a b s t r a c t

Background Vascular access is a critical component of emergency department (ED) care. ultrasound guided placement of pe- ripheral intravenous (USIV) catheters is increasingly common. However, USIV are thought to suffer from reduced durability and higher complication rates. extended dwell catheters (EDC) are long peripheral IVs placed under combined ultrasound and wire guidance. The goal of this study is to compare dwell times and complication rates of EDC to standard peripheral USIV.

Methods

We performed a retrospective cohort study at a tertiary care adult ED comparing IV placements during a 17-month period (8/1/2018-12/31/2019), stratified by standard USIV versus EDC. The primary outcome was catheter dwell time and secondary outcomes included need for inpatient vascular access team (VAST) consulta- tion, peripherally inserted central catheter (PICC) insertions, and radiocontrast extravasations. Multivariable Cox regression time-to-event analyses were used to evaluate dwell times, adjusting for age, gender, BMI and end- stage renal disease.

Results 359 EDC and 4190 standard USIV were included for analysis. Most USIV (95.6%) and EDC (98.3%) were placed by ED technicians trained in ultrasound vascular access. EDC median dwell time (5.9 days [95%CI: 5.1-6.7]) exceeded standard USIV (3.8 days [95% CI: 3.6-4.0]). Patients with EDC placed in the ED required less VAST con- sultation (0.84 vs 0.99 charges/encounter), had similar rates of PICC line use (8.0% vs 8.4% of encounters) and had no radiocontrast extravasation events. Multivariable Cox regression demonstrated Survival benefit (longer dwell time) favoring EDC (HR 0.70 [95%CI 0.60-0.81]).

Conclusion Use of EDC results in longer dwell time and reduces subsequent use of vascular access resources, while maintain- ing low complication rates. EDC demonstrate superior durability which may justify their selection over standard USIV in some patients.

(C) 2021

1. Introduction
   1. Background

* Corresponding author at: Taubman Center B1-380A, 1500 E. Medical Center Dr., SPC 5305, Ann Arbor, MI 48109, USA.

E-mail address: [email protected] (C.M. Fung).

When Intravenous access is required, difficult access compli- cates approximately 10% of all emergency department (ED) visits and can lead to delays in urgent diagnostic testing or treatment [1]. Ultra- sound guided IV (USIV) placement is increasingly common and for many patients can reduce the number of insertion attempts or the

https://doi.org/10.1016/j.ajem.2021.05.005

0735-6757/(C) 2021

need for central venous access [2]. However, USIV have also been found to have higher rates of premature failure and thus significantly lower dwell times when compared to standard IVs placed with physical land- marks [3,4]. USIV are also thought to be associated with higher rates of complications including extravasation of computed tomography (CT) iodinated Contrast agents [5,6]. This has led to blanket policies at some institutions prohibiting the use of USIVs for power injection and admin- istration of potentially irritant or vesicant medicines [7]. Extended dwell catheters (EDC) are long length, ultrasound guided peripheral catheter systems with the addition of wire guidance that are approved for dwell times of up to 29 days. Use of a combined ultrasound and wire guidance system, similar to other Seldinger technique systems for vas- cular access, is thought to enhance catheter safety via reduction in ve- nous trauma during placement. Additionally, the longer length of EDC (typically 6-10 cm) compared to standard long USIV (4-6 cm) contrib- utes to their durability [8]. EDC are also different from midline catheters as they are typically inserted below the antecubital fossa, generally have a shorter catheter length, and their final placement does not approach the central veins [9]. EDC use however, has been limited by their cost relative to standard USIV as well as additional training and time re- quired for their insertion. Lack of evidence as to their durability and re- liability has also limited their use.

• 1. Objectives

Our medical center implemented an ED difficult vascular access team led by paramedic trained technicians beginning in 2008 and began utilizing EDCs in 2018. The purpose of this study is to evaluate the comparative durability of EDC to standard USIV, by assessing differ- ences in dwell times and complication rates between catheters.

1. Methods
   1. Study design

This study is a retrospective cohort study utilizing Electronic Medical Record data to evaluate the durability and complication rates of periph- eral IVs placed in the emergency department with ultrasound guidance. The study timeframe spanned 17-months (08/01/2018-12/31/2019). In the two months prior to this period, the Endurance Extended Dwell Pe- ripheral Catheter System (Fig. S1, Teleflex, Morrisvile, NC) was intro- duced in our department in 20 and 22 gauge diameters as well as 6 and 8 cm lengths for both sizes. These EDC were inserted using standard short axis ultrasound guidance with a wire guide system. During the study period, standard USIV were also inserted with similar ultrasound guidance technique. These catheters were BD Insyte Autoguard IV cath- eters (Becton, Dickinson and Co., Franklin Lakes, NJ) available in 14 to 24 gauge sizes and lengths from 2 to 4.5 cm. Both catheters are standard polyurethane plastic IV catheters.

• 1. Setting, participants and IV insertion

The study was conducted at an urban, university affiliated, teaching hospital with 106,418 adult ED visits and 41,081 patients admitted from the ED during the study period (Fig. 1). Our institutional IRB approved this study. General peripheral IV access insertion may be performed by nursing staff, paramedic trained ED technicians or physicians and physician assistants. The ED difficult IV access (DIVA) team, consisting of a subset of ED technicians, is available at all hours and by department policy may be called to place ultrasound guided IVs in cases of multiple prior attempts by the bedside nurse or known difficult IV access. All DIVA team inserters were highly experienced with more than 1000 pe- ripheral IV placements each. The DIVA team training program began in 2008 and well established prior to the study timeframe. Physician and nurse IV inserters had variable levels of experience but consisted of a tiny fraction of operators placing IVs using ultrasound guidance in this

Fig. 1. Study flow diagram. All ultrasound guided intravenous catheters (USIV) placed between 8/1/2018 and 12/31/2019 were utilized in the primary analysis. Placement of either standard USIV or extended dwell catheters (EDC) was considered difficult intravenous access.

study. All IVs placed in patients with ultrasound guidance and admitted from the ED during the study period were selected for inclusion and none were excluded from the initial analysis. Selection of EDC or stan- dard USIV was at the discretion of the inserter but EDC selection gener- ally favors patients with suspected very difficult IV access. However, both groups are considered to have difficult IV access. Departmental policy for DIVA team members mandated insertion of EDC in the fore- arm. Other inserters (primarily physicians) could place EDC at other sites. Standard USIV site selection was at the discretion of the inserter. For all inserters, departmental policy is for insertion to a depth no more than one third of the catheter length.

• 1. Data collection and variables

Our institution utilizes the Epic electronic medical record (EMR, Epic Systems, Verona, WI) and standard lines, drains and airways flowsheet charting system for all peripheral IVs placed in the hospital and this data is routinely collected during the course of care. The Epic Clarity data warehouse is also utilized to store clinical information as an Oracle da- tabase allowing direct queries for variables entered into the EMR as structured data. Variables from initial placement including time, performing provider, type of IV, length, size, guidance used, number of insertion attempts, and site were extracted directly from Epic Clarity. Removal variables including time, removal reason, and complications were also extracted. Both insertion and removal variables are main- tained as a quality analytics report at our institution. Using unique en- counter and patient identifiers, these data were joined to variables

regarding the ED visit and hospitalization as well as patient characteris- tics which were also directly extracted from our institution’s research data warehouse. Manual chart review was not performed and data ex- traction, cleaning and management was performed by the first author who was not blinded to the hypothesis of this study. These data in- cluded records of inpatient vascular access services team (VAST) charges, peripherally inserted central catheter (PICC) line insertion, computed tomography (CT) contrast injections and extravasations. All data cleaning and exploratory analyses were performed in Tableau or Tableau Prep Builder (Tableau Software, Seattle, WA). The primary ex- posure variable in this study was the type of peripheral IV placed (USIV or EDC) and each peripheral IV was considered a single observation.

• 1. Determination of peripheral IV failure

The primary outcome of this study was time to IV catheter failure and was defined as removal prior to 8 h before discharge from the hos- pital. In cases where no specific removal time was documented, the IV was determined to have remained in place without failure till discharge if charting in the IV maintenance or infusions flowsheets remained. In cases where neither removal time nor maintenance or infusion docu- mentation occurred during the hospitalization, the IV was considered to have failed early; similar to those with dwell time of less than 4 h.

• 1. Study size

No a priori sample size estimate was performed as the study time frame was determined by convenience. However, when evaluated post hoc based on a relative 9:1 ratio of standard USIV to EDC placement during the two-month trial period, a desired beta of 0.2 and alpha of 0.05, we estimated that 1751 IV failure events were required for ade- quate power when dwell time was analyzed by survival in a Cox propor- tional hazards model with a hazard ratio of 0.8 favoring EDC.

• 1. Statistical analysis

Descriptive statistics were calculated on all variables of interest. Counts with percentages are included for all categorical variables and medians with interquartile ranges for all continuous variables. Differ- ences in patient characteristics that were continuous variables were assessed for statistical significance using two sample t-tests and for cat- egorical variables the chi-squared or Fisher’s exact test were used where appropriate. Kaplan-Meier survival analysis was used to visualize sur- vival curves and estimate time to IV failure. survival curves were evalu- ated for statistically significant differences using the log rank test. Multivariable Cox proportional hazards regression was used to evaluate time to IV failure. For all survival analyses, IVs with a dwell time within 8 h before hospital discharge were right censored and this was consid- ered the primary endpoint. Covariates in the multivariable model were determined a priori as known predictors of IV failure or difficult IV access and consisted of end-stage renal disease status, gender, age, and body mass index (BMI). intravenous drug use (IVDU) status was se- lected a priori as well but not included in the final analysis due to ex- tremely low prevalence of IVDU in our cohort. Early IV failures, defined as within 4 h of placement, were excluded from the final multi- variable model for two reasons First, these were considered to be a re- flection of placement technique rather than catheter durability. Second, removal from the final Cox model significantly improved the proportionality of hazards. Additionally, vein location was not used in our model due to standard of care placement of EDC in the forearm ver- sus multiple potential sites for standard USIV. However, subgroup anal- yses were conducted limiting analyses to only observations obtained from forearm IVs. In all analyses, each IV was treated as an individual observation ignoring the effects of clustering by patient or encounter. All statistical analyses were performed in RStudio version 1.2.5

(RStudio, Boston, MA) with R version 3.6.2 (The R Foundation, Vienna, Austria).

1. Results
   1. Participants

Patient and Encounter characteristics are detailed in Table 1 and Fig. 1. A total of 4549 ultrasound guided IVs were placed during the study period and 359 of these IVs were EDC. There were no significant differences between groups in age, gender, race, BMI and comorbidities. ESRD patients were more likely to receive an EDC. IVDU was extremely low prevalence in this cohort. Of note, black and female patients were enriched in this cohort relative to the ED population during the study period where black patients represented approximately 17% of encoun- ters and females were approximately 54%.

• 1. IV characteristics

IV characteristics are detailed in Table 2. The majority of IVs in this cohort were placed by ED technicians (98.3% EDC and 95.6% USIV) and on the first attempt (92.8% EDC and 91.8% USIV). There were no signif- icant differences between groups for operator or first pass insertion suc- cess rate. Catheter length, size and placement location were different between groups by protocol.

• 1. Peripheral IV dwell time

In univariate survival analyses, median catheter dwell time was 3.8 days (95% CI 3.6-4.0) for standard USIV and 5.9 days (95% CI 5.1-6.7) for EDC (Table 3 and Fig. 2A). A higher percentage (48.2% vs. 42.7%) of EDC survived to discharge and a lower percentage (EDC: 1.1% vs. USIV: 4.2%) failed within 4 h after placement and a higher percentage of EDCs (9.5% vs. 6.6%) did not have a removal time specifically docu- mented. A multivariable Cox proportional hazards model was used to estimate an adjusted median dwell time of 4.0 days (95% CI 3.8-4.1) for standard USIV and 5.9 days (95% CI 5.1-6.7) for EDC (Fig. 2B). EDC demonstrated significantly improved survival with a HR 0.70 (95% CI 0.60-0.81) when adjusted for ESRD, gender, BMI and age (Fig. 3). Older age and male gender were also significantly associated with

Table 1

Patient characteristics.
	USIV	EDC	p-Valuea
Peripheral IVs, n (%)	4190 (92.1%)	359 (7.9%)
ED Encounters, n (%)	3668 (91.3%)	348 (8.7%)
Patients, n (%)	2669 (89.0%)	330 (11.0%)
Max PIV per ED encounter, n	4	2
Max encounters per patient, n	20	3
Max PIV per patient, n	24	3
Female, n (% of PIVs)	2634 (62.9%)	230 (64.1%)	0.69
Race
White, n (% of PIVs)	2800 (66.8%)	246 (68.5%)
Black, n (% of PIVs)	1214 (29.0%)	95 (26.5%)
Asian, n (% of PIVs)	59 (1.4%)	9 (2.5%)
Other, n (% of PIVs)	94 (2.2%)	5 (1.4%)
Not Reported, n (% of PIVs)	23 (0.1%)	4 (1.1%)
Age, Median years (IQR)	57.9	57.7	0.58
	(44.2-69.5)	(41.7-71.5)
BMI, Median kg/m2 (IQR)b	28.7	29.4	0.35
	(23.7-35.4)	(24.0-36.9)
Charlson comorbidity score, Median	4 (1-7)	4 (2-7)
(IQR) ESRD, n (% of PIVs)	536 (12.8%)	64 (17.8%)	0.01
IVDU, n (% of PIVs)c	23 (0.5%)	1 (0.3%)

^a p-Value is reported only for a priori selected covariates in subsequent analyses.

^b BMI not documented in 1 patient encounter (1 PIV).

^c IVDU status not reported in 77 patient encounters (89 PIVs).

Table 2

PIV insertion characteristics.

USIV EDC p-Value

shorter dwell times and increased complications [7]. Of note, in a small randomized controlled trial, EDC were also shown to have supe- rior dwell times to standard USIV (4.04 vs 1.25 days median survival)

Inserting Provider, n (% of Gr ED Technician	oup) 4007 (95.6%)	353 (98.3%)	0.82a	[8]. While our results are similar, key differences between our study and Bahl et al. [8] are differences in inserting operator and significantly
Physician	60 (1.4%)	5 (1.4%)		larger sample size in our study (>4000 USIV and 359 EDC versus 33 and

Nurse 6 (0.1%)

Not Documented 117 (2.8%) 1 (0.3%)

Insertion Attempts, (% of Group) 1	3846 (91.8%)	333 (92.8%)	0.62b
2	224 (5.4%)	19 (5.3%)
3	25 (0.6%)	4 (1.1%)
4	3 (0.1%)
Not Documented	92 (2.2%)	3 (0.8%)

Catheter Length, n (% of Group)

<1.5 in 250 (6.0%)

>1.5 in 3687 (88.0%)

6 cm 313 (87.2%)

8 cm 34 (9.5%)

10 cm 3 (1.1%)

Not Documented 253 (6.0%) 9 (1.4%) Catheter Gauge, n (% of Group)

14 g 2 (0.1%)

16 g	31 (0.7%)
18 g	769 (18.3%)	11 (3.1%)
20 g	3300 (78.8%)	165 (46.0%)

37 respectively). Similarly, in this study and in others evaluating USIV to non-ultrasound guided IVs [3,10], the dwell times of standard USIV were approximately 2 days suggesting a significant difference between our baseline USIV survival and that of other centers. Several important differences in our study that may explain this discrepancy are inclusion of only patients admitted to the hospital in our study, a long history of the use of USIV in our ED, and maintenance after admission of all pe- ripheral IVs by a dedicated vascular access team. Additionally, our study’s relatively long mean hospital length of stay (approximately 5 days) may have influenced dwell times. The etiology of increased dwell time of EDC is likely multifactorial and due in part to these factors as well as the longer catheter length which has been previously shown to improve catheter dwell time [10]. In addition to increased dwell time, we have also shown that EDC catheters result in fewer vascular access charges per encounter, without a significant increase in other complica- tions including CT contrast extravasation, and similar differences in re-

22 g	64 (1.5%)	177 (49.3%)	moval reason or PICC line insertions while maintaining a high level of
24 g	1 (0%)		first pass insertion success. Interestingly, in our multivariable model
Not Documented	26 (0.6%)	6 (1.7%)	for time to IV failure, ESRD status exhibited a trend towards reduced

Location, n (% of Group)

Neck 2 (0.1%)

Upper Arm 1225 (29.3%) 13 (3.6%)

hazard of failure which reached statistical significance when examining only forearm placed IVs (Fig. 3 and Fig. S2). The etiology of this effect is

Antecubital Fossa	387 (9.2%)	1 (0.3%)	unclear although likely due to increased efforts by all patient care teams
Brachial	165 (3.9%)		to preserve IV access and longer lengths of stay in ESRD patients. Age
Forearm	2092 (49.9%)	302 (84.1%)	also exhibited this protective effect in our model which we believe is
Wrist Hand Lower Leg	193 (4.6%) 10 (0.2%) 35 (0.8%)	31 (8.6%) 7 (2.0%)	due to similar reasons. Overall, our results demonstrate improved dura- bility of EDC compared to standard USIV even in our setting where base-
Foot or Ankle	2 (0.1%)		line dwell time is high and complication rates are relatively low. Given
Other Not Documented	9 (0.2%) 70 (1.7%)	2 (0.6%) 3 (0.8%)	increasing popularity of midline catheter use both in admitted and ED patients [11,12], as well as increasing emphasis on appropriate selection

^a ED technician placed versus other.

^b First pass success versus more than 1 attempt.

decreased risk of failure. In this model, we removed early failure (dwell time < 4 h) observations and one observation was missing BMI data which was also subject to listwise deletion. All covariates included the final model did not demonstrate collinearity and the final model did not violate the proportional hazards assumption. Subgroup analysis of IVs placed in the forearm and exclusion of short (<1.5 in.) standard USIV did not significantly alter these results. These data are presented in the supplemental figures (Figs. S2 and S3).

• 1. IV outcomes“>Other peripheral IV outcomes

There were fewer VAST charges per encounter in the EDC group (0.84 charges/encounter in EDC versus 0.99/encounter in USIV, p < 0.0001, Table 3). In approximately 45% of total observations, a removal reason was documented (Chi-sq p-value = 0.313). Documented re- moval reasons (Table 4) were similar between USIV and EDC groups but not analyzed for statistical significance due to limited numbers of observations in the EDC group. First insertion attempt success was also similar between groups. Finally, there were no documented CT con- trast extravasations in the EDC group and 2 in the USIV group.

1. Discussion

Our results demonstrate that EDC placements, primarily performed by paramedic trained ED technicians, have superior dwell time to stan- dard USIV even after adjustment for age, gender, BMI and ESRD status. Prior studies have called into question the durability of USIV citing

of vascular access to match duration and type of infusion therapy [9], EDC may represent a new alternative in the spectrum of vascular access devices.

• 1. Limitations

Our study has several limitations. First, measurement of dwell time was calculated as days between documented placement and removal. However, in approximately 9.5% of EDC and 6.6% of standard USIV no re- moval time was documented. These observations continued to have other IV maintenance or medication charting and were treated as if they had survival to hospital discharge. The survival analysis with these observations removed did not result in significant changes (Fig. S4). Secondly, while the vast majority of EDC and USIV are placed by ED technicians in our DIVA team, a handful were placed by physi- cians or nurses. These observations represent a tiny fraction of total IVs in this cohort and thus were not expected to modify results. Prior studies have also demonstrated that placement location can adversely affect dwell times of ED placed peripheral IVs [13]. By departmental pol- icy, EDC are placed almost exclusively in the forearm while USIV have more heterogeneity in site selection (Table 3). Similarly, there is more heterogeneity in catheter size of USIV when compared to EDC which are predominantly 20 g or 22 g. DIVA team members use their discretion based on clinical judgement when selecting EDC or standard USIV which likely introduces selection bias based on the perceived need for longer term peripheral IV access or difficulty of IV insertion. This may be the reason for enrichment of ESRD patients in the EDC group and lon- ger observed hospital length of stay. Prior studies have demonstrated that insertion depth is a significant predictor for IV failure and, in our in- stitution insertion depth from skin to target vein is not documented. Thus, this variable could not be analyzed. However, the postulated

Fig. 2. Unadjusted and adjusted Kaplan-Meier survival curves. A) Unadjusted survival analysis curve with all USIV and EDC in the cohort included for analysis. Risk table shown below.

B) Adjusted survival curve where early PIV failures (<4 h) are excluded and covariate means (age 56.7 and BMI 30.8) or reference condition (absence of ESRD and female) are fixed. Dotted lines indicate estimated median survival.

mechanism for failure of IVs placed in deeper veins is due to short cath- eter length actually inserted into the vein which is less likely in our de- partment given policy limiting insertion depth to no more than one third the catheter length and use of predominantly long (4.5 cm) IV catheters in the standard USIV group. Finally, in our multivariable anal- ysis, early IV failures (dwell time < 4 h) were removed because these

failures were considered to be failed insertion technique rather than re- flective of catheter durability and their removal from analysis improved proportionality of hazards in the Cox model. A higher percentage of early failures occurred in the USIV group compared to EDC (4.4% versus 1.1%, Table 3) and our overall results were not altered by retention of these observations in the multivariable analysis.

Fig. 3. Forest Plot for Hazard of Peripheral IV (PIV) Failure. A multivariable Cox proportional hazards model was constructed using time to failure as the outcome of interest and PIV type, ESRD status, Female gender, BMI and age (per 10 years) as predictor variables. Extended dwell catheter (EDC) and higher age predicted longer catheter dwell time. Female gender predicted shorter catheter dwell time.

Table 3					Table 4
PIV outcomes.					Removal reason.
	USIV	EDC	p-Value			USIV	EDC
Median dwell time, Days (95% CI)a	3.8	5.9	<0.0001		Catheter damage, n (%)	165 (3.9%)	13 (3.6%)
	(3.6-4.0)	(5.1-6.7)			Change in patient condition, n (%)	35 (0.8%)	3 (0.8%)
Median hospital LOS, Days (IQR)	4.5	5.2	0.018		Change in site condition, n (%)	108 (2.6%)	9 (2.5%)
	(2.4-8.1)	(3.0-9.5)			Damage to catheter, n (%)	79 (1.9%)	5 (1.4%)
PIV survival to discharge, n (%)b	1791	173	0.046		Drainage, n (%)	69 (1.6%)	7 (1.9%)
	(42.7%)	(48.2%)			Infiltrated, n (%)	275 (6.6%)	13 (3.6%)
PIV failure within 4 h, n (%)	177 (4.2%)	4 (1.1%)	0.002		Occluded, n (%)	292 (7.0%)	30 (8.4%)
PIV removal time not documentedc	277 (6.6%)	34 (9.5%)	0.039		Other, n (%)	270 (6.4%)	28 (7.8%)
VAST PIV charges, n (#/encounter)	3643	291 (0.84)	<0.0001		Patient discharged, n (%)	212 (5.1%)	17 (4.7%)
	(0.99)				Per order, n (%)	520 (12.4%)	51 (14.2%)
Encounters with zero VAST charges, n (% of	2289	227	0.298		Per patient/family request, n (%)	157 (3.7%)	5 (1.4%)
encounters)	(62.4%)	(65.2%)			Planned change, n (%)	20 (0.5%)	4 (1.1%)
PICC line insertions, n (% of encounters)	309 (8.4%)	28 (8.0%)	0.808		Unplanned by other, n (%)	33 (0.8%)	1 (0.3%)
CT contrast injections, n	2490	214			Unplanned by patient, n (%)	133 (3.2%)	7 (1.9%)
CT contrast extravasations, n (% of injections)	2 (0.08%)	0 (0%)	1.00		Not documented, n (%)	1822 (43.5%)	166 (46.2%)

^a Median dwell time was estimated using the Kaplan-Meier method.

^b Survival to discharge defined as survival to within 8 h of discharge.

^c For PIVs without documented removal time, hospital discharge time was substituted.

• 1. Generalizability

There are also several factors that limit the generalizability of our findings. First, in our institution all ED placed IVs are maintained by a dedicated vascular access team after admission. This is unlike other in- stitutions where bedside nurses or medical assistants may be responsi- ble for IV maintenance. Also, by institutional policy, Blood draws from existing peripheral IVs are prohibited after patients leave the ED. These two factors may contribute to longer dwell times seen at our in- stitution relative to others. Our institution is also a tertiary care center with a high prevalence of difficult access including ESRD patients which may limit applicability to other centers where difficult IV access is less common. As a result of this, our department has employed a para- medic trained, ED technician run DIVA team since 2008 and all current members of this team are highly experienced. At other centers where

physicians or nurses primarily perform ultrasound guided IV access these results may not be applicable. Finally, the retrospective nature of this study and reliance on EMR flowsheet documentation may be inher- ent limitations on generalizability.

1. Conclusions

Extended dwell peripheral IVs are a superior alternative to standard long IV catheters with improved dwell time and a very low rate of com- plications when placed by emergency department technicians.

Supplementary data to this article can be found online at https://doi. org/10.1016/j.ajem.2021.05.005.

Prior presentations

Components of this work have previously been presented at the SAEM 2020 Annual Meeting.

Funding sources/disclosures

CF has received research support unrelated to this work from the Na- tional Heart, Lung, and Blood Institute (1K12HL133304). NT has re- ceived research support unrelated to this work from GE Healthcare Inc.

Declaration of Competing Interest

JT reports that has received payment from Teleflex Inc. as a clinical educator.

Acknowledgements

The authors acknowledge and thank the ED technicians, nurses, phy- sicians and physician assistants of the Department of Emergency Medi- cine at Michigan Medicine. The authors also acknowledge and thank the Vascular Access Services Team at Michigan Medicine.

References

1. Davis EM, Feinsmith S, Amick AE, Sell J, McDonald V, Trinquero P, et al. Difficult in- travenous access in the emergency department: Performance and impact of ultrasound-guided IV insertion performed by nurses. Am J Emerg Med. 2020. https://doi.org/10.1016/j.ajem.2020.11.013.
2. Au AK, Rotte MJ, Grzybowski RJ, Ku BS, Fields JM. Decrease in Central venous catheter placement due to use of ultrasound guidance for peripheral intravenous catheters. Am J Emerg Med. 2012;30(9):1950-4. https://doi.org/10.1016/J.AJEM.2012.04.016.
3. Adhikari S, Blaivas M, Morrison D, Lander L. Comparison of infection rates among ultrasound-guided versus traditionally placed peripheral intravenous lines. J Ultra- sound Med. 2010;29(5):741-7. https://doi.org/10.7863/jum.2010.29.5.741.
4. Dargin JM, Rebholz CM, Lowenstein RA, Mitchell PM, Feldman JA. Ultrasonography- guided peripheral intravenous catheter survival in ED patients with difficult access. Am J Emerg Med. 2010;28(1):1-7. https://doi.org/10.1016/j.ajem.2008.09.001.
5. Rupp JD, Ferre RM, Boyd JS, Dearing E, McNaughton CD, Liu D, et al. Extravasation risk using ultrasound-guided peripheral intravenous catheters for computed tomog- raphy contrast administration. Jang TB, ed. Acad Emerg Med. 2016;23(8):918-21. https://doi.org/10.1111/acem.13000.
6. Favot M, Gallien J, Malik A, Kasten A, Wells R, Ehrman R. Contrast extravasation as a complication of emergency nurse-performed ultrasound-guided peripheral intrave- nous catheter placement. J Emerg Nurs. 2019;45(5):512-6. https://doi.org/10.1016/ j.jen.2019.05.016.
7. Witting MD, Moayedi S, Dunning K, Babin LS, Cogan BM. Power injection through ultrasound-guided intravenous lines: safety and efficacy under an institutional pro- tocol. J Emerg Med. 2017;52:16-22. https://doi.org/10.1016/j.jemermed.2016.09. 017 Elsevier USA.
8. Bahl A, Hang B, Brackney A, Joseph S, Karabon P, Mohammad A, et al. Standard long

IV catheters versus extended dwell catheters: a randomized comparison of ultrasound-guided catheter survival. Am J Emerg Med. 2019;37(4):715-21. https://doi.org/10.1016/j.ajem.2018.07.031.

1. Chopra V, Flanders SA, Saint S, Woller SC, O’Grady NP, Safdar N, et al. The Michigan Appropriateness Guide for Intravenous Catheters (MAGIC): results from a multispecialty panel using the RAND/UCLA appropriateness method. Ann Intern Med. 2015;163(6_Supplement):S1. https://doi.org/10.7326/M15-0744.
2. Pandurangadu AV, Tucker J, Brackney AR, Bahl A. Ultrasound-guided intravenous catheter survival impacted by amount of catheter residing in the vein. Emerg Med

J. 2018;35(9):550-5. https://doi.org/10.1136/emermed-2017-206803.

1. Chopra V, Kaatz S, Swaminathan L, Boldenow T, Snyder A, Burris R, et al. Variation in use and outcomes related to midline catheters: results from a multicentre pilot study. BMJ Qual Saf. 2019;28(9):714-20. https://doi.org/10.1136/bmjqs-2018- 008554.
2. Spiegel RJ, Eraso D, Leibner E, Thode H, Morley EJ, Weingart S. The utility of midline intravenous catheters in critically ill emergency department patients. Ann Emerg Med. 2020;75(4):538-45. https://doi.org/10.1016/j.annemergmed.2019.09.018.
3. Fields JM, Dean AJ, Todman RW, Au AK, Anderson KL, Ku BS, et al. The effect of vessel depth, diameter, and location on ultrasound-guided peripheral intravenous catheter longevity. Am J Emerg Med. 2012;30(7):1134-40. https://doi.org/10.1016/j.ajem. 2011.07.027.