catheter survival. Fig. 2 reveals a notable difference in catheter survival between males and females, with females exhibiting lower survival, suggesting that female sex may be a risk factor for catheter survival. In Fig. 3, it is observed that patients with diabetes have slightly lower cumulative survival compared to those without diabetes, implying that diabetes may be associated with an increased risk of phlebitis. Lastly, Fig. 4 shows that arterial hypertension does not appear to significantly impact catheter survival, as the survival curves for patients with and without hypertension are similar. These results suggest that factors such as sex and diabetes may influence catheter survival, while arterial hypertension does not seem to be a significant factor.

Fig. 1.

Probability of remaining phlebitis-free over the 15-day follow-up period, overall and stratified by sex, the presence or absence of diabetes, and the presence or absence of hypertension.

Fig. 2.

Probability of remaining phlebitis-free over the 15-day follow-up period, stratified by the presence or absence of obesity, the presence or absence of cancer, categorised age, and catheter gauge.

Fig. 3.

Probability of remaining phlebitis-free over the 15-day follow-up period, stratified by the accessories used, the anatomical insertion site of the PIVC, and the number of medications administered through the PIVC.

Fig. 4.

Probability of remaining phlebitis-free over the 15-day follow-up period, stratified by dressing type, the placement of securing strips relative to the PIVC insertion site, and the number of hospital beds.

Table 3 presents the results of the Mantel-Cox log-rank test. Table 4 shows the hazard ratio (HR), 95 % confidence intervals (CI), and p-values for each clinical and sociodemographic variable analysed using univariate Cox proportional hazards regression models.

Table 3.

Mantel-cox log-rank test.

Variable	χ²	p-value
Sex	7.76	0.005*
Diabetes	1.43	0.119
Hypertension	0.74	0.387
Neoplasia	0.12	0.726
Obesity	0.17	0.682
Categorised age	11.31	0.011*
Catheter gauge	22.25	<0.000*
Number of drugs	67.48	<0.000*
Insertion site	44.94	<0.000*
Accessories	16.12	0.024*
Dressing type	36.72	<0.000*
Strips	0.02	0.991
Hospital beds number	74.91	<0.000*
*: statistically significant value (p < 0.05)

Table 4.

Hazard ratio (HR), 95 % CI, and p-value for clinical and sociodemographic variables analyzed in relation to phlebitis incidence.

Variable	HR	95 % CI	p-value
Sex
Male	Reference category
Female	1.168	1.045 to 1.305	0.006*
Diabetes	0.904	0.795 to 1.029	0.126
Hypertension	0.953	0.853 to 1.065	0.397
Obesity	1.034	0.878 to 1.218	0.687
Neoplastic disease	1.029	0.877 to 1.207	0.730
Age
Under 50	Reference category
50 to 64	1.185	0.935 to 1.502	0.161
65 to 79	1.236	1.003 to 1.533	0.044*
80 or older	1.017	0.820 to 1.261	0.878
Catheter gauge
16G	Reference category
18G	0.616	0.149 to 2.552	0.504
20G	1.273	0.317 to 5.106	0.733
22G	1.291	0.322 to 5.173	0.718
24G	1.087	0.260 to 4.534	0.909
Number of drugs
0 drugs	Reference category
1 drug	1.387	1.229 to 1.564	<0.000*
2 drugs	1.944	1.622 to 2.329	<0.000*
3 drugs	1.528	1.058 to 2.206	0.024*
4 or more drugs	1.635	1.122 to 3.287	0.013*
Insertion site
Forearm	Reference category
Arm	0.696	0.570 to 0.849	<0.000*
Cubital fossa	0.658	0.537 to 0.805	<0.000*
Lower limbs	0.348	0.08 to 0.619	<0.000*
Hand	0.687	0.588 to 0.803	<0.000*
Wrist	0.988	0.109 to 1.210	0.912
Other	0.779	0.109 to 5.541	0.809
Hospital beds
200 beds or fewer	Reference category
201 to 500 beds	0.818	0.704 to 0.942	0.008*
501 to 1000 beds	1.466	1.267 to 1.697	<0.000*
Over 1000 beds	1.024	0.836 to 1.254	0.821
Accessories
Extension + three-way stopcock	Reference category
Three-way stopcock + bioconnector	1.169		0.485
Extension	0.913		0.442
Extension + bioconnector + three-way stopcock	1.337		0.015*
Extension + bioconnector	1.097		0.342
Bioconnector	0.975		0.804
Three-way stopcock	0.947		0.697
Cap	0.845		0.502
Dressing type
Gauze	Reference category
Non-bordered polyurethane	1.359	0.552 to 3.344	0.504
Partially bordered polyurethane	0.718	0.298 to 1.732	0.461
Fully bordered polyurethane	0.814	0.337 to 1.967	0.648
Strips
Away from insertion site	Reference category
Over insertion site	1.012	0.826 to 1.242	0.904
No strips	0.968	0.796 to 1.244	0.968
*: statistically significant value (p < 0.05)

Table 5 displays the final predictive model derived from Cox proportional hazards regression, selected from all possible equations. This model demonstrated the lowest AIC (21,638.7) and BIC (21,758.1) and the highest Harrell's C-index (0.813). The Schoenfeld residual test yielded a result of 0.196 (z = 1.01; p = 0.317), confirming the proportionality assumption of the estimated model. The analysis of the squared linear predictor produced a result of 0.099 (z = −1.19; p = 0.191), supporting the linearity assumption of the final model.

Table 5.

Hazard ratio (HR), 95 % CI, and p-value of the final estimated model.

Variable	HR	95 % CI	p-value
Sex
Male	Reference category
Female	1.204	1.077 to 1.346	0.001*
Number of drugs
0 drugs	Reference category
1 drug	1.379	1.223 to 1.557	<0.000*
2 drugs	1.991	1.661 to 2.387	<0.000*
3 drugs	1.558	1.078 to 2.249	0.018*
4 or more drugs	1.784	1.125 to 2.315	0.024*
Hospital beds number
200 beds or fewer	Reference category
201 to 500 beds	0.789	0.679 to 0.916	0.002*
501 to 1000 beds	1.444	1.247 to 1.673	0.024*
Over 1000 beds	1.036	0.846 to 1.269	0.736
*: statistically significant value (p < 0.05)

4. Discussion

In the present study, the cumulative incidence of phlebitis was found to be 9.43 %. This value exceeds the 5 % threshold recommended by the Infusion Nursing Society (Gorski et al., 2021). The variability in phlebitis incidence is considerable across different studies: in the PREBACP study (Blanco-Mavillard et al., 2021), conducted in seven Spanish hospitals, the cumulative incidence ranged between 13.43 % and 16.66 %. In the multicentric CATHEVAL study (Miliani et al., 2017) conducted in France, the cumulative incidence of phlebitis was 4.1 %. In the study by Simões et al. (2022) in Brazil, the cumulative incidence of phlebitis was reported as 6.1 %. Liu et al. (2022) calculated a cumulative incidence of 10.5 % in their study, whereas Göransson et al. (2017), using various measurement tools on the same population, reported a cumulative incidence ranging from 11.7 % to 33.6 % (Göransson et al., 2017).

In the present study, most PIVCs examined had a Maddox grade of 2 or 3, while higher Maddox grades constituted a minimal proportion of the total. This finding is consistent with the work of Simin (2019), where the highest number of phlebitis cases corresponded to a Maddox grade of 3 (Simin et al., 2019), and with the research by Liu et al. (2022), where most PIVCs exhibited a Maddox grade of 1 or 2 (Liu et al., 2022). This trend could be attributed to a higher level of early recognition of phlebitis signs by nurses during the assessment of PIVC conditions.

In the present study, it was observed that the risk of phlebitis was highest between 48 and 96 h post-PIVC insertion, while this risk decreased after 96 h. This finding aligns with the report by Miliani et al. (2017), which indicated that the main adverse events related to PIVC insertion occurred within the first four days. Liu et al. (2022) further specified that the time period between 24 and 72 h post-insertion was when the risk of adverse events was greatest. Wei et al. (2019) concluded that the risk of phlebitis decreased after 38 h post-insertion.

These observations differ from those reported by Cicolini et al. (2014), who indicated that the probability of phlebitis increased proportionally with the duration since PIVC insertion, reaching its peak after 96 h of device dwell time. Nevertheless, some studies (Rickard et al., 2012) have shown no benefit from routine PIVC replacement (e.g., every 72–96 h) and have suggested that replacing PIVCs based on clinical indications (such as phlebitis or extravasation) is safe and spares patients the pain associated with a new PIVC insertion. According to the 2024 INS Standards of Practice, PIVCs should be removed when clinically indicated, predicated on accurate and consistent vascular access device assessment based on patient and infusate risk, strict adherence to Aseptic Non Touch Technique (ANTT®) principles, and early recognition and management of complications (Gidaro et al., 2025). Liu et al. (2022) recommend replacing a PIVC with a midline catheter or a central venous catheter if the dwell time is expected to exceed seven days.

In this study, female sex emerged as a predictor of PIVC-associated phlebitis. This finding is consistent with other studies (Jiménez-Martínez et al., 2024; Martin-Loeches et al., 2019; Ozger et al., 2021; Webster et al., 2019). This may be due to women having smaller caliber blood vessels compared to men, which could result in greater endothelial damage (Marsh et al., 2018). Additionally, differences in hormonal distribution and adipose tissue between sexes may contribute to this increased risk in women (Fox-Rawlings et al., 2018).

In our study, we observed a slight decrease in cumulative survival among diabetic patients. A systematic review that included thirty-five studies (20,697 catheters used in 15,791 patients) found that although two studies identified diabetes mellitus as a risk factor for PIVC‑associated phlebitis, the meta‑regression did not show a statistically significant association between the proportion of diabetic patients and the overall incidence of phlebitis (MC = 0.354; p = 0.538) (Lv y Zhang, 2020). Nevertheless, vascular access guidelines recognize that diabetes increases cannulation difficulty and susceptibility to local inflammatory reactions (Nickel et al., 2024).

According to our results, the presence of arterial hypertension did not affect PIVC survival. Likewise, a recent systematic review on peripheral catheter failure and infection found no statistical association between comorbidities such as hypertension and device failure rates (Marsh et al., 2024). This suggests that, unlike diabetes or sex, hypertension is not a determining factor for phlebitis or catheter occlusion in hospital settings.

Another predictor of PIVC-associated phlebitis risk is the site of catheter insertion. In the present study, the forearm and wrist were the anatomical regions with the highest risk. PIVCs inserted in the antecubital fossa demonstrated a lower and statistically significant risk compared to those inserted in the forearm or wrist, as also reported by Cicolini et al. (2014) (Cicolini et al., 2014). This finding differs from Carr's study (2018), which concluded that PIVCs inserted in the antecubital fossa had a higher risk of adverse events compared to those inserted in the arm or forearm (Carr et al., 2018). Similarly, the study by Miliani et al. (2017) found a higher risk of phlebitis in the antecubital fossa compared to veins on the dorsum of the hand (Miliani et al., 2017). Simões et al. (2022) identified the dorsum of the hand as the insertion site with the highest risk of developing PIVC-associated phlebitis.

Several studies in neonatal and pediatric populations suggest that catheter insertions in the lower limbs may be less prone to phlebitis than those in the upper limbs. In a cohort of neonates, Elmekkawi et al. (2018) observed significantly fewer catheter‑related complications—including phlebitis—in lower‑extremity PICCs, with a relative risk of 0.66 (95 % CI: 0.48–0.90; p = 0.007) compared to upper‑limb insertions (Elmekkawi et al., 2018). Although data on adult peripheral IV cannulas are more limited, Wrightson (2013) hypothesized that the larger caliber of lower‑limb veins and reduced joint movement around the ankle and foot lead to less endothelial trauma and inflammatory activation, thereby mitigating the development of phlebitis (Wrightson, 2013). Our finding of a markedly lower hazard of phlebitis in lower‑limb PIVCs (HR 0.348; 95 % CI: 0.080–0.619; p < 0.001) aligns with these mechanistic and clinical observations, although further prospective studies in adult populations are warranted to confirm these site‑specific effects

No statistically significant differences were identified regarding catheter gauge, similar to the findings of Simões et al. (2022). However, the Infusion Nurses Society (INS) recommends using the smallest gauge catheter (Gidaro et al., 2025). This is because smaller gauge catheters provide more cushioning space around them, resulting in less mechanical or chemical trauma and reduced irritation to the inner walls of the veins.

In this study, it was observed that the risk of developing phlebitis increases progressively with age, with this risk being statistically significant among individuals aged 65 to 79 years. A similar finding was reported by Cernuda Martínez et al. (2024), where the highest incidence of phlebitis was observed in individuals aged 50 to 79 years. This can be explained by two factors: firstly, ageing leads to changes in the skin and subcutaneous tissue, making PIVC insertion more likely to induce adverse effects (Ascoli et al., 2012). Secondly, the ageing process results in a decline in immune response (immunosenescence), which increases the risk of developing PIVC-associated phlebitis (Cernuda Martínez et al., 2024). However, the study by Simões et al. (2022) did not find a statistically significant relationship between age and phlebitis.

In our cohort, the only accessory configuration associated with a significantly increased phlebitis hazard was the combination of extension set + bioconnector + three‑way stopcock (HR 1.337; p = 0.015). This finding is concordant with in vitro data showing that each additional connector point induces local mechanical stress and micro‑turbulence at the catheter hub, exacerbating endothelial injury and promoting phlebitis (Lv y Zhang, 2020). Moreover, a recent systematic review and meta‑analysis comparing integrated short peripheral intravenous cannulas—which feature preassembled extension lines and connectors—with composite assemblies of separate extension sets, needleless connectors, and stopcocks reported a trend toward lower thrombophlebitis rates for the integrated devices (RR 0.91; 95 % CI 0.78–1.07; p = 0.25) (Gidaro et al., 2025). Together, these data underscore the importance of minimizing catheter‑hub complexity to reduce PIVC‑related inflammation and phlebitis risk.

Two additional characteristics were identified as predictors of elevated phlebitis risk due to their statistically significant associations: larger hospital size (determined by the number of beds) and the number of medications administered through the PIVC (with risk increasing as the number of medications rises).

The relationship between hospital size and the incidence of PIVC-associated phlebitis is an intriguing and complex one. Our study identified hospital size as a significant predictive factor for phlebitis, and this finding is supported by recent research. Previous results from the Flebitis Zero project found that hospitals with >500 beds had a higher incidence of phlebitis, with an odds ratio of 1.507 (p < 0.001) (Cernuda Martínez et al., 2024). This suggests that larger hospitals may face unique challenges in managing PIVC-related complications. Several factors could contribute to this association. Larger hospitals often have more complex patient populations and higher patient turnover, which may increase the risk of phlebitis due to more frequent catheter insertions and replacements. Additionally, larger hospitals may have more diverse nursing staff, potentially leading to variations in catheter care practices. Furthermore, the higher volume of patients in larger hospitals could strain resources, leading to less frequent monitoring of catheter sites and delayed recognition of phlebitis (da Silva Furlan et al., 2024). Similarly, studies of healthcare‑associated infections (HCAI) have shown that hospitals with larger bed capacity exhibit higher rates of device‑related complications, alluding to greater care complexity, procedural volume, and resource strain (Hixson et al., 2024; Iancu et al., 2023). In a nationwide survey of Brazilian hospitals, HCAI prevalence was significantly higher in large facilities (≥200 beds; 13.5 %) compared to medium (50–199 beds; 7.7 %) and small (<50 beds; 5.5 %) hospitals (Fortaleza et al., 2017).

Irritating fluids and medications, such as antibiotics, vasoactive drugs, or anti-haemorrhagic agents, increase the risk of PIVC phlebitis (Marsh et al., 2018; Wei et al., 2019). Chen et al. (2022) suggested that the infusion of irritating fluids, daily infusion duration, and the type of flushing solution are independent predictors of adverse PIVC events, including phlebitis (Chen et al., 2022).

In our study, we observed that administration of two drugs was associated with a higher risk of phlebitis (HR 1.944) compared with four or more agents. The apparently paradoxical finding of a higher hazard ratio for phlebitis with two drugs compared to four or more may reflect the overriding influence of drug irritancy and infusion practices rather than merely the count of medications. In vitro work by Drouet et al. (2021) demonstrated that mechanical trauma from sequential infusions through the same PIVC—simulating multidrug Y‑site administration—can itself exacerbate endothelial injury and promote phlebitis, independent of solute concentration fluctuations (Drouet et al., 2021). Conversely, clinical studies illustrate that rigorous flushing protocols markedly reduce phlebitis risk: in a prospective cohort of patients receiving amphotericin B, Ahimbisibwe et al. (2019) reported that intercalated 1 L normal‑saline flushes after each dose cut phlebitis incidence despite high irritant load (Ahimbisibwe et al., 2019), while a cluster‑randomized trial by Bertolino et al. (2012) showed that intermittent saline (or low‑dose heparin) locks reduced catheter‑related phlebitis and occlusion compared with no flush (Bertolino et al., 2012). It is therefore plausible that patients receiving four or more agents also underwent more frequent non‑irritant flushes or saline locks, diluting residual irritants and lowering overall phlebitis risk relative to those receiving only two irritant infusions without adequate flushing. Future work should dissect the relative contributions of flush volume, timing, and drug irritancy to optimize PIVC management.

4.1. Limitations

This study has some limitations. The monitoring of PIVCs was conducted by different nurses, which may have introduced variability in the interpretation of Maddox grades. Only the number of medications administered through the PIVC was considered, without accounting for the specific types of medications used. The type of phlebitis (mechanical, infectious, or chemical) was not evaluated. The material from which the PIVCs were manufactured was not considered. Possible concomitant signs of phlebitis, such as pain, were not assessed. Lastly, it was not determined whether the phlebitis occurred in the first PIVC inserted or in subsequent PIVCs. This study did not account for patients with prior or recent receipt of a PIVC. A key limitation of this study is the exclusion of the most critically ill patients—those managed in emergency departments, pediatric wards, and intensive care units—who may have distinct risk profiles, thereby potentially limiting the generalizability of our findings to these high‑acuity populations.

5. Conclusions

The main predictive factors for the development of PIVC-associated phlebitis were female sex, the number of infused medications, patient age, hospital size, and the dwell time of the PIVC. In this regard, PIVCs that do not show signs of phlebitis within 96 h can remain in use for a longer duration without an increased risk of its occurrence. All these characteristics should be considered to reduce the risk of PIVC-associated phlebitis.

In the present study, it is not possible to determine the suitability of using the antecubital fossa as an optimal insertion site, highlighting the need for further research to clarify this aspect.

Funding: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Flebitis Zero Group (in alphabetical order): Ainhoa Bolinaga Gullón (Hospital de Mendaro), Alejandra Tiburcio Domínguez (Hospital de Zafra), Alexandra Lucia Moisé (Pius Hospital de Valls), Alexandra Xavier Aller (Hospital Carmen y Severo Ochoa), Alfredo Cano Reyes (Hospital General Universitario Reina Sofía de Murcia), Alicia Peña Rodríguez (Fundación Hospital Avilés), Almudena Quintás Viqueira (Hospital Universitario La Paz), Amanda Acosta Arrocha (Hospital Doctor José Molina Orosa), Ana Belén Arredondo Provecho (Complejo Asistencial Universitario de León), Ana Belén Diez Álvarez (Hospital Universitario Central de Asturias), Ana Belén Lozano (Complejo Asistencial de Zamora), Ana Belén Martín Díaz (Hospital General de Fuerteventura), Ana Belén Moreno López (Hospital Universitario de Jaén), Ana Gómez Rodríguez (Hospital Universitario Infanta Elena), Ana Mª Rodríguez Núñez (Hospital Universitario de Basurto), Ana Mª Rus Moreno (Hospital Universitario de Jaén), Ana Teresa Domínguez Martin (Complejo Hospitalario Universitario de Cáceres), Ana Álvarez Morales (Hospital Universitario de Fuenlabrada), Andrés Jumilla Burugorría (Hospital General de la Defensa de Zaragoza), Anna Besolí Codina (Consorci Sanitari de Vic), Arantza Arroitajauregi Gutierrez (Hospital de Górliz), Águeda Barcina Valle (Hospital Santiago Apóstol), Ángela Carmen García Carro (Hospital Universitario Río Hortega), Bárbara Álvarez Martínez (Hospital Universitario de Móstoles), Beatrice Bejerano Ramos (Hospital Virgen de las Montañas), Beatriz Marín González (Hospital Quirónsalud Córdoba), Beatriz Pacho Marín (Hospital de Santa Marina), Beatriz de Llano Rivera (Gerencia de Asistencia Sanitaria del Bierzo), Berta Gorlat Sánchez (Hospital Universitario Virgen de las Nieves), Camino Del Río Pisabarro (Hospital Universitario Donostia), Carlos Núñez Ortiz (Hospital Universitario de La Línea de La Concepción), Carmelo Villafanca Renes (Hospital Universitario de Burgos), Carmen Arribas del Cid (Complejo Asistencial de Ávila), Carmen Martínez Ortega (Hospital Valle del Nalón), Celia Herrán Bergado (Hospital Santiago Apóstol), Cesar M Gómez Álvarez (Hospital Virgen de las Montañas), Claudia Ruiz-Huerta García de Viedma (Hospital Central de la Cruz Roja), Conchita Hernández Magide (Parc Taulí Hospital Universitari), Cristina Berbel Bertolo (Hospital Quirónsalud Barcelona), Cristina Díaz Garzón Marín (Hospital del Bidasoa), Cristina López Díaz (Hospital Universitario de Salamanca), Cristina de la Rosa Pastor (Hospital Clínico Universitario de Valladolid), César Otero Gutiérrez (Hospital Universitario de Fuenlabrada), David Pineda Fernández (Hospital Universitari Sagrat Cor), David Sanjuán Hernández (Hospital General de La Palma), Delia Milagros Gallardo Ferrer (Hospital Universitario de Jerez de la Frontera), Dolors Mas Rubio (Althaia Xarxa Asistencial Universitària de Manresa), Dolors Pintado Ferreño (Complejo Hospitalario Moisès Broggi), Dácil Rosario de León Pérez (Hospital General de Fuerteventura), Elena Sancho Sena (Complejo Hospitalario de Navarra), Elena Vidal Díez (Hospital de Mataró), Encarna Maraver Bermúdez (Althaia Xarxa Asistencial Universitària de Manresa), Encarna Moreno Castañeda (Parc Sanitari Sant Joan de Déu), Estefanía López Cabrera (Hospital Universitario Reina Sofía de Córdoba), Ester Marco Juan (Hospital VITHAS Medimar), Estrella Diego Blanco (Hospital de Cabueñes), Eva Martín Gil (Complejo Asistencial Universitario de Soria), Eva Redón Ruiz (Hospital Universitari Mollet), Eva Vidarte Uriarte (Hospital García Orcoyen), Faustino Gonzalez Menéndez (Hospital Universitario Marqués de Valdecilla), Francisca Matas Aguilera (Hospital Público de Montilla), Gemma Alonso Fernández (Centro Médico de Asturias), Glòria Noguer Padrosa (Hospital d'Olot i Comarcal de la Garrotxa), Greyce de Jesús Ferreira (Hospital Universitario de Poniente), Ignacio Enríquez de Salamanca Holzinger (Hospital POVISA), Inés Narrillos Martin (Hospital Universitario La Paz), Inmaculada Alonso Araujo (Hospital Universitario Virgen del Rocío), Irene Aznar Vázquez (Hospital San Juan de Dios de Zaragoza), Jaume Fernández Roige (Hospital Universitario de la Luz), Javier Lozano García (Hospital Universitario de Burgos), Javier Silva Contreras (Hospital Virgen de la Luz), Joaquín López-Contreras González (Hospital de la Santa Creu i Sant Pau), José Ángel Franco Romero (Hospital Universitario Juan Ramón Jiménez), José Manuel García Garrido (Hospital Comarcal de Inca), Josefa Segura Riquelme (Hospital Universitario de Torrecárdenas), Josefina Arana Basterrechea (Hospital San Eloy), José Antonio Sánchez García (Hospital Ciudad de Coria), José Luis Mendoza García (Hospital Universitario de Torrevieja), Juan Carlos Durán Alonso (Hospital San Juan Grande-San Juan de Dios), Juan del Oso Morán (Hospital Universitario Vall d'Hebrón), Judit Santamaria Rodríguez (Hospital Universitari d'Igualada), Judtih Baldris Catafau (Hospital Universitari General de Catalunya), Kevin Bliek Bueno (Hospital Universitario Príncipe de Asturias), Kizkitza Lasa Garmendia (Hospital de Zumárraga), Lourdes Bosch Navarro (Fundació Hospital de l’Esperit Sant), Lourdes Sangrador Calzada (Complejo Asistencial Universitario de Palencia), Lucrecia López González (Complejo Hospitalario Moisès Broggi), Mª Angels Fernández Labrada (Fundació Hospital de l’Esperit Sant), Mª José Zapico Baragaño (Hospital Vital Álvarez Buylla).

CRediT authorship contribution statement

José Antonio Cernuda-Martínez: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. José Luis Cobo-Sánchez: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Eva María Alarcón-Duque: Writing – review & editing, Writing – original draft, Conceptualization. Esther Moreno-Rubio: Writing – review & editing, Writing – original draft, Formal analysis, Conceptualization. María Belén Suárez-Mier: Writing – review & editing, Writing – original draft, Formal analysis, Conceptualization. María del Camino del Río-Pisabarro: Writing – review & editing, Writing – original draft, Methodology, Conceptualization. Marta Ferraz-Torres: Writing – review & editing, Writing – original draft, Investigation, Formal analysis, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.