A Descriptive Overview of Deviations From an Institutional Peripheral Vasopressor Protocol and Its Effects on Safety Across an Academic Health System

Abstract

OBJECTIVES:

This study characterized deviations from an institutional peripheral vasopressor protocol and assessed associated extravasation events.

DESIGN:

Retrospective observational study of electronic medical records data.

SETTING:

Ten hospitals in an academic health system from October 2020 to May 2023.

PATIENTS:

All admitted adult patients who received vasopressor infusions.

INTERVENTIONS:

Implementation of a peripheral vasopressor protocol in October 2020.

MEASUREMENTS AND MAIN RESULTS:

Three thousand five hundred eighteen patients were identified as having received peripheral vasopressors from October 2020 to May 2023, with 1258 patients suspected of having a deviation from the protocol. One hundred sixty-two patients were randomly selected for chart review, and 73 patients had confirmed deviation. Sixty-three (86.3%) exceeded the maximum allowed dose, 23 had multiple simultaneously infusing peripheral vasopressors, 13 had a peripheral vasopressor infusing longer than 24 hours, and 13 received peripheral vasopressin. Most deviations (79.4%) occurred due to worsening shock while waiting for central access. Extravasation occurred in 10 patients (13.7%), with three patients requiring phentolamine treatment. Fifty-six patients (76.7%) subsequently required central access for continued vasopressor support. In a secondary analysis of the original source population, 47 of 3518 patients (1.3%) received phentolamine for peripheral vasopressor extravasation, of whom the majority (72.4%) were protocol concordant at the time of extravasation.

CONCLUSIONS:

Deviations from a peripheral vasopressor protocol primarily involved exceeding the protocol dose limit in the setting of worsening shock while awaiting central access. Clinically significant extravasations were infrequent even among patients with protocol deviations and most treated extravasations occurred in protocol-concordant patients, supporting the overall safety of peripheral vasopressors.

KEY POINTS

Question: What are the types and causes of deviations from an institutional peripheral vasopressor protocol, and do peripheral vasopressors remain safe despite these deviations?

Findings: In a retrospective cohort, most protocol deviations involved exceeding the maximum allowed dose in the setting of worsening shock. Among patients with deviations, 4.1% required phentolamine for extravasation treatment, and most treated extravasations across the health system occurred in protocol-concordant patients.

Meaning: Deviations from an institutional peripheral vasopressor protocol were driven primarily by clinical urgency, and clinically significant extravasation was rare, supporting the safety of peripheral vasopressors.

Vasopressors are used in the ICU to support patients with shock. Peripheral administration is increasingly common and generally considered safe (1–6), reducing time to vasopressor initiation (7), central venous catheter (CVC) utilization (8), and central line-associated blood infections (CLABSIs) (9). Consequently, the 2021 Surviving Sepsis guidelines recommend prompt initiation of peripheral vasopressors in septic shock (10).

While peripheral vasopressor use is increasingly adopted, data are lacking on protocol adherence, why deviations occur, and their impact on safety. Protocols have been developed to standardize use and minimize adverse events, but they vary widely in allowed vasopressors, dosing, duration, monitoring requirements, and extravasation protocols (1, 6, 9, 11–13). Our academic health system implemented a peripheral vasopressor protocol in October 2020 (Supplemental Table 1, https://links.lww.com/CCX/B675) (4) to reduce the risk of extravasation while preserving the benefits of early peripheral initiation. No protocol was previously in place, creating ambiguity around acceptable and safe peripheral vasopressor use and leading to provider variation. We sought to characterize peripheral vasopressor use that deviated from this protocol. Our secondary objective was to assess extravasation events requiring phentolamine administration.

METHODS

This was a retrospective study of adult patients who received any peripheral vasopressor infusion across our 10-hospital academic health system between October 2020 and May 2023. Patients were identified using a Learning Health System database as previously described (4, 14, 15). Route of vasopressor infusion was determined by the presence of central access and cross-referenced against a pharmacy dispensing database. Peripheral vasopressor use was limited to a single vasopressor for a maximum duration of 24 hours, with no peripheral administration of vasopressin. Peripheral IV (PIV) and dosing requirements are described in the Supplement (https://links.lww.com/CCX/B675). Vasopressors studied included norepinephrine, epinephrine, vasopressin, phenylephrine, dopamine, and dobutamine. Deviations were defined as exceeding protocol dose or duration limits, infusion of multiple peripheral vasopressors simultaneously, or peripheral administration of vasopressin.

A weighted random sample representing 12.8% of the patients suspected of having a protocol deviation was selected for manual chart abstraction. Sample weights were based on hospital volume of peripheral vasopressor use, with selection proportional to each hospital to ensure representation from all hospital types (based on Vizient cohorts) (16). Sampling weights were applied only at the selection stage; the primary analysis did not include post hoc adjustments. Data abstracted included baseline patient characteristics, hospital admission characteristics, vasopressor data, PIV characteristics, and extravasation events. These data were compared with the protocol to confirm whether deviations had occurred. Chart review then characterized the reasons underlying each deviation.

Analysis of Phentolamine Usage

This secondary analysis of a subset from the original source population identified all patients who received phentolamine for peripheral vasopressor extravasation regardless of protocol concordance. Extravasation event characteristics, vasopressor data, and protocol adherence were collected. Extravasations before hospital arrival, in the operating room, or associated with bolus vasopressor administration were excluded.

Statistical Analysis

Study data were collected and managed using Research Electronic Data Capture tools hosted at the University of Minnesota (UMN). Continuous variables were presented as median and interquartile range (IQR) and categorical variables as numbers and percentages. Statistical analysis was performed with R Statistical Software (v4.2.7; R Core Team, 2024) (17). The primary analysis used descriptive statistics only. Secondary analysis used Wilcoxon rank-sum test for continuous variables and Fisher exact test for categorical variables. This study was deemed exempt by the UMN Institutional Review Board (STUDY00019070).

RESULTS

Deviation Cohort

We identified 3518 patients who received peripheral vasopressor infusions from October 2020 to May 2023. Of 1258 patients suspected of having a possible deviation from the protocol, 162 patients (12.8%) were randomly selected for manual chart abstraction. Of these, 73 patients were identified as having a true deviation (Fig. 1). Sixty-five patients (89.0%) received norepinephrine, 13 (17.8%) received vasopressin, 10 (13.7%) received phenylephrine, 6 (8.2%) received epinephrine, 2 (2.7%) received dopamine, and 1 (1.4%) received dobutamine peripherally (Table 1). The median maximum dose for norepinephrine was 0.2 µg/kg/min (IQR, 0.14–0.33 µg/kg/min). The median duration of peripheral vasopressor infusion was 5.6 hours (IQR, 3.0–15.7 hr). Fifty-six of 73 patients (76.7%) ultimately required central access for ongoing vasopressor therapy.

Figure 1.

Study flow diagram showing the identification of patients who had deviated from the peripheral vasopressor protocol. Using an algorithm with an electronic health record (EHR) database, 162 patients who were suspected of deviating from the protocol were randomly selected and manually reviewed. Seventy-three of those patients had confirmed deviations. Each deviation from the protocol was characterized and 25 of those patients had multiple deviations. Among the 89 patients who did not have a true deviation or were not included based on the inclusion criteria, the discrepancies between the algorithm’s suspected deviations and the manual chart review were characterized. ED = emergency department, OR = operating room.

TABLE 1.

Patient, Hospital, and Infusion Characteristics

Patient characteristics
Age (yr)	68 (56–77)
Male	38 (52%)
Elixhauser score	12 (3–20.5)
Laboratory-based Acute Physiology Score, version 2 admit score	172 (140–212)
Hospital stay characteristics
Length of stay (d)	10 (3.5–19.5)
ICU length of stay (d)	4 (1.5–6.0)
Mortality	38%
Location of vasopressor initiation
ICU/ward	50 (68.5%)
Emergency department	13 (17.8%)
Operating room	10 (13.7%)
Hospital typea
Academic	26 (35.6%)
Large, specialized	18 (24.7%)
Complex care	15 (20.5%)
Community	14 (19.2%)
Peripheral vasopressor characteristics
Number of patients (median maximum dose)
Norepinephrine, µg/kg/min	65 (0.2)
Vasopressin, U/hr	13 (2.4)
Phenylephrine, µg/kg/min	10 (1.5)
Epinephrine, µg/kg/min	6 (0.265)
Dopamine, µg/kg/min	2 (12.5)
Dobutamine, µg/kg/min	1 (3.0)
Median duration (hr)	5.6 (3.0–15.7)
IV characteristics
Gauge
16	2 (2.7%)
18	31 (42.5%)
20	17 (23.3%)
22	5 (6.8%)
Other	18 (24.6%)
Location
Forearm	48 (65.8%)
Upper arm	7 (9.6%)
Hand	3 (4.1%)
Wrist	2 (2.7%)
Antecubital fossa	1 (1.4%)
Other	12 (16.4%)
Multiple peripheral IVs required	17 (23.3%)
Extravasation events	10 (13.7%)
Central access required	56 (76.7%)

^aHospital type categorized based on Vizient cohorts (16).

Data are presented as n (%) or median (interquartile range), unless otherwise specified.

About one-third of patients (34.2%) had more than one deviation, resulting in a total of 137 deviations. Sixty-three patients (86.3%) exceeded the protocol dose limit, 23 (31.5%) had multiple peripheral vasopressors infusing simultaneously, 13 (17.8%) had peripheral vasopressor infusions lasting longer than 24 hours, and 13 (17.8%) had vasopressin infusing peripherally. Patients also had multiple reasons for protocol deviations. Fifty-eight patients (79.4%) had their deviations due to worsening shock while awaiting central access, 7 (9.5%) were in the end-of-life phase, 4 (5.5%) declined central access, 3 (4.1%) had intermittent vasopressor needs, and 15 (20.5%) had other or unknown reasons. Other reasons included the inability to obtain central access, active bacteremia with a history of CLABSIs, and expectation of minimal vasopressor requirement. Among these 73 patients with protocol deviations, 10 (13.7%) had reported extravasation with 3 (4.1%) receiving phentolamine as an antidote.

Extravasation Cohort

Of the original cohort of 3518 patients receiving any peripheral vasopressors regardless of protocol adherence, 47 received phentolamine for extravasation of infusing peripheral vasopressors (Supplemental Table 2, https://links.lww.com/CCX/B675). Of these, 39 patients (83.0%) received norepinephrine with a median dose of 0.05 µg/kg/min, eight received phenylephrine, two received epinephrine, and one received vasopressin. Among the 47 patients treated with phentolamine for peripheral vasopressor extravasation, 13 (27.6%) had protocol deviations at the time of extravasation, whereas 72.4% were protocol concordant. Of the 13 patients with deviations (patients could have more than one deviation type), 10 exceeded the maximum allowed dose, three had more than one vasopressor infusing peripherally, two exceeded the 24-hour limit, and one had vasopressin infusing peripherally. Infiltration grade was documented for 37 of 47 extravasation events using the Infusion Nurses Society Infiltration Scale (18), with the majority graded two or lower. One patient required surgical consultation for tissue necrosis.

DISCUSSION

In this multicenter study, deviations from a peripheral vasopressor protocol were mostly driven by clinical urgency and were rarely associated with clinically significant harm. Deviations most often occurred due to exceeding the protocol dose limit while awaiting placement of central access in the setting of worsening shock. While there was a higher rate of extravasation compared with the existing literature (13.7% vs. 3–5%) (1, 3, 5, 6), this was not unexpected, as our confirmed protocol deviation cohort was intentionally enriched for a higher-risk vasopressor exposure, including higher than recommended doses, prolonged infusion durations, and/or the use of multiple vasopressors concurrently. When limiting the analysis to extravasations only occurring in the confirmed deviation cohort and requiring treatment with phentolamine, the observed extravasation rate was 4.1%, comparable to the 3–5% referenced above. Notably, these studies varied in how extravasations were identified and treated, limiting direct comparison. A recent systematic review found a pooled extravasation rate of 2.2% and included both studies that had no standardized protocol for extravasation monitoring and studies that treated every extravasation with phentolamine (19).

In our secondary analysis of a subset from the original source population (all patients receiving peripheral vasopressors), 1.3% of patients received phentolamine for extravasation and the majority of cases (72.4%) were protocol concordant at the time of extravasation. These findings suggest that while protocol deviations may increase suspected extravasation events, clinically significant events remain uncommon, even in these higher-risk scenarios. This raises the question of whether dose and duration limits are effective at minimizing extravasation risk and the true clinical impact of reported extravasations.

In the deviation cohort, most patients (86.3%) deviated due to exceeding the maximum dosage. The median maximum norepinephrine dose was 0.2 µg/kg/min, which is closer to the limit in other published protocols (6, 9, 13). The second most common deviation was infusion of multiple peripheral vasopressors simultaneously, often including vasopressin, which is discouraged peripherally due to its highly vesicant properties and extravasation risk without an antidote (20, 21). Nearly all patients who deviated due to worsening shock ultimately required CVC placement, suggesting earlier placement may have been warranted and that nonadherence was often driven by clinical necessity. Last, despite a median vasopressor infusion duration of 5.6 hours, 17.8% of deviations were due to exceeding the 24-hour limit for peripheral vasopressors, a similar threshold to other protocols (11). Often, these patients were on low-dose intermittent vasopressors or had their infusion switched to a different PIV, two scenarios not explicitly addressed in our protocol. This led to clinical uncertainty as to whether temporary stoppage of vasopressors or switching PIVs reset the clock on the 24-hour limit. While both situations were treated as deviations for this study, these ambiguous situations are unlikely to be unique to our institution and highlight implementation challenges inherent to rigid time- and dose-based protocols in dynamic clinical settings.

Recently published protocols have supported the safety and potential benefits of peripheral vasopressors, including avoiding CVC placement and decreasing CLABSIs (10). However, these protocols vary widely in dosing, duration, PIV characteristics, and monitoring requirements. Few also report on adherence, making it difficult to determine whether outcomes reflect protocol design or implementation. This variation in practice was observed in a secondary analysis of the Crystalloid Liberal vs Early Vasopressors in Sepsis (CLOVERS) trial, which found study site was the only factor independently associated with the route of vasopressor initiation (22). Our institution implemented a relatively conservative (with respect to dose and duration) protocol previously found to have a low rate of extravasation requiring phentolamine of 0.8% (4). Here, we add additional context that even among patients with protocol deviations, the extravasation rate of 4.1% did not result in increased clinically significant events. Informed by these findings, our institution liberalized its protocol to allow for increased dosage (i.e., 0.2 µg/kg/min for norepinephrine) and duration (48 hr). This real-world example demonstrates how systematic evaluation of protocol deviations can inform safe, evidence-based refinements to guidelines for broader implementation.

Limitations include the retrospective single-health-system design, reliance on nursing documentation, challenges with retrospectively inferring clinician decision-making, and unstandardized extravasation documentation, leading us to focus on medication deviations. While weighted sampling ensured representation across hospital types, sample sizes were small, limiting extrapolation of deviation rates and outcomes.

CONCLUSIONS

Deviations from a peripheral vasopressor protocol primarily involved exceeding the protocol dose limit in the setting of worsening shock while awaiting central access. Clinically significant extravasations were infrequent, even among patients with protocol deviations, supporting the overall safety of peripheral vasopressors.

Supplementary Material

cc9-8-e1469-s001.pdf ^{(139.2KB, pdf)}